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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1548 - (show annotations) (download)
Tue Apr 14 17:02:30 2015 UTC (3 days, 22 hours ago) by ph10
File size: 277120 byte(s)
Documentation and tidies preparatory to 8.37 release.

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

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