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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1540 - (show annotations) (download)
Sun Mar 29 17:41:16 2015 UTC (26 hours, 22 minutes ago) by ph10
File size: 275975 byte(s)
Fix possessive quantifier after group containing subroutine call.

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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12