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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1563 - (show annotations)
Mon Jun 8 17:55:54 2015 UTC (4 years, 4 months ago) by ph10
File size: 279345 byte(s)
Add integer overflow check to (?n) code.
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
46 Version 8.37 28-April-2015
47 --------------------------
48
49 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
50 for those parentheses to be closed with whatever has been captured so far.
51 However, it was failing to mark any other groups between the hightest
52 capture so far and the currrent group as "unset". Thus, the ovector for
53 those groups contained whatever was previously there. An example is the
54 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
55
56 2. If an assertion condition was quantified with a minimum of zero (an odd
57 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
58
59 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
60 unrecognized modifier, a crash could occur.
61
62 4. An attempt to do global matching in pcretest with a zero-length ovector
63 caused a crash.
64
65 5. Fixed a memory leak during matching that could occur for a subpattern
66 subroutine call (recursive or otherwise) if the number of captured groups
67 that had to be saved was greater than ten.
68
69 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
70 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
71 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
72
73 7. A UTF pattern containing a "not" match of a non-ASCII character and a
74 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
75
76 8. When a pattern is compiled, it remembers the highest back reference so that
77 when matching, if the ovector is too small, extra memory can be obtained to
78 use instead. A conditional subpattern whose condition is a check on a
79 capture having happened, such as, for example in the pattern
80 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
81 setting the highest backreference number. This mattered only if pcre_exec()
82 was called with an ovector that was too small to hold the capture, and there
83 was no other kind of back reference (a situation which is probably quite
84 rare). The effect of the bug was that the condition was always treated as
85 FALSE when the capture could not be consulted, leading to a incorrect
86 behaviour by pcre_exec(). This bug has been fixed.
87
88 9. A reference to a duplicated named group (either a back reference or a test
89 for being set in a conditional) that occurred in a part of the pattern where
90 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
91 to be incorrectly calculated, leading to overwriting.
92
93 10. A mutually recursive set of back references such as (\2)(\1) caused a
94 segfault at study time (while trying to find the minimum matching length).
95 The infinite loop is now broken (with the minimum length unset, that is,
96 zero).
97
98 11. If an assertion that was used as a condition was quantified with a minimum
99 of zero, matching went wrong. In particular, if the whole group had
100 unlimited repetition and could match an empty string, a segfault was
101 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
102 assertions to be quantified, but not if they are being used as conditions,
103 so the above pattern is faulted by Perl. PCRE has now been changed so that
104 it also rejects such patterns.
105
106 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
107 failed to allow the zero-repeat case if pcre2_exec() was called with an
108 ovector too small to capture the group.
109
110 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
111 Red Hat Product Security:
112
113 (a) A crash if /K and /F were both set with the option to save the compiled
114 pattern.
115
116 (b) Another crash if the option to print captured substrings in a callout
117 was combined with setting a null ovector, for example \O\C+ as a subject
118 string.
119
120 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
121 forward reference repeated a large (but limited) number of times within a
122 repeated outer group that has a zero minimum quantifier, caused incorrect
123 code to be compiled, leading to the error "internal error:
124 previously-checked referenced subpattern not found" when an incorrect
125 memory address was read. This bug was reported as "heap overflow",
126 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
127 CVE-2015-2325.
128
129 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
130 call within a group that also contained a recursive back reference caused
131 incorrect code to be compiled. This bug was reported as "heap overflow",
132 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
133 number CVE-2015-2326.
134
135 24. Computing the size of the JIT read-only data in advance has been a source
136 of various issues, and new ones are still appear unfortunately. To fix
137 existing and future issues, size computation is eliminated from the code,
138 and replaced by on-demand memory allocation.
139
140 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
141 adjacent to the end of the range, and the range contained characters with
142 more than one other case, caused incorrect behaviour when compiled in UTF
143 mode. In that example, the range a-j was left out of the class.
144
145 26. Fix JIT compilation of conditional blocks, which assertion
146 is converted to (*FAIL). E.g: /(?(?!))/.
147
148 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
149 discovered by the LLVM fuzzer.
150
151 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
152 when this assertion was used as a condition, for example (?(?!)a|b). In
153 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
154 error about an unsupported item.
155
156 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
157 possessification code could take exponential time to complete. A recursion
158 depth limit of 1000 has been imposed to limit the resources used by this
159 optimization.
160
161 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
162 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
163 because \S ensures they are all in the class. The code for doing this was
164 interacting badly with the code for computing the amount of space needed to
165 compile the pattern, leading to a buffer overflow. This bug was discovered
166 by the LLVM fuzzer.
167
168 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
169 other kinds of group caused stack overflow at compile time. This bug was
170 discovered by the LLVM fuzzer.
171
172 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
173 between a subroutine call and its quantifier was incorrectly compiled,
174 leading to buffer overflow or other errors. This bug was discovered by the
175 LLVM fuzzer.
176
177 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
178 assertion after (?(. The code was failing to check the character after
179 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
180 was discovered by the LLVM fuzzer.
181
182 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
183 a fixed maximum following a group that contains a subroutine reference was
184 incorrectly compiled and could trigger buffer overflow. This bug was
185 discovered by the LLVM fuzzer.
186
187 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
188 caused a stack overflow instead of the diagnosis of a non-fixed length
189 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
190
191 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
192 (e.g. /(?<=\Ka)/) could make pcregrep loop.
193
194 37. There was a similar problem to 36 in pcretest for global matches.
195
196 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
197 and a subsequent item in the pattern caused a non-match, backtracking over
198 the repeated \X did not stop, but carried on past the start of the subject,
199 causing reference to random memory and/or a segfault. There were also some
200 other cases where backtracking after \C could crash. This set of bugs was
201 discovered by the LLVM fuzzer.
202
203 39. The function for finding the minimum length of a matching string could take
204 a very long time if mutual recursion was present many times in a pattern,
205 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
206 method has been implemented. This infelicity was discovered by the LLVM
207 fuzzer.
208
209 40. Static linking against the PCRE library using the pkg-config module was
210 failing on missing pthread symbols.
211
212
213 Version 8.36 26-September-2014
214 ------------------------------
215
216 1. Got rid of some compiler warnings in the C++ modules that were shown up by
217 -Wmissing-field-initializers and -Wunused-parameter.
218
219 2. The tests for quantifiers being too big (greater than 65535) were being
220 applied after reading the number, and stupidly assuming that integer
221 overflow would give a negative number. The tests are now applied as the
222 numbers are read.
223
224 3. Tidy code in pcre_exec.c where two branches that used to be different are
225 now the same.
226
227 4. The JIT compiler did not generate match limit checks for certain
228 bracketed expressions with quantifiers. This may lead to exponential
229 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
230 issue should be resolved now.
231
232 5. Fixed an issue, which occures when nested alternatives are optimized
233 with table jumps.
234
235 6. Inserted two casts and changed some ints to size_t in the light of some
236 reported 64-bit compiler warnings (Bugzilla 1477).
237
238 7. Fixed a bug concerned with zero-minimum possessive groups that could match
239 an empty string, which sometimes were behaving incorrectly in the
240 interpreter (though correctly in the JIT matcher). This pcretest input is
241 an example:
242
243 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
244 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
245
246 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
247 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
248 for an empty string was breaking the inner loop and carrying on at a lower
249 level, when possessive repeated groups should always return to a higher
250 level as they have no backtrack points in them. The empty string test now
251 occurs at the outer level.
252
253 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
254 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
255
256 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
257 doesn't).
258
259 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
260 to hold a pattern to be incorrectly computed (too small) when there were
261 named back references to duplicated names. This could cause "internal
262 error: code overflow" or "double free or corruption" or other memory
263 handling errors.
264
265 11. When named subpatterns had the same prefixes, back references could be
266 confused. For example, in this pattern:
267
268 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
269
270 the reference to 'Name' was incorrectly treated as a reference to a
271 duplicate name.
272
273 12. A pattern such as /^s?c/mi8 where the optional character has more than
274 one "other case" was incorrectly compiled such that it would only try to
275 match starting at "c".
276
277 13. When a pattern starting with \s was studied, VT was not included in the
278 list of possible starting characters; this should have been part of the
279 8.34/18 patch.
280
281 14. If a character class started [\Qx]... where x is any character, the class
282 was incorrectly terminated at the ].
283
284 15. If a pattern that started with a caseless match for a character with more
285 than one "other case" was studied, PCRE did not set up the starting code
286 unit bit map for the list of possible characters. Now it does. This is an
287 optimization improvement, not a bug fix.
288
289 16. The Unicode data tables have been updated to Unicode 7.0.0.
290
291 17. Fixed a number of memory leaks in pcregrep.
292
293 18. Avoid a compiler warning (from some compilers) for a function call with
294 a cast that removes "const" from an lvalue by using an intermediate
295 variable (to which the compiler does not object).
296
297 19. Incorrect code was compiled if a group that contained an internal recursive
298 back reference was optional (had quantifier with a minimum of zero). This
299 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
300 caused segmentation faults because of stack overflows at compile time.
301
302 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
303 group that is quantified with an indefinite repeat, caused a compile-time
304 loop which used up all the system stack and provoked a segmentation fault.
305 This was not the same bug as 19 above.
306
307 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
308 Patch by Mike Frysinger.
309
310
311 Version 8.35 04-April-2014
312 --------------------------
313
314 1. A new flag is set, when property checks are present in an XCLASS.
315 When this flag is not set, PCRE can perform certain optimizations
316 such as studying these XCLASS-es.
317
318 2. The auto-possessification of character sets were improved: a normal
319 and an extended character set can be compared now. Furthermore
320 the JIT compiler optimizes more character set checks.
321
322 3. Got rid of some compiler warnings for potentially uninitialized variables
323 that show up only when compiled with -O2.
324
325 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
326 of a match later then the end of the match. The pcretest program was not
327 handling the case sensibly - it was outputting from the start to the next
328 binary zero. It now reports this situation in a message, and outputs the
329 text from the end to the start.
330
331 5. Fast forward search is improved in JIT. Instead of the first three
332 characters, any three characters with fixed position can be searched.
333 Search order: first, last, middle.
334
335 6. Improve character range checks in JIT. Characters are read by an inprecise
336 function now, which returns with an unknown value if the character code is
337 above a certain threshold (e.g: 256). The only limitation is that the value
338 must be bigger than the threshold as well. This function is useful when
339 the characters above the threshold are handled in the same way.
340
341 7. The macros whose names start with RAWUCHAR are placeholders for a future
342 mode in which only the bottom 21 bits of 32-bit data items are used. To
343 make this more memorable for those maintaining the code, the names have
344 been changed to start with UCHAR21, and an extensive comment has been added
345 to their definition.
346
347 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
348 to the export list in Makefile.am (they were accidentally omitted from the
349 8.34 tarball).
350
351 9. The informational output from pcretest used the phrase "starting byte set"
352 which is inappropriate for the 16-bit and 32-bit libraries. As the output
353 for "first char" and "need char" really means "non-UTF-char", I've changed
354 "byte" to "char", and slightly reworded the output. The documentation about
355 these values has also been (I hope) clarified.
356
357 10. Another JIT related optimization: use table jumps for selecting the correct
358 backtracking path, when more than four alternatives are present inside a
359 bracket.
360
361 11. Empty match is not possible, when the minimum length is greater than zero,
362 and there is no \K in the pattern. JIT should avoid empty match checks in
363 such cases.
364
365 12. In a caseless character class with UCP support, when a character with more
366 than one alternative case was not the first character of a range, not all
367 the alternative cases were added to the class. For example, s and \x{17f}
368 are both alternative cases for S: the class [RST] was handled correctly,
369 but [R-T] was not.
370
371 13. The configure.ac file always checked for pthread support when JIT was
372 enabled. This is not used in Windows, so I have put this test inside a
373 check for the presence of windows.h (which was already tested for).
374
375 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
376 The algorithm provides a way to skip certain starting offsets, and usually
377 faster than linear prefix searches.
378
379 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
380 as for 'fr_FR' and 'french'. For some reason, however, it then used the
381 Windows-specific input and output files, which have 'french' screwed in.
382 So this could never have worked. One of the problems with locales is that
383 they aren't always the same. I have now updated RunTest so that it checks
384 the output of the locale test (test 3) against three different output
385 files, and it allows the test to pass if any one of them matches. With luck
386 this should make the test pass on some versions of Solaris where it was
387 failing. Because of the uncertainty, the script did not used to stop if
388 test 3 failed; it now does. If further versions of a French locale ever
389 come to light, they can now easily be added.
390
391 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
392 there was a message during ./configure, but it did not stop. This now
393 provokes an error. The invalid example in README has been corrected.
394 If a value less than the minimum is given, the minimum value has always
395 been used, but now a warning is given.
396
397 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
398 was a bug in the test system, which is now fixed. Also, the list of various
399 configurations that are tested for each release did not have one with both
400 16/32 bits and --enable-bar-anycrlf. It now does.
401
402 18. pcretest was missing "-C bsr" for displaying the \R default setting.
403
404 19. Little endian PowerPC systems are supported now by the JIT compiler.
405
406 20. The fast forward newline mechanism could enter to an infinite loop on
407 certain invalid UTF-8 input. Although we don't support these cases
408 this issue can be fixed by a performance optimization.
409
410 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
411 not take account if existing stack usage. There is now a new global
412 variable called pcre_stack_guard that can be set to point to an external
413 function to check stack availability. It is called at the start of
414 processing every parenthesized group.
415
416 22. A typo in the code meant that in ungreedy mode the max/min qualifier
417 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
418 not match "ab".
419
420 23. When UTF was disabled, the JIT program reported some incorrect compile
421 errors. These messages are silenced now.
422
423 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
424 compiler.
425
426 25. Change all the temporary files used in RunGrepTest to be different to those
427 used by RunTest so that the tests can be run simultaneously, for example by
428 "make -j check".
429
430
431 Version 8.34 15-December-2013
432 -----------------------------
433
434 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
435 executable memory. Patch inspired by Carsten Klein.
436
437 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
438 this macro is never tested and has no effect, because the work to support
439 coverage involves only compiling and linking options and special targets in
440 the Makefile. The comment in config.h implied that defining the macro would
441 enable coverage support, which is totally false. There was also support for
442 setting this macro in the CMake files (my fault, I just copied it from
443 configure). SUPPORT_GCOV has now been removed.
444
445 3. Make a small performance improvement in strlen16() and strlen32() in
446 pcretest.
447
448 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
449 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
450 these cases). There was also one in pcretest.c.
451
452 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
453
454 6. In UTF mode, the code for checking whether a group could match an empty
455 string (which is used for indefinitely repeated groups to allow for
456 breaking an infinite loop) was broken when the group contained a repeated
457 negated single-character class with a character that occupied more than one
458 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
459 UTF-8 mode). The effect was undefined: the group might or might not be
460 deemed as matching an empty string, or the program might have crashed.
461
462 7. The code for checking whether a group could match an empty string was not
463 recognizing that \h, \H, \v, \V, and \R must match a character.
464
465 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
466 an empty string. If it can, pcretest shows this in its information output.
467
468 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
469 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
470 matched by pcre_exec() without using JIT:
471
472 (a) If the rest of the pattern did not match after a maximal run of
473 grapheme clusters, the code for backing up to try with fewer of them
474 did not always back up over a full grapheme when characters that do not
475 have the modifier quality were involved, e.g. Hangul syllables.
476
477 (b) If the match point in a subject started with modifier character, and
478 there was no match, the code could incorrectly back up beyond the match
479 point, and potentially beyond the first character in the subject,
480 leading to a segfault or an incorrect match result.
481
482 10. A conditional group with an assertion condition could lead to PCRE
483 recording an incorrect first data item for a match if no other first data
484 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
485 first data item, and therefore matched "ca" after "c" instead of at the
486 start.
487
488 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
489 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
490
491 12. The source of pcregrep now includes z/OS-specific code so that it can be
492 compiled for z/OS as part of the special z/OS distribution.
493
494 13. Added the -T and -TM options to pcretest.
495
496 14. The code in pcre_compile.c for creating the table of named capturing groups
497 has been refactored. Instead of creating the table dynamically during the
498 actual compiling pass, the information is remembered during the pre-compile
499 pass (on the stack unless there are more than 20 named groups, in which
500 case malloc() is used) and the whole table is created before the actual
501 compile happens. This has simplified the code (it is now nearly 150 lines
502 shorter) and prepared the way for better handling of references to groups
503 with duplicate names.
504
505 15. A back reference to a named subpattern when there is more than one of the
506 same name now checks them in the order in which they appear in the pattern.
507 The first one that is set is used for the reference. Previously only the
508 first one was inspected. This change makes PCRE more compatible with Perl.
509
510 16. Unicode character properties were updated from Unicode 6.3.0.
511
512 17. The compile-time code for auto-possessification has been refactored, based
513 on a patch by Zoltan Herczeg. It now happens after instead of during
514 compilation. The code is cleaner, and more cases are handled. The option
515 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
516 options in pcretest are provided to set it. It can also be set by
517 (*NO_AUTO_POSSESS) at the start of a pattern.
518
519 18. The character VT has been added to the default ("C" locale) set of
520 characters that match \s and are generally treated as white space,
521 following this same change in Perl 5.18. There is now no difference between
522 "Perl space" and "POSIX space". Whether VT is treated as white space in
523 other locales depends on the locale.
524
525 19. The code for checking named groups as conditions, either for being set or
526 for being recursed, has been refactored (this is related to 14 and 15
527 above). Processing unduplicated named groups should now be as fast at
528 numerical groups, and processing duplicated groups should be faster than
529 before.
530
531 20. Two patches to the CMake build system, by Alexander Barkov:
532
533 (1) Replace the "source" command by "." in CMakeLists.txt because
534 "source" is a bash-ism.
535
536 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
537 without these the CMake build does not work on Solaris.
538
539 21. Perl has changed its handling of \8 and \9. If there is no previously
540 encountered capturing group of those numbers, they are treated as the
541 literal characters 8 and 9 instead of a binary zero followed by the
542 literals. PCRE now does the same.
543
544 22. Following Perl, added \o{} to specify codepoints in octal, making it
545 possible to specify values greater than 0777 and also making them
546 unambiguous.
547
548 23. Perl now gives an error for missing closing braces after \x{... instead of
549 treating the string as literal. PCRE now does the same.
550
551 24. RunTest used to grumble if an inappropriate test was selected explicitly,
552 but just skip it when running all tests. This make it awkward to run ranges
553 of tests when one of them was inappropriate. Now it just skips any
554 inappropriate tests, as it always did when running all tests.
555
556 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
557 character types such as \d or \w, too many callouts were inserted, and the
558 data that they returned was rubbish.
559
560 26. In UCP mode, \s was not matching two of the characters that Perl matches,
561 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
562 were matched by \h. The code has now been refactored so that the lists of
563 the horizontal and vertical whitespace characters used for \h and \v (which
564 are defined only in one place) are now also used for \s.
565
566 27. Add JIT support for the 64 bit TileGX architecture.
567 Patch by Jiong Wang (Tilera Corporation).
568
569 28. Possessive quantifiers for classes (both explicit and automatically
570 generated) now use special opcodes instead of wrapping in ONCE brackets.
571
572 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
573 (because it's meaningless), this was not happening when PCRE_CASELESS was
574 set. Not wrong, but inefficient.
575
576 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
577 properties for \w, \d, etc) is present in a test regex. Otherwise if the
578 test contains no characters greater than 255, Perl doesn't realise it
579 should be using Unicode semantics.
580
581 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
582 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
583 does in Unicode mode.
584
585 32. Added the "forbid" facility to pcretest so that putting tests into the
586 wrong test files can sometimes be quickly detected.
587
588 33. There is now a limit (default 250) on the depth of nesting of parentheses.
589 This limit is imposed to control the amount of system stack used at compile
590 time. It can be changed at build time by --with-parens-nest-limit=xxx or
591 the equivalent in CMake.
592
593 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
594 errors. Perl warns for these when in warning mode, but PCRE has no facility
595 for giving warnings.
596
597 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
598 However, this was not working for (?!) because it is optimized to (*FAIL),
599 for which PCRE does not allow quantifiers. The optimization is now disabled
600 when a quantifier follows (?!). I can't see any use for this, but it makes
601 things uniform.
602
603 36. Perl no longer allows group names to start with digits, so I have made this
604 change also in PCRE. It simplifies the code a bit.
605
606 37. In extended mode, Perl ignores spaces before a + that indicates a
607 possessive quantifier. PCRE allowed a space before the quantifier, but not
608 before the possessive +. It now does.
609
610 38. The use of \K (reset reported match start) within a repeated possessive
611 group such as (a\Kb)*+ was not working.
612
613 40. Document that the same character tables must be used at compile time and
614 run time, and that the facility to pass tables to pcre_exec() and
615 pcre_dfa_exec() is for use only with saved/restored patterns.
616
617 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
618 features for Builds with MSVC:
619
620 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
621 builds). The .rc files can be used to set FileDescription and many other
622 attributes.
623
624 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
625 This allows higher-level build scripts which want .pdb files to avoid
626 hard-coding the exact files needed."
627
628 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
629 mean "start of word" and "end of word", respectively, as a transition aid.
630
631 43. A minimizing repeat of a class containing codepoints greater than 255 in
632 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
633 compiled to use the heap for recursion.
634
635 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
636 is configured.
637
638
639 Version 8.33 28-May-2013
640 ------------------------
641
642 1. Added 'U' to some constants that are compared to unsigned integers, to
643 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
644 variables that are added to signed variables, to ensure the result is
645 signed and can be negated.
646
647 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
648 CMake config files.
649
650 3. Revise the creation of config.h.generic so that all boolean macros are
651 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
652 overriding via -D on the command line possible.
653
654 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
655 to unsigned int is reported to make a quite noticeable speed difference in
656 a specific Windows environment. Testing on Linux did also appear to show
657 some benefit (and it is clearly not harmful). Also fixed the definition of
658 Xop which should be unsigned.
659
660 5. Related to (4), changing the definition of the intermediate variable cc
661 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
662 improvements.
663
664 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
665 unnecessary spaces.
666
667 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
668 and later.
669
670 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
671
672 9. Optimizing fast_forward_start_bits in JIT.
673
674 10. Adding support for callouts in JIT, and fixing some issues revealed
675 during this work. Namely:
676
677 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
678
679 (b) Minimum length was not checked before the matching is started.
680
681 11. The value of capture_last that is passed to callouts was incorrect in some
682 cases when there was a capture on one path that was subsequently abandoned
683 after a backtrack. Also, the capture_last value is now reset after a
684 recursion, since all captures are also reset in this case.
685
686 12. The interpreter no longer returns the "too many substrings" error in the
687 case when an overflowing capture is in a branch that is subsequently
688 abandoned after a backtrack.
689
690 13. In the pathological case when an offset vector of size 2 is used, pcretest
691 now prints out the matched string after a yield of 0 or 1.
692
693 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
694 Only supported by the JIT compiler at the moment.
695
696 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
697
698 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
699 offset of the starting point of the matching process, provided the offsets
700 vector is large enough.
701
702 17. The \A escape now records a lookbehind value of 1, though its execution
703 does not actually inspect the previous character. This is to ensure that,
704 in partial multi-segment matching, at least one character from the old
705 segment is retained when a new segment is processed. Otherwise, if there
706 are no lookbehinds in the pattern, \A might match incorrectly at the start
707 of a new segment.
708
709 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
710
711 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
712 gives some modest performance improvement in 8-bit mode.
713
714 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
715 be expressed in certain programming languages using Universal Character
716 Names.
717
718 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
719 which points out that "non characters" are not "characters that may not
720 appear in Unicode strings" but rather "characters that are reserved for
721 internal use and have only local meaning".
722
723 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
724 there was a conditional group that depended on an assertion, if the
725 assertion was false, the callout that immediately followed the alternation
726 in the condition was skipped when pcre_exec() was used for matching.
727
728 23. Allow an explicit callout to be inserted before an assertion that is the
729 condition for a conditional group, for compatibility with automatic
730 callouts, which always insert a callout at this point.
731
732 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
733 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
734
735 25. (*PRUNE) is now supported by the JIT compiler.
736
737 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
738
739 27. Fix the case where there are two or more SKIPs with arguments that may be
740 ignored.
741
742 28. (*SKIP) is now supported by the JIT compiler.
743
744 29. (*THEN) is now supported by the JIT compiler.
745
746 30. Update RunTest with additional test selector options.
747
748 31. The way PCRE handles backtracking verbs has been changed in two ways.
749
750 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
751 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
752 backtracking. In some cases this makes it more Perl-compatible, but Perl's
753 rather obscure rules do not always do the same thing.
754
755 (2) Previously, backtracking verbs were confined within assertions. This is
756 no longer the case for positive assertions, except for (*ACCEPT). Again,
757 this sometimes improves Perl compatibility, and sometimes does not.
758
759 32. A number of tests that were in test 2 because Perl did things differently
760 have been moved to test 1, because either Perl or PCRE has changed, and
761 these tests are now compatible.
762
763 32. Backtracking control verbs are now handled in the same way in JIT and
764 interpreter.
765
766 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
767 contained a forward subroutine reference caused a compile error.
768
769 34. Auto-detect and optimize limited repetitions in JIT.
770
771 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
772 blocking (*UTF) etc.
773
774 36. In the interpreter, maximizing pattern repetitions for characters and
775 character types now use tail recursion, which reduces stack usage.
776
777 37. The value of the max lookbehind was not correctly preserved if a compiled
778 and saved regex was reloaded on a host of different endianness.
779
780 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
781 of the compiled pattern block, expand the flags field from 16 to 32 bits
782 because it was almost full.
783
784 39. Try madvise first before posix_madvise.
785
786 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
787 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
788 original change locked out all matches of empty strings. This has been
789 changed so that one match of an empty string per line is recognized.
790 Subsequent searches on the same line (for colouring or for --only-matching,
791 for example) do not recognize empty strings.
792
793 41. Applied a user patch to fix a number of spelling mistakes in comments.
794
795 42. Data lines longer than 65536 caused pcretest to crash.
796
797 43. Clarified the data type for length and startoffset arguments for pcre_exec
798 and pcre_dfa_exec in the function-specific man pages, where they were
799 explicitly stated to be in bytes, never having been updated. I also added
800 some clarification to the pcreapi man page.
801
802 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
803 a segmentation fault.
804
805
806 Version 8.32 30-November-2012
807 -----------------------------
808
809 1. Improved JIT compiler optimizations for first character search and single
810 character iterators.
811
812 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
813 Patch by Daniel Richard G.
814
815 3. Single character iterator optimizations in the JIT compiler.
816
817 4. Improved JIT compiler optimizations for character ranges.
818
819 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
820 Reported by Giuseppe D'Angelo.
821
822 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
823 of a line, was being set incorrectly in cases where .* appeared inside
824 atomic brackets at the start of a pattern, or where there was a subsequent
825 *PRUNE or *SKIP.
826
827 7. Improved instruction cache flush for POWER/PowerPC.
828 Patch by Daniel Richard G.
829
830 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
831 grep:
832
833 (a) There is now no limit to the number of patterns to be matched.
834
835 (b) An error is given if a pattern is too long.
836
837 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
838 are now supported.
839
840 (d) --exclude-from and --include-from (multiple use) have been added.
841
842 (e) Exclusions and inclusions now apply to all files and directories, not
843 just to those obtained from scanning a directory recursively.
844
845 (f) Multiple uses of -f and --file-list are now supported.
846
847 (g) In a Windows environment, the default for -d has been changed from
848 "read" (the GNU grep default) to "skip", because otherwise the presence
849 of a directory in the file list provokes an error.
850
851 (h) The documentation has been revised and clarified in places.
852
853 9. Improve the matching speed of capturing brackets.
854
855 10. Changed the meaning of \X so that it now matches a Unicode extended
856 grapheme cluster.
857
858 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
859 out POSIX threads when JIT support is configured.
860
861 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
862
863 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
864 structure could go wrong in environments where size_t is not the same size
865 as int.
866
867 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
868 set.
869
870 15. The EBCDIC support had decayed; later updates to the code had included
871 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
872 general tidy up of EBCDIC-related issues, and the documentation was also
873 not quite right. There is now a test that can be run on ASCII systems to
874 check some of the EBCDIC-related things (but is it not a full test).
875
876 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
877 in a small tidy to the code.
878
879 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
880
881 18. If the --only-matching (-o) option in pcregrep is specified multiple
882 times, each one causes appropriate output. For example, -o1 -o2 outputs the
883 substrings matched by the 1st and 2nd capturing parentheses. A separating
884 string can be specified by --om-separator (default empty).
885
886 19. Improving the first n character searches.
887
888 20. Turn case lists for horizontal and vertical white space into macros so that
889 they are defined only once.
890
891 21. This set of changes together give more compatible Unicode case-folding
892 behaviour for characters that have more than one other case when UCP
893 support is available.
894
895 (a) The Unicode property table now has offsets into a new table of sets of
896 three or more characters that are case-equivalent. The MultiStage2.py
897 script that generates these tables (the pcre_ucd.c file) now scans
898 CaseFolding.txt instead of UnicodeData.txt for character case
899 information.
900
901 (b) The code for adding characters or ranges of characters to a character
902 class has been abstracted into a generalized function that also handles
903 case-independence. In UTF-mode with UCP support, this uses the new data
904 to handle characters with more than one other case.
905
906 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
907 whose other case is greater than 256 are now correctly matched
908 caselessly. Previously, the high codepoint matched the low one, but not
909 vice versa.
910
911 (d) The processing of \h, \H, \v, and \ in character classes now makes use
912 of the new class addition function, using character lists defined as
913 macros alongside the case definitions of 20 above.
914
915 (e) Caseless back references now work with characters that have more than
916 one other case.
917
918 (f) General caseless matching of characters with more than one other case
919 is supported.
920
921 22. Unicode character properties were updated from Unicode 6.2.0
922
923 23. Improved CMake support under Windows. Patch by Daniel Richard G.
924
925 24. Add support for 32-bit character strings, and UTF-32
926
927 25. Major JIT compiler update (code refactoring and bugfixing).
928 Experimental Sparc 32 support is added.
929
930 26. Applied a modified version of Daniel Richard G's patch to create
931 pcre.h.generic and config.h.generic by "make" instead of in the
932 PrepareRelease script.
933
934 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
935 pcre_compile.c when checking for a zero character.
936
937 28. Introducing a native interface for JIT. Through this interface, the compiled
938 machine code can be directly executed. The purpose of this interface is to
939 provide fast pattern matching, so several sanity checks are not performed.
940 However, feature tests are still performed. The new interface provides
941 1.4x speedup compared to the old one.
942
943 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
944 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
945 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
946
947 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
948 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
949 the "old" RFC 2279). Instead, it ended up passing a negative length to
950 pcre_exec().
951
952 31. Add support for GCC's visibility feature to hide internal functions.
953
954 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
955 "unknown -C option" after outputting 0 or 1.
956
957 33. There is now support for generating a code coverage report for the test
958 suite in environments where gcc is the compiler and lcov is installed. This
959 is mainly for the benefit of the developers.
960
961 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
962 unaddressable using valgrind annotations, allowing valgrind to detect
963 invalid memory accesses. This is mainly for the benefit of the developers.
964
965 25. (*UTF) can now be used to start a pattern in any of the three libraries.
966
967 26. Give configure error if --enable-cpp but no C++ compiler found.
968
969
970 Version 8.31 06-July-2012
971 -------------------------
972
973 1. Fixing a wrong JIT test case and some compiler warnings.
974
975 2. Removed a bashism from the RunTest script.
976
977 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
978 to unsigned type, result still unsigned" that was given by an MS compiler
979 on encountering the code "-sizeof(xxx)".
980
981 4. Partial matching support is added to the JIT compiler.
982
983 5. Fixed several bugs concerned with partial matching of items that consist
984 of more than one character:
985
986 (a) /^(..)\1/ did not partially match "aba" because checking references was
987 done on an "all or nothing" basis. This also applied to repeated
988 references.
989
990 (b) \R did not give a hard partial match if \r was found at the end of the
991 subject.
992
993 (c) \X did not give a hard partial match after matching one or more
994 characters at the end of the subject.
995
996 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
997 a partial match for the string "\r".
998
999 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1000 a partial match for a CR character at the end of the subject string.
1001
1002 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1003 running pcretest, the text "(JIT)" added to the output whenever JIT is
1004 actually used to run the match.
1005
1006 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1007 or /S+[+] with a digit between 1 and 7.
1008
1009 8. OP_NOT now supports any UTF character not just single-byte ones.
1010
1011 9. (*MARK) control verb is now supported by the JIT compiler.
1012
1013 10. The command "./RunTest list" lists the available tests without actually
1014 running any of them. (Because I keep forgetting what they all are.)
1015
1016 11. Add PCRE_INFO_MAXLOOKBEHIND.
1017
1018 12. Applied a (slightly modified) user-supplied patch that improves performance
1019 when the heap is used for recursion (compiled with --disable-stack-for-
1020 recursion). Instead of malloc and free for each heap frame each time a
1021 logical recursion happens, frames are retained on a chain and re-used where
1022 possible. This sometimes gives as much as 30% improvement.
1023
1024 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1025 call.
1026
1027 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1028
1029 15. It is now possible to link pcretest with libedit as an alternative to
1030 libreadline.
1031
1032 16. (*COMMIT) control verb is now supported by the JIT compiler.
1033
1034 17. The Unicode data tables have been updated to Unicode 6.1.0.
1035
1036 18. Added --file-list option to pcregrep.
1037
1038 19. Added binary file support to pcregrep, including the -a, --binary-files,
1039 -I, and --text options.
1040
1041 20. The madvise function is renamed for posix_madvise for QNX compatibility
1042 reasons. Fixed by Giuseppe D'Angelo.
1043
1044 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1045
1046 22. Changed the option for creating symbolic links for 16-bit man pages from
1047 -s to -sf so that re-installing does not cause issues.
1048
1049 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1050
1051 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1052 matches in certain environments (the workspace was not being correctly
1053 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1054 some of the workspace data at the beginning of a restart.
1055
1056 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1057 was not doing so when it should - probably a typo introduced by SVN 528
1058 (change 8.10/14).
1059
1060 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1061 \w+ when the character tables indicated that \x{c4} was a word character.
1062 There were several related cases, all because the tests for doing a table
1063 lookup were testing for characters less than 127 instead of 255.
1064
1065 27. If a pattern contains capturing parentheses that are not used in a match,
1066 their slots in the ovector are set to -1. For those that are higher than
1067 any matched groups, this happens at the end of processing. In the case when
1068 there were back references that the ovector was too small to contain
1069 (causing temporary malloc'd memory to be used during matching), and the
1070 highest capturing number was not used, memory off the end of the ovector
1071 was incorrectly being set to -1. (It was using the size of the temporary
1072 memory instead of the true size.)
1073
1074 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1075 ovector size, it uses memory at the end of the block that it has got.
1076
1077 29. Check for an overlong MARK name and give an error at compile time. The
1078 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1079
1080 30. JIT compiler update.
1081
1082 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1083 Rill for the patch.
1084
1085 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1086
1087 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1088
1089 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1090 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1091 was enabled.
1092
1093 35. Improve JIT code generation for greedy plus quantifier.
1094
1095 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1096 1 to "aa" instead of to an empty string. The bug affected repeated groups
1097 that could potentially match an empty string.
1098
1099 37. Optimizing single character iterators in JIT.
1100
1101 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1102 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1103 codepoints that are too big for the mode are faulted, and in a UTF mode,
1104 disallowed codepoints are also faulted.
1105
1106 39. If PCRE was compiled with UTF support, in three places in the DFA
1107 matcher there was code that should only have been obeyed in UTF mode, but
1108 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1109 processing when bytes with values greater than 127 were present. In 16-bit
1110 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1111 both cases the values are those that cannot be the first data item in a UTF
1112 character. The three items that might have provoked this were recursions,
1113 possessively repeated groups, and atomic groups.
1114
1115 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1116 and pcregrep, because some OS require shared objects to be explicitly
1117 passed to ld, causing the link step to fail if they are not.
1118
1119 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1120 mode, patterns that started with \h* or \R* might be incorrectly matched.
1121
1122
1123 Version 8.30 04-February-2012
1124 -----------------------------
1125
1126 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1127 name is defined in ctype.h.
1128
1129 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1130 only in quite long subpatterns.
1131
1132 3. Removed the function pcre_info(), which has been obsolete and deprecated
1133 since it was replaced by pcre_fullinfo() in February 2000.
1134
1135 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1136 match a (*MARK), and the match failed at the start of the subject, a
1137 reference to memory before the start of the subject could occur. This bug
1138 was introduced by fix 17 of release 8.21.
1139
1140 5. A reference to an unset group with zero minimum repetition was giving
1141 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1142 /(another)?(\1?)test/ matched against "hello world test". This bug was
1143 introduced in release 8.13.
1144
1145 6. Add support for 16-bit character strings (a large amount of work involving
1146 many changes and refactorings).
1147
1148 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1149 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1150 from a file.
1151
1152 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1153 rounding is not applied in this particular case).
1154
1155 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1156 if they appear, or are escaped, in patterns.
1157
1158 10. Get rid of a number of -Wunused-but-set-variable warnings.
1159
1160 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1161 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1162 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1163 also returns the mark "x". This bug applied to capturing parentheses,
1164 non-capturing parentheses, and atomic parentheses. It also applied to some
1165 assertions.
1166
1167 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1168 information out of configure.ac instead of relying on pcre.h.generic, which
1169 is not stored in the repository.
1170
1171 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1172 -lreadline.
1173
1174 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1175
1176 15. Applied Graycode's patch to put the top-level frame on the stack rather
1177 than the heap when not using the stack for recursion. This gives a
1178 performance improvement in many cases when recursion is not deep.
1179
1180 16. Experimental code added to "pcretest -C" to output the stack frame size.
1181
1182
1183 Version 8.21 12-Dec-2011
1184 ------------------------
1185
1186 1. Updating the JIT compiler.
1187
1188 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1189 are added as well.
1190
1191 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1192 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1193 calling _pcre_jit_exec. Some extra comments are added.
1194
1195 4. (*MARK) settings inside atomic groups that do not contain any capturing
1196 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1197 was introduced by change 18 for 8.20.
1198
1199 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1200 ECMA-262 standard.
1201
1202 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1203 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1204 This bug was probably introduced by change 9 of 8.13.
1205
1206 7. While fixing 6 above, I noticed that a number of other items were being
1207 incorrectly rejected as "not fixed length". This arose partly because newer
1208 opcodes had not been added to the fixed-length checking code. I have (a)
1209 corrected the bug and added tests for these items, and (b) arranged for an
1210 error to occur if an unknown opcode is encountered while checking for fixed
1211 length instead of just assuming "not fixed length". The items that were
1212 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1213 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1214 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1215
1216 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1217 being incorrectly compiled and would have given unpredicatble results.
1218
1219 9. A possessively repeated subpattern with minimum repeat count greater than
1220 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1221 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1222 the first (A) could occur when it should not.
1223
1224 10. Add a cast and remove a redundant test from the code.
1225
1226 11. JIT should use pcre_malloc/pcre_free for allocation.
1227
1228 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1229 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1230 is anything other than /usr, -L is still shown).
1231
1232 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1233
1234 14. Perl does not support \N without a following name in a [] class; PCRE now
1235 also gives an error.
1236
1237 15. If a forward reference was repeated with an upper limit of around 2000,
1238 it caused the error "internal error: overran compiling workspace". The
1239 maximum number of forward references (including repeats) was limited by the
1240 internal workspace, and dependent on the LINK_SIZE. The code has been
1241 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1242 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1243 of around 200,000 forward references. While doing this, I also speeded up
1244 the filling in of repeated forward references.
1245
1246 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1247 incorrectly expecting the subject to contain another "a" after the start.
1248
1249 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1250 in the match, the SKIP should be ignored. This was not happening; instead
1251 the SKIP was being treated as NOMATCH. For patterns such as
1252 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1253 tested.
1254
1255 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1256 now much more compatible with Perl, in particular in cases where the result
1257 is a non-match for a non-anchored pattern. For example, if
1258 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1259 "m", where previously it did not return a name. A side effect of this
1260 change is that for partial matches, the last encountered mark name is
1261 returned, as for non matches. A number of tests that were previously not
1262 Perl-compatible have been moved into the Perl-compatible test files. The
1263 refactoring has had the pleasing side effect of removing one argument from
1264 the match() function, thus reducing its stack requirements.
1265
1266 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1267 subsequent uses of /S (without +) incorrectly behaved like /S+.
1268
1269 21. Retrieve executable code size support for the JIT compiler and fixing
1270 some warnings.
1271
1272 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1273 not work when the shorter character appeared right at the end of the
1274 subject string.
1275
1276 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1277 systems.
1278
1279 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1280 output it when the /M option is used in pcretest.
1281
1282 25. The CheckMan script was not being included in the distribution. Also, added
1283 an explicit "perl" to run Perl scripts from the PrepareRelease script
1284 because this is reportedly needed in Windows.
1285
1286 26. If study data was being save in a file and studying had not found a set of
1287 "starts with" bytes for the pattern, the data written to the file (though
1288 never used) was taken from uninitialized memory and so caused valgrind to
1289 complain.
1290
1291 27. Updated RunTest.bat as provided by Sheri Pierce.
1292
1293 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1294
1295 29. Computation of memory usage for the table of capturing group names was
1296 giving an unnecessarily large value.
1297
1298
1299 Version 8.20 21-Oct-2011
1300 ------------------------
1301
1302 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1303 a POSIX class. After further experiments with Perl, which convinced me that
1304 Perl has bugs and confusions, a closing square bracket is no longer allowed
1305 in a POSIX name. This bug also affected patterns with classes that started
1306 with full stops.
1307
1308 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1309 captured substring, but while checking the failing first alternative,
1310 substring 1 is temporarily captured. If the output vector supplied to
1311 pcre_exec() was not big enough for this capture, the yield of the function
1312 was still zero ("insufficient space for captured substrings"). This cannot
1313 be totally fixed without adding another stack variable, which seems a lot
1314 of expense for a edge case. However, I have improved the situation in cases
1315 such as /(a)(b)x|abc/ matched against "abc", where the return code
1316 indicates that fewer than the maximum number of slots in the ovector have
1317 been set.
1318
1319 3. Related to (2) above: when there are more back references in a pattern than
1320 slots in the output vector, pcre_exec() uses temporary memory during
1321 matching, and copies in the captures as far as possible afterwards. It was
1322 using the entire output vector, but this conflicts with the specification
1323 that only 2/3 is used for passing back captured substrings. Now it uses
1324 only the first 2/3, for compatibility. This is, of course, another edge
1325 case.
1326
1327 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1328 main code base, and can be used by building with --enable-jit. When this is
1329 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1330 runtime --no-jit option is given.
1331
1332 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1333 ovector, the return from the function was zero, implying that there were
1334 other matches that did not fit. The correct "exactly full" value is now
1335 returned.
1336
1337 6. If a subpattern that was called recursively or as a subroutine contained
1338 (*PRUNE) or any other control that caused it to give a non-standard return,
1339 invalid errors such as "Error -26 (nested recursion at the same subject
1340 position)" or even infinite loops could occur.
1341
1342 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1343 computing the minimum length on reaching *ACCEPT, and so ended up with the
1344 wrong value of 1 rather than 0. Further investigation indicates that
1345 computing a minimum subject length in the presence of *ACCEPT is difficult
1346 (think back references, subroutine calls), and so I have changed the code
1347 so that no minimum is registered for a pattern that contains *ACCEPT.
1348
1349 8. If (*THEN) was present in the first (true) branch of a conditional group,
1350 it was not handled as intended. [But see 16 below.]
1351
1352 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1353 Sheri Pierce.
1354
1355 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1356 the first byte in a match must be "a".
1357
1358 11. Change 17 for 8.13 increased the recursion depth for patterns like
1359 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1360 pattern contains any instances of (*THEN). If it does not, the old
1361 optimizations are restored. It would be nice to do this on a per-group
1362 basis, but at the moment that is not feasible.
1363
1364 12. In some environments, the output of pcretest -C is CRLF terminated. This
1365 broke RunTest's code that checks for the link size. A single white space
1366 character after the value is now allowed for.
1367
1368 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1369 For "fr", it uses the Windows-specific input and output files.
1370
1371 14. If (*THEN) appeared in a group that was called recursively or as a
1372 subroutine, it did not work as intended. [But see next item.]
1373
1374 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1375 pattern fragments (but not containing any | characters). If A and B are
1376 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1377 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1378 D. In other words, Perl considers parentheses that do not contain any |
1379 characters to be part of a surrounding alternative, whereas PCRE was
1380 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1381 differently. PCRE now behaves in the same way as Perl, except in the case
1382 of subroutine/recursion calls such as (?1) which have in any case always
1383 been different (but PCRE had them first :-).
1384
1385 16. Related to 15 above: Perl does not treat the | in a conditional group as
1386 creating alternatives. Such a group is treated in the same way as an
1387 ordinary group without any | characters when processing (*THEN). PCRE has
1388 been changed to match Perl's behaviour.
1389
1390 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1391 RunGrepTest script failed.
1392
1393 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1394 inevitable for groups that contain captures, but it can lead to a lot of
1395 stack use in large patterns. The old behaviour has been restored for atomic
1396 groups that do not contain any capturing parentheses.
1397
1398 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1399 suppress the check for a minimum subject length at run time. (If it was
1400 given to pcre_exec() or pcre_dfa_exec() it did work.)
1401
1402 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1403 fail to work when decoding hex characters in data strings in EBCDIC
1404 environments.
1405
1406 21. It appears that in at least one Mac OS environment, the isxdigit() function
1407 is implemented as a macro that evaluates to its argument more than once,
1408 contravening the C 90 Standard (I haven't checked a later standard). There
1409 was an instance in pcretest which caused it to go wrong when processing
1410 \x{...} escapes in subject strings. The has been rewritten to avoid using
1411 things like p++ in the argument of isxdigit().
1412
1413
1414 Version 8.13 16-Aug-2011
1415 ------------------------
1416
1417 1. The Unicode data tables have been updated to Unicode 6.0.0.
1418
1419 2. Two minor typos in pcre_internal.h have been fixed.
1420
1421 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1422 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1423 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1424
1425 4. There were a number of related bugs in the code for matching backrefences
1426 caselessly in UTF-8 mode when codes for the characters concerned were
1427 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1428 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1429 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1430 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1431 2-byte code at the end of the subject (it thought there wasn't enough data
1432 left).
1433
1434 5. Comprehensive information about what went wrong is now returned by
1435 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1436 as the output vector has at least 2 elements. The offset of the start of
1437 the failing character and a reason code are placed in the vector.
1438
1439 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1440 now returned is for the first byte of the failing character, instead of the
1441 last byte inspected. This is an incompatible change, but I hope it is small
1442 enough not to be a problem. It makes the returned offset consistent with
1443 pcre_exec() and pcre_dfa_exec().
1444
1445 7. pcretest now gives a text phrase as well as the error number when
1446 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1447 failure, the offset and reason code are output.
1448
1449 8. When \R was used with a maximizing quantifier it failed to skip backwards
1450 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1451 back over a single character (\n). This seems wrong (because it treated the
1452 two characters as a single entity when going forwards), conflicts with the
1453 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1454 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1455 has been changed.
1456
1457 9. Some internal refactoring has changed the processing so that the handling
1458 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1459 time (the PCRE_DOTALL option was changed this way some time ago: version
1460 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1461 which was always a bit of a fudge. It also means that there is one less
1462 argument for the match() function, which reduces its stack requirements
1463 slightly. This change also fixes an incompatibility with Perl: the pattern
1464 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1465
1466 10. More internal refactoring has drastically reduced the number of recursive
1467 calls to match() for possessively repeated groups such as (abc)++ when
1468 using pcre_exec().
1469
1470 11. While implementing 10, a number of bugs in the handling of groups were
1471 discovered and fixed:
1472
1473 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1474 (a|)*(?1) gave a compile-time internal error.
1475 ((a|)+)+ did not notice that the outer group could match an empty string.
1476 (^a|^)+ was not marked as anchored.
1477 (.*a|.*)+ was not marked as matching at start or after a newline.
1478
1479 12. Yet more internal refactoring has removed another argument from the match()
1480 function. Special calls to this function are now indicated by setting a
1481 value in a variable in the "match data" data block.
1482
1483 13. Be more explicit in pcre_study() instead of relying on "default" for
1484 opcodes that mean there is no starting character; this means that when new
1485 ones are added and accidentally left out of pcre_study(), testing should
1486 pick them up.
1487
1488 14. The -s option of pcretest has been documented for ages as being an old
1489 synonym of -m (show memory usage). I have changed it to mean "force study
1490 for every regex", that is, assume /S for every regex. This is similar to -i
1491 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1492 using it. It makes it easier to run collections of tests with and without
1493 study enabled, and thereby test pcre_study() more easily. All the standard
1494 tests are now run with and without -s (but some patterns can be marked as
1495 "never study" - see 20 below).
1496
1497 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1498 restoration of the capturing data to the outer values was not happening
1499 correctly.
1500
1501 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1502 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1503 pattern had matched an empty string, and so incorrectly returned a no
1504 match.
1505
1506 17. There was optimizing code for the last branch of non-capturing parentheses,
1507 and also for the obeyed branch of a conditional subexpression, which used
1508 tail recursion to cut down on stack usage. Unfortunately, now that there is
1509 the possibility of (*THEN) occurring in these branches, tail recursion is
1510 no longer possible because the return has to be checked for (*THEN). These
1511 two optimizations have therefore been removed. [But see 8.20/11 above.]
1512
1513 18. If a pattern containing \R was studied, it was assumed that \R always
1514 matched two bytes, thus causing the minimum subject length to be
1515 incorrectly computed because \R can also match just one byte.
1516
1517 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1518 was incorrectly computed.
1519
1520 20. If /S is present twice on a test pattern in pcretest input, it now
1521 *disables* studying, thereby overriding the use of -s on the command line
1522 (see 14 above). This is necessary for one or two tests to keep the output
1523 identical in both cases.
1524
1525 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1526 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1527
1528 22. When an atomic group that contained a capturing parenthesis was
1529 successfully matched, but the branch in which it appeared failed, the
1530 capturing was not being forgotten if a higher numbered group was later
1531 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1532 group 1 to "a", when in fact it should be unset. This applied to multi-
1533 branched capturing and non-capturing groups, repeated or not, and also to
1534 positive assertions (capturing in negative assertions does not happen
1535 in PCRE) and also to nested atomic groups.
1536
1537 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1538 subject after a captured substring, to make it easier to tell which of a
1539 number of identical substrings has been captured.
1540
1541 24. The way atomic groups are processed by pcre_exec() has been changed so that
1542 if they are repeated, backtracking one repetition now resets captured
1543 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1544 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1545 "aaa". Previously, it would have been "a". As part of this code
1546 refactoring, the way recursive calls are handled has also been changed.
1547
1548 25. If an assertion condition captured any substrings, they were not passed
1549 back unless some other capturing happened later. For example, if
1550 (?(?=(a))a) was matched against "a", no capturing was returned.
1551
1552 26. When studying a pattern that contained subroutine calls or assertions,
1553 the code for finding the minimum length of a possible match was handling
1554 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1555 group 1 called group 2 while simultaneously a separate group 2 called group
1556 1). A stack overflow occurred in this case. I have fixed this by limiting
1557 the recursion depth to 10.
1558
1559 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1560 Fortmann. This supports explicit test numbers on the command line, and has
1561 argument validation and error reporting.
1562
1563 28. An instance of \X with an unlimited repeat could fail if at any point the
1564 first character it looked at was a mark character.
1565
1566 29. Some minor code refactoring concerning Unicode properties and scripts
1567 should reduce the stack requirement of match() slightly.
1568
1569 30. Added the '=' option to pcretest to check the setting of unused capturing
1570 slots at the end of the pattern, which are documented as being -1, but are
1571 not included in the return count.
1572
1573 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1574 compiled something random. Now it gives a compile-time error (as does
1575 Perl).
1576
1577 32. A *MARK encountered during the processing of a positive assertion is now
1578 recorded and passed back (compatible with Perl).
1579
1580 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1581 had alternative anchored branches, the search for a second match in a line
1582 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1583 matched the line "0102" twice. The same bug affected patterns that started
1584 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1585 twice.
1586
1587 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1588 does, and because of capturing effects, quantifying parenthesized
1589 assertions may at times be useful. Quantifiers are now allowed for
1590 parenthesized assertions.
1591
1592 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1593
1594 36. \g was being checked for fancy things in a character class, when it should
1595 just be a literal "g".
1596
1597 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1598 appearance of a nested POSIX class supersedes an apparent external class.
1599 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1600 unescaped square brackets may also appear as part of class names. For
1601 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1602 more like Perl. (But see 8.20/1 above.)
1603
1604 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1605 was because it thought it was \N{name}, which is not supported).
1606
1607 39. Add minix to OS list not supporting the -S option in pcretest.
1608
1609 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1610 cannot analyze patterns in sufficient detail to catch mutual recursions
1611 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1612 subgroup is called recursively as a subpattern for a second time at the
1613 same position in the subject string. In previous releases this might have
1614 been caught by the recursion limit, or it might have run out of stack.
1615
1616 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1617 happen only once. PCRE was, however incorrectly giving a compile time error
1618 "recursive call could loop indefinitely" because it cannot analyze the
1619 pattern in sufficient detail. The compile time test no longer happens when
1620 PCRE is compiling a conditional subpattern, but actual runaway loops are
1621 now caught at runtime (see 40 above).
1622
1623 42. It seems that Perl allows any characters other than a closing parenthesis
1624 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1625 has been changed to be the same.
1626
1627 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1628 as not to get warnings when autogen.sh is called. Also changed
1629 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1630
1631 44. To help people who use pcregrep to scan files containing exceedingly long
1632 lines, the following changes have been made:
1633
1634 (a) The default value of the buffer size parameter has been increased from
1635 8K to 20K. (The actual buffer used is three times this size.)
1636
1637 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1638 PCRE is built.
1639
1640 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1641 to be set at run time.
1642
1643 (d) Numerical values in pcregrep options can be followed by K or M, for
1644 example --buffer-size=50K.
1645
1646 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1647 given and the return code is set to 2.
1648
1649 45. Add a pointer to the latest mark to the callout data block.
1650
1651 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1652 partial match of an empty string instead of no match. This was specific to
1653 the use of ".".
1654
1655 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1656 complete match instead of a partial match. This bug was dependent on both
1657 the PCRE_UTF8 and PCRE_DOTALL options being set.
1658
1659 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1660 starting byte set, because \b was not being ignored.
1661
1662
1663 Version 8.12 15-Jan-2011
1664 ------------------------
1665
1666 1. Fixed some typos in the markup of the man pages, and wrote a script that
1667 checks for such things as part of the documentation building process.
1668
1669 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1670 --match-limit and --recursion-limit options (added for 8.11). In
1671 particular, this made one of the standard tests fail. (The integer value
1672 went into the wrong half of a long int.)
1673
1674 3. If the --colour option was given to pcregrep with -v (invert match), it
1675 did strange things, either producing crazy output, or crashing. It should,
1676 of course, ignore a request for colour when reporting lines that do not
1677 match.
1678
1679 4. Another pcregrep bug caused similar problems if --colour was specified with
1680 -M (multiline) and the pattern match finished with a line ending.
1681
1682 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1683 matched in multiline mode, the following line was shown as part of the
1684 match. This seems wrong, so I have changed it.
1685
1686 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1687 the check for further matches in the same line (so they could be coloured)
1688 to overrun the end of the current line. If another match was found, it was
1689 incorrectly shown (and then shown again when found in the next line).
1690
1691 7. If pcregrep was compiled under Windows, there was a reference to the
1692 function pcregrep_exit() before it was defined. I am assuming this was
1693 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1694 reported by a user. I've moved the definition above the reference.
1695
1696
1697 Version 8.11 10-Dec-2010
1698 ------------------------
1699
1700 1. (*THEN) was not working properly if there were untried alternatives prior
1701 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1702 backtracked to try for "b" instead of moving to the next alternative branch
1703 at the same level (in this case, to look for "c"). The Perl documentation
1704 is clear that when (*THEN) is backtracked onto, it goes to the "next
1705 alternative in the innermost enclosing group".
1706
1707 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1708 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1709 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1710 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1711 (*THEN).
1712
1713 3. If \s appeared in a character class, it removed the VT character from
1714 the class, even if it had been included by some previous item, for example
1715 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1716 of \s, but is part of the POSIX "space" class.)
1717
1718 4. A partial match never returns an empty string (because you can always
1719 match an empty string at the end of the subject); however the checking for
1720 an empty string was starting at the "start of match" point. This has been
1721 changed to the "earliest inspected character" point, because the returned
1722 data for a partial match starts at this character. This means that, for
1723 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1724 (previously it gave "no match").
1725
1726 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1727 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1728 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1729 has an implication that the given string is incomplete (because a partial
1730 match is preferred over a full match). For this reason, these items now
1731 give a partial match in this situation. [Aside: previously, the one case
1732 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1733 match rather than a full match, which was wrong by the old rules, but is
1734 now correct.]
1735
1736 6. There was a bug in the handling of #-introduced comments, recognized when
1737 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1738 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1739 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1740 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1741 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1742 places in pcre_compile().
1743
1744 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1745 comments when looking ahead for named forward references to subpatterns,
1746 the only newline sequence it recognized was NL. It now handles newlines
1747 according to the set newline convention.
1748
1749 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1750 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1751 cater for a lack of strerror(). These oversights have been fixed.
1752
1753 9. Added --match-limit and --recursion-limit to pcregrep.
1754
1755 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1756
1757 11. When the -o option was used, pcregrep was setting a return code of 1, even
1758 when matches were found, and --line-buffered was not being honoured.
1759
1760 12. Added an optional parentheses number to the -o and --only-matching options
1761 of pcregrep.
1762
1763 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1764 can match an empty string. The code to do it in pcretest and pcredemo
1765 needed fixing:
1766
1767 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1768 only one byte after an empty string match just before CRLF (this case
1769 just got forgotten; "any" and "anycrlf" were OK).
1770
1771 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1772 mode when an empty string match preceded an ASCII character followed by
1773 a non-ASCII character. (The code for advancing by one character rather
1774 than one byte was nonsense.)
1775
1776 (c) The pcredemo.c sample program did not have any code at all to handle
1777 the cases when CRLF is a valid newline sequence.
1778
1779 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1780 as a starting offset was within the subject string. There is now a new
1781 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1782 negative or greater than the length of the string. In order to test this,
1783 pcretest is extended to allow the setting of negative starting offsets.
1784
1785 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1786 starting offset points to the beginning of a UTF-8 character was
1787 unnecessarily clumsy. I tidied it up.
1788
1789 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1790 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1791
1792 17. Nobody had reported that the --include_dir option, which was added in
1793 release 7.7 should have been called --include-dir (hyphen, not underscore)
1794 for compatibility with GNU grep. I have changed it to --include-dir, but
1795 left --include_dir as an undocumented synonym, and the same for
1796 --exclude-dir, though that is not available in GNU grep, at least as of
1797 release 2.5.4.
1798
1799 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1800 characters from a string of bytes) have been redefined so as not to use
1801 loops, in order to improve performance in some environments. At the same
1802 time, I abstracted some of the common code into auxiliary macros to save
1803 repetition (this should not affect the compiled code).
1804
1805 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1806 compile-time error is now given if \c is not followed by an ASCII
1807 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1808 different, and any byte value is allowed.)
1809
1810 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1811 START_OPTIMIZE option, which is now allowed at compile time - but just
1812 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1813 to pcregrep and other applications that have no direct access to PCRE
1814 options. The new /Y option in pcretest sets this option when calling
1815 pcre_compile().
1816
1817 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1818 back references. Groups containing recursive back references were forced to
1819 be atomic by that change, but in the case of named groups, the amount of
1820 memory required was incorrectly computed, leading to "Failed: internal
1821 error: code overflow". This has been fixed.
1822
1823 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1824 pcretest.c, to avoid build problems in some Borland environments.
1825
1826
1827 Version 8.10 25-Jun-2010
1828 ------------------------
1829
1830 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1831 THEN.
1832
1833 2. (*ACCEPT) was not working when inside an atomic group.
1834
1835 3. Inside a character class, \B is treated as a literal by default, but
1836 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1837 causes the error). The code is unchanged, but I tidied the documentation.
1838
1839 4. Inside a character class, PCRE always treated \R and \X as literals,
1840 whereas Perl faults them if its -w option is set. I have changed PCRE so
1841 that it faults them when PCRE_EXTRA is set.
1842
1843 5. Added support for \N, which always matches any character other than
1844 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1845
1846 6. When compiling pcregrep with newer versions of gcc which may have
1847 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1848 declared with attribute warn_unused_result" were given. Just casting the
1849 result to (void) does not stop the warnings; a more elaborate fudge is
1850 needed. I've used a macro to implement this.
1851
1852 7. Minor change to pcretest.c to avoid a compiler warning.
1853
1854 8. Added four artifical Unicode properties to help with an option to make
1855 \s etc use properties (see next item). The new properties are: Xan
1856 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1857
1858 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1859 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1860 this option. Modified pcretest to add /W to test this facility. Added
1861 REG_UCP to make it available via the POSIX interface.
1862
1863 10. Added --line-buffered to pcregrep.
1864
1865 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1866 studied, and the match started with a letter with a code point greater than
1867 127 whose first byte was different to the first byte of the other case of
1868 the letter, the other case of this starting letter was not recognized
1869 (#976).
1870
1871 12. If a pattern that was studied started with a repeated Unicode property
1872 test, for example, \p{Nd}+, there was the theoretical possibility of
1873 setting up an incorrect bitmap of starting bytes, but fortunately it could
1874 not have actually happened in practice until change 8 above was made (it
1875 added property types that matched character-matching opcodes).
1876
1877 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1878 possible starting bytes for non-anchored patterns.
1879
1880 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1881 \R, and also a number of cases that involve Unicode properties, both
1882 explicit and implicit when PCRE_UCP is set.
1883
1884 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1885 input, it could crash or give wrong results if characters with values
1886 greater than 0xc0 were present in the subject string. (Detail: it assumed
1887 UTF-8 input when processing these items.)
1888
1889 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1890 size_t is 64-bit (#991).
1891
1892 17. Added a check for running out of memory when PCRE is compiled with
1893 --disable-stack-for-recursion (#990).
1894
1895 18. If the last data line in a file for pcretest does not have a newline on
1896 the end, a newline was missing in the output.
1897
1898 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1899 less than 128) in its various bitmaps. However, there is a facility for
1900 generating tables according to the current locale when PCRE is compiled. It
1901 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1902 characters, are recognized by isspace() and therefore were getting set in
1903 these tables, and indeed these tables seem to approximate to ISO 8859. This
1904 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1905 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1906 which of course cannot start UTF-8 characters. I have changed the code so
1907 that only real ASCII characters (less than 128) and the correct starting
1908 bytes for UTF-8 encodings are set for characters greater than 127 when in
1909 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1910 altogether.)
1911
1912 20. Added the /T option to pcretest so as to be able to run tests with non-
1913 standard character tables, thus making it possible to include the tests
1914 used for 19 above in the standard set of tests.
1915
1916 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1917 reference to a subpattern the other side of a comment that contains an
1918 opening parenthesis caused either an internal compiling error, or a
1919 reference to the wrong subpattern.
1920
1921
1922 Version 8.02 19-Mar-2010
1923 ------------------------
1924
1925 1. The Unicode data tables have been updated to Unicode 5.2.0.
1926
1927 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1928 configured.
1929
1930 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1931 original author of that file, following a query about its status.
1932
1933 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1934 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1935
1936 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1937 quantifier applied to a forward-referencing subroutine call, could compile
1938 incorrect code or give the error "internal error: previously-checked
1939 referenced subpattern not found".
1940
1941 6. Both MS Visual Studio and Symbian OS have problems with initializing
1942 variables to point to external functions. For these systems, therefore,
1943 pcre_malloc etc. are now initialized to local functions that call the
1944 relevant global functions.
1945
1946 7. There were two entries missing in the vectors called coptable and poptable
1947 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1948 I've fixed the data, and added a kludgy way of testing at compile time that
1949 the lengths are correct (equal to the number of opcodes).
1950
1951 8. Following on from 7, I added a similar kludge to check the length of the
1952 eint vector in pcreposix.c.
1953
1954 9. Error texts for pcre_compile() are held as one long string to avoid too
1955 much relocation at load time. To find a text, the string is searched,
1956 counting zeros. There was no check for running off the end of the string,
1957 which could happen if a new error number was added without updating the
1958 string.
1959
1960 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1961
1962 11. \K was not working if it appeared in an atomic group or in a group that
1963 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1964 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1965 the assertion is positive, but not if it is negative.
1966
1967 12. Change 11 fortuitously reduced the size of the stack frame used in the
1968 "match()" function of pcre_exec.c by one pointer. Forthcoming
1969 implementation of support for (*MARK) will need an extra pointer on the
1970 stack; I have reserved it now, so that the stack frame size does not
1971 decrease.
1972
1973 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1974 item in branch that calls a recursion is a subroutine call - as in the
1975 second branch in the above example - was incorrectly given the compile-
1976 time error "recursive call could loop indefinitely" because pcre_compile()
1977 was not correctly checking the subroutine for matching a non-empty string.
1978
1979 14. The checks for overrunning compiling workspace could trigger after an
1980 overrun had occurred. This is a "should never occur" error, but it can be
1981 triggered by pathological patterns such as hundreds of nested parentheses.
1982 The checks now trigger 100 bytes before the end of the workspace.
1983
1984 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1985
1986
1987 Version 8.01 19-Jan-2010
1988 ------------------------
1989
1990 1. If a pattern contained a conditional subpattern with only one branch (in
1991 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1992 computed the wrong minimum data length (which is of course zero for such
1993 subpatterns). This could cause incorrect "no match" results.
1994
1995 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1996 the pattern is reset in the first branch, pcre_compile() failed with
1997 "internal error: code overflow at offset...". This happened only when
1998 the reset was to the original external option setting. (An optimization
1999 abstracts leading options settings into an external setting, which was the
2000 cause of this.)
2001
2002 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2003 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2004 assertion pattern did not match (meaning that the assertion was true), it
2005 was incorrectly treated as false if the SKIP had been reached during the
2006 matching. This also applied to assertions used as conditions.
2007
2008 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2009 assertion subpattern, including such a pattern used as a condition,
2010 unpredictable results occurred, instead of the error return
2011 PCRE_ERROR_DFA_UITEM.
2012
2013 5. The C++ GlobalReplace function was not working like Perl for the special
2014 situation when an empty string is matched. It now does the fancy magic
2015 stuff that is necessary.
2016
2017 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2018 removed. (These were left over from very, very early versions of PCRE.)
2019
2020 7. Some cosmetic changes to the code to make life easier when compiling it
2021 as part of something else:
2022
2023 (a) Change DEBUG to PCRE_DEBUG.
2024
2025 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2026 called "current" as "current_branch", to prevent a collision with the
2027 Linux macro when compiled as a kernel module.
2028
2029 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2030 prevent a collision with the Linux macro when compiled as a kernel
2031 module.
2032
2033 8. In pcre_compile() there are some checks for integer overflows that used to
2034 cast potentially large values to (double). This has been changed to that
2035 when building, a check for int64_t is made, and if it is found, it is used
2036 instead, thus avoiding the use of floating point arithmetic. (There is no
2037 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2038 double.
2039
2040 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2041 2005 (difference between two addresses compared to an unsigned value).
2042
2043 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2044 custom one, because of the following reported problem in Windows:
2045
2046 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2047 under Win32.
2048 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2049 therefore missing the function definition.
2050 - The compiler thus generates a "C" signature for the test function.
2051 - The linker fails to find the "C" function.
2052 - PCRE fails to configure if asked to do so against libbz2.
2053
2054 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2055 messages were output:
2056
2057 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2058 rerunning libtoolize, to keep the correct libtool macros in-tree.
2059 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2060
2061 I have done both of these things.
2062
2063 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2064 most of the time, it *can* run out if it is given a pattern that contains a
2065 runaway infinite recursion. I updated the discussion in the pcrestack man
2066 page.
2067
2068 13. Now that we have gone to the x.xx style of version numbers, the minor
2069 version may start with zero. Using 08 or 09 is a bad idea because users
2070 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2071 interpreted as invalid octal numbers. I've updated the previous comment in
2072 configure.ac, and also added a check that gives an error if 08 or 09 are
2073 used.
2074
2075 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2076 causing partial matching to fail when the end of the subject matched \W
2077 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2078
2079 15. There were some discrepancies between the declarations in pcre_internal.h
2080 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2081 their definitions. The declarations used "const uschar *" and the
2082 definitions used USPTR. Even though USPTR is normally defined as "const
2083 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2084 reported that: "This difference in casting confuses some C++ compilers, for
2085 example, SunCC recognizes above declarations as different functions and
2086 generates broken code for hbpcre." I have changed the declarations to use
2087 USPTR.
2088
2089 16. GNU libtool is named differently on some systems. The autogen.sh script now
2090 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2091 (FreeBSD).
2092
2093 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2094 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2095 comment: "Figure out how to create a longlong from a string: strtoll and
2096 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2097 instance, but it only takes 2 args instead of 3!"
2098
2099 18. A subtle bug concerned with back references has been fixed by a change of
2100 specification, with a corresponding code fix. A pattern such as
2101 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2102 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2103 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2104 same bug. Such groups have to be quantified to be useful, or contained
2105 inside another quantified group. (If there's no repetition, the reference
2106 can never match.) The problem arises because, having left the group and
2107 moved on to the rest of the pattern, a later failure that backtracks into
2108 the group uses the captured value from the final iteration of the group
2109 rather than the correct earlier one. I have fixed this in PCRE by forcing
2110 any group that contains a reference to itself to be an atomic group; that
2111 is, there cannot be any backtracking into it once it has completed. This is
2112 similar to recursive and subroutine calls.
2113
2114
2115 Version 8.00 19-Oct-09
2116 ----------------------
2117
2118 1. The table for translating pcre_compile() error codes into POSIX error codes
2119 was out-of-date, and there was no check on the pcre_compile() error code
2120 being within the table. This could lead to an OK return being given in
2121 error.
2122
2123 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2124 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2125 in a Windows environment.
2126
2127 3. The pcregrep --count option prints the count for each file even when it is
2128 zero, as does GNU grep. However, pcregrep was also printing all files when
2129 --files-with-matches was added. Now, when both options are given, it prints
2130 counts only for those files that have at least one match. (GNU grep just
2131 prints the file name in this circumstance, but including the count seems
2132 more useful - otherwise, why use --count?) Also ensured that the
2133 combination -clh just lists non-zero counts, with no names.
2134
2135 4. The long form of the pcregrep -F option was incorrectly implemented as
2136 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2137 but it seems right to fix it, and I didn't think it was worth preserving
2138 the old behaviour.
2139
2140 5. The command line items --regex=pattern and --regexp=pattern were not
2141 recognized by pcregrep, which required --regex pattern or --regexp pattern
2142 (with a space rather than an '='). The man page documented the '=' forms,
2143 which are compatible with GNU grep; these now work.
2144
2145 6. No libpcreposix.pc file was created for pkg-config; there was just
2146 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2147
2148 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2149 when UCP support is not needed, by modifying the Python script that
2150 generates it from Unicode data files. This should not matter if the module
2151 is correctly used as a library, but I received one complaint about 50K of
2152 unwanted data. My guess is that the person linked everything into his
2153 program rather than using a library. Anyway, it does no harm.
2154
2155 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2156 was a minimum greater than 1 for a wide character in a possessive
2157 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2158 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2159 character. Chaos in the form of incorrect output or a compiling loop could
2160 result.
2161
2162 9. The restrictions on what a pattern can contain when partial matching is
2163 requested for pcre_exec() have been removed. All patterns can now be
2164 partially matched by this function. In addition, if there are at least two
2165 slots in the offset vector, the offset of the earliest inspected character
2166 for the match and the offset of the end of the subject are set in them when
2167 PCRE_ERROR_PARTIAL is returned.
2168
2169 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2170 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2171 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2172 and may be more useful for multi-segment matching.
2173
2174 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2175 used to be given if ever the end of the subject was reached; now it is
2176 given only if matching could not proceed because another character was
2177 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2178 string "Z", which now yields "no match" instead of "partial match". In the
2179 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2180 final character ended with (*FAIL).
2181
2182 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2183 if the pattern had a "must contain" character that was already found in the
2184 earlier partial match, unless partial matching was again requested. For
2185 example, with the pattern /dog.(body)?/, the "must contain" character is
2186 "g". If the first part-match was for the string "dog", restarting with
2187 "sbody" failed. This bug has been fixed.
2188
2189 13. The string returned by pcre_dfa_exec() after a partial match has been
2190 changed so that it starts at the first inspected character rather than the
2191 first character of the match. This makes a difference only if the pattern
2192 starts with a lookbehind assertion or \b or \B (\K is not supported by
2193 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2194 matching functions compatible, and I think it's the right thing to do.
2195
2196 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2197 so that the demonstration program is easily available in environments where
2198 PCRE has not been installed from source.
2199
2200 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2201 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2202 library.
2203
2204 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2205 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2206 is not the first non-POSIX option to be added. Clearly some people find
2207 these options useful.
2208
2209 17. If a caller to the POSIX matching function regexec() passes a non-zero
2210 value for nmatch with a NULL value for pmatch, the value of
2211 nmatch is forced to zero.
2212
2213 18. RunGrepTest did not have a test for the availability of the -u option of
2214 the diff command, as RunTest does. It now checks in the same way as
2215 RunTest, and also checks for the -b option.
2216
2217 19. If an odd number of negated classes containing just a single character
2218 interposed, within parentheses, between a forward reference to a named
2219 subpattern and the definition of the subpattern, compilation crashed with
2220 an internal error, complaining that it could not find the referenced
2221 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2222 [The bug was that it was starting one character too far in when skipping
2223 over the character class, thus treating the ] as data rather than
2224 terminating the class. This meant it could skip too much.]
2225
2226 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2227 /g option in pcretest when the pattern contains \K, which makes it possible
2228 to have an empty string match not at the start, even when the pattern is
2229 anchored. Updated pcretest and pcredemo to use this option.
2230
2231 21. If the maximum number of capturing subpatterns in a recursion was greater
2232 than the maximum at the outer level, the higher number was returned, but
2233 with unset values at the outer level. The correct (outer level) value is
2234 now given.
2235
2236 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2237 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2238 make it do so. The string so far is captured, making this feature
2239 compatible with Perl.
2240
2241 23. The tests have been re-organized, adding tests 11 and 12, to make it
2242 possible to check the Perl 5.10 features against Perl 5.10.
2243
2244 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2245 pattern matches a fixed length string. PCRE did not allow this; now it
2246 does. Neither allows recursion.
2247
2248 25. I finally figured out how to implement a request to provide the minimum
2249 length of subject string that was needed in order to match a given pattern.
2250 (It was back references and recursion that I had previously got hung up
2251 on.) This code has now been added to pcre_study(); it finds a lower bound
2252 to the length of subject needed. It is not necessarily the greatest lower
2253 bound, but using it to avoid searching strings that are too short does give
2254 some useful speed-ups. The value is available to calling programs via
2255 pcre_fullinfo().
2256
2257 26. While implementing 25, I discovered to my embarrassment that pcretest had
2258 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2259 study optimizations had never been tested with that matching function.
2260 Oops. What is worse, even when it was passed study data, there was a bug in
2261 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2262 were also very few tests of studied patterns with pcre_dfa_exec().
2263
2264 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2265 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2266 on the other side of the coin, they are no longer allowed to have different
2267 names, because these cannot be distinguished in PCRE, and this has caused
2268 confusion. (This is a difference from Perl.)
2269
2270 28. When duplicate subpattern names are present (necessarily with different
2271 numbers, as required by 27 above), and a test is made by name in a
2272 conditional pattern, either for a subpattern having been matched, or for
2273 recursion in such a pattern, all the associated numbered subpatterns are
2274 tested, and the overall condition is true if the condition is true for any
2275 one of them. This is the way Perl works, and is also more like the way
2276 testing by number works.
2277
2278
2279 Version 7.9 11-Apr-09
2280 ---------------------
2281
2282 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2283 (pcretest), all targets were linked against these libraries. This included
2284 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2285 libraries. This caused unwanted dependencies to be created. This problem
2286 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2287 pcretest is linked with readline.
2288
2289 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2290 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2291 moved outside it again, because FALSE and TRUE are already defined in AIX,
2292 but BOOL is not.
2293
2294 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2295 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2296
2297 4. The pcregrep documentation said spaces were inserted as well as colons (or
2298 hyphens) following file names and line numbers when outputting matching
2299 lines. This is not true; no spaces are inserted. I have also clarified the
2300 wording for the --colour (or --color) option.
2301
2302 5. In pcregrep, when --colour was used with -o, the list of matching strings
2303 was not coloured; this is different to GNU grep, so I have changed it to be
2304 the same.
2305
2306 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2307 each matching line was coloured. Now it goes on to look for further matches
2308 of any of the test patterns, which is the same behaviour as GNU grep.
2309
2310 7. A pattern that could match an empty string could cause pcregrep to loop; it
2311 doesn't make sense to accept an empty string match in pcregrep, so I have
2312 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2313 seems to be how GNU grep behaves. [But see later change 40 for release
2314 8.33.]
2315
2316 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2317 start or after a newline", because the conditional assertion was not being
2318 correctly handled. The rule now is that both the assertion and what follows
2319 in the first alternative must satisfy the test.
2320
2321 9. If auto-callout was enabled in a pattern with a conditional group whose
2322 condition was an assertion, PCRE could crash during matching, both with
2323 pcre_exec() and pcre_dfa_exec().
2324
2325 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2326 used for matching.
2327
2328 11. Unicode property support in character classes was not working for
2329 characters (bytes) greater than 127 when not in UTF-8 mode.
2330
2331 12. Added the -M command line option to pcretest.
2332
2333 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2334
2335 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2336
2337 16. Added comments and documentation about mis-use of no_arg in the C++
2338 wrapper.
2339
2340 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2341 from Martin Jerabek that uses macro names for all relevant character and
2342 string constants.
2343
2344 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2345 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2346 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2347 these, but not everybody uses configure.
2348
2349 19. A conditional group that had only one branch was not being correctly
2350 recognized as an item that could match an empty string. This meant that an
2351 enclosing group might also not be so recognized, causing infinite looping
2352 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2353 with the subject "ab", where knowledge that the repeated group can match
2354 nothing is needed in order to break the loop.
2355
2356 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2357 exec(), but without supplying a callout function, matching went wrong.
2358
2359 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2360 leak if the size of the offset vector was greater than 30. When the vector
2361 is smaller, the saved offsets during recursion go onto a local stack
2362 vector, but for larger vectors malloc() is used. It was failing to free
2363 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2364 error, in fact).
2365
2366 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2367 heapframe that is used only when UTF-8 support is enabled. This caused no
2368 problem, but was untidy.
2369
2370 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2371 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2372 included within another project.
2373
2374 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2375 slightly modified by me:
2376
2377 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2378 not building pcregrep.
2379
2380 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2381 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2382
2383 25. Forward references, both numeric and by name, in patterns that made use of
2384 duplicate group numbers, could behave incorrectly or give incorrect errors,
2385 because when scanning forward to find the reference group, PCRE was not
2386 taking into account the duplicate group numbers. A pattern such as
2387 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2388
2389 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2390 the feature of a custom pointer more persuasive (as requested by a user).
2391
2392 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2393 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2394 pre-defined.
2395
2396 28. Added support for (*UTF8) at the start of a pattern.
2397
2398 29. Arrange for flags added by the "release type" setting in CMake to be shown
2399 in the configuration summary.
2400
2401
2402 Version 7.8 05-Sep-08
2403 ---------------------
2404
2405 1. Replaced UCP searching code with optimized version as implemented for Ad
2406 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2407 stage table and inline lookup instead of a function, giving speed ups of 2
2408 to 5 times on some simple patterns that I tested. Permission was given to
2409 distribute the MultiStage2.py script that generates the tables (it's not in
2410 the tarball, but is in the Subversion repository).
2411
2412 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2413 scripts.
2414
2415 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2416 a group with a zero qualifier. The result of the study could be incorrect,
2417 or the function might crash, depending on the pattern.
2418
2419 4. Caseless matching was not working for non-ASCII characters in back
2420 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2421 It now works when Unicode Property Support is available.
2422
2423 5. In pcretest, an escape such as \x{de} in the data was always generating
2424 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2425 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2426 truncation.
2427
2428 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2429
2430 7. Added two (int) casts to pcregrep when printing the difference of two
2431 pointers, in case they are 64-bit values.
2432
2433 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2434 test 2 if it fails.
2435
2436 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2437 and a #define of that name to empty if it is not externally set. This is to
2438 allow users of MSVC to set it if necessary.
2439
2440 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2441 the convenience functions in the pcre_get.c source file.
2442
2443 11. An option change at the start of a pattern that had top-level alternatives
2444 could cause overwriting and/or a crash. This command provoked a crash in
2445 some environments:
2446
2447 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2448
2449 This potential security problem was recorded as CVE-2008-2371.
2450
2451 12. For a pattern where the match had to start at the beginning or immediately
2452 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2453 pcre_dfa_exec() could read past the end of the passed subject if there was
2454 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2455 pcretest so that it places the subject at the end of its malloc-ed buffer.
2456
2457 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2458 exec() might read past the end of the data buffer in UTF-8 mode.
2459
2460 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2461 the data contained the byte 0x85 as part of a UTF-8 character within its
2462 first line. This applied both to normal and DFA matching.
2463
2464 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2465 /^[^d]*?$/8 failed to match "abc".
2466
2467 16. Added a missing copyright notice to pcrecpp_internal.h.
2468
2469 17. Make it more clear in the documentation that values returned from
2470 pcre_exec() in ovector are byte offsets, not character counts.
2471
2472 18. Tidied a few places to stop certain compilers from issuing warnings.
2473
2474 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2475 supplied by Stefan Weber. I made a further small update for 7.8 because
2476 there is a change of source arrangements: the pcre_searchfuncs.c module is
2477 replaced by pcre_ucd.c.
2478
2479
2480 Version 7.7 07-May-08
2481 ---------------------
2482
2483 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2484 a string to a long long, pretend we don't even have a long long." This is
2485 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2486
2487 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2488 pre-7.6 versions, which defined a global no_arg variable instead of putting
2489 it in the RE class. (See also #8 below.)
2490
2491 3. Remove a line of dead code, identified by coverity and reported by Nuno
2492 Lopes.
2493
2494 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2495
2496 (1) The include/exclude patterns were being applied to the whole pathnames
2497 of files, instead of just to the final components.
2498
2499 (2) If there was more than one level of directory, the subdirectories were
2500 skipped unless they satisfied the include/exclude conditions. This is
2501 inconsistent with GNU grep (and could even be seen as contrary to the
2502 pcregrep specification - which I improved to make it absolutely clear).
2503 The action now is always to scan all levels of directory, and just
2504 apply the include/exclude patterns to regular files.
2505
2506 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2507 --exclude_dir in the tests to avoid scanning .svn directories.
2508
2509 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2510 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2511 doesn't support NULs in patterns.
2512
2513 7. Added some missing "const"s to declarations of static tables in
2514 pcre_compile.c and pcre_dfa_exec.c.
2515
2516 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2517 caused by fix #2 above. (Subsequently also a second patch to fix the
2518 first patch. And a third patch - this was a messy problem.)
2519
2520 9. Applied Craig's patch to remove the use of push_back().
2521
2522 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2523 matching function regexec().
2524
2525 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2526 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2527 references. PCRE supports relative numbers with this syntax (I don't think
2528 Oniguruma does).
2529
2530 12. Previously, a group with a zero repeat such as (...){0} was completely
2531 omitted from the compiled regex. However, this means that if the group
2532 was called as a subroutine from elsewhere in the pattern, things went wrong
2533 (an internal error was given). Such groups are now left in the compiled
2534 pattern, with a new opcode that causes them to be skipped at execution
2535 time.
2536
2537 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2538 to the way PCRE behaves:
2539
2540 (a) A lone ] character is dis-allowed (Perl treats it as data).
2541
2542 (b) A back reference to an unmatched subpattern matches an empty string
2543 (Perl fails the current match path).
2544
2545 (c) A data ] in a character class must be notated as \] because if the
2546 first data character in a class is ], it defines an empty class. (In
2547 Perl it is not possible to have an empty class.) The empty class []
2548 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2549 The negative empty class [^] matches any one character, independently
2550 of the DOTALL setting.
2551
2552 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2553 non-existent subpattern following a character class starting with ']' and
2554 containing () gave an internal compiling error instead of "reference to
2555 non-existent subpattern". Fortunately, when the pattern did exist, the
2556 compiled code was correct. (When scanning forwards to check for the
2557 existence of the subpattern, it was treating the data ']' as terminating
2558 the class, so got the count wrong. When actually compiling, the reference
2559 was subsequently set up correctly.)
2560
2561 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2562 it was being rejected as not supported by pcre_dfa_exec(), even though
2563 other assertions are supported. I have made pcre_dfa_exec() support
2564 (*FAIL).
2565
2566 16. The implementation of 13c above involved the invention of a new opcode,
2567 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2568 cannot be changed at match time, I realized I could make a small
2569 improvement to matching performance by compiling OP_ALLANY instead of
2570 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2571 on the OP_ANY path.
2572
2573 17. Compiling pcretest on Windows with readline support failed without the
2574 following two fixes: (1) Make the unistd.h include conditional on
2575 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2576
2577 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2578 ncurses library to be included for pcretest when ReadLine support is
2579 requested, but also to allow for it to be overridden. This patch came from
2580 Daniel Bergström.
2581
2582 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2583 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2584 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2585 spotting this.
2586
2587
2588 Version 7.6 28-Jan-08
2589 ---------------------
2590
2591 1. A character class containing a very large number of characters with
2592 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2593 overflow.
2594
2595 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2596 HAVE_LONG_LONG is not defined.
2597
2598 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2599 bring them up to date and include new features. This patch includes:
2600
2601 - Fixed PH's badly added libz and libbz2 support.
2602 - Fixed a problem with static linking.
2603 - Added pcredemo. [But later removed - see 7 below.]
2604 - Fixed dftables problem and added an option.
2605 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2606 HAVE_LONG_LONG.
2607 - Added readline support for pcretest.
2608 - Added an listing of the option settings after cmake has run.
2609
2610 4. A user submitted a patch to Makefile that makes it easy to create
2611 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2612 Makefile.am that cause it to include this special target, without
2613 affecting anything else. Note that the same mingw target plus all
2614 the other distribution libraries and programs are now supported
2615 when configuring with CMake (see 6 below) instead of with
2616 Configure/Make.
2617
2618 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2619 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2620 exported in the Windows port". It has not yet been confirmed that the patch
2621 solves the problem, but it does no harm.
2622
2623 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2624 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2625 with CMake, and also correct the comment about stack recursion.
2626
2627 7. Remove the automatic building of pcredemo from the ./configure system and
2628 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2629 of a program that users should build themselves after PCRE is installed, so
2630 building it automatically is not really right. What is more, it gave
2631 trouble in some build environments.
2632
2633 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2634
2635
2636 Version 7.5 10-Jan-08
2637 ---------------------
2638
2639 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2640 values in parens when parsing an RE using the C++ wrapper."
2641
2642 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2643 Characters greater than 255 were excluded from the class instead of being
2644 included.
2645
2646 3. The same bug as (2) above applied to negated POSIX classes such as
2647 [:^space:].
2648
2649 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2650 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2651 I have changed it.
2652
2653 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2654 first named subpattern) and a construct such as (?&a) would reference the
2655 first named subpattern whose name started with "a" (in other words, the
2656 length check was missing). Both these problems are fixed. "Subpattern name
2657 expected" is now given for (?&) (a zero-length name), and this patch also
2658 makes it give the same error for \k'' (previously it complained that that
2659 was a reference to a non-existent subpattern).
2660
2661 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2662 this is right because (?- can be followed by option settings as well as by
2663 digits. I have, however, made the messages clearer.
2664
2665 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2666 than the number used in the conditional) now cause a compile-time error.
2667 This is actually not compatible with Perl, which accepts such patterns, but
2668 treats the conditional as always being FALSE (as PCRE used to), but it
2669 seems to me that giving a diagnostic is better.
2670
2671 8. Change "alphameric" to the more common word "alphanumeric" in comments
2672 and messages.
2673
2674 9. Fix two occurrences of "backslash" in comments that should have been
2675 "backspace".
2676
2677 10. Remove two redundant lines of code that can never be obeyed (their function
2678 was moved elsewhere).
2679
2680 11. The program that makes PCRE's Unicode character property table had a bug
2681 which caused it to generate incorrect table entries for sequences of
2682 characters that have the same character type, but are in different scripts.
2683 It amalgamated them into a single range, with the script of the first of
2684 them. In other words, some characters were in the wrong script. There were
2685 thirteen such cases, affecting characters in the following ranges:
2686
2687 U+002b0 - U+002c1
2688 U+0060c - U+0060d
2689 U+0061e - U+00612
2690 U+0064b - U+0065e
2691 U+0074d - U+0076d
2692 U+01800 - U+01805
2693 U+01d00 - U+01d77
2694 U+01d9b - U+01dbf
2695 U+0200b - U+0200f
2696 U+030fc - U+030fe
2697 U+03260 - U+0327f
2698 U+0fb46 - U+0fbb1
2699 U+10450 - U+1049d
2700
2701 12. The -o option (show only the matching part of a line) for pcregrep was not
2702 compatible with GNU grep in that, if there was more than one match in a
2703 line, it showed only the first of them. It now behaves in the same way as
2704 GNU grep.
2705
2706 13. If the -o and -v options were combined for pcregrep, it printed a blank
2707 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2708 does the same. The return code can be used to tell if there were any
2709 non-matching lines.
2710
2711 14. Added --file-offsets and --line-offsets to pcregrep.
2712
2713 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2714 infinitely looping recursion. The bug was that positive lookaheads were not
2715 being skipped when checking for a possible empty match (negative lookaheads
2716 and both kinds of lookbehind were skipped).
2717
2718 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2719 inclusion of <windows.h> to before rather than after the definition of
2720 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2721
2722 17. Specifying a possessive quantifier with a specific limit for a Unicode
2723 character property caused pcre_compile() to compile bad code, which led at
2724 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2725 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2726 caused the error; without that there was no problem.
2727
2728 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2729
2730 19. Added --enable-pcretest-libreadline.
2731
2732 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2733 RE::GlobalReplace(). As a result, the number of replacements returned was
2734 double what it should be. I removed one of the increments, but Craig sent a
2735 later patch that removed the other one (the right fix) and added unit tests
2736 that check the return values (which was not done before).
2737
2738 21. Several CMake things:
2739
2740 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2741 the names libpcre and libpcreposix, not just pcre and pcreposix.
2742
2743 (2) The above change means that pcretest and pcregrep are now correctly
2744 linked with the newly-built libraries, not previously installed ones.
2745
2746 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2747
2748 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2749 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2750 UTF-8 newline character). The key issue is that the pattern starts .*;
2751 this means that the match must be either at the beginning, or after a
2752 newline. The bug was in the code for advancing after a failed match and
2753 checking that the new position followed a newline. It was not taking
2754 account of UTF-8 characters correctly.
2755
2756 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2757 character classes. PCRE was not treating the sequence [:...:] as a
2758 character class unless the ... were all letters. Perl, however, seems to
2759 allow any characters between [: and :], though of course it rejects as
2760 unknown any "names" that contain non-letters, because all the known class
2761 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2762 for example, whereas PCRE did not - it did not recognize a POSIX character
2763 class. This seemed a bit dangerous, so the code has been changed to be
2764 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2765 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2766 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2767 Perl does, and where it didn't before.
2768
2769 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2770 Windows environments %n is disabled by default.
2771
2772
2773 Version 7.4 21-Sep-07
2774 ---------------------
2775
2776 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2777 means that a class such as [\s] counted as "explicit reference to CR or
2778 LF". That isn't really right - the whole point of the change was to try to
2779 help when there was an actual mention of one of the two characters. So now
2780 the change happens only if \r or \n (or a literal CR or LF) character is
2781 encountered.
2782
2783 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2784 of both had grown to the point where there were only 3 bits left.
2785 Fortunately, there was spare space in the data structure, and so I have
2786 moved the internal flags into a new 16-bit field to free up more option
2787 bits.
2788
2789 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2790 but did not set the internal JCHANGED flag - either of these is enough to
2791 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2792 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2793 start sets both bits.
2794
2795 4. Added options (at build time, compile time, exec time) to change \R from
2796 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2797
2798 5. doc/pcresyntax.html was missing from the distribution.
2799
2800 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2801 compatibility, even though it is no longer used.
2802
2803 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2804 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2805 windows.h file is present (where different names are used). [This was
2806 reversed later after testing - see 16 below.]
2807
2808 8. Changed all #include <config.h> to #include "config.h". There were also
2809 some further <pcre.h> cases that I changed to "pcre.h".
2810
2811 9. When pcregrep was used with the --colour option, it missed the line ending
2812 sequence off the lines that it output.
2813
2814 10. It was pointed out to me that arrays of string pointers cause lots of
2815 relocations when a shared library is dynamically loaded. A technique of
2816 using a single long string with a table of offsets can drastically reduce
2817 these. I have refactored PCRE in four places to do this. The result is
2818 dramatic:
2819
2820 Originally: 290
2821 After changing UCP table: 187
2822 After changing error message table: 43
2823 After changing table of "verbs" 36
2824 After changing table of Posix names 22
2825
2826 Thanks to the folks working on Gregex for glib for this insight.
2827
2828 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2829 unicode-properties was also set.
2830
2831 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2832
2833 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2834 checked only for CRLF.
2835
2836 14. Added casts to pcretest.c to avoid compiler warnings.
2837
2838 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2839
2840 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2841 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2842 entirely. This removes changes made in 7 above.
2843
2844 17. The CMake files have been updated, and there is now more information about
2845 building with CMake in the NON-UNIX-USE document.
2846
2847
2848 Version 7.3 28-Aug-07
2849 ---------------------
2850
2851 1. In the rejigging of the build system that eventually resulted in 7.1, the
2852 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2853 brackets there is not right, since it causes compilers to look for an
2854 installed pcre.h, not the version that is in the source that is being
2855 compiled (which of course may be different). I have changed it back to:
2856
2857 #include "pcre.h"
2858
2859 I have a vague recollection that the change was concerned with compiling in
2860 different directories, but in the new build system, that is taken care of
2861 by the VPATH setting the Makefile.
2862
2863 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2864 when the subject happened to end in the byte 0x85 (e.g. if the last
2865 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2866 characters but of course it shouldn't be taken as a newline when it is part
2867 of another character. The bug was that, for an unlimited repeat of . in
2868 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2869 characters when looking for a newline.
2870
2871 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2872
2873 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2874 in debug output.
2875
2876 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2877 long printing in the pcrecpp unittest when running under MinGW.
2878
2879 6. ESC_K was left out of the EBCDIC table.
2880
2881 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2882 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2883 limit also applies to "virtual nesting" when a pattern is recursive, and in
2884 this case 1000 isn't so big. I have been able to remove this limit at the
2885 expense of backing off one optimization in certain circumstances. Normally,
2886 when pcre_exec() would call its internal match() function recursively and
2887 immediately return the result unconditionally, it uses a "tail recursion"
2888 feature to save stack. However, when a subpattern that can match an empty
2889 string has an unlimited repetition quantifier, it no longer makes this
2890 optimization. That gives it a stack frame in which to save the data for
2891 checking that an empty string has been matched. Previously this was taken
2892 from the 1000-entry workspace that had been reserved. So now there is no
2893 explicit limit, but more stack is used.
2894
2895 8. Applied Daniel's patches to solve problems with the import/export magic
2896 syntax that is required for Windows, and which was going wrong for the
2897 pcreposix and pcrecpp parts of the library. These were overlooked when this
2898 problem was solved for the main library.
2899
2900 9. There were some crude static tests to avoid integer overflow when computing
2901 the size of patterns that contain repeated groups with explicit upper
2902 limits. As the maximum quantifier is 65535, the maximum group length was
2903 set at 30,000 so that the product of these two numbers did not overflow a
2904 32-bit integer. However, it turns out that people want to use groups that
2905 are longer than 30,000 bytes (though not repeat them that many times).
2906 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2907 made it possible to implement the integer overflow checks in a much more
2908 dynamic way, which I have now done. The artificial limitation on group
2909 length has been removed - we now have only the limit on the total length of
2910 the compiled pattern, which depends on the LINK_SIZE setting.
2911
2912 10. Fixed a bug in the documentation for get/copy named substring when
2913 duplicate names are permitted. If none of the named substrings are set, the
2914 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2915 empty string.
2916
2917 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2918 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2919 because the ] is interpreted as the first data character and the
2920 terminating ] is not found. PCRE has been made compatible with Perl in this
2921 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2922 cause memory overwriting.
2923
2924 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2925 string has been matched (to stop an infinite loop). It was not recognizing
2926 a conditional subpattern that could match an empty string if that
2927 subpattern was within another subpattern. For example, it looped when
2928 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2929 condition was not nested. This bug has been fixed.
2930
2931 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2932 past the start of the subject in the presence of bytes with the top bit
2933 set, for example "\x8aBCD".
2934
2935 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2936 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2937
2938 14. Optimized (?!) to (*FAIL).
2939
2940 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2941 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2942 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2943 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2944 does: it's just the validity check that is more restrictive.
2945
2946 16. Inserted checks for integer overflows during escape sequence (backslash)
2947 processing, and also fixed erroneous offset values for syntax errors during
2948 backslash processing.
2949
2950 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2951 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2952
2953 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2954 caused an overrun.
2955
2956 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2957 something other than just ASCII characters) inside a group that had an
2958 unlimited repeat caused a loop at compile time (while checking to see
2959 whether the group could match an empty string).
2960
2961 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2962 [\P{Any}] did so. (Error in the code for printing property names.)
2963
2964 21. An orphan \E inside a character class could cause a crash.
2965
2966 22. A repeated capturing bracket such as (A)? could cause a wild memory
2967 reference during compilation.
2968
2969 23. There are several functions in pcre_compile() that scan along a compiled
2970 expression for various reasons (e.g. to see if it's fixed length for look
2971 behind). There were bugs in these functions when a repeated \p or \P was
2972 present in the pattern. These operators have additional parameters compared
2973 with \d, etc, and these were not being taken into account when moving along
2974 the compiled data. Specifically:
2975
2976 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2977 length.
2978
2979 (b) An item such as \pL+ within a repeated group could cause crashes or
2980 loops.
2981
2982 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2983 "reference to non-existent subpattern" error.
2984
2985 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2986
2987 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2988 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2989
2990 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2991
2992 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2993 character were causing crashes (broken optimization).
2994
2995 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2996 \p or \P) caused a compile-time loop.
2997
2998 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2999 break. For example, the unstudied pattern [\r\n]A does not match the string
3000 "\r\nA" because change 7.0/46 below moves the current point on by two
3001 characters after failing to match at the start. However, the pattern \nA
3002 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3003 the same is true. There doesn't seem any very clean way out of this, but
3004 what I have chosen to do makes the common cases work: PCRE now takes note
3005 of whether there can be an explicit match for \r or \n anywhere in the
3006 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3007 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3008 pattern has explicit CR or LF references.
3009
3010 29. Added (*CR) etc for changing newline setting at start of pattern.
3011
3012
3013 Version 7.2 19-Jun-07
3014 ---------------------
3015
3016 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3017 which is apparently normally available under Windows.
3018
3019 2. Re-jig the pcregrep tests with different newline settings in an attempt
3020 to make them independent of the local environment's newline setting.
3021
3022 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3023
3024 4. Some of the "internals" tests were previously cut out when the link size
3025 was not 2, because the output contained actual offsets. The recent new
3026 "Z" feature of pcretest means that these can be cut out, making the tests
3027 usable with all link sizes.
3028
3029 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3030 stack recursion. This gives a massive performance boost under BSD, but just
3031 a small improvement under Linux. However, it saves one field in the frame
3032 in all cases.
3033
3034 6. Added more features from the forthcoming Perl 5.10:
3035
3036 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3037 recursion call. It refers to the nth most recently opened parentheses.
3038
3039 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3040 to be opened parentheses.
3041
3042 (c) Conditions that refer to capturing parentheses can be specified
3043 relatively, for example, (?(-2)... or (?(+3)...
3044
3045 (d) \K resets the start of the current match so that everything before
3046 is not part of it.
3047
3048 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3049
3050 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3051 reference syntax.
3052
3053 (g) (?| introduces a group in which the numbering of parentheses in each
3054 alternative starts with the same number.
3055
3056 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3057
3058 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3059 PCRE_INFO_JCHANGED.
3060
3061 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3062 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3063 for detecting groups that can match an empty string.
3064
3065 9. A pattern with a very large number of alternatives (more than several
3066 hundred) was running out of internal workspace during the pre-compile
3067 phase, where pcre_compile() figures out how much memory will be needed. A
3068 bit of new cunning has reduced the workspace needed for groups with
3069 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3070 workspace instead of running out of the 4096 that are available.
3071
3072 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3073
3074 11. Applied patch from Google to remove an optimization that didn't quite work.
3075 The report of the bug said:
3076
3077 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3078 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3079 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3080
3081 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3082 it matched the wrong number of bytes.
3083
3084
3085 Version 7.1 24-Apr-07
3086 ---------------------
3087
3088 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3089 that is more "standard", making use of automake and other Autotools. There
3090 is some re-arrangement of the files and adjustment of comments consequent
3091 on this.
3092
3093 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3094 for recursive directory scanning broke on some systems because the files
3095 are not scanned in any specific order and on different systems the order
3096 was different. A call to "sort" has been inserted into RunGrepTest for the
3097 approprate test as a short-term fix. In the longer term there may be an
3098 alternative.
3099
3100 3. I had an email from Eric Raymond about problems translating some of PCRE's
3101 man pages to HTML (despite the fact that I distribute HTML pages, some
3102 people do their own conversions for various reasons). The problems
3103 concerned the use of low-level troff macros .br and .in. I have therefore
3104 removed all such uses from the man pages (some were redundant, some could
3105 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3106 HTML has been updated to handle .nf/.fi and to complain if it encounters
3107 .br or .in.
3108
3109 4. Updated comments in configure.ac that get placed in config.h.in and also
3110 arranged for config.h to be included in the distribution, with the name
3111 config.h.generic, for the benefit of those who have to compile without
3112 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3113
3114 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3115 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3116 makevp.bat for the current PCRE, using the additional files
3117 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3118
3119 6. A Windows user reported a minor discrepancy with test 2, which turned out
3120 to be caused by a trailing space on an input line that had got lost in his
3121 copy. The trailing space was an accident, so I've just removed it.
3122
3123 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3124 that is needed.
3125
3126 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3127 as "const" (a) because they are and (b) because it helps the PHP
3128 maintainers who have recently made a script to detect big data structures
3129 in the php code that should be moved to the .rodata section. I remembered
3130 to update Builducptable as well, so it won't revert if ucptable.h is ever
3131 re-created.
3132
3133 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3134 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3135 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3136 support is not required. This saves 1.5-2K of code, which is important in
3137 some applications.
3138
3139 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3140 so as not to refer to the tables, even though these functions will never be
3141 called when UTF-8 support is disabled. Otherwise there are problems with a
3142 shared library.
3143
3144 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3145
3146 (a) It was defining its arguments as char * instead of void *.
3147
3148 (b) It was assuming that all moves were upwards in memory; this was true
3149 a long time ago when I wrote it, but is no longer the case.
3150
3151 The emulated memove() is provided for those environments that have neither
3152 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3153 is clearly not the case, as these two bugs were recently reported.
3154
3155 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3156 and Detrail to create the HTML documentation, the .txt form of the man
3157 pages, and it removes trailing spaces from listed files. It also creates
3158 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3159 case, it wraps all the #defines with #ifndefs. This script should be run
3160 before "make dist".
3161
3162 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3163 with Unicode property support.
3164
3165 (a) For a maximizing quantifier, if the two different cases of the
3166 character were of different lengths in their UTF-8 codings (there are
3167 some cases like this - I found 11), and the matching function had to
3168 back up over a mixture of the two cases, it incorrectly assumed they
3169 were both the same length.
3170
3171 (b) When PCRE was configured to use the heap rather than the stack for
3172 recursion during matching, it was not correctly preserving the data for
3173 the other case of a UTF-8 character when checking ahead for a match
3174 while processing a minimizing repeat. If the check also involved
3175 matching a wide character, but failed, corruption could cause an
3176 erroneous result when trying to check for a repeat of the original
3177 character.
3178
3179 13. Some tidying changes to the testing mechanism:
3180
3181 (a) The RunTest script now detects the internal link size and whether there
3182 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3183 values substituted by "configure". (The RunGrepTest script already did
3184 this for UTF-8.) The configure.ac script no longer substitutes the
3185 relevant variables.
3186
3187 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3188 with length and offset values. This means that the output is different
3189 for different internal link sizes. Test 2 is skipped for link sizes
3190 other than 2 because of this, bypassing the problem. Unfortunately,
3191 there was also a test in test 3 (the locale tests) that used /B and
3192 failed for link sizes other than 2. Rather than cut the whole test out,
3193 I have added a new /Z option to pcretest that replaces the length and
3194 offset values with spaces. This is now used to make test 3 independent
3195 of link size. (Test 2 will be tidied up later.)
3196
3197 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3198 segmentation fault instead of returning the appropriate error message.
3199
3200 15. In multiline mode when the newline sequence was set to "any", the pattern
3201 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3202 This doesn't seem right; it now treats the CRLF combination as the line
3203 ending, and so does not match in that case. It's only a pattern such as ^$
3204 that would hit this one: something like ^ABC$ would have failed after \r
3205 and then tried again after \r\n.
3206
3207 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3208 in an attempt to make files that differ only in their line terminators
3209 compare equal. This works on Linux.
3210
3211 17. Under certain error circumstances pcregrep might try to free random memory
3212 as it exited. This is now fixed, thanks to valgrind.
3213
3214 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3215 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3216 was because its rules for how to advance for /g after matching an empty
3217 string at the end of a line did not allow for this case. They now check for
3218 it specially.
3219
3220 20. pcretest is supposed to handle patterns and data of any length, by
3221 extending its buffers when necessary. It was getting this wrong when the
3222 buffer for a data line had to be extended.
3223
3224 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3225 CRLF as a newline sequence.
3226
3227 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3228 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3229 I have nevertheless tidied it up.
3230
3231 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3232
3233 24. Added a man page for pcre-config.
3234
3235
3236 Version 7.0 19-Dec-06
3237 ---------------------
3238
3239 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3240 moving to gcc 4.1.1.
3241
3242 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3243 sys/time.h, which is documented as needed for this function. It doesn't
3244 seem to matter on Linux, but it showed up on some releases of OS X.
3245
3246 3. It seems that there are systems where bytes whose values are greater than
3247 127 match isprint() in the "C" locale. The "C" locale should be the
3248 default when a C program starts up. In most systems, only ASCII printing
3249 characters match isprint(). This difference caused the output from pcretest
3250 to vary, making some of the tests fail. I have changed pcretest so that:
3251
3252 (a) When it is outputting text in the compiled version of a pattern, bytes
3253 other than 32-126 are always shown as hex escapes.
3254
3255 (b) When it is outputting text that is a matched part of a subject string,
3256 it does the same, unless a different locale has been set for the match
3257 (using the /L modifier). In this case, it uses isprint() to decide.
3258
3259 4. Fixed a major bug that caused incorrect computation of the amount of memory
3260 required for a compiled pattern when options that changed within the
3261 pattern affected the logic of the preliminary scan that determines the
3262 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3263 that the computed length was too small. The symptoms of this bug were
3264 either the PCRE error "internal error: code overflow" from pcre_compile(),
3265 or a glibc crash with a message such as "pcretest: free(): invalid next
3266 size (fast)". Examples of patterns that provoked this bug (shown in
3267 pcretest format) are:
3268
3269 /(?-x: )/x
3270 /(?x)(?-x: \s*#\s*)/
3271 /((?i)[\x{c0}])/8
3272 /(?i:[\x{c0}])/8
3273
3274 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3275 is now done differently.
3276
3277 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3278 wrapper classes; (b) implement a new function in the C++ scanner that is
3279 more efficient than the old way of doing things because it avoids levels of
3280 recursion in the regex matching; (c) add a paragraph to the documentation
3281 for the FullMatch() function.
3282
3283 6. The escape sequence \n was being treated as whatever was defined as
3284 "newline". Not only was this contrary to the documentation, which states
3285 that \n is character 10 (hex 0A), but it also went horribly wrong when
3286 "newline" was defined as CRLF. This has been fixed.
3287
3288 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3289 was being set to -1 for the "end of line" case (supposedly a value that no
3290 character can have). Though this value is never used (the check for end of
3291 line is "zero bytes in current character"), it caused compiler complaints.
3292 I've changed it to 0xffffffff.
3293
3294 8. In pcre_version.c, the version string was being built by a sequence of
3295 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3296 string (as it is for production releases) called a macro with an empty
3297 argument. The C standard says the result of this is undefined. The gcc
3298 compiler treats it as an empty string (which was what was wanted) but it is
3299 reported that Visual C gives an error. The source has been hacked around to
3300 avoid this problem.
3301
3302 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3303 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3304 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3305 of them did).
3306
3307 10. Originally, pcretest opened its input and output without "b"; then I was
3308 told that "b" was needed in some environments, so it was added for release
3309 5.0 to both the input and output. (It makes no difference on Unix-like
3310 systems.) Later I was told that it is wrong for the input on Windows. I've
3311 now abstracted the modes into two macros, to make it easier to fiddle with
3312 them, and removed "b" from the input mode under Windows.
3313
3314 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3315
3316 12. Added -help and --help to pcretest as an official way of being reminded
3317 of the options.
3318
3319 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3320 and pcrecpp.cc because they annoy compilers at high warning levels.
3321
3322 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3323
3324 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3325 scripts are not C programs :-) and which was not noticed because it works
3326 on Linux.
3327
3328 16. pcretest is supposed to handle any length of pattern and data line (as one
3329 line or as a continued sequence of lines) by extending its input buffer if
3330 necessary. This feature was broken for very long pattern lines, leading to
3331 a string of junk being passed to pcre_compile() if the pattern was longer
3332 than about 50K.
3333
3334 17. I have done a major re-factoring of the way pcre_compile() computes the
3335 amount of memory needed for a compiled pattern. Previously, there was code
3336 that made a preliminary scan of the pattern in order to do this. That was
3337 OK when PCRE was new, but as the facilities have expanded, it has become
3338 harder and harder to keep it in step with the real compile phase, and there
3339 have been a number of bugs (see for example, 4 above). I have now found a
3340 cunning way of running the real compile function in a "fake" mode that
3341 enables it to compute how much memory it would need, while actually only
3342 ever using a few hundred bytes of working memory and without too many
3343 tests of the mode. This should make future maintenance and development
3344 easier. A side effect of this work is that the limit of 200 on the nesting
3345 depth of parentheses has been removed (though this was never a serious
3346 limitation, I suspect). However, there is a downside: pcre_compile() now
3347 runs more slowly than before (30% or more, depending on the pattern). I
3348 hope this isn't a big issue. There is no effect on runtime performance.
3349
3350 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3351 newline (only possible for the last line of a file) and it was a
3352 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3353
3354 19. Added additional timing features to pcretest. (1) The -tm option now times
3355 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3356 separate command line item, by a number that specifies the number of
3357 repeats to use when timing. The default is 50000; this gives better
3358 precision, but takes uncomfortably long for very large patterns.
3359
3360 20. Extended pcre_study() to be more clever in cases where a branch of a
3361 subpattern has no definite first character. For example, (a*|b*)[cd] would
3362 previously give no result from pcre_study(). Now it recognizes that the
3363 first character must be a, b, c, or d.
3364
3365 21. There was an incorrect error "recursive call could loop indefinitely" if
3366 a subpattern (or the entire pattern) that was being tested for matching an
3367 empty string contained only one non-empty item after a nested subpattern.
3368 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3369 incorrectly, because the \d was being skipped in the check.
3370
3371 22. The pcretest program now has a new pattern option /B and a command line
3372 option -b, which is equivalent to adding /B to every pattern. This causes
3373 it to show the compiled bytecode, without the additional information that
3374 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3375 is the same as /B/I).
3376
3377 23. A new optimization is now able automatically to treat some sequences such
3378 as a*b as a*+b. More specifically, if something simple (such as a character
3379 or a simple class like \d) has an unlimited quantifier, and is followed by
3380 something that cannot possibly match the quantified thing, the quantifier
3381 is automatically "possessified".
3382
3383 24. A recursive reference to a subpattern whose number was greater than 39
3384 went wrong under certain circumstances in UTF-8 mode. This bug could also
3385 have affected the operation of pcre_study().
3386
3387 25. Realized that a little bit of performance could be had by replacing
3388 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3389
3390 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3391
3392 27. Possessive quantifiers such as a++ were previously implemented by turning
3393 them into atomic groups such as ($>a+). Now they have their own opcodes,
3394 which improves performance. This includes the automatically created ones
3395 from 23 above.
3396
3397 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3398 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3399 the first matched character to be a colon. This applied both to named and
3400 numbered groups.
3401
3402 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3403
3404 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3405 building PCRE on Windows easier, so I have included it in the distribution.
3406
3407 31. There is now a check in pcretest against a ridiculously large number being
3408 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3409 loop, the loop is abandoned.
3410
3411 32. Forward references to subpatterns in conditions such as (?(2)...) where
3412 subpattern 2 is defined later cause pcre_compile() to search forwards in
3413 the pattern for the relevant set of parentheses. This search went wrong
3414 when there were unescaped parentheses in a character class, parentheses
3415 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3416
3417 33. "Subroutine" calls and backreferences were previously restricted to
3418 referencing subpatterns earlier in the regex. This restriction has now
3419 been removed.
3420
3421 34. Added a number of extra features that are going to be in Perl 5.10. On the
3422 whole, these are just syntactic alternatives for features that PCRE had
3423 previously implemented using the Python syntax or my own invention. The
3424 other formats are all retained for compatibility.
3425
3426 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3427 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3428 also .NET compatible.
3429
3430 (b) A recursion or subroutine call to a named group can now be defined as
3431 (?&name) as well as (?P>name).
3432
3433 (c) A backreference to a named group can now be defined as \k<name> or
3434 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3435 5.10, are also .NET compatible.
3436
3437 (d) A conditional reference to a named group can now use the syntax
3438 (?(<name>) or (?('name') as well as (?(name).
3439
3440 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3441 groups (named and numbered) that are never evaluated inline, but can be
3442 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3443 is always false. There may be only one alternative in such a group.
3444
3445 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3446 as the simple (?(R). The condition is true only if the most recent
3447 recursion is that of the given number or name. It does not search out
3448 through the entire recursion stack.
3449
3450 (g) The escape \gN or \g{N} has been added, where N is a positive or
3451 negative number, specifying an absolute or relative reference.
3452
3453 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3454 some "unreachable code" warnings.
3455
3456 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3457 things, this adds five new scripts.
3458
3459 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3460 There were also incompatibilities regarding the handling of \Q..\E inside
3461 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3462 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3463
3464 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3465 matches an empty string, and forcibly breaks the loop. There were bugs in
3466 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3467 against aaaa the result was just "a" rather than "aaaa", for example. Two
3468 separate and independent bugs (that affected different cases) have been
3469 fixed.
3470
3471 39. Refactored the code to abolish the use of different opcodes for small
3472 capturing bracket numbers. This is a tidy that I avoided doing when I
3473 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3474 The new approach is not only tidier, it makes it possible to reduce the
3475 memory needed to fix the previous bug (38).
3476
3477 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3478 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3479 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3480 mode.
3481
3482 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3483 report.
3484
3485 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3486 copy construction and assignment in the C++ wrapper.
3487
3488 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3489 couple of bugs that could have given wrong results in the "--newline=crlf"
3490 case.
3491
3492 44. Added a number of casts and did some reorganization of signed/unsigned int
3493 variables following suggestions from Dair Grant. Also renamed the variable
3494 "this" as "item" because it is a C++ keyword.
3495
3496 45. Arranged for dftables to add
3497
3498 #include "pcre_internal.h"
3499
3500 to pcre_chartables.c because without it, gcc 4.x may remove the array
3501 definition from the final binary if PCRE is built into a static library and
3502 dead code stripping is activated.
3503
3504 46. For an unanchored pattern, if a match attempt fails at the start of a
3505 newline sequence, and the newline setting is CRLF or ANY, and the next two
3506 characters are CRLF, advance by two characters instead of one.
3507
3508
3509 Version 6.7 04-Jul-06
3510 ---------------------
3511
3512 1. In order to handle tests when input lines are enormously long, pcretest has
3513 been re-factored so that it automatically extends its buffers when
3514 necessary. The code is crude, but this _is_ just a test program. The
3515 default size has been increased from 32K to 50K.
3516
3517 2. The code in pcre_study() was using the value of the re argument before
3518 testing it for NULL. (Of course, in any sensible call of the function, it
3519 won't be NULL.)
3520
3521 3. The memmove() emulation function in pcre_internal.h, which is used on
3522 systems that lack both memmove() and bcopy() - that is, hardly ever -
3523 was missing a "static" storage class specifier.
3524
3525 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3526 containing an extended class (one that cannot be represented by a bitmap
3527 because it contains high-valued characters or Unicode property items, e.g.
3528 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3529 pattern, but PCRE should not loop if you do not (it no longer does).
3530 [Detail: two cases were found: (a) a repeated subpattern containing an
3531 extended class; (b) a recursive reference to a subpattern that followed a
3532 previous extended class. It wasn't skipping over the extended class
3533 correctly when UTF-8 mode was not set.]
3534
3535 5. A negated single-character class was not being recognized as fixed-length
3536 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3537 compile error "lookbehind assertion is not fixed length".
3538
3539 6. The RunPerlTest auxiliary script was showing an unexpected difference
3540 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3541 write a Perl script that can interpret lines of an input file either as
3542 byte characters or as UTF-8, which is what "perltest" was being required to
3543 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3544 can't do is switch easily at run time between having the "use utf8;" pragma
3545 or not. In the end, I fudged it by using the RunPerlTest script to insert
3546 "use utf8;" explicitly for the UTF-8 tests.
3547
3548 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3549 the end of the subject string, contrary to the documentation and to what
3550 Perl does. This was true of both matching functions. Now it matches only at
3551 the start of the subject and immediately after *internal* newlines.
3552
3553 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3554 a pointer to an int instead of a pointer to an unsigned long int. This
3555 caused problems on 64-bit systems.
3556
3557 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3558 instance of the 'standard' template library not being so standard".
3559
3560 10. There was no check on the number of named subpatterns nor the maximum
3561 length of a subpattern name. The product of these values is used to compute
3562 the size of the memory block for a compiled pattern. By supplying a very
3563 long subpattern name and a large number of named subpatterns, the size
3564 computation could be caused to overflow. This is now prevented by limiting
3565 the length of names to 32 characters, and the number of named subpatterns
3566 to 10,000.
3567
3568 11. Subpatterns that are repeated with specific counts have to be replicated in
3569 the compiled pattern. The size of memory for this was computed from the
3570 length of the subpattern and the repeat count. The latter is limited to
3571 65535, but there was no limit on the former, meaning that integer overflow
3572 could in principle occur. The compiled length of a repeated subpattern is
3573 now limited to 30,000 bytes in order to prevent this.
3574
3575 12. Added the optional facility to have named substrings with the same name.
3576
3577 13. Added the ability to use a named substring as a condition, using the
3578 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3579 are numbers (not recommended). Forward references are permitted.
3580
3581 14. Added forward references in named backreferences (if you see what I mean).
3582
3583 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3584 pattern could run off the end of the subject. For example, the pattern
3585 "(?s)(.{1,5})"8 did this with the subject "ab".
3586
3587 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3588 PCRE_CASELESS was set when matching characters that were quantified with ?
3589 or *.
3590
3591 17. A character class other than a single negated character that had a minimum
3592 but no maximum quantifier - for example [ab]{6,} - was not handled
3593 correctly by pce_dfa_exec(). It would match only one character.
3594
3595 18. A valid (though odd) pattern that looked like a POSIX character
3596 class but used an invalid character after [ (for example [[,abc,]]) caused
3597 pcre_compile() to give the error "Failed: internal error: code overflow" or
3598 in some cases to crash with a glibc free() error. This could even happen if
3599 the pattern terminated after [[ but there just happened to be a sequence of
3600 letters, a binary zero, and a closing ] in the memory that followed.
3601
3602 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3603 over the years. Originally (before any Unicode support), just the bottom 8
3604 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3605 output from "man perlunicode" includes this:
3606
3607 The regular expression compiler produces polymorphic opcodes. That
3608 is, the pattern adapts to the data and automatically switches to
3609 the Unicode character scheme when presented with Unicode data--or
3610 instead uses a traditional byte scheme when presented with byte
3611 data.
3612
3613 Sadly, a wide octal escape does not cause a switch, and in a string with
3614 no other multibyte characters, these octal escapes are treated as before.
3615 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3616 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3617 Unicode string.
3618
3619 I have not perpetrated such confusion in PCRE. Up till now, it took just
3620 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3621 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3622 translate to the appropriate multibyte character.
3623
3624 29. Applied some refactoring to reduce the number of warnings from Microsoft
3625 and Borland compilers. This has included removing the fudge introduced
3626 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3627 a warning about an unused variable.
3628
3629 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3630 characters since release 4.0, because Perl (from release 5.004) does not.
3631 [Or at least, is documented not to: some releases seem to be in conflict
3632 with the documentation.] However, when a pattern was studied with
3633 pcre_study() and all its branches started with \s, PCRE still included VT
3634 as a possible starting character. Of course, this did no harm; it just
3635 caused an unnecessary match attempt.
3636
3637 22. Removed a now-redundant internal flag bit that recorded the fact that case
3638 dependency changed within the pattern. This was once needed for "required
3639 byte" processing, but is no longer used. This recovers a now-scarce options
3640 bit. Also moved the least significant internal flag bit to the most-
3641 significant bit of the word, which was not previously used (hangover from
3642 the days when it was an int rather than a uint) to free up another bit for
3643 the future.
3644
3645 23. Added support for CRLF line endings as well as CR and LF. As well as the
3646 default being selectable at build time, it can now be changed at runtime
3647 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3648 specify that it is scanning data with non-default line endings.
3649
3650 24. Changed the definition of CXXLINK to make it agree with the definition of
3651 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3652
3653 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3654 recursions. This makes a big different to stack usage for some patterns.
3655
3656 26. If a subpattern containing a named recursion or subroutine reference such
3657 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3658 the space required for the compiled pattern went wrong and gave too small a
3659 value. Depending on the environment, this could lead to "Failed: internal
3660 error: code overflow at offset 49" or "glibc detected double free or
3661 corruption" errors.
3662
3663 27. Applied patches from Google (a) to support the new newline modes and (b) to
3664 advance over multibyte UTF-8 characters in GlobalReplace.
3665
3666 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3667 difference for some implementation of PCRE in some Windows version.
3668
3669 29. Added some extra testing facilities to pcretest:
3670
3671 \q<number> in a data line sets the "match limit" value
3672 \Q<number> in a data line sets the "match recursion limt" value
3673 -S <number> sets the stack size, where <number> is in megabytes
3674
3675 The -S option isn't available for Windows.
3676
3677
3678 Version 6.6 06-Feb-06
3679 ---------------------
3680
3681 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3682 in pcreposix.h. I have copied the definition from pcre.h.
3683
3684 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3685 because pcre.h is no longer a built file.
3686
3687 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3688 not normally included in the compiled code.
3689
3690
3691 Version 6.5 01-Feb-06
3692 ---------------------
3693
3694 1. When using the partial match feature with pcre_dfa_exec(), it was not
3695 anchoring the second and subsequent partial matches at the new starting
3696 point. This could lead to incorrect results. For example, with the pattern
3697 /1234/, partially matching against "123" and then "a4" gave a match.
3698
3699 2. Changes to pcregrep:
3700
3701 (a) All non-match returns from pcre_exec() were being treated as failures
3702 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3703 error message is output. Some extra information is given for the
3704 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3705 probably the only errors that are likely to be caused by users (by
3706 specifying a regex that has nested indefinite repeats, for instance).
3707 If there are more than 20 of these errors, pcregrep is abandoned.
3708
3709 (b) A binary zero was treated as data while matching, but terminated the
3710 output line if it was written out. This has been fixed: binary zeroes
3711 are now no different to any other data bytes.
3712
3713 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3714 used to set a locale for matching. The --locale=xxxx long option has
3715 been added (no short equivalent) to specify a locale explicitly on the
3716 pcregrep command, overriding the environment variables.
3717
3718 (d) When -B was used with -n, some line numbers in the output were one less
3719 than they should have been.
3720
3721 (e) Added the -o (--only-matching) option.
3722
3723 (f) If -A or -C was used with -c (count only), some lines of context were
3724 accidentally printed for the final match.
3725
3726 (g) Added the -H (--with-filename) option.
3727
3728 (h) The combination of options -rh failed to suppress file names for files
3729 that were found from directory arguments.
3730
3731 (i) Added the -D (--devices) and -d (--directories) options.
3732
3733 (j) Added the -F (--fixed-strings) option.
3734
3735 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3736
3737 (l) Added the --colo(u)r option.
3738
3739 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3740 is not present by default.
3741
3742 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3743 items such as (?R) or (?1), when the recursion could match a number of
3744 alternatives. If it matched one of the alternatives, but subsequently,
3745 outside the recursion, there was a failure, the code tried to back up into
3746 the recursion. However, because of the way PCRE is implemented, this is not
3747 possible, and the result was an incorrect result from the match.
3748
3749 In order to prevent this happening, the specification of recursion has
3750 been changed so that all such subpatterns are automatically treated as
3751 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3752
3753 4. I had overlooked the fact that, in some locales, there are characters for
3754 which isalpha() is true but neither isupper() nor islower() are true. In
3755 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3756 and ordfeminine) are like this. This affected the treatment of \w and \W
3757 when they appeared in character classes, but not when they appeared outside
3758 a character class. The bit map for "word" characters is now created
3759 separately from the results of isalnum() instead of just taking it from the
3760 upper, lower, and digit maps. (Plus the underscore character, of course.)
3761
3762 5. The above bug also affected the handling of POSIX character classes such as
3763 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3764 permanent tables. Instead, the bit maps for such a class were previously
3765 created as the appropriate unions of the upper, lower, and digit bitmaps.
3766 Now they are created by subtraction from the [[:word:]] class, which has
3767 its own bitmap.
3768
3769 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3770 It is created by subtracting the vertical space characters (\x09, \x0a,
3771 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3772 subtraction was done in the overall bitmap for a character class, meaning
3773 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3774 be recognized. This bug has been fixed.
3775
3776 7. Patches from the folks at Google:
3777
3778 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3779 real life, but is still worth protecting against".
3780
3781 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3782 regular expressions".
3783
3784 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3785 have it.
3786
3787 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3788 "configure" and the latter not, in order to fix a problem somebody had
3789 with compiling the Arg class on HP-UX.
3790
3791 (e) Improve the error-handling of the C++ wrapper a little bit.
3792
3793 (f) New tests for checking recursion limiting.
3794
3795 8. The pcre_memmove() function, which is used only if the environment does not
3796 have a standard memmove() function (and is therefore rarely compiled),
3797 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3798 returning a result (though PCRE never actually uses the result).
3799
3800 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3801 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3802 returned instead of calling malloc() with an overflowing number that would
3803 most likely cause subsequent chaos.
3804
3805 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3806
3807 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3808 with this option is matched, the nmatch and pmatch options of regexec() are
3809 ignored.
3810
3811 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3812 provided in case anyone wants to the the POSIX interface with UTF-8
3813 strings.
3814
3815 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3816 C++ linking (needed for some HP-UX environments).
3817
3818 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3819 (unused parameter) and in the pcre_printint() function (omitted "default"
3820 switch label when the default is to do nothing).
3821
3822 15. Added some code to make it possible, when PCRE is compiled as a C++
3823 library, to replace subject pointers for pcre_exec() with a smart pointer
3824 class, thus making it possible to process discontinuous strings.
3825
3826 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3827 much the same function. They were added by different people who were trying
3828 to make PCRE easy to compile on non-Unix systems. It has been suggested
3829 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3830 for compiling on Windows systems. I have therefore replaced it with
3831 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3832 defaults to "extern" for C or "extern C" for C++, which works fine on
3833 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3834 SCOPE with something explicit in config.h. In addition:
3835
3836 (a) pcreposix.h still had just "extern" instead of either of these macros;
3837 I have replaced it with PCRE_DATA_SCOPE.
3838
3839 (b) Functions such as _pcre_xclass(), which are internal to the library,
3840 but external in the C sense, all had PCRE_EXPORT in their definitions.
3841 This is apparently wrong for the Windows case, so I have removed it.
3842 (It makes no difference on Unix-like systems.)
3843
3844 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3845 of recursive calls to match(). This is different to MATCH_LIMIT because
3846 that limits the total number of calls to match(), not all of which increase
3847 the depth of recursion. Limiting the recursion depth limits the amount of
3848 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3849 when PCRE is compiled, and changed at run time. A patch from Google adds
3850 this functionality to the C++ interface.
3851
3852 18. Changes to the handling of Unicode character properties:
3853
3854 (a) Updated the table to Unicode 4.1.0.
3855
3856 (b) Recognize characters that are not in the table as "Cn" (undefined).
3857
3858 (c) I revised the way the table is implemented to a much improved format
3859 which includes recognition of ranges. It now supports the ranges that
3860 are defined in UnicodeData.txt, and it also amalgamates other
3861 characters into ranges. This has reduced the number of entries in the
3862 table from around 16,000 to around 3,000, thus reducing its size
3863 considerably. I realized I did not need to use a tree structure after
3864 all - a binary chop search is just as efficient. Having reduced the
3865 number of entries, I extended their size from 6 bytes to 8 bytes to
3866 allow for more data.
3867
3868 (d) Added support for Unicode script names via properties such as \p{Han}.
3869
3870 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3871 matching that character.
3872
3873 20. When matching a repeated Unicode property with a minimum greater than zero,
3874 (for example \pL{2,}), PCRE could look past the end of the subject if it
3875 reached it while seeking the minimum number of characters. This could
3876 happen only if some of the characters were more than one byte long, because
3877 there is a check for at least the minimum number of bytes.
3878
3879 21. Refactored the implementation of \p and \P so as to be more general, to
3880 allow for more different types of property in future. This has changed the
3881 compiled form incompatibly. Anybody with saved compiled patterns that use
3882 \p or \P will have to recompile them.
3883
3884 22. Added "Any" and "L&" to the supported property types.
3885
3886 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3887 but give a compile time error if the value is greater than 0xff.
3888
3889 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3890 accidentally not being installed or uninstalled.
3891
3892 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3893 made were to insert the current release number. This seemed silly, because
3894 it made things harder for people building PCRE on systems that don't run
3895 "configure". I have turned pcre.h into a distributed file, no longer built
3896 by "configure", with the version identification directly included. There is
3897 no longer a pcre.h.in file.
3898
3899 However, this change necessitated a change to the pcre-config script as
3900 well. It is built from pcre-config.in, and one of the substitutions was the
3901 release number. I have updated configure.ac so that ./configure now finds
3902 the release number by grepping pcre.h.
3903
3904 26. Added the ability to run the tests under valgrind.
3905
3906
3907 Version 6.4 05-Sep-05
3908 ---------------------
3909
3910 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3911 "--" to be printed when multiple files were scanned, even when none of the
3912 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3913 consider it to be a bug, and have restored the previous behaviour.
3914
3915 2. A couple of code tidies to get rid of compiler warnings.
3916
3917 3. The pcretest program used to cheat by referring to symbols in the library
3918 whose names begin with _pcre_. These are internal symbols that are not
3919 really supposed to be visible externally, and in some environments it is
3920 possible to suppress them. The cheating is now confined to including
3921 certain files from the library's source, which is a bit cleaner.
3922
3923 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3924 file's purpose clearer.
3925
3926 5. Reorganized pcre_ucp_findchar().
3927
3928
3929 Version 6.3 15-Aug-05
3930 ---------------------
3931
3932 1. The file libpcre.pc.in did not have general read permission in the tarball.
3933
3934 2. There were some problems when building without C++ support:
3935
3936 (a) If C++ support was not built, "make install" and "make test" still
3937 tried to test it.
3938
3939 (b) There were problems when the value of CXX was explicitly set. Some
3940 changes have been made to try to fix these, and ...
3941
3942 (c) --disable-cpp can now be used to explicitly disable C++ support.
3943
3944 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3945 backslash in a target when C++ was disabled. This confuses some
3946 versions of "make", apparently. Using an intermediate variable solves
3947 this. (Same for CPP_LOBJ.)
3948
3949 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3950 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3951 necessary on certain architectures.
3952
3953 4. Added a setting of -export-symbols-regex to the link command to remove
3954 those symbols that are exported in the C sense, but actually are local
3955 within the library, and not documented. Their names all begin with
3956 "_pcre_". This is not a perfect job, because (a) we have to except some
3957 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3958 available (and never for static libraries). I have made a note to try to
3959 find a way round (a) in the future.
3960
3961
3962 Version 6.2 01-Aug-05
3963 ---------------------
3964
3965 1. There was no test for integer overflow of quantifier values. A construction
3966 such as {1111111111111111} would give undefined results. What is worse, if
3967 a minimum quantifier for a parenthesized subpattern overflowed and became
3968 negative, the calculation of the memory size went wrong. This could have
3969 led to memory overwriting.
3970
3971 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3972
3973 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3974 operating environments where this matters.
3975
3976 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3977 PCRE options from within the C++ wrapper.
3978
3979 5. Named capturing subpatterns were not being correctly counted when a pattern
3980 was compiled. This caused two problems: (a) If there were more than 100
3981 such subpatterns, the calculation of the memory needed for the whole
3982 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3983 back references of the form \12, where the number was greater than 9, were
3984 not recognized as back references, even though there were sufficient
3985 previous subpatterns.
3986
3987 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3988 versions of gcc, e.g. 2.95.4.
3989
3990
3991 Version 6.1 21-Jun-05
3992 ---------------------
3993
3994 1. There was one reference to the variable "posix" in pcretest.c that was not
3995 surrounded by "#if !defined NOPOSIX".
3996
3997 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3998 the cross-check on the old pcre_info() function, for the benefit of the
3999 cut-down version of PCRE that is currently imported into Exim.
4000
4001 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4002 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4003 patterns that start (?i) but I don't think that matters. The use of (?i) is
4004 just an example; this all applies to the other options as well.
4005
4006 4. Since libtool seems to echo the compile commands it is issuing, the output
4007 from "make" can be reduced a bit by putting "@" in front of each libtool
4008 compile command.
4009
4010 5. Patch from the folks at Google for configure.in to be a bit more thorough
4011 in checking for a suitable C++ installation before trying to compile the
4012 C++ stuff. This should fix a reported problem when a compiler was present,
4013 but no suitable headers.
4014
4015 6. The man pages all had just "PCRE" as their title. I have changed them to
4016 be the relevant file name. I have also arranged that these names are
4017 retained in the file doc/pcre.txt, which is a concatenation in text format
4018 of all the man pages except the little individual ones for each function.
4019
4020 7. The NON-UNIX-USE file had not been updated for the different set of source
4021 files that come with release 6. I also added a few comments about the C++
4022 wrapper.
4023
4024
4025 Version 6.0 07-Jun-05
4026 ---------------------
4027
4028 1. Some minor internal re-organization to help with my DFA experiments.
4029
4030 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4031 didn't matter for the library itself when fully configured, but did matter
4032 when compiling without UCP support, or within Exim, where the ucp files are
4033 not imported.
4034
4035 3. Refactoring of the library code to split up the various functions into
4036 different source modules. The addition of the new DFA matching code (see
4037 below) to a single monolithic source would have made it really too
4038 unwieldy, quite apart from causing all the code to be include in a
4039 statically linked application, when only some functions are used. This is
4040 relevant even without the DFA addition now that patterns can be compiled in
4041 one application and matched in another.
4042
4043 The downside of splitting up is that there have to be some external
4044 functions and data tables that are used internally in different modules of
4045 the library but which are not part of the API. These have all had their
4046 names changed to start with "_pcre_" so that they are unlikely to clash
4047 with other external names.
4048
4049 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4050 a different (DFA) algorithm. Although it is slower than the original
4051 function, it does have some advantages for certain types of matching
4052 problem.
4053
4054 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4055 including restarting after a partial match.
4056
4057 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4058 defined when compiling for Windows was sent to me. I have put it into the
4059 code, though I have no means of testing or verifying it.
4060
4061 7. Added the pcre_refcount() auxiliary function.
4062
4063 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4064 match before or at the first newline in the subject string. In pcretest,
4065 the /f option on a pattern can be used to set this.
4066
4067 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4068 would behave wrongly. This has been present in PCRE since release 4.0.
4069
4070 10. A number of changes to the pcregrep command:
4071
4072 (a) Refactored how -x works; insert ^(...)$ instead of setting
4073 PCRE_ANCHORED and checking the length, in preparation for adding
4074 something similar for -w.
4075
4076 (b) Added the -w (match as a word) option.
4077
4078 (c) Refactored the way lines are read and buffered so as to have more
4079 than one at a time available.
4080
4081 (d) Implemented a pcregrep test script.
4082
4083 (e) Added the -M (multiline match) option. This allows patterns to match
4084 over several lines of the subject. The buffering ensures that at least
4085 8K, or the rest of the document (whichever is the shorter) is available
4086 for matching (and similarly the previous 8K for lookbehind assertions).
4087
4088 (f) Changed the --help output so that it now says
4089
4090 -w, --word-regex(p)
4091
4092 instead of two lines, one with "regex" and the other with "regexp"
4093 because that confused at least one person since the short forms are the
4094 same. (This required a bit of code, as the output is generated
4095 automatically from a table. It wasn't just a text change.)
4096
4097 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4098 option but starts with a hyphen. Could be a pattern or a path name
4099 starting with a hyphen, for instance.
4100
4101 (h) "-" can be given as a file name to represent stdin.
4102
4103 (i) When file names are being printed, "(standard input)" is used for
4104 the standard input, for compatibility with GNU grep. Previously
4105 "<stdin>" was used.
4106
4107 (j) The option --label=xxx can be used to supply a name to be used for
4108 stdin when file names are being printed. There is no short form.
4109
4110 (k) Re-factored the options decoding logic because we are going to add
4111 two more options that take data. Such options can now be given in four
4112 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4113
4114 (l) Added the -A, -B, and -C options for requesting that lines of context
4115 around matches be printed.
4116
4117 (m) Added the -L option to print the names of files that do not contain
4118 any matching lines, that is, the complement of -l.
4119
4120 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4121 continue to scan other files.
4122
4123 (o) The -s option was incorrectly implemented. For compatibility with other
4124 greps, it now suppresses the error message for a non-existent or non-
4125 accessible file (but not the return code). There is a new option called
4126 -q that suppresses the output of matching lines, which was what -s was
4127 previously doing.
4128
4129 (p) Added --include and --exclude options to specify files for inclusion
4130 and exclusion when recursing.
4131
4132 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4133 Hopefully, it now does.
4134
4135 12. Missing cast in pcre_study().
4136
4137 13. Added an "uninstall" target to the makefile.
4138
4139 14. Replaced "extern" in the function prototypes in Makefile.in with
4140 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4141 world, but is set differently for Windows.
4142
4143 15. Added a second compiling function called pcre_compile2(). The only
4144 difference is that it has an extra argument, which is a pointer to an
4145 integer error code. When there is a compile-time failure, this is set
4146 non-zero, in addition to the error test pointer being set to point to an
4147 error message. The new argument may be NULL if no error number is required
4148 (but then you may as well call pcre_compile(), which is now just a
4149 wrapper). This facility is provided because some applications need a
4150 numeric error indication, but it has also enabled me to tidy up the way
4151 compile-time errors are handled in the POSIX wrapper.
4152
4153 16. Added VPATH=.libs to the makefile; this should help when building with one
4154 prefix path and installing with another. (Or so I'm told by someone who
4155 knows more about this stuff than I do.)
4156
4157 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4158 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4159 match everything, including newlines. This is not POSIX-compatible, but
4160 somebody wanted the feature. From pcretest it can be activated by using
4161 both the P and the s flags.
4162
4163 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4164
4165 19. libpcre.pc was being incorrectly installed as executable.
4166
4167 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4168 it now also looks for '\r' so that it will work unmodified on Windows.
4169
4170 21. Added Google's contributed C++ wrapper to the distribution.
4171
4172 22. Added some untidy missing memory free() calls in pcretest, to keep
4173 Electric Fence happy when testing.
4174
4175
4176
4177 Version 5.0 13-Sep-04
4178 ---------------------
4179
4180 1. Internal change: literal characters are no longer packed up into items
4181 containing multiple characters in a single byte-string. Each character
4182 is now matched using a separate opcode. However, there may be more than one
4183 byte in the character in UTF-8 mode.
4184
4185 2. The pcre_callout_block structure has two new fields: pattern_position and
4186 next_item_length. These contain the offset in the pattern to the next match
4187 item, and its length, respectively.
4188
4189 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4190 insertion of callouts before each pattern item. Added the /C option to
4191 pcretest to make use of this.
4192
4193 4. On the advice of a Windows user, the lines
4194
4195 #if defined(_WIN32) || defined(WIN32)
4196 _setmode( _fileno( stdout ), 0x8000 );
4197 #endif /* defined(_WIN32) || defined(WIN32) */
4198
4199 have been added to the source of pcretest. This apparently does useful
4200 magic in relation to line terminators.
4201
4202 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4203 for the benefit of those environments where the "b" makes a difference.
4204
4205 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4206 to know about it. I have put a hack into configure.in that adds in code
4207 to set GCC=yes if CC=icc. This seems to end up at a point in the
4208 generated configure script that is early enough to affect the setting of
4209 compiler options, which is what is needed, but I have no means of testing
4210 whether it really works. (The user who reported this had patched the
4211 generated configure script, which of course I cannot do.)
4212
4213 LATER: After change 22 below (new libtool files), the configure script
4214 seems to know about icc (and also ecc). Therefore, I have commented out
4215 this hack in configure.in.
4216
4217 7. Added support for pkg-config (2 patches were sent in).
4218
4219 8. Negated POSIX character classes that used a combination of internal tables
4220 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4221 [[:^ascii]]. Typically, they would match almost any characters. The other
4222 POSIX classes were not broken in this way.
4223
4224 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4225 to find the match, as PCRE was deluded into thinking that the match had to
4226 start at the start point or following a newline. The same bug applied to
4227 patterns with negative forward assertions or any backward assertions
4228 preceding ".*" at the start, unless the pattern required a fixed first
4229 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4230
4231 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4232 starting at the last subject character, bytes beyond the end of the subject
4233 string were read.
4234
4235 11. Renamed the variable "class" as "classbits" to make life easier for C++
4236 users. (Previously there was a macro definition, but it apparently wasn't
4237 enough.)
4238
4239 12. Added the new field "tables" to the extra data so that tables can be passed
4240 in at exec time, or the internal tables can be re-selected. This allows
4241 a compiled regex to be saved and re-used at a later time by a different
4242 program that might have everything at different addresses.
4243
4244 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4245 -R library as well as a -L library.
4246
4247 14. The debugging options of pcretest (-d on the command line or D on a
4248 pattern) showed incorrect output for anything following an extended class
4249 that contained multibyte characters and which was followed by a quantifier.
4250
4251 15. Added optional support for general category Unicode character properties
4252 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4253 support. It adds about 90K to the size of the library. The meanings of the
4254 inbuilt class escapes such as \d and \s have NOT been changed.
4255
4256 16. Updated pcredemo.c to include calls to free() to release the memory for the
4257 compiled pattern.
4258
4259 17. The generated file chartables.c was being created in the source directory
4260 instead of in the building directory. This caused the build to fail if the
4261 source directory was different from the building directory, and was
4262 read-only.
4263
4264 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4265 file. No doubt somebody will tell me if they don't make sense... Also added
4266 Dan Mooney's comments about building on OpenVMS.
4267
4268 19. Added support for partial matching via the PCRE_PARTIAL option for
4269 pcre_exec() and the \P data escape in pcretest.
4270
4271 20. Extended pcretest with 3 new pattern features:
4272
4273 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4274 write the compiled pattern to the file whose name is "rest-of-line".
4275 This is a straight binary dump of the data, with the saved pointer to
4276 the character tables forced to be NULL. The study data, if any, is
4277 written too. After writing, pcretest reads a new pattern.
4278
4279 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4280 compiled pattern from the given file. There must not be any
4281 occurrences of "<" in the file name (pretty unlikely); if there are,
4282 pcretest will instead treat the initial "<" as a pattern delimiter.
4283 After reading in the pattern, pcretest goes on to read data lines as
4284 usual.
4285
4286 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4287 and 16-bit fields in a compiled pattern, to simulate a pattern that
4288 was compiled on a host of opposite endianness.
4289
4290 21. The pcre-exec() function can now cope with patterns that were compiled on
4291 hosts of opposite endianness, with this restriction:
4292
4293 As for any compiled expression that is saved and used later, the tables
4294 pointer field cannot be preserved; the extra_data field in the arguments
4295 to pcre_exec() should be used to pass in a tables address if a value
4296 other than the default internal tables were used at compile time.
4297
4298 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4299 now diagnosed as an error. Previously, most of the time, a negative number
4300 would have been treated as zero, but if in addition "ovector" was passed as
4301 NULL, a crash could occur.
4302
4303 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4304 new versions from the libtool 1.5 distribution (the last one is a copy of
4305 a file called libtool.m4). This seems to have fixed the need to patch
4306 "configure" to support Darwin 1.3 (which I used to do). However, I still
4307 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4308 workstation).
4309
4310 24. Changed the PCRE licence to be the more standard "BSD" licence.
4311
4312
4313 Version 4.5 01-Dec-03
4314 ---------------------
4315
4316 1. There has been some re-arrangement of the code for the match() function so
4317 that it can be compiled in a version that does not call itself recursively.
4318 Instead, it keeps those local variables that need separate instances for
4319 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4320 needs to "recurse". Keeping track of where control must go is done by means
4321 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4322 hide most of the details from the main code, and operates only if
4323 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4324 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4325 operating.
4326
4327 To make it easier for callers to provide specially tailored get/free
4328 functions for this usage, two new functions, pcre_stack_malloc, and
4329 pcre_stack_free, are used. They are always called in strict stacking order,
4330 and the size of block requested is always the same.
4331
4332 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4333 PCRE has been compiled to use the stack or the heap for recursion. The
4334 -C option of pcretest uses this to show which version is compiled.
4335
4336 A new data escape \S, is added to pcretest; it causes the amounts of store
4337 obtained and freed by both kinds of malloc/free at match time to be added
4338 to the output.
4339
4340 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4341 what's available on my current Linux desktop machine.
4342
4343 3. When matching a UTF-8 string, the test for a valid string at the start has
4344 been extended. If start_offset is not zero, PCRE now checks that it points
4345 to a byte that is the start of a UTF-8 character. If not, it returns
4346 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4347 this is necessary because there may be backward assertions in the pattern.
4348 When matching the same subject several times, it may save resources to use
4349 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4350
4351 4. The code for checking the validity of UTF-8 strings has been tightened so
4352 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4353 containing "overlong sequences".
4354
4355 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4356 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4357 should have been "&", but it just so happened that all the cases this let
4358 through by mistake were picked up later in the function.
4359
4360 6. I had used a variable called "isblank" - this is a C99 function, causing
4361 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4362
4363 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4364 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4365 so the version is shown; (c) stop "make test" echoing "./RunTest".
4366
4367 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4368
4369 9. The prototype for memmove() for systems that don't have it was using
4370 size_t, but the inclusion of the header that defines size_t was later. I've
4371 moved the #includes for the C headers earlier to avoid this.
4372
4373 10. Added some adjustments to the code to make it easier to compiler on certain
4374 special systems:
4375
4376 (a) Some "const" qualifiers were missing.
4377 (b) Added the macro EXPORT before all exported functions; by default this
4378 is defined to be empty.
4379 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4380 that it reads its output file name as an argument instead of writing
4381 to the standard output and assuming this can be redirected.
4382
4383 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4384 class containing characters with values greater than 255, PCRE compilation
4385 went into a loop.
4386
4387 12. A recursive reference to a subpattern that was within another subpattern
4388 that had a minimum quantifier of zero caused PCRE to crash. For example,
4389 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4390 recursion. If the recursively-called subpattern itself had a zero repeat,
4391 that was OK.
4392
4393 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4394 buffer into which it was copied (for escape processing) was still set at
4395 1024, so long lines caused crashes.
4396
4397 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4398 "internal error: code overflow...". This applied to any character class
4399 that was followed by a possessive quantifier.
4400
4401 15. Modified the Makefile to add libpcre.la as a prerequisite for
4402 libpcreposix.la because I was told this is needed for a parallel build to
4403 work.
4404
4405 16. If a pattern that contained .* following optional items at the start was
4406 studied, the wrong optimizing data was generated, leading to matching
4407 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4408 matching string must start with a or b or c. The correct conclusion for
4409 this pattern is that a match can start with any character.
4410
4411
4412 Version 4.4 13-Aug-03
4413 ---------------------
4414
4415 1. In UTF-8 mode, a character class containing characters with values between
4416 127 and 255 was not handled correctly if the compiled pattern was studied.
4417 In fixing this, I have also improved the studying algorithm for such
4418 classes (slightly).
4419
4420 2. Three internal functions had redundant arguments passed to them. Removal
4421 might give a very teeny performance improvement.
4422
4423 3. Documentation bug: the value of the capture_top field in a callout is *one
4424 more than* the number of the hightest numbered captured substring.
4425
4426 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4427 in incorrectly linking with a previously installed version. They now link
4428 explicitly with libpcre.la.
4429
4430 5. configure.in no longer needs to recognize Cygwin specially.
4431
4432 6. A problem in pcre.in for Windows platforms is fixed.
4433
4434 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4435 pcretest, it used to include the size of the study block as part of its
4436 output. Unfortunately, the structure contains a field that has a different
4437 size on different hardware architectures. This meant that the tests that
4438 showed this size failed. As the block is currently always of a fixed size,
4439 this information isn't actually particularly useful in pcretest output, so
4440 I have just removed it.
4441
4442 8. Three pre-processor statements accidentally did not start in column 1.
4443 Sadly, there are *still* compilers around that complain, even though
4444 standard C has not required this for well over a decade. Sigh.
4445
4446 9. In pcretest, the code for checking callouts passed small integers in the
4447 callout_data field, which is a void * field. However, some picky compilers
4448 complained about the casts involved for this on 64-bit systems. Now
4449 pcretest passes the address of the small integer instead, which should get
4450 rid of the warnings.
4451
4452 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4453 both compile and run time, and gives an error if an invalid UTF-8 sequence
4454 is found. There is a option for disabling this check in cases where the
4455 string is known to be correct and/or the maximum performance is wanted.
4456
4457 11. In response to a bug report, I changed one line in Makefile.in from
4458
4459 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4460 to
4461 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4462
4463 to look similar to other lines, but I have no way of telling whether this
4464 is the right thing to do, as I do not use Windows. No doubt I'll get told
4465 if it's wrong...
4466
4467
4468 Version 4.3 21-May-03
4469 ---------------------
4470
4471 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4472 Makefile.
4473
4474 2. Some refactoring to improve the quality of the code:
4475
4476 (i) The utf8_table... variables are now declared "const".
4477
4478 (ii) The code for \cx, which used the "case flipping" table to upper case
4479 lower case letters, now just substracts 32. This is ASCII-specific,
4480 but the whole concept of \cx is ASCII-specific, so it seems
4481 reasonable.
4482
4483 (iii) PCRE was using its character types table to recognize decimal and
4484 hexadecimal digits in the pattern. This is silly, because it handles
4485 only 0-9, a-f, and A-F, but the character types table is locale-
4486 specific, which means strange things might happen. A private
4487 table is now used for this - though it costs 256 bytes, a table is
4488 much faster than multiple explicit tests. Of course, the standard
4489 character types table is still used for matching digits in subject
4490 strings against \d.
4491
4492 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4493 ending in _t are). So I've renamed it as ESC_tee.
4494
4495 3. The first argument for regexec() in the POSIX wrapper should have been
4496 defined as "const".
4497
4498 4. Changed pcretest to use malloc() for its buffers so that they can be
4499 Electric Fenced for debugging.
4500
4501 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4502 to read one or more bytes before the start of the subject string. Often this
4503 had no effect on PCRE's behaviour, but in some circumstances it could
4504 provoke a segmentation fault.
4505
4506 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4507 to try to read one or more bytes before the start of the subject string.
4508
4509 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4510 UTF-8 support could misbehave in various ways if the subject string
4511 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4512 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4513 back over UTF-8 characters.)
4514
4515
4516 Version 4.2 14-Apr-03
4517 ---------------------
4518
4519 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4520
4521 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4522 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4523 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4524 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4525 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4526 and BUILD_EXEEXT
4527 Note: automatic setting of the BUILD variables is not yet working
4528 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4529 compile-time but not at link-time
4530 [LINK]: use for linking executables only
4531 make different versions for Windows and non-Windows
4532 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4533 libraries
4534 [LINK_FOR_BUILD]: new variable
4535 [OBJEXT]: use throughout
4536 [EXEEXT]: use throughout
4537 <winshared>: new target
4538 <wininstall>: new target
4539 <dftables.o>: use native compiler
4540 <dftables>: use native linker
4541 <install>: handle Windows platform correctly
4542 <clean>: ditto
4543 <check>: ditto
4544 copy DLL to top builddir before testing
4545
4546 As part of these changes, -no-undefined was removed again. This was reported
4547 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4548 in any case.
4549
4550 3. Some tidies to get rid of compiler warnings:
4551
4552 . In the match_data structure, match_limit was an unsigned long int, whereas
4553 match_call_count was an int. I've made them both unsigned long ints.
4554
4555 . In pcretest the fact that a const uschar * doesn't automatically cast to
4556 a void * provoked a warning.
4557
4558 . Turning on some more compiler warnings threw up some "shadow" variables
4559 and a few more missing casts.
4560
4561 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4562 option, a class that contained a single character with a value between 128
4563 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4564
4565 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4566 option, a class that contained several characters, but with at least one
4567 whose value was between 128 and 255 caused PCRE to crash.
4568
4569
4570 Version 4.1 12-Mar-03
4571 ---------------------
4572
4573 1. Compiling with gcc -pedantic found a couple of places where casts were
4574 needed, and a string in dftables.c that was longer than standard compilers are
4575 required to support.
4576
4577 2. Compiling with Sun's compiler found a few more places where the code could
4578 be tidied up in order to avoid warnings.
4579
4580 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4581 first of these names is deprecated in the latest Autoconf in favour of the name
4582 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4583 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4584 analogy I have changed it to CFLAGS_FOR_BUILD.
4585
4586 4. Added -no-undefined to the linking command in the Makefile, because this is
4587 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4588 linking step for the pcreposix library.
4589
4590 5. PCRE was failing to diagnose the case of two named groups with the same
4591 name.
4592
4593 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4594 literal character that is needed in the subject for a match, and scans along to
4595 ensure that it is present before embarking on the full matching process. This
4596 saves time in cases of nested unlimited repeats that are never going to match.
4597 Problem: the scan can take a lot of time if the subject is very long (e.g.
4598 megabytes), thus penalizing straightforward matches. It is now done only if the
4599 amount of subject to be scanned is less than 1000 bytes.
4600
4601 7. A lesser problem with the same optimization is that it was recording the
4602 first character of an anchored pattern as "needed", thus provoking a search
4603 right along the subject, even when the first match of the pattern was going to
4604 fail. The "needed" character is now not set for anchored patterns, unless it
4605 follows something in the pattern that is of non-fixed length. Thus, it still
4606 fulfils its original purpose of finding quick non-matches in cases of nested
4607 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4608
4609
4610 Version 4.0 17-Feb-03
4611 ---------------------
4612
4613 1. If a comment in an extended regex that started immediately after a meta-item
4614 extended to the end of string, PCRE compiled incorrect data. This could lead to
4615 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4616
4617 2. Moved to autoconf 2.53 and libtool 1.4.2.
4618
4619 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4620 the special perltest8 script is no longer needed - all the tests can be run
4621 from a single perltest script.
4622
4623 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4624 by \s. It has now been removed in PCRE. This means it isn't recognized as
4625 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4626 class [:space:] *does* include VT, thereby creating a mess.
4627
4628 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4629 space and tab.
4630
4631 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4632 its new features into the main test script, reducing the number of scripts.
4633
4634 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4635 were backward compatible, and made the (?i) apply to the whole pattern, as if
4636 /i were given. Now it behaves more logically, and applies the option setting
4637 only to what follows. PCRE has been changed to follow suit. However, if it
4638 finds options settings right at the start of the pattern, it extracts them into
4639 the global options, as before. Thus, they show up in the info data.
4640
4641 8. Added support for the \Q...\E escape sequence. Characters in between are
4642 treated as literals. This is slightly different from Perl in that $ and @ are
4643 also handled as literals inside the quotes. In Perl, they will cause variable
4644 interpolation. Note the following examples:
4645
4646 Pattern PCRE matches Perl matches
4647
4648 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4649 \Qabc\$xyz\E abc\$xyz abc\$xyz
4650 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4651
4652 For compatibility with Perl, \Q...\E sequences are recognized inside character
4653 classes as well as outside them.
4654
4655 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4656 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4657 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4658 signed/unsigned warnings.
4659
4660 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4661 option for pcretest, so I've replaced it by a simple function that does just
4662 that job.
4663
4664 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4665 "pcregrep -".
4666
4667 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4668 Java package. This provides some syntactic sugar for simple cases of what my
4669 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4670 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4671 item, you can use this simplified notation. Note that only makes sense with
4672 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4673 greediness, whatever the setting of the PCRE_UNGREEDY option.
4674
4675 13. A change of greediness default within a pattern was not taking effect at
4676 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4677 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4678 was abstracted outside.
4679
4680 14. PCRE now supports the \G assertion. It is true when the current matching
4681 position is at the start point of the match. This differs from \A when the
4682 starting offset is non-zero. Used with the /g option of pcretest (or similar
4683 code), it works in the same way as it does for Perl's /g option. If all
4684 alternatives of a regex begin with \G, the expression is anchored to the start
4685 match position, and the "anchored" flag is set in the compiled expression.
4686
4687 15. Some bugs concerning the handling of certain option changes within patterns
4688 have been fixed. These applied to options other than (?ims). For example,
4689 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4690 been the other way round. Some of this was related to change 7 above.
4691
4692 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4693 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4694 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4695 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4696
4697 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4698 mode. Unlike ".", it always matches newline, whatever the setting of
4699 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4700 assertions. Perl allows it, but it doesn't (in general) work because it can't
4701 calculate the length of the lookbehind. At least, that's the case for Perl
4702 5.8.0 - I've been told they are going to document that it doesn't work in
4703 future.
4704
4705 18. Added an error diagnosis for escapes that PCRE does not support: these are
4706 \L, \l, \N, \P, \p, \U, \u, and \X.
4707
4708 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4709 reading past the end of the pattern in cases such as /[abcd/.
4710
4711 20. PCRE was getting more memory than necessary for patterns with classes that
4712 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4713
4714 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4715 compiling PCRE for use with Virtual Pascal.
4716
4717 22. Small fix to the Makefile to make it work properly if the build is done
4718 outside the source tree.
4719
4720 23. Added a new extension: a condition to go with recursion. If a conditional
4721 subpattern starts with (?(R) the "true" branch is used if recursion has
4722 happened, whereas the "false" branch is used only at the top level.
4723
4724 24. When there was a very long string of literal characters (over 255 bytes
4725 without UTF support, over 250 bytes with UTF support), the computation of how
4726 much memory was required could be incorrect, leading to segfaults or other
4727 strange effects.
4728
4729 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4730 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4731 there was a subsequent back reference to those brackets. This meant that, for
4732 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4733 possible to check for precisely this case. All we can do is abandon the
4734 optimization if .* occurs inside capturing brackets when there are any back
4735 references whatsoever. (See below for a better fix that came later.)
4736
4737 26. The handling of the optimization for finding the first character of a
4738 non-anchored pattern, and for finding a character that is required later in the
4739 match were failing in some cases. This didn't break the matching; it just
4740 failed to optimize when it could. The way this is done has been re-implemented.
4741
4742 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4743
4744 28. Added a new feature that provides some of the functionality that Perl
4745 provides with (?{...}). The facility is termed a "callout". The way it is done
4746 in PCRE is for the caller to provide an optional function, by setting
4747 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4748 global variable. By default it is unset, which disables all calling out. To get
4749 the function called, the regex must include (?C) at appropriate points. This
4750 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4751 This provides a means of identifying different callout points. When PCRE
4752 reaches such a point in the regex, if pcre_callout has been set, the external
4753 function is called. It is provided with data in a structure called
4754 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4755 matching continues; if it returns a non-zero value, the match at the current
4756 point fails. However, backtracking will occur if possible. [This was changed
4757 later and other features added - see item 49 below.]
4758
4759 29. pcretest is upgraded to test the callout functionality. It provides a
4760 callout function that displays information. By default, it shows the start of
4761 the match and the current position in the text. There are some new data escapes
4762 to vary what happens:
4763
4764 \C+ in addition, show current contents of captured substrings
4765 \C- do not supply a callout function
4766 \C!n return 1 when callout number n is reached
4767 \C!n!m return 1 when callout number n is reached for the mth time
4768
4769 30. If pcregrep was called with the -l option and just a single file name, it
4770 output "<stdin>" if a match was found, instead of the file name.
4771
4772 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4773 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4774 pcre_exec(). This saves a malloc/free per call. The default value of
4775 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4776 when configuring.
4777
4778 32. The default maximum size of a compiled pattern is 64K. There have been a
4779 few cases of people hitting this limit. The code now uses macros to handle the
4780 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4781 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4782 configuring. Tests 2 and 5 work only with 2-byte links because they output
4783 debugging information about compiled patterns.
4784
4785 33. Internal code re-arrangements:
4786
4787 (a) Moved the debugging function for printing out a compiled regex into
4788 its own source file (printint.c) and used #include to pull it into
4789 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4790 separate copies.
4791
4792 (b) Defined the list of op-code names for debugging as a macro in
4793 internal.h so that it is next to the definition of the opcodes.
4794
4795 (c) Defined a table of op-code lengths for simpler skipping along compiled
4796 code. This is again a macro in internal.h so that it is next to the
4797 definition of the opcodes.
4798
4799 34. Added support for recursive calls to individual subpatterns, along the
4800 lines of Robin Houston's patch (but implemented somewhat differently).
4801
4802 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4803 allow it to read and process whole directories in Win32. This code was
4804 contributed by Lionel Fourquaux; it has not been tested by me.
4805
4806 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4807 used to name a group. Names consist of alphanumerics and underscores, and must
4808 be unique. Back references use the syntax (?P=name) and recursive calls use
4809 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4810 numbers. The function pcre_fullinfo() can be used after compilation to extract
4811 a name/number map. There are three relevant calls:
4812
4813 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4814 PCRE_INFO_NAMECOUNT yields the number of entries
4815 PCRE_INFO_NAMETABLE yields a pointer to the map.
4816
4817 The map is a vector of fixed-size entries. The size of each entry depends on
4818 the length of the longest name used. The first two bytes of each entry are the
4819 group number, most significant byte first. There follows the corresponding
4820 name, zero terminated. The names are in alphabetical order.
4821
4822 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4823 case instead of 255. Making it the same both with and without UTF-8 support
4824 means that the same test output works with both.
4825
4826 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4827 calling malloc() with a zero argument.
4828
4829 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4830 optimization. I've improved things by keeping a bitmap of backreferences with
4831 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4832 fact referenced, the optimization can be applied. It is unlikely that a
4833 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4834 the match to follow \n) will appear inside brackets with a number greater than
4835 31, but if it does, any back reference > 31 suppresses the optimization.
4836
4837 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4838 of disabling numbered capturing parentheses. Any opening parenthesis that is
4839 not followed by ? behaves as if it were followed by ?: but named parentheses
4840 can still be used for capturing (and they will acquire numbers in the usual
4841 way).
4842
4843 41. Redesigned the return codes from the match() function into yes/no/error so
4844 that errors can be passed back from deep inside the nested calls. A malloc
4845 failure while inside a recursive subpattern call now causes the
4846 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4847
4848 42. It is now possible to set a limit on the number of times the match()
4849 function is called in a call to pcre_exec(). This facility makes it possible to
4850 limit the amount of recursion and backtracking, though not in a directly
4851 obvious way, because the match() function is used in a number of different
4852 circumstances. The count starts from zero for each position in the subject
4853 string (for non-anchored patterns). The default limit is, for compatibility, a
4854 large number, namely 10 000 000. You can change this in two ways:
4855
4856 (a) When configuring PCRE before making, you can use --with-match-limit=n
4857 to set a default value for the compiled library.
4858
4859 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4860 a different value is set. See 45 below.
4861
4862 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4863
4864 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4865 of things that can be changed at compile time. The first argument specifies
4866 what is wanted and the second points to where the information is to be placed.
4867 The current list of available information is:
4868
4869 PCRE_CONFIG_UTF8
4870
4871 The output is an integer that is set to one if UTF-8 support is available;
4872 otherwise it is set to zero.
4873
4874 PCRE_CONFIG_NEWLINE
4875
4876 The output is an integer that it set to the value of the code that is used for
4877 newline. It is either LF (10) or CR (13).
4878
4879 PCRE_CONFIG_LINK_SIZE
4880
4881 The output is an integer that contains the number of bytes used for internal
4882 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4883
4884 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4885
4886 The output is an integer that contains the threshold above which the POSIX
4887 interface uses malloc() for output vectors. See item 31 above.
4888
4889 PCRE_CONFIG_MATCH_LIMIT
4890
4891 The output is an unsigned integer that contains the default limit of the number
4892 of match() calls in a pcre_exec() execution. See 42 above.
4893
4894 44. pcretest has been upgraded by the addition of the -C option. This causes it
4895 to extract all the available output from the new pcre_config() function, and to
4896 output it. The program then exits immediately.
4897
4898 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4899 order to support additional features. One way would have been to define
4900 pcre_exec2() (for example) with extra arguments, but this would not have been
4901 extensible, and would also have required all calls to the original function to
4902 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4903 is used for passing in "extra" data from pcre_study().
4904
4905 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4906 contains the following fields:
4907
4908 flags a bitmap indicating which of the following fields are set
4909 study_data opaque data from pcre_study()
4910 match_limit a way of specifying a limit on match() calls for a specific
4911 call to pcre_exec()
4912 callout_data data for callouts (see 49 below)
4913
4914 The flag bits are also defined in pcre.h, and are
4915
4916 PCRE_EXTRA_STUDY_DATA
4917 PCRE_EXTRA_MATCH_LIMIT
4918 PCRE_EXTRA_CALLOUT_DATA
4919
4920 The pcre_study() function now returns one of these new pcre_extra blocks, with
4921 the actual study data pointed to by the study_data field, and the
4922 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4923 before. That is, this change is entirely upwards-compatible and requires no
4924 change to existing code.
4925
4926 If you want to pass in additional data to pcre_exec(), you can either place it
4927 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4928 block.
4929
4930 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4931 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4932 times with different match limits, until it finds the minimum value needed for
4933 pcre_exec() to complete. The value is then output. This can be instructive; for
4934 most simple matches the number is quite small, but for pathological cases it
4935 gets very large very quickly.
4936
4937 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4938 returns the size of the data block pointed to by the study_data field in a
4939 pcre_extra block, that is, the value that was passed as the argument to
4940 pcre_malloc() when PCRE was getting memory in which to place the information
4941 created by pcre_study(). The fourth argument should point to a size_t variable.
4942 pcretest has been extended so that this information is shown after a successful
4943 pcre_study() call when information about the compiled regex is being displayed.
4944
4945 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4946 because what follows is always an absolute path. (Later: it turns out that this
4947 is more than cosmetic for MinGW, because it doesn't like empty path
4948 components.)
4949
4950 49. Some changes have been made to the callout feature (see 28 above):
4951
4952 (i) A callout function now has three choices for what it returns:
4953
4954 0 => success, carry on matching
4955 > 0 => failure at this point, but backtrack if possible
4956 < 0 => serious error, return this value from pcre_exec()
4957
4958 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4959 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4960 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4961 use by callout functions. It will never be used by PCRE itself.
4962
4963 (ii) The pcre_extra structure (see 45 above) has a void * field called
4964 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4965 pcre_callout_block structure has a field of the same name. The contents of
4966 the field passed in the pcre_extra structure are passed to the callout
4967 function in the corresponding field in the callout block. This makes it
4968 easier to use the same callout-containing regex from multiple threads. For
4969 testing, the pcretest program has a new data escape
4970
4971 \C*n pass the number n (may be negative) as callout_data
4972
4973 If the callout function in pcretest receives a non-zero value as
4974 callout_data, it returns that value.
4975
4976 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4977 there were some redundant $(CFLAGS) in commands that are now specified as
4978 $(LINK), which already includes $(CFLAGS).
4979
4980 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4981 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4982 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4983 one-byte characters throughout. Note that case-insensitive matching applies
4984 only to characters whose values are less than 256. PCRE doesn't support the
4985 notion of cases for higher-valued characters.
4986
4987 (i) A character class whose characters are all within 0-255 is handled as
4988 a bit map, and the map is inverted for negative classes. Previously, a
4989 character > 255 always failed to match such a class; however it should
4990 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4991
4992 (ii) A negated character class with a single character < 255 is coded as
4993 "not this character" (OP_NOT). This wasn't working properly when the test
4994 character was multibyte, either singly or repeated.
4995
4996 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4997 mode, for example: \x{100}{2,3}.
4998
4999 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5000 singly or repeated) now correctly test multibyte characters. However,
5001 PCRE doesn't recognize any characters with values greater than 255 as
5002 digits, spaces, or word characters. Such characters always match \D, \S,
5003 and \W, and never match \d, \s, or \w.
5004
5005 (v) Classes may now contain characters and character ranges with values
5006 greater than 255. For example: [ab\x{100}-\x{400}].
5007
5008 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5009 PCRE in UTF-8 mode.
5010
5011 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5012 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5013 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5014 value.)
5015
5016 53. The single man page has become too large. I have therefore split it up into
5017 a number of separate man pages. These also give rise to individual HTML pages;
5018 these are now put in a separate directory, and there is an index.html page that
5019 lists them all. Some hyperlinking between the pages has been installed.
5020
5021 54. Added convenience functions for handling named capturing parentheses.
5022
5023 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5024 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5025 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5026 are faulted.
5027
5028 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5029 calling configure. These values are used when compiling the dftables.c program
5030 which is run to generate the source of the default character tables. They
5031 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5032 you will need to set these values.
5033
5034 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5035
5036
5037 Version 3.9 02-Jan-02
5038 ---------------------
5039
5040 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5041
5042 2. If --disable-static was given, the building process failed when trying to
5043 build pcretest and pcregrep. (For some reason it was using libtool to compile
5044 them, which is not right, as they aren't part of the library.)
5045
5046
5047 Version 3.8 18-Dec-01
5048 ---------------------
5049
5050 1. The experimental UTF-8 code was completely screwed up. It was packing the
5051 bytes in the wrong order. How dumb can you get?
5052
5053
5054 Version 3.7 29-Oct-01
5055 ---------------------
5056
5057 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5058 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5059 this didn't happen under Solaris 8, where I normally test things.
5060
5061 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5062 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5063 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5064 this fix an hour or so after the initial 3.7 release.)
5065
5066
5067 Version 3.6 23-Oct-01
5068 ---------------------
5069
5070 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5071 offsets passed as NULL with zero offset count.
5072
5073 2. The config.guess and config.sub files had not been updated when I moved to
5074 the latest autoconf.
5075
5076
5077 Version 3.5 15-Aug-01
5078 ---------------------
5079
5080 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5081 had been forgotten.
5082
5083 2. By using declared but undefined structures, we can avoid using "void"
5084 definitions in pcre.h while keeping the internal definitions of the structures
5085 private.
5086
5087 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5088 user point of view, this means that both static and shared libraries are built
5089 by default, but this can be individually controlled. More of the work of
5090 handling this static/shared cases is now inside libtool instead of PCRE's make
5091 file.
5092
5093 4. The pcretest utility is now installed along with pcregrep because it is
5094 useful for users (to test regexs) and by doing this, it automatically gets
5095 relinked by libtool. The documentation has been turned into a man page, so
5096 there are now .1, .txt, and .html versions in /doc.
5097
5098 5. Upgrades to pcregrep:
5099 (i) Added long-form option names like gnu grep.
5100 (ii) Added --help to list all options with an explanatory phrase.
5101 (iii) Added -r, --recursive to recurse into sub-directories.
5102 (iv) Added -f, --file to read patterns from a file.
5103
5104 6. pcre_exec() was referring to its "code" argument before testing that
5105 argument for NULL (and giving an error if it was NULL).
5106
5107 7. Upgraded Makefile.in to allow for compiling in a different directory from
5108 the source directory.
5109
5110 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5111 options bits, the pointer it was passed was to an int instead of to an unsigned
5112 long int. This mattered only on 64-bit systems.
5113
5114 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5115 generated) instead of pcre.in, which it its source. Also made the same change
5116 in several of the .c files.
5117
5118 10. A new release of gcc defines printf() as a macro, which broke pcretest
5119 because it had an ifdef in the middle of a string argument for printf(). Fixed
5120 by using separate calls to printf().
5121
5122 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5123 script, to force use of CR or LF instead of \n in the source. On non-Unix
5124 systems, the value can be set in config.h.
5125
5126 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5127 absolute limit. Changed the text of the error message to make this clear, and
5128 likewise updated the man page.
5129
5130 13. The limit of 99 on the number of capturing subpatterns has been removed.
5131 The new limit is 65535, which I hope will not be a "real" limit.
5132
5133
5134 Version 3.4 22-Aug-00
5135 ---------------------
5136
5137 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5138
5139 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5140
5141
5142 Version 3.3 01-Aug-00
5143 ---------------------
5144
5145 1. If an octal character was given, but the value was greater than \377, it
5146 was not getting masked to the least significant bits, as documented. This could
5147 lead to crashes in some systems.
5148
5149 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5150 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5151
5152 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5153 These just pass their arguments on to (pcre_free)(), but they are provided
5154 because some uses of PCRE bind it to non-C systems that can call its functions,
5155 but cannot call free() or pcre_free() directly.
5156
5157 4. Add "make test" as a synonym for "make check". Corrected some comments in
5158 the Makefile.
5159
5160 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5161 Makefile.
5162
5163 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5164 command called pgrep for grepping around the active processes.
5165
5166 7. Added the beginnings of support for UTF-8 character strings.
5167
5168 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5169 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5170 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5171 out for the ar command.)
5172
5173
5174 Version 3.2 12-May-00
5175 ---------------------
5176
5177 This is purely a bug fixing release.
5178
5179 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5180 of ZA. This was just one example of several cases that could provoke this bug,
5181 which was introduced by change 9 of version 2.00. The code for breaking
5182 infinite loops after an iteration that matches an empty string was't working
5183 correctly.
5184
5185 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5186 when matched against abbab (for example). After matching an empty string, it
5187 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5188 caused it to match further down the string than it should.
5189
5190 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5191 was there because it doesn't seem to be needed, and it causes trouble on some
5192 systems, as it is not a Standard C header. It has been removed.
5193
5194 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5195 were reported on the Macintosh. The changes were from
5196
5197 while ((c = *(++ptr)) != 0 && c != '\n');
5198 to
5199 while ((c = *(++ptr)) != 0 && c != '\n') ;
5200
5201 Totally extraordinary, but if that's what it takes...
5202
5203 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5204 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5205 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5206 assumes the way PCRE uses memmove() (always moving upwards).
5207
5208 6. PCRE is being used in one environment where strchr() is not available. There
5209 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5210 faster code anyway.
5211
5212
5213 Version 3.1 09-Feb-00
5214 ---------------------
5215
5216 The only change in this release is the fixing of some bugs in Makefile.in for
5217 the "install" target:
5218
5219 (1) It was failing to install pcreposix.h.
5220
5221 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5222
5223
5224 Version 3.0 01-Feb-00
5225 ---------------------
5226
5227 1. Add support for the /+ modifier to perltest (to output $` like it does in
5228 pcretest).
5229
5230 2. Add support for the /g modifier to perltest.
5231
5232 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5233 matches null strings.
5234
5235 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5236 pattern. Perl treats empty patterns specially - it reuses the most recent
5237 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5238 effect.
5239
5240 5. The POSIX interface was broken in that it was just handing over the POSIX
5241 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5242 required a bigger vector, with some working space on the end. This means that
5243 the POSIX wrapper now has to get and free some memory, and copy the results.
5244
5245 6. Added some simple autoconf support, placing the test data and the
5246 documentation in separate directories, re-organizing some of the
5247 information files, and making it build pcre-config (a GNU standard). Also added
5248 libtool support for building PCRE as a shared library, which is now the
5249 default.
5250
5251 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5252 09 are not valid octal constants. Single digits will be used for minor values
5253 less than 10.
5254
5255 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5256 existing programs that set these in the POSIX interface can use PCRE without
5257 modification.
5258
5259 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5260 return all that pcre_info() returns, plus additional data. The pcre_info()
5261 function is retained for compatibility, but is considered to be obsolete.
5262
5263 10. Added experimental recursion feature (?R) to handle one common case that
5264 Perl 5.6 will be able to do with (?p{...}).
5265
5266 11. Added support for POSIX character classes like [:alpha:], which Perl is
5267 adopting.
5268
5269
5270 Version 2.08 31-Aug-99
5271 ----------------------
5272
5273 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5274 trying to match at the startoffset position, but instead was moving forward to
5275 the next newline as if a previous match had failed.
5276
5277 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5278 and could get into a loop if a null string was matched other than at the start
5279 of the subject.
5280
5281 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5282 be distinguished at compile time, and for completeness also added PCRE_DATE.
5283
5284 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5285 in GnuWin32 environments.
5286
5287
5288 Version 2.07 29-Jul-99
5289 ----------------------
5290
5291 1. The documentation is now supplied in plain text form and HTML as well as in
5292 the form of man page sources.
5293
5294 2. C++ compilers don't like assigning (void *) values to other pointer types.
5295 In particular this affects malloc(). Although there is no problem in Standard
5296 C, I've put in casts to keep C++ compilers happy.
5297
5298 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5299 should be (const char *).
5300
5301 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5302 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5303 However, I haven't made this a standard facility. The documentation doesn't
5304 mention it, and the Makefile doesn't support it.
5305
5306 5. The Makefile now contains an "install" target, with editable destinations at
5307 the top of the file. The pcretest program is not installed.
5308
5309 6. pgrep -V now gives the PCRE version number and date.
5310
5311 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5312 causing the entire string to be ignored, instead of just the last character.
5313
5314 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5315 non-matching string, it can take a very, very long time, even for strings of
5316 quite modest length, because of the nested recursion. PCRE now does better in
5317 some of these cases. It does this by remembering the last required literal
5318 character in the pattern, and pre-searching the subject to ensure it is present
5319 before running the real match. In other words, it applies a heuristic to detect
5320 some types of certain failure quickly, and in the above example, if presented
5321 with a string that has no trailing " it gives "no match" very quickly.
5322
5323 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5324 other alternatives are tried instead.
5325
5326
5327 Version 2.06 09-Jun-99
5328 ----------------------
5329
5330 1. Change pcretest's output for amount of store used to show just the code
5331 space, because the remainder (the data block) varies in size between 32-bit and
5332 64-bit systems.
5333
5334 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5335 start matching at. This allows lookbehinds to work when searching for multiple
5336 occurrences in a string.
5337
5338 3. Added additional options to pcretest for testing multiple occurrences:
5339
5340 /+ outputs the rest of the string that follows a match
5341 /g loops for multiple occurrences, using the new startoffset argument
5342 /G loops for multiple occurrences by passing an incremented pointer
5343
5344 4. PCRE wasn't doing the "first character" optimization for patterns starting
5345 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5346 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5347 the letter 'x'. On long subject strings, this gives a significant speed-up.
5348
5349
5350 Version 2.05 21-Apr-99
5351 ----------------------
5352
5353 1. Changed the type of magic_number from int to long int so that it works
5354 properly on 16-bit systems.
5355
5356 2. Fixed a bug which caused patterns starting with .* not to work correctly
5357 when the subject string contained newline characters. PCRE was assuming
5358 anchoring for such patterns in all cases, which is not correct because .* will
5359 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5360 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5361 must be retried after every newline in the subject.
5362
5363
5364 Version 2.04 18-Feb-99
5365 ----------------------
5366
5367 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5368 computation of the store needed to hold the pattern was incorrect (too large).
5369 If such patterns were nested a few deep, this could multiply and become a real
5370 problem.
5371
5372 2. Added /M option to pcretest to show the memory requirement of a specific
5373 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5374
5375 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5376 compiled in such a way that the backtracking after subsequent failure was
5377 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5378 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5379
5380
5381 Version 2.03 02-Feb-99
5382 ----------------------
5383
5384 1. Fixed typo and small mistake in man page.
5385
5386 2. Added 4th condition (GPL supersedes if conflict) and created separate
5387 LICENCE file containing the conditions.
5388
5389 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5390 Perl, that is the internal \ allows the delimiter to be included in the
5391 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5392 the final delimiter, add \ to the end of the pattern (to test the error).
5393
5394 4. Added the convenience functions for extracting substrings after a successful
5395 match. Updated pcretest to make it able to test these functions.
5396
5397
5398 Version 2.02 14-Jan-99
5399 ----------------------
5400
5401 1. Initialized the working variables associated with each extraction so that
5402 their saving and restoring doesn't refer to uninitialized store.
5403
5404 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5405 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5406 fix the problem.
5407
5408 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5409 calls, and wasn't printing the correct value for compiling calls. Increased the
5410 default value of LOOPREPEAT, and the number of significant figures in the
5411 times.
5412
5413 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5414
5415 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5416 a building problem on Windows NT with a FAT file system.
5417
5418
5419 Version 2.01 21-Oct-98
5420 ----------------------
5421
5422 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5423 to character tables built by pcre_maketables() in the current locale. If NULL
5424 is passed, the default tables are used.
5425
5426
5427 Version 2.00 24-Sep-98
5428 ----------------------
5429
5430 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5431 it any more.
5432
5433 2. Allow quantification of (?>) groups, and make it work correctly.
5434
5435 3. The first character computation wasn't working for (?>) groups.
5436
5437 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5438 end of the subject) and add 5.005's \z, which really does match only at the
5439 very end of the subject.
5440
5441 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5442
5443 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5444 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5445 localized options. All options to pcre_study() were also removed.
5446
5447 7. Add other new features from 5.005:
5448
5449 $(?<= positive lookbehind
5450 $(?<! negative lookbehind
5451 (?imsx-imsx) added the unsetting capability
5452 such a setting is global if at outer level; local otherwise
5453 (?imsx-imsx:) non-capturing groups with option setting
5454 (?(cond)re|re) conditional pattern matching
5455
5456 A backreference to itself in a repeated group matches the previous
5457 captured string.
5458
5459 8. General tidying up of studying (both automatic and via "study")
5460 consequential on the addition of new assertions.
5461
5462 9. As in 5.005, unlimited repeated groups that could match an empty substring
5463 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5464 runtime if any iteration does actually match an empty substring.
5465
5466 10. Include the RunTest script in the distribution.
5467
5468 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5469 discrepancies, some of which were old and were also with respect to 5.004. They
5470 have now been fixed.
5471
5472
5473 Version 1.09 28-Apr-98
5474 ----------------------
5475
5476 1. A negated single character class followed by a quantifier with a minimum
5477 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5478 program crashes, or just wrong answers. This did not apply to negated classes
5479 containing more than one character, or to minima other than one.
5480
5481
5482 Version 1.08 27-Mar-98
5483 ----------------------
5484
5485 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5486
5487 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5488 latter must appear before anything that relies on it in the pattern.
5489
5490
5491 Version 1.07 16-Feb-98
5492 ----------------------
5493
5494 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5495 repeat of a potentially empty string).
5496
5497
5498 Version 1.06 23-Jan-98
5499 ----------------------
5500
5501 1. Added Markus Oberhumer's little patches for C++.
5502
5503 2. Literal strings longer than 255 characters were broken.
5504
5505
5506 Version 1.05 23-Dec-97
5507 ----------------------
5508
5509 1. Negated character classes containing more than one character were failing if
5510 PCRE_CASELESS was set at run time.
5511
5512
5513 Version 1.04 19-Dec-97
5514 ----------------------
5515
5516 1. Corrected the man page, where some "const" qualifiers had been omitted.
5517
5518 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5519 input syntax.
5520
5521 3. Fixed memory leak which occurred when a regex with back references was
5522 matched with an offsets vector that wasn't big enough. The temporary memory
5523 that is used in this case wasn't being freed if the match failed.
5524
5525 4. Tidied pcretest to ensure it frees memory that it gets.
5526
5527 5. Temporary memory was being obtained in the case where the passed offsets
5528 vector was exactly big enough.
5529
5530 6. Corrected definition of offsetof() from change 5 below.
5531
5532 7. I had screwed up change 6 below and broken the rules for the use of
5533 setjmp(). Now fixed.
5534
5535
5536 Version 1.03 18-Dec-97
5537 ----------------------
5538
5539 1. A erroneous regex with a missing opening parenthesis was correctly
5540 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5541 on some systems.
5542
5543 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5544 it was reported that one broken compiler failed on the former because "code" is
5545 also an independent variable.
5546
5547 3. The erroneous regex a[]b caused an array overrun reference.
5548
5549 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5550 fail on data ending with that character. (It was going on too far, and checking
5551 the next character, typically a binary zero.) This was specific to the
5552 optimized code for single-character negative classes.
5553
5554 5. Added a contributed patch from the TIN world which does the following:
5555
5556 + Add an undef for memmove, in case the the system defines a macro for it.
5557
5558 + Add a definition of offsetof(), in case there isn't one. (I don't know
5559 the reason behind this - offsetof() is part of the ANSI standard - but
5560 it does no harm).
5561
5562 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5563 most of the places where whitespace preceded '#'. I have given up and
5564 allowed the remaining 2 cases to be at the margin.
5565
5566 + Rename some variables in pcre to eliminate shadowing. This seems very
5567 pedantic, but does no harm, of course.
5568
5569 6. Moved the call to setjmp() into its own function, to get rid of warnings
5570 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5571
5572 7. Constructs such as \d{8,} were compiling into the equivalent of
5573 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5574 outcome, but in this particular case used more store than had been allocated,
5575 which caused the bug to be discovered because it threw up an internal error.
5576
5577 8. The debugging code in both pcre and pcretest for outputting the compiled
5578 form of a regex was going wrong in the case of back references followed by
5579 curly-bracketed repeats.
5580
5581
5582 Version 1.02 12-Dec-97
5583 ----------------------
5584
5585 1. Typos in pcre.3 and comments in the source fixed.
5586
5587 2. Applied a contributed patch to get rid of places where it used to remove
5588 'const' from variables, and fixed some signed/unsigned and uninitialized
5589 variable warnings.
5590
5591 3. Added the "runtest" target to Makefile.
5592
5593 4. Set default compiler flag to -O2 rather than just -O.
5594
5595
5596 Version 1.01 19-Nov-97
5597 ----------------------
5598
5599 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5600 like /([ab]*)*/, that is, for classes with more than one character in them.
5601
5602 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5603 as /((?>a*))*/ (a PCRE_EXTRA facility).
5604
5605
5606 Version 1.00 18-Nov-97
5607 ----------------------
5608
5609 1. Added compile-time macros to support systems such as SunOS4 which don't have
5610 memmove() or strerror() but have other things that can be used instead.
5611
5612 2. Arranged that "make clean" removes the executables.
5613
5614
5615 Version 0.99 27-Oct-97
5616 ----------------------
5617
5618 1. Fixed bug in code for optimizing classes with only one character. It was
5619 initializing a 32-byte map regardless, which could cause it to run off the end
5620 of the memory it had got.
5621
5622 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5623
5624
5625 Version 0.98 22-Oct-97
5626 ----------------------
5627
5628 1. Fixed bug in code for handling temporary memory usage when there are more
5629 back references than supplied space in the ovector. This could cause segfaults.
5630
5631
5632 Version 0.97 21-Oct-97
5633 ----------------------
5634
5635 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5636
5637 2. Optimized negated single characters not to use a bit map.
5638
5639 3. Brought error texts together as macro definitions; clarified some of them;
5640 fixed one that was wrong - it said "range out of order" when it meant "invalid
5641 escape sequence".
5642
5643 4. Changed some char * arguments to const char *.
5644
5645 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5646
5647 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5648 pcretest.
5649
5650
5651 Version 0.96 16-Oct-97
5652 ----------------------
5653
5654 1. Added a simple "pgrep" utility to the distribution.
5655
5656 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5657 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5658 where "ddd" means "one or more decimal digits".
5659
5660 3. Fixed serious bug. If a pattern had a back reference, but the call to
5661 pcre_exec() didn't supply a large enough ovector to record the related
5662 identifying subpattern, the match always failed. PCRE now remembers the number
5663 of the largest back reference, and gets some temporary memory in which to save
5664 the offsets during matching if necessary, in order to ensure that
5665 backreferences always work.
5666
5667 4. Increased the compatibility with Perl in a number of ways:
5668
5669 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5670 to request this handling. The option can be set at compile or exec time.
5671
5672 (b) $ matches before a terminating newline by default; an option
5673 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5674 mode). The option can be set at compile or exec time.
5675
5676 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5677 the same as Perl's. If the decimal number it represents is less than 10
5678 or there aren't that many previous left capturing parentheses, an octal
5679 escape is read. Inside a character class, it's always an octal escape,
5680 even if it is a single digit.
5681
5682 (d) An escaped but undefined alphabetic character is taken as a literal,
5683 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5684 escapes.
5685
5686 (e) {0} is now permitted. (The previous item is removed from the compiled
5687 pattern).
5688
5689 5. Changed all the names of code files so that the basic parts are no longer
5690 than 10 characters, and abolished the teeny "globals.c" file.
5691
5692 6. Changed the handling of character classes; they are now done with a 32-byte
5693 bit map always.
5694
5695 7. Added the -d and /D options to pcretest to make it possible to look at the
5696 internals of compilation without having to recompile pcre.
5697
5698
5699 Version 0.95 23-Sep-97
5700 ----------------------
5701
5702 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5703 \x20 at the start of a run of normal characters. These were being treated as
5704 real characters, instead of the source characters being re-checked.
5705
5706
5707 Version 0.94 18-Sep-97
5708 ----------------------
5709
5710 1. The functions are now thread-safe, with the caveat that the global variables
5711 containing pointers to malloc() and free() or alternative functions are the
5712 same for all threads.
5713
5714 2. Get pcre_study() to generate a bitmap of initial characters for non-
5715 anchored patterns when this is possible, and use it if passed to pcre_exec().
5716
5717
5718 Version 0.93 15-Sep-97
5719 ----------------------
5720
5721 1. /(b)|(:+)/ was computing an incorrect first character.
5722
5723 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5724 but not actually doing anything yet.
5725
5726 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5727 as Perl does (e.g. [-az] or [az-]).
5728
5729 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5730 all possible positions.
5731
5732 5. Split up into different modules to avoid including unneeded functions in a
5733 compiled binary. However, compile and exec are still in one module. The "study"
5734 function is split off.
5735
5736 6. The character tables are now in a separate module whose source is generated
5737 by an auxiliary program - but can then be edited by hand if required. There are
5738 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5739 toupper() in the code.
5740
5741 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5742 make them global. Abolish the function for setting them, as the caller can now
5743 set them directly.
5744
5745
5746 Version 0.92 11-Sep-97
5747 ----------------------
5748
5749 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5750 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5751
5752 2. Caseless matching was not working in character classes if the characters in
5753 the pattern were in upper case.
5754
5755 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5756
5757 4. Make PCRE_ANCHORED public and accept as a compile option.
5758
5759 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5760 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5761 pass them.
5762
5763 6. Give an error if bad option bits passed at compile or run time.
5764
5765 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5766 pcretest to cause it to pass that flag.
5767
5768 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5769 options, and the first character, if set.
5770
5771 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5772
5773
5774 Version 0.91 10-Sep-97
5775 ----------------------
5776
5777 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5778 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5779
5780 2. PCRE was looping on encountering an indefinitely repeated back reference to
5781 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5782 Perl does - treats the match as successful.
5783
5784 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5