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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1558 - (show annotations)
Fri May 15 17:17:03 2015 UTC (4 years, 5 months ago) by ph10
File size: 278357 byte(s)
Error occurred while calculating annotation data.
Fix buffer overflow for named recursive back reference when the name is 
duplicated.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.38 xx-xxx-xxxx
5 ------------------------
6
7 1. If a group that contained a recursive back reference also contained a
8 forward reference subroutine call followed by a non-forward-reference
9 subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
10 compile correct code, leading to undefined behaviour or an internally
11 detected error. This bug was discovered by the LLVM fuzzer.
12
13 2. Quantification of certain items (e.g. atomic back references) could cause
14 incorrect code to be compiled when recursive forward references were
15 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
16 This bug was discovered by the LLVM fuzzer.
17
18 3. A repeated conditional group whose condition was a reference by name caused
19 a buffer overflow if there was more than one group with the given name.
20 This bug was discovered by the LLVM fuzzer.
21
22 4. A recursive back reference by name within a group that had the same name as
23 another group caused a buffer overflow. For example:
24 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
25
26
27 Version 8.37 28-April-2015
28 --------------------------
29
30 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
31 for those parentheses to be closed with whatever has been captured so far.
32 However, it was failing to mark any other groups between the hightest
33 capture so far and the currrent group as "unset". Thus, the ovector for
34 those groups contained whatever was previously there. An example is the
35 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
36
37 2. If an assertion condition was quantified with a minimum of zero (an odd
38 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
39
40 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
41 unrecognized modifier, a crash could occur.
42
43 4. An attempt to do global matching in pcretest with a zero-length ovector
44 caused a crash.
45
46 5. Fixed a memory leak during matching that could occur for a subpattern
47 subroutine call (recursive or otherwise) if the number of captured groups
48 that had to be saved was greater than ten.
49
50 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
51 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
52 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
53
54 7. A UTF pattern containing a "not" match of a non-ASCII character and a
55 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
56
57 8. When a pattern is compiled, it remembers the highest back reference so that
58 when matching, if the ovector is too small, extra memory can be obtained to
59 use instead. A conditional subpattern whose condition is a check on a
60 capture having happened, such as, for example in the pattern
61 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
62 setting the highest backreference number. This mattered only if pcre_exec()
63 was called with an ovector that was too small to hold the capture, and there
64 was no other kind of back reference (a situation which is probably quite
65 rare). The effect of the bug was that the condition was always treated as
66 FALSE when the capture could not be consulted, leading to a incorrect
67 behaviour by pcre_exec(). This bug has been fixed.
68
69 9. A reference to a duplicated named group (either a back reference or a test
70 for being set in a conditional) that occurred in a part of the pattern where
71 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
72 to be incorrectly calculated, leading to overwriting.
73
74 10. A mutually recursive set of back references such as (\2)(\1) caused a
75 segfault at study time (while trying to find the minimum matching length).
76 The infinite loop is now broken (with the minimum length unset, that is,
77 zero).
78
79 11. If an assertion that was used as a condition was quantified with a minimum
80 of zero, matching went wrong. In particular, if the whole group had
81 unlimited repetition and could match an empty string, a segfault was
82 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
83 assertions to be quantified, but not if they are being used as conditions,
84 so the above pattern is faulted by Perl. PCRE has now been changed so that
85 it also rejects such patterns.
86
87 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
88 failed to allow the zero-repeat case if pcre2_exec() was called with an
89 ovector too small to capture the group.
90
91 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
92 Red Hat Product Security:
93
94 (a) A crash if /K and /F were both set with the option to save the compiled
95 pattern.
96
97 (b) Another crash if the option to print captured substrings in a callout
98 was combined with setting a null ovector, for example \O\C+ as a subject
99 string.
100
101 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
102 forward reference repeated a large (but limited) number of times within a
103 repeated outer group that has a zero minimum quantifier, caused incorrect
104 code to be compiled, leading to the error "internal error:
105 previously-checked referenced subpattern not found" when an incorrect
106 memory address was read. This bug was reported as "heap overflow",
107 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
108 CVE-2015-2325.
109
110 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
111 call within a group that also contained a recursive back reference caused
112 incorrect code to be compiled. This bug was reported as "heap overflow",
113 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
114 number CVE-2015-2326.
115
116 24. Computing the size of the JIT read-only data in advance has been a source
117 of various issues, and new ones are still appear unfortunately. To fix
118 existing and future issues, size computation is eliminated from the code,
119 and replaced by on-demand memory allocation.
120
121 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
122 adjacent to the end of the range, and the range contained characters with
123 more than one other case, caused incorrect behaviour when compiled in UTF
124 mode. In that example, the range a-j was left out of the class.
125
126 26. Fix JIT compilation of conditional blocks, which assertion
127 is converted to (*FAIL). E.g: /(?(?!))/.
128
129 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
130 discovered by the LLVM fuzzer.
131
132 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
133 when this assertion was used as a condition, for example (?(?!)a|b). In
134 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
135 error about an unsupported item.
136
137 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
138 possessification code could take exponential time to complete. A recursion
139 depth limit of 1000 has been imposed to limit the resources used by this
140 optimization.
141
142 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
143 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
144 because \S ensures they are all in the class. The code for doing this was
145 interacting badly with the code for computing the amount of space needed to
146 compile the pattern, leading to a buffer overflow. This bug was discovered
147 by the LLVM fuzzer.
148
149 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
150 other kinds of group caused stack overflow at compile time. This bug was
151 discovered by the LLVM fuzzer.
152
153 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
154 between a subroutine call and its quantifier was incorrectly compiled,
155 leading to buffer overflow or other errors. This bug was discovered by the
156 LLVM fuzzer.
157
158 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
159 assertion after (?(. The code was failing to check the character after
160 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
161 was discovered by the LLVM fuzzer.
162
163 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
164 a fixed maximum following a group that contains a subroutine reference was
165 incorrectly compiled and could trigger buffer overflow. This bug was
166 discovered by the LLVM fuzzer.
167
168 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
169 caused a stack overflow instead of the diagnosis of a non-fixed length
170 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
171
172 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
173 (e.g. /(?<=\Ka)/) could make pcregrep loop.
174
175 37. There was a similar problem to 36 in pcretest for global matches.
176
177 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
178 and a subsequent item in the pattern caused a non-match, backtracking over
179 the repeated \X did not stop, but carried on past the start of the subject,
180 causing reference to random memory and/or a segfault. There were also some
181 other cases where backtracking after \C could crash. This set of bugs was
182 discovered by the LLVM fuzzer.
183
184 39. The function for finding the minimum length of a matching string could take
185 a very long time if mutual recursion was present many times in a pattern,
186 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
187 method has been implemented. This infelicity was discovered by the LLVM
188 fuzzer.
189
190 40. Static linking against the PCRE library using the pkg-config module was
191 failing on missing pthread symbols.
192
193
194 Version 8.36 26-September-2014
195 ------------------------------
196
197 1. Got rid of some compiler warnings in the C++ modules that were shown up by
198 -Wmissing-field-initializers and -Wunused-parameter.
199
200 2. The tests for quantifiers being too big (greater than 65535) were being
201 applied after reading the number, and stupidly assuming that integer
202 overflow would give a negative number. The tests are now applied as the
203 numbers are read.
204
205 3. Tidy code in pcre_exec.c where two branches that used to be different are
206 now the same.
207
208 4. The JIT compiler did not generate match limit checks for certain
209 bracketed expressions with quantifiers. This may lead to exponential
210 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
211 issue should be resolved now.
212
213 5. Fixed an issue, which occures when nested alternatives are optimized
214 with table jumps.
215
216 6. Inserted two casts and changed some ints to size_t in the light of some
217 reported 64-bit compiler warnings (Bugzilla 1477).
218
219 7. Fixed a bug concerned with zero-minimum possessive groups that could match
220 an empty string, which sometimes were behaving incorrectly in the
221 interpreter (though correctly in the JIT matcher). This pcretest input is
222 an example:
223
224 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
225 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
226
227 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
228 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
229 for an empty string was breaking the inner loop and carrying on at a lower
230 level, when possessive repeated groups should always return to a higher
231 level as they have no backtrack points in them. The empty string test now
232 occurs at the outer level.
233
234 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
235 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
236
237 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
238 doesn't).
239
240 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
241 to hold a pattern to be incorrectly computed (too small) when there were
242 named back references to duplicated names. This could cause "internal
243 error: code overflow" or "double free or corruption" or other memory
244 handling errors.
245
246 11. When named subpatterns had the same prefixes, back references could be
247 confused. For example, in this pattern:
248
249 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
250
251 the reference to 'Name' was incorrectly treated as a reference to a
252 duplicate name.
253
254 12. A pattern such as /^s?c/mi8 where the optional character has more than
255 one "other case" was incorrectly compiled such that it would only try to
256 match starting at "c".
257
258 13. When a pattern starting with \s was studied, VT was not included in the
259 list of possible starting characters; this should have been part of the
260 8.34/18 patch.
261
262 14. If a character class started [\Qx]... where x is any character, the class
263 was incorrectly terminated at the ].
264
265 15. If a pattern that started with a caseless match for a character with more
266 than one "other case" was studied, PCRE did not set up the starting code
267 unit bit map for the list of possible characters. Now it does. This is an
268 optimization improvement, not a bug fix.
269
270 16. The Unicode data tables have been updated to Unicode 7.0.0.
271
272 17. Fixed a number of memory leaks in pcregrep.
273
274 18. Avoid a compiler warning (from some compilers) for a function call with
275 a cast that removes "const" from an lvalue by using an intermediate
276 variable (to which the compiler does not object).
277
278 19. Incorrect code was compiled if a group that contained an internal recursive
279 back reference was optional (had quantifier with a minimum of zero). This
280 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
281 caused segmentation faults because of stack overflows at compile time.
282
283 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
284 group that is quantified with an indefinite repeat, caused a compile-time
285 loop which used up all the system stack and provoked a segmentation fault.
286 This was not the same bug as 19 above.
287
288 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
289 Patch by Mike Frysinger.
290
291
292 Version 8.35 04-April-2014
293 --------------------------
294
295 1. A new flag is set, when property checks are present in an XCLASS.
296 When this flag is not set, PCRE can perform certain optimizations
297 such as studying these XCLASS-es.
298
299 2. The auto-possessification of character sets were improved: a normal
300 and an extended character set can be compared now. Furthermore
301 the JIT compiler optimizes more character set checks.
302
303 3. Got rid of some compiler warnings for potentially uninitialized variables
304 that show up only when compiled with -O2.
305
306 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
307 of a match later then the end of the match. The pcretest program was not
308 handling the case sensibly - it was outputting from the start to the next
309 binary zero. It now reports this situation in a message, and outputs the
310 text from the end to the start.
311
312 5. Fast forward search is improved in JIT. Instead of the first three
313 characters, any three characters with fixed position can be searched.
314 Search order: first, last, middle.
315
316 6. Improve character range checks in JIT. Characters are read by an inprecise
317 function now, which returns with an unknown value if the character code is
318 above a certain threshold (e.g: 256). The only limitation is that the value
319 must be bigger than the threshold as well. This function is useful when
320 the characters above the threshold are handled in the same way.
321
322 7. The macros whose names start with RAWUCHAR are placeholders for a future
323 mode in which only the bottom 21 bits of 32-bit data items are used. To
324 make this more memorable for those maintaining the code, the names have
325 been changed to start with UCHAR21, and an extensive comment has been added
326 to their definition.
327
328 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
329 to the export list in Makefile.am (they were accidentally omitted from the
330 8.34 tarball).
331
332 9. The informational output from pcretest used the phrase "starting byte set"
333 which is inappropriate for the 16-bit and 32-bit libraries. As the output
334 for "first char" and "need char" really means "non-UTF-char", I've changed
335 "byte" to "char", and slightly reworded the output. The documentation about
336 these values has also been (I hope) clarified.
337
338 10. Another JIT related optimization: use table jumps for selecting the correct
339 backtracking path, when more than four alternatives are present inside a
340 bracket.
341
342 11. Empty match is not possible, when the minimum length is greater than zero,
343 and there is no \K in the pattern. JIT should avoid empty match checks in
344 such cases.
345
346 12. In a caseless character class with UCP support, when a character with more
347 than one alternative case was not the first character of a range, not all
348 the alternative cases were added to the class. For example, s and \x{17f}
349 are both alternative cases for S: the class [RST] was handled correctly,
350 but [R-T] was not.
351
352 13. The configure.ac file always checked for pthread support when JIT was
353 enabled. This is not used in Windows, so I have put this test inside a
354 check for the presence of windows.h (which was already tested for).
355
356 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
357 The algorithm provides a way to skip certain starting offsets, and usually
358 faster than linear prefix searches.
359
360 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
361 as for 'fr_FR' and 'french'. For some reason, however, it then used the
362 Windows-specific input and output files, which have 'french' screwed in.
363 So this could never have worked. One of the problems with locales is that
364 they aren't always the same. I have now updated RunTest so that it checks
365 the output of the locale test (test 3) against three different output
366 files, and it allows the test to pass if any one of them matches. With luck
367 this should make the test pass on some versions of Solaris where it was
368 failing. Because of the uncertainty, the script did not used to stop if
369 test 3 failed; it now does. If further versions of a French locale ever
370 come to light, they can now easily be added.
371
372 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
373 there was a message during ./configure, but it did not stop. This now
374 provokes an error. The invalid example in README has been corrected.
375 If a value less than the minimum is given, the minimum value has always
376 been used, but now a warning is given.
377
378 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
379 was a bug in the test system, which is now fixed. Also, the list of various
380 configurations that are tested for each release did not have one with both
381 16/32 bits and --enable-bar-anycrlf. It now does.
382
383 18. pcretest was missing "-C bsr" for displaying the \R default setting.
384
385 19. Little endian PowerPC systems are supported now by the JIT compiler.
386
387 20. The fast forward newline mechanism could enter to an infinite loop on
388 certain invalid UTF-8 input. Although we don't support these cases
389 this issue can be fixed by a performance optimization.
390
391 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
392 not take account if existing stack usage. There is now a new global
393 variable called pcre_stack_guard that can be set to point to an external
394 function to check stack availability. It is called at the start of
395 processing every parenthesized group.
396
397 22. A typo in the code meant that in ungreedy mode the max/min qualifier
398 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
399 not match "ab".
400
401 23. When UTF was disabled, the JIT program reported some incorrect compile
402 errors. These messages are silenced now.
403
404 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
405 compiler.
406
407 25. Change all the temporary files used in RunGrepTest to be different to those
408 used by RunTest so that the tests can be run simultaneously, for example by
409 "make -j check".
410
411
412 Version 8.34 15-December-2013
413 -----------------------------
414
415 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
416 executable memory. Patch inspired by Carsten Klein.
417
418 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
419 this macro is never tested and has no effect, because the work to support
420 coverage involves only compiling and linking options and special targets in
421 the Makefile. The comment in config.h implied that defining the macro would
422 enable coverage support, which is totally false. There was also support for
423 setting this macro in the CMake files (my fault, I just copied it from
424 configure). SUPPORT_GCOV has now been removed.
425
426 3. Make a small performance improvement in strlen16() and strlen32() in
427 pcretest.
428
429 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
430 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
431 these cases). There was also one in pcretest.c.
432
433 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
434
435 6. In UTF mode, the code for checking whether a group could match an empty
436 string (which is used for indefinitely repeated groups to allow for
437 breaking an infinite loop) was broken when the group contained a repeated
438 negated single-character class with a character that occupied more than one
439 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
440 UTF-8 mode). The effect was undefined: the group might or might not be
441 deemed as matching an empty string, or the program might have crashed.
442
443 7. The code for checking whether a group could match an empty string was not
444 recognizing that \h, \H, \v, \V, and \R must match a character.
445
446 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
447 an empty string. If it can, pcretest shows this in its information output.
448
449 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
450 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
451 matched by pcre_exec() without using JIT:
452
453 (a) If the rest of the pattern did not match after a maximal run of
454 grapheme clusters, the code for backing up to try with fewer of them
455 did not always back up over a full grapheme when characters that do not
456 have the modifier quality were involved, e.g. Hangul syllables.
457
458 (b) If the match point in a subject started with modifier character, and
459 there was no match, the code could incorrectly back up beyond the match
460 point, and potentially beyond the first character in the subject,
461 leading to a segfault or an incorrect match result.
462
463 10. A conditional group with an assertion condition could lead to PCRE
464 recording an incorrect first data item for a match if no other first data
465 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
466 first data item, and therefore matched "ca" after "c" instead of at the
467 start.
468
469 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
470 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
471
472 12. The source of pcregrep now includes z/OS-specific code so that it can be
473 compiled for z/OS as part of the special z/OS distribution.
474
475 13. Added the -T and -TM options to pcretest.
476
477 14. The code in pcre_compile.c for creating the table of named capturing groups
478 has been refactored. Instead of creating the table dynamically during the
479 actual compiling pass, the information is remembered during the pre-compile
480 pass (on the stack unless there are more than 20 named groups, in which
481 case malloc() is used) and the whole table is created before the actual
482 compile happens. This has simplified the code (it is now nearly 150 lines
483 shorter) and prepared the way for better handling of references to groups
484 with duplicate names.
485
486 15. A back reference to a named subpattern when there is more than one of the
487 same name now checks them in the order in which they appear in the pattern.
488 The first one that is set is used for the reference. Previously only the
489 first one was inspected. This change makes PCRE more compatible with Perl.
490
491 16. Unicode character properties were updated from Unicode 6.3.0.
492
493 17. The compile-time code for auto-possessification has been refactored, based
494 on a patch by Zoltan Herczeg. It now happens after instead of during
495 compilation. The code is cleaner, and more cases are handled. The option
496 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
497 options in pcretest are provided to set it. It can also be set by
498 (*NO_AUTO_POSSESS) at the start of a pattern.
499
500 18. The character VT has been added to the default ("C" locale) set of
501 characters that match \s and are generally treated as white space,
502 following this same change in Perl 5.18. There is now no difference between
503 "Perl space" and "POSIX space". Whether VT is treated as white space in
504 other locales depends on the locale.
505
506 19. The code for checking named groups as conditions, either for being set or
507 for being recursed, has been refactored (this is related to 14 and 15
508 above). Processing unduplicated named groups should now be as fast at
509 numerical groups, and processing duplicated groups should be faster than
510 before.
511
512 20. Two patches to the CMake build system, by Alexander Barkov:
513
514 (1) Replace the "source" command by "." in CMakeLists.txt because
515 "source" is a bash-ism.
516
517 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
518 without these the CMake build does not work on Solaris.
519
520 21. Perl has changed its handling of \8 and \9. If there is no previously
521 encountered capturing group of those numbers, they are treated as the
522 literal characters 8 and 9 instead of a binary zero followed by the
523 literals. PCRE now does the same.
524
525 22. Following Perl, added \o{} to specify codepoints in octal, making it
526 possible to specify values greater than 0777 and also making them
527 unambiguous.
528
529 23. Perl now gives an error for missing closing braces after \x{... instead of
530 treating the string as literal. PCRE now does the same.
531
532 24. RunTest used to grumble if an inappropriate test was selected explicitly,
533 but just skip it when running all tests. This make it awkward to run ranges
534 of tests when one of them was inappropriate. Now it just skips any
535 inappropriate tests, as it always did when running all tests.
536
537 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
538 character types such as \d or \w, too many callouts were inserted, and the
539 data that they returned was rubbish.
540
541 26. In UCP mode, \s was not matching two of the characters that Perl matches,
542 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
543 were matched by \h. The code has now been refactored so that the lists of
544 the horizontal and vertical whitespace characters used for \h and \v (which
545 are defined only in one place) are now also used for \s.
546
547 27. Add JIT support for the 64 bit TileGX architecture.
548 Patch by Jiong Wang (Tilera Corporation).
549
550 28. Possessive quantifiers for classes (both explicit and automatically
551 generated) now use special opcodes instead of wrapping in ONCE brackets.
552
553 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
554 (because it's meaningless), this was not happening when PCRE_CASELESS was
555 set. Not wrong, but inefficient.
556
557 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
558 properties for \w, \d, etc) is present in a test regex. Otherwise if the
559 test contains no characters greater than 255, Perl doesn't realise it
560 should be using Unicode semantics.
561
562 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
563 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
564 does in Unicode mode.
565
566 32. Added the "forbid" facility to pcretest so that putting tests into the
567 wrong test files can sometimes be quickly detected.
568
569 33. There is now a limit (default 250) on the depth of nesting of parentheses.
570 This limit is imposed to control the amount of system stack used at compile
571 time. It can be changed at build time by --with-parens-nest-limit=xxx or
572 the equivalent in CMake.
573
574 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
575 errors. Perl warns for these when in warning mode, but PCRE has no facility
576 for giving warnings.
577
578 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
579 However, this was not working for (?!) because it is optimized to (*FAIL),
580 for which PCRE does not allow quantifiers. The optimization is now disabled
581 when a quantifier follows (?!). I can't see any use for this, but it makes
582 things uniform.
583
584 36. Perl no longer allows group names to start with digits, so I have made this
585 change also in PCRE. It simplifies the code a bit.
586
587 37. In extended mode, Perl ignores spaces before a + that indicates a
588 possessive quantifier. PCRE allowed a space before the quantifier, but not
589 before the possessive +. It now does.
590
591 38. The use of \K (reset reported match start) within a repeated possessive
592 group such as (a\Kb)*+ was not working.
593
594 40. Document that the same character tables must be used at compile time and
595 run time, and that the facility to pass tables to pcre_exec() and
596 pcre_dfa_exec() is for use only with saved/restored patterns.
597
598 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
599 features for Builds with MSVC:
600
601 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
602 builds). The .rc files can be used to set FileDescription and many other
603 attributes.
604
605 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
606 This allows higher-level build scripts which want .pdb files to avoid
607 hard-coding the exact files needed."
608
609 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
610 mean "start of word" and "end of word", respectively, as a transition aid.
611
612 43. A minimizing repeat of a class containing codepoints greater than 255 in
613 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
614 compiled to use the heap for recursion.
615
616 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
617 is configured.
618
619
620 Version 8.33 28-May-2013
621 ------------------------
622
623 1. Added 'U' to some constants that are compared to unsigned integers, to
624 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
625 variables that are added to signed variables, to ensure the result is
626 signed and can be negated.
627
628 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
629 CMake config files.
630
631 3. Revise the creation of config.h.generic so that all boolean macros are
632 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
633 overriding via -D on the command line possible.
634
635 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
636 to unsigned int is reported to make a quite noticeable speed difference in
637 a specific Windows environment. Testing on Linux did also appear to show
638 some benefit (and it is clearly not harmful). Also fixed the definition of
639 Xop which should be unsigned.
640
641 5. Related to (4), changing the definition of the intermediate variable cc
642 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
643 improvements.
644
645 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
646 unnecessary spaces.
647
648 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
649 and later.
650
651 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
652
653 9. Optimizing fast_forward_start_bits in JIT.
654
655 10. Adding support for callouts in JIT, and fixing some issues revealed
656 during this work. Namely:
657
658 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
659
660 (b) Minimum length was not checked before the matching is started.
661
662 11. The value of capture_last that is passed to callouts was incorrect in some
663 cases when there was a capture on one path that was subsequently abandoned
664 after a backtrack. Also, the capture_last value is now reset after a
665 recursion, since all captures are also reset in this case.
666
667 12. The interpreter no longer returns the "too many substrings" error in the
668 case when an overflowing capture is in a branch that is subsequently
669 abandoned after a backtrack.
670
671 13. In the pathological case when an offset vector of size 2 is used, pcretest
672 now prints out the matched string after a yield of 0 or 1.
673
674 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
675 Only supported by the JIT compiler at the moment.
676
677 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
678
679 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
680 offset of the starting point of the matching process, provided the offsets
681 vector is large enough.
682
683 17. The \A escape now records a lookbehind value of 1, though its execution
684 does not actually inspect the previous character. This is to ensure that,
685 in partial multi-segment matching, at least one character from the old
686 segment is retained when a new segment is processed. Otherwise, if there
687 are no lookbehinds in the pattern, \A might match incorrectly at the start
688 of a new segment.
689
690 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
691
692 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
693 gives some modest performance improvement in 8-bit mode.
694
695 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
696 be expressed in certain programming languages using Universal Character
697 Names.
698
699 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
700 which points out that "non characters" are not "characters that may not
701 appear in Unicode strings" but rather "characters that are reserved for
702 internal use and have only local meaning".
703
704 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
705 there was a conditional group that depended on an assertion, if the
706 assertion was false, the callout that immediately followed the alternation
707 in the condition was skipped when pcre_exec() was used for matching.
708
709 23. Allow an explicit callout to be inserted before an assertion that is the
710 condition for a conditional group, for compatibility with automatic
711 callouts, which always insert a callout at this point.
712
713 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
714 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
715
716 25. (*PRUNE) is now supported by the JIT compiler.
717
718 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
719
720 27. Fix the case where there are two or more SKIPs with arguments that may be
721 ignored.
722
723 28. (*SKIP) is now supported by the JIT compiler.
724
725 29. (*THEN) is now supported by the JIT compiler.
726
727 30. Update RunTest with additional test selector options.
728
729 31. The way PCRE handles backtracking verbs has been changed in two ways.
730
731 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
732 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
733 backtracking. In some cases this makes it more Perl-compatible, but Perl's
734 rather obscure rules do not always do the same thing.
735
736 (2) Previously, backtracking verbs were confined within assertions. This is
737 no longer the case for positive assertions, except for (*ACCEPT). Again,
738 this sometimes improves Perl compatibility, and sometimes does not.
739
740 32. A number of tests that were in test 2 because Perl did things differently
741 have been moved to test 1, because either Perl or PCRE has changed, and
742 these tests are now compatible.
743
744 32. Backtracking control verbs are now handled in the same way in JIT and
745 interpreter.
746
747 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
748 contained a forward subroutine reference caused a compile error.
749
750 34. Auto-detect and optimize limited repetitions in JIT.
751
752 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
753 blocking (*UTF) etc.
754
755 36. In the interpreter, maximizing pattern repetitions for characters and
756 character types now use tail recursion, which reduces stack usage.
757
758 37. The value of the max lookbehind was not correctly preserved if a compiled
759 and saved regex was reloaded on a host of different endianness.
760
761 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
762 of the compiled pattern block, expand the flags field from 16 to 32 bits
763 because it was almost full.
764
765 39. Try madvise first before posix_madvise.
766
767 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
768 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
769 original change locked out all matches of empty strings. This has been
770 changed so that one match of an empty string per line is recognized.
771 Subsequent searches on the same line (for colouring or for --only-matching,
772 for example) do not recognize empty strings.
773
774 41. Applied a user patch to fix a number of spelling mistakes in comments.
775
776 42. Data lines longer than 65536 caused pcretest to crash.
777
778 43. Clarified the data type for length and startoffset arguments for pcre_exec
779 and pcre_dfa_exec in the function-specific man pages, where they were
780 explicitly stated to be in bytes, never having been updated. I also added
781 some clarification to the pcreapi man page.
782
783 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
784 a segmentation fault.
785
786
787 Version 8.32 30-November-2012
788 -----------------------------
789
790 1. Improved JIT compiler optimizations for first character search and single
791 character iterators.
792
793 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
794 Patch by Daniel Richard G.
795
796 3. Single character iterator optimizations in the JIT compiler.
797
798 4. Improved JIT compiler optimizations for character ranges.
799
800 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
801 Reported by Giuseppe D'Angelo.
802
803 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
804 of a line, was being set incorrectly in cases where .* appeared inside
805 atomic brackets at the start of a pattern, or where there was a subsequent
806 *PRUNE or *SKIP.
807
808 7. Improved instruction cache flush for POWER/PowerPC.
809 Patch by Daniel Richard G.
810
811 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
812 grep:
813
814 (a) There is now no limit to the number of patterns to be matched.
815
816 (b) An error is given if a pattern is too long.
817
818 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
819 are now supported.
820
821 (d) --exclude-from and --include-from (multiple use) have been added.
822
823 (e) Exclusions and inclusions now apply to all files and directories, not
824 just to those obtained from scanning a directory recursively.
825
826 (f) Multiple uses of -f and --file-list are now supported.
827
828 (g) In a Windows environment, the default for -d has been changed from
829 "read" (the GNU grep default) to "skip", because otherwise the presence
830 of a directory in the file list provokes an error.
831
832 (h) The documentation has been revised and clarified in places.
833
834 9. Improve the matching speed of capturing brackets.
835
836 10. Changed the meaning of \X so that it now matches a Unicode extended
837 grapheme cluster.
838
839 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
840 out POSIX threads when JIT support is configured.
841
842 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
843
844 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
845 structure could go wrong in environments where size_t is not the same size
846 as int.
847
848 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
849 set.
850
851 15. The EBCDIC support had decayed; later updates to the code had included
852 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
853 general tidy up of EBCDIC-related issues, and the documentation was also
854 not quite right. There is now a test that can be run on ASCII systems to
855 check some of the EBCDIC-related things (but is it not a full test).
856
857 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
858 in a small tidy to the code.
859
860 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
861
862 18. If the --only-matching (-o) option in pcregrep is specified multiple
863 times, each one causes appropriate output. For example, -o1 -o2 outputs the
864 substrings matched by the 1st and 2nd capturing parentheses. A separating
865 string can be specified by --om-separator (default empty).
866
867 19. Improving the first n character searches.
868
869 20. Turn case lists for horizontal and vertical white space into macros so that
870 they are defined only once.
871
872 21. This set of changes together give more compatible Unicode case-folding
873 behaviour for characters that have more than one other case when UCP
874 support is available.
875
876 (a) The Unicode property table now has offsets into a new table of sets of
877 three or more characters that are case-equivalent. The MultiStage2.py
878 script that generates these tables (the pcre_ucd.c file) now scans
879 CaseFolding.txt instead of UnicodeData.txt for character case
880 information.
881
882 (b) The code for adding characters or ranges of characters to a character
883 class has been abstracted into a generalized function that also handles
884 case-independence. In UTF-mode with UCP support, this uses the new data
885 to handle characters with more than one other case.
886
887 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
888 whose other case is greater than 256 are now correctly matched
889 caselessly. Previously, the high codepoint matched the low one, but not
890 vice versa.
891
892 (d) The processing of \h, \H, \v, and \ in character classes now makes use
893 of the new class addition function, using character lists defined as
894 macros alongside the case definitions of 20 above.
895
896 (e) Caseless back references now work with characters that have more than
897 one other case.
898
899 (f) General caseless matching of characters with more than one other case
900 is supported.
901
902 22. Unicode character properties were updated from Unicode 6.2.0
903
904 23. Improved CMake support under Windows. Patch by Daniel Richard G.
905
906 24. Add support for 32-bit character strings, and UTF-32
907
908 25. Major JIT compiler update (code refactoring and bugfixing).
909 Experimental Sparc 32 support is added.
910
911 26. Applied a modified version of Daniel Richard G's patch to create
912 pcre.h.generic and config.h.generic by "make" instead of in the
913 PrepareRelease script.
914
915 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
916 pcre_compile.c when checking for a zero character.
917
918 28. Introducing a native interface for JIT. Through this interface, the compiled
919 machine code can be directly executed. The purpose of this interface is to
920 provide fast pattern matching, so several sanity checks are not performed.
921 However, feature tests are still performed. The new interface provides
922 1.4x speedup compared to the old one.
923
924 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
925 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
926 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
927
928 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
929 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
930 the "old" RFC 2279). Instead, it ended up passing a negative length to
931 pcre_exec().
932
933 31. Add support for GCC's visibility feature to hide internal functions.
934
935 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
936 "unknown -C option" after outputting 0 or 1.
937
938 33. There is now support for generating a code coverage report for the test
939 suite in environments where gcc is the compiler and lcov is installed. This
940 is mainly for the benefit of the developers.
941
942 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
943 unaddressable using valgrind annotations, allowing valgrind to detect
944 invalid memory accesses. This is mainly for the benefit of the developers.
945
946 25. (*UTF) can now be used to start a pattern in any of the three libraries.
947
948 26. Give configure error if --enable-cpp but no C++ compiler found.
949
950
951 Version 8.31 06-July-2012
952 -------------------------
953
954 1. Fixing a wrong JIT test case and some compiler warnings.
955
956 2. Removed a bashism from the RunTest script.
957
958 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
959 to unsigned type, result still unsigned" that was given by an MS compiler
960 on encountering the code "-sizeof(xxx)".
961
962 4. Partial matching support is added to the JIT compiler.
963
964 5. Fixed several bugs concerned with partial matching of items that consist
965 of more than one character:
966
967 (a) /^(..)\1/ did not partially match "aba" because checking references was
968 done on an "all or nothing" basis. This also applied to repeated
969 references.
970
971 (b) \R did not give a hard partial match if \r was found at the end of the
972 subject.
973
974 (c) \X did not give a hard partial match after matching one or more
975 characters at the end of the subject.
976
977 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
978 a partial match for the string "\r".
979
980 (e) When newline was set to CRLF, the metacharacter "." did not recognize
981 a partial match for a CR character at the end of the subject string.
982
983 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
984 running pcretest, the text "(JIT)" added to the output whenever JIT is
985 actually used to run the match.
986
987 7. Individual JIT compile options can be set in pcretest by following -s+[+]
988 or /S+[+] with a digit between 1 and 7.
989
990 8. OP_NOT now supports any UTF character not just single-byte ones.
991
992 9. (*MARK) control verb is now supported by the JIT compiler.
993
994 10. The command "./RunTest list" lists the available tests without actually
995 running any of them. (Because I keep forgetting what they all are.)
996
997 11. Add PCRE_INFO_MAXLOOKBEHIND.
998
999 12. Applied a (slightly modified) user-supplied patch that improves performance
1000 when the heap is used for recursion (compiled with --disable-stack-for-
1001 recursion). Instead of malloc and free for each heap frame each time a
1002 logical recursion happens, frames are retained on a chain and re-used where
1003 possible. This sometimes gives as much as 30% improvement.
1004
1005 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1006 call.
1007
1008 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1009
1010 15. It is now possible to link pcretest with libedit as an alternative to
1011 libreadline.
1012
1013 16. (*COMMIT) control verb is now supported by the JIT compiler.
1014
1015 17. The Unicode data tables have been updated to Unicode 6.1.0.
1016
1017 18. Added --file-list option to pcregrep.
1018
1019 19. Added binary file support to pcregrep, including the -a, --binary-files,
1020 -I, and --text options.
1021
1022 20. The madvise function is renamed for posix_madvise for QNX compatibility
1023 reasons. Fixed by Giuseppe D'Angelo.
1024
1025 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1026
1027 22. Changed the option for creating symbolic links for 16-bit man pages from
1028 -s to -sf so that re-installing does not cause issues.
1029
1030 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1031
1032 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1033 matches in certain environments (the workspace was not being correctly
1034 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1035 some of the workspace data at the beginning of a restart.
1036
1037 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1038 was not doing so when it should - probably a typo introduced by SVN 528
1039 (change 8.10/14).
1040
1041 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1042 \w+ when the character tables indicated that \x{c4} was a word character.
1043 There were several related cases, all because the tests for doing a table
1044 lookup were testing for characters less than 127 instead of 255.
1045
1046 27. If a pattern contains capturing parentheses that are not used in a match,
1047 their slots in the ovector are set to -1. For those that are higher than
1048 any matched groups, this happens at the end of processing. In the case when
1049 there were back references that the ovector was too small to contain
1050 (causing temporary malloc'd memory to be used during matching), and the
1051 highest capturing number was not used, memory off the end of the ovector
1052 was incorrectly being set to -1. (It was using the size of the temporary
1053 memory instead of the true size.)
1054
1055 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1056 ovector size, it uses memory at the end of the block that it has got.
1057
1058 29. Check for an overlong MARK name and give an error at compile time. The
1059 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1060
1061 30. JIT compiler update.
1062
1063 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1064 Rill for the patch.
1065
1066 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1067
1068 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1069
1070 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1071 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1072 was enabled.
1073
1074 35. Improve JIT code generation for greedy plus quantifier.
1075
1076 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1077 1 to "aa" instead of to an empty string. The bug affected repeated groups
1078 that could potentially match an empty string.
1079
1080 37. Optimizing single character iterators in JIT.
1081
1082 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1083 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1084 codepoints that are too big for the mode are faulted, and in a UTF mode,
1085 disallowed codepoints are also faulted.
1086
1087 39. If PCRE was compiled with UTF support, in three places in the DFA
1088 matcher there was code that should only have been obeyed in UTF mode, but
1089 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1090 processing when bytes with values greater than 127 were present. In 16-bit
1091 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1092 both cases the values are those that cannot be the first data item in a UTF
1093 character. The three items that might have provoked this were recursions,
1094 possessively repeated groups, and atomic groups.
1095
1096 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1097 and pcregrep, because some OS require shared objects to be explicitly
1098 passed to ld, causing the link step to fail if they are not.
1099
1100 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1101 mode, patterns that started with \h* or \R* might be incorrectly matched.
1102
1103
1104 Version 8.30 04-February-2012
1105 -----------------------------
1106
1107 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1108 name is defined in ctype.h.
1109
1110 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1111 only in quite long subpatterns.
1112
1113 3. Removed the function pcre_info(), which has been obsolete and deprecated
1114 since it was replaced by pcre_fullinfo() in February 2000.
1115
1116 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1117 match a (*MARK), and the match failed at the start of the subject, a
1118 reference to memory before the start of the subject could occur. This bug
1119 was introduced by fix 17 of release 8.21.
1120
1121 5. A reference to an unset group with zero minimum repetition was giving
1122 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1123 /(another)?(\1?)test/ matched against "hello world test". This bug was
1124 introduced in release 8.13.
1125
1126 6. Add support for 16-bit character strings (a large amount of work involving
1127 many changes and refactorings).
1128
1129 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1130 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1131 from a file.
1132
1133 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1134 rounding is not applied in this particular case).
1135
1136 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1137 if they appear, or are escaped, in patterns.
1138
1139 10. Get rid of a number of -Wunused-but-set-variable warnings.
1140
1141 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1142 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1143 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1144 also returns the mark "x". This bug applied to capturing parentheses,
1145 non-capturing parentheses, and atomic parentheses. It also applied to some
1146 assertions.
1147
1148 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1149 information out of configure.ac instead of relying on pcre.h.generic, which
1150 is not stored in the repository.
1151
1152 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1153 -lreadline.
1154
1155 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1156
1157 15. Applied Graycode's patch to put the top-level frame on the stack rather
1158 than the heap when not using the stack for recursion. This gives a
1159 performance improvement in many cases when recursion is not deep.
1160
1161 16. Experimental code added to "pcretest -C" to output the stack frame size.
1162
1163
1164 Version 8.21 12-Dec-2011
1165 ------------------------
1166
1167 1. Updating the JIT compiler.
1168
1169 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1170 are added as well.
1171
1172 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1173 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1174 calling _pcre_jit_exec. Some extra comments are added.
1175
1176 4. (*MARK) settings inside atomic groups that do not contain any capturing
1177 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1178 was introduced by change 18 for 8.20.
1179
1180 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1181 ECMA-262 standard.
1182
1183 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1184 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1185 This bug was probably introduced by change 9 of 8.13.
1186
1187 7. While fixing 6 above, I noticed that a number of other items were being
1188 incorrectly rejected as "not fixed length". This arose partly because newer
1189 opcodes had not been added to the fixed-length checking code. I have (a)
1190 corrected the bug and added tests for these items, and (b) arranged for an
1191 error to occur if an unknown opcode is encountered while checking for fixed
1192 length instead of just assuming "not fixed length". The items that were
1193 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1194 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1195 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1196
1197 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1198 being incorrectly compiled and would have given unpredicatble results.
1199
1200 9. A possessively repeated subpattern with minimum repeat count greater than
1201 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1202 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1203 the first (A) could occur when it should not.
1204
1205 10. Add a cast and remove a redundant test from the code.
1206
1207 11. JIT should use pcre_malloc/pcre_free for allocation.
1208
1209 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1210 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1211 is anything other than /usr, -L is still shown).
1212
1213 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1214
1215 14. Perl does not support \N without a following name in a [] class; PCRE now
1216 also gives an error.
1217
1218 15. If a forward reference was repeated with an upper limit of around 2000,
1219 it caused the error "internal error: overran compiling workspace". The
1220 maximum number of forward references (including repeats) was limited by the
1221 internal workspace, and dependent on the LINK_SIZE. The code has been
1222 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1223 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1224 of around 200,000 forward references. While doing this, I also speeded up
1225 the filling in of repeated forward references.
1226
1227 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1228 incorrectly expecting the subject to contain another "a" after the start.
1229
1230 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1231 in the match, the SKIP should be ignored. This was not happening; instead
1232 the SKIP was being treated as NOMATCH. For patterns such as
1233 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1234 tested.
1235
1236 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1237 now much more compatible with Perl, in particular in cases where the result
1238 is a non-match for a non-anchored pattern. For example, if
1239 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1240 "m", where previously it did not return a name. A side effect of this
1241 change is that for partial matches, the last encountered mark name is
1242 returned, as for non matches. A number of tests that were previously not
1243 Perl-compatible have been moved into the Perl-compatible test files. The
1244 refactoring has had the pleasing side effect of removing one argument from
1245 the match() function, thus reducing its stack requirements.
1246
1247 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1248 subsequent uses of /S (without +) incorrectly behaved like /S+.
1249
1250 21. Retrieve executable code size support for the JIT compiler and fixing
1251 some warnings.
1252
1253 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1254 not work when the shorter character appeared right at the end of the
1255 subject string.
1256
1257 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1258 systems.
1259
1260 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1261 output it when the /M option is used in pcretest.
1262
1263 25. The CheckMan script was not being included in the distribution. Also, added
1264 an explicit "perl" to run Perl scripts from the PrepareRelease script
1265 because this is reportedly needed in Windows.
1266
1267 26. If study data was being save in a file and studying had not found a set of
1268 "starts with" bytes for the pattern, the data written to the file (though
1269 never used) was taken from uninitialized memory and so caused valgrind to
1270 complain.
1271
1272 27. Updated RunTest.bat as provided by Sheri Pierce.
1273
1274 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1275
1276 29. Computation of memory usage for the table of capturing group names was
1277 giving an unnecessarily large value.
1278
1279
1280 Version 8.20 21-Oct-2011
1281 ------------------------
1282
1283 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1284 a POSIX class. After further experiments with Perl, which convinced me that
1285 Perl has bugs and confusions, a closing square bracket is no longer allowed
1286 in a POSIX name. This bug also affected patterns with classes that started
1287 with full stops.
1288
1289 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1290 captured substring, but while checking the failing first alternative,
1291 substring 1 is temporarily captured. If the output vector supplied to
1292 pcre_exec() was not big enough for this capture, the yield of the function
1293 was still zero ("insufficient space for captured substrings"). This cannot
1294 be totally fixed without adding another stack variable, which seems a lot
1295 of expense for a edge case. However, I have improved the situation in cases
1296 such as /(a)(b)x|abc/ matched against "abc", where the return code
1297 indicates that fewer than the maximum number of slots in the ovector have
1298 been set.
1299
1300 3. Related to (2) above: when there are more back references in a pattern than
1301 slots in the output vector, pcre_exec() uses temporary memory during
1302 matching, and copies in the captures as far as possible afterwards. It was
1303 using the entire output vector, but this conflicts with the specification
1304 that only 2/3 is used for passing back captured substrings. Now it uses
1305 only the first 2/3, for compatibility. This is, of course, another edge
1306 case.
1307
1308 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1309 main code base, and can be used by building with --enable-jit. When this is
1310 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1311 runtime --no-jit option is given.
1312
1313 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1314 ovector, the return from the function was zero, implying that there were
1315 other matches that did not fit. The correct "exactly full" value is now
1316 returned.
1317
1318 6. If a subpattern that was called recursively or as a subroutine contained
1319 (*PRUNE) or any other control that caused it to give a non-standard return,
1320 invalid errors such as "Error -26 (nested recursion at the same subject
1321 position)" or even infinite loops could occur.
1322
1323 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1324 computing the minimum length on reaching *ACCEPT, and so ended up with the
1325 wrong value of 1 rather than 0. Further investigation indicates that
1326 computing a minimum subject length in the presence of *ACCEPT is difficult
1327 (think back references, subroutine calls), and so I have changed the code
1328 so that no minimum is registered for a pattern that contains *ACCEPT.
1329
1330 8. If (*THEN) was present in the first (true) branch of a conditional group,
1331 it was not handled as intended. [But see 16 below.]
1332
1333 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1334 Sheri Pierce.
1335
1336 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1337 the first byte in a match must be "a".
1338
1339 11. Change 17 for 8.13 increased the recursion depth for patterns like
1340 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1341 pattern contains any instances of (*THEN). If it does not, the old
1342 optimizations are restored. It would be nice to do this on a per-group
1343 basis, but at the moment that is not feasible.
1344
1345 12. In some environments, the output of pcretest -C is CRLF terminated. This
1346 broke RunTest's code that checks for the link size. A single white space
1347 character after the value is now allowed for.
1348
1349 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1350 For "fr", it uses the Windows-specific input and output files.
1351
1352 14. If (*THEN) appeared in a group that was called recursively or as a
1353 subroutine, it did not work as intended. [But see next item.]
1354
1355 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1356 pattern fragments (but not containing any | characters). If A and B are
1357 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1358 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1359 D. In other words, Perl considers parentheses that do not contain any |
1360 characters to be part of a surrounding alternative, whereas PCRE was
1361 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1362 differently. PCRE now behaves in the same way as Perl, except in the case
1363 of subroutine/recursion calls such as (?1) which have in any case always
1364 been different (but PCRE had them first :-).
1365
1366 16. Related to 15 above: Perl does not treat the | in a conditional group as
1367 creating alternatives. Such a group is treated in the same way as an
1368 ordinary group without any | characters when processing (*THEN). PCRE has
1369 been changed to match Perl's behaviour.
1370
1371 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1372 RunGrepTest script failed.
1373
1374 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1375 inevitable for groups that contain captures, but it can lead to a lot of
1376 stack use in large patterns. The old behaviour has been restored for atomic
1377 groups that do not contain any capturing parentheses.
1378
1379 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1380 suppress the check for a minimum subject length at run time. (If it was
1381 given to pcre_exec() or pcre_dfa_exec() it did work.)
1382
1383 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1384 fail to work when decoding hex characters in data strings in EBCDIC
1385 environments.
1386
1387 21. It appears that in at least one Mac OS environment, the isxdigit() function
1388 is implemented as a macro that evaluates to its argument more than once,
1389 contravening the C 90 Standard (I haven't checked a later standard). There
1390 was an instance in pcretest which caused it to go wrong when processing
1391 \x{...} escapes in subject strings. The has been rewritten to avoid using
1392 things like p++ in the argument of isxdigit().
1393
1394
1395 Version 8.13 16-Aug-2011
1396 ------------------------
1397
1398 1. The Unicode data tables have been updated to Unicode 6.0.0.
1399
1400 2. Two minor typos in pcre_internal.h have been fixed.
1401
1402 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1403 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1404 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1405
1406 4. There were a number of related bugs in the code for matching backrefences
1407 caselessly in UTF-8 mode when codes for the characters concerned were
1408 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1409 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1410 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1411 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1412 2-byte code at the end of the subject (it thought there wasn't enough data
1413 left).
1414
1415 5. Comprehensive information about what went wrong is now returned by
1416 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1417 as the output vector has at least 2 elements. The offset of the start of
1418 the failing character and a reason code are placed in the vector.
1419
1420 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1421 now returned is for the first byte of the failing character, instead of the
1422 last byte inspected. This is an incompatible change, but I hope it is small
1423 enough not to be a problem. It makes the returned offset consistent with
1424 pcre_exec() and pcre_dfa_exec().
1425
1426 7. pcretest now gives a text phrase as well as the error number when
1427 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1428 failure, the offset and reason code are output.
1429
1430 8. When \R was used with a maximizing quantifier it failed to skip backwards
1431 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1432 back over a single character (\n). This seems wrong (because it treated the
1433 two characters as a single entity when going forwards), conflicts with the
1434 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1435 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1436 has been changed.
1437
1438 9. Some internal refactoring has changed the processing so that the handling
1439 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1440 time (the PCRE_DOTALL option was changed this way some time ago: version
1441 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1442 which was always a bit of a fudge. It also means that there is one less
1443 argument for the match() function, which reduces its stack requirements
1444 slightly. This change also fixes an incompatibility with Perl: the pattern
1445 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1446
1447 10. More internal refactoring has drastically reduced the number of recursive
1448 calls to match() for possessively repeated groups such as (abc)++ when
1449 using pcre_exec().
1450
1451 11. While implementing 10, a number of bugs in the handling of groups were
1452 discovered and fixed:
1453
1454 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1455 (a|)*(?1) gave a compile-time internal error.
1456 ((a|)+)+ did not notice that the outer group could match an empty string.
1457 (^a|^)+ was not marked as anchored.
1458 (.*a|.*)+ was not marked as matching at start or after a newline.
1459
1460 12. Yet more internal refactoring has removed another argument from the match()
1461 function. Special calls to this function are now indicated by setting a
1462 value in a variable in the "match data" data block.
1463
1464 13. Be more explicit in pcre_study() instead of relying on "default" for
1465 opcodes that mean there is no starting character; this means that when new
1466 ones are added and accidentally left out of pcre_study(), testing should
1467 pick them up.
1468
1469 14. The -s option of pcretest has been documented for ages as being an old
1470 synonym of -m (show memory usage). I have changed it to mean "force study
1471 for every regex", that is, assume /S for every regex. This is similar to -i
1472 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1473 using it. It makes it easier to run collections of tests with and without
1474 study enabled, and thereby test pcre_study() more easily. All the standard
1475 tests are now run with and without -s (but some patterns can be marked as
1476 "never study" - see 20 below).
1477
1478 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1479 restoration of the capturing data to the outer values was not happening
1480 correctly.
1481
1482 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1483 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1484 pattern had matched an empty string, and so incorrectly returned a no
1485 match.
1486
1487 17. There was optimizing code for the last branch of non-capturing parentheses,
1488 and also for the obeyed branch of a conditional subexpression, which used
1489 tail recursion to cut down on stack usage. Unfortunately, now that there is
1490 the possibility of (*THEN) occurring in these branches, tail recursion is
1491 no longer possible because the return has to be checked for (*THEN). These
1492 two optimizations have therefore been removed. [But see 8.20/11 above.]
1493
1494 18. If a pattern containing \R was studied, it was assumed that \R always
1495 matched two bytes, thus causing the minimum subject length to be
1496 incorrectly computed because \R can also match just one byte.
1497
1498 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1499 was incorrectly computed.
1500
1501 20. If /S is present twice on a test pattern in pcretest input, it now
1502 *disables* studying, thereby overriding the use of -s on the command line
1503 (see 14 above). This is necessary for one or two tests to keep the output
1504 identical in both cases.
1505
1506 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1507 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1508
1509 22. When an atomic group that contained a capturing parenthesis was
1510 successfully matched, but the branch in which it appeared failed, the
1511 capturing was not being forgotten if a higher numbered group was later
1512 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1513 group 1 to "a", when in fact it should be unset. This applied to multi-
1514 branched capturing and non-capturing groups, repeated or not, and also to
1515 positive assertions (capturing in negative assertions does not happen
1516 in PCRE) and also to nested atomic groups.
1517
1518 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1519 subject after a captured substring, to make it easier to tell which of a
1520 number of identical substrings has been captured.
1521
1522 24. The way atomic groups are processed by pcre_exec() has been changed so that
1523 if they are repeated, backtracking one repetition now resets captured
1524 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1525 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1526 "aaa". Previously, it would have been "a". As part of this code
1527 refactoring, the way recursive calls are handled has also been changed.
1528
1529 25. If an assertion condition captured any substrings, they were not passed
1530 back unless some other capturing happened later. For example, if
1531 (?(?=(a))a) was matched against "a", no capturing was returned.
1532
1533 26. When studying a pattern that contained subroutine calls or assertions,
1534 the code for finding the minimum length of a possible match was handling
1535 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1536 group 1 called group 2 while simultaneously a separate group 2 called group
1537 1). A stack overflow occurred in this case. I have fixed this by limiting
1538 the recursion depth to 10.
1539
1540 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1541 Fortmann. This supports explicit test numbers on the command line, and has
1542 argument validation and error reporting.
1543
1544 28. An instance of \X with an unlimited repeat could fail if at any point the
1545 first character it looked at was a mark character.
1546
1547 29. Some minor code refactoring concerning Unicode properties and scripts
1548 should reduce the stack requirement of match() slightly.
1549
1550 30. Added the '=' option to pcretest to check the setting of unused capturing
1551 slots at the end of the pattern, which are documented as being -1, but are
1552 not included in the return count.
1553
1554 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1555 compiled something random. Now it gives a compile-time error (as does
1556 Perl).
1557
1558 32. A *MARK encountered during the processing of a positive assertion is now
1559 recorded and passed back (compatible with Perl).
1560
1561 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1562 had alternative anchored branches, the search for a second match in a line
1563 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1564 matched the line "0102" twice. The same bug affected patterns that started
1565 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1566 twice.
1567
1568 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1569 does, and because of capturing effects, quantifying parenthesized
1570 assertions may at times be useful. Quantifiers are now allowed for
1571 parenthesized assertions.
1572
1573 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1574
1575 36. \g was being checked for fancy things in a character class, when it should
1576 just be a literal "g".
1577
1578 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1579 appearance of a nested POSIX class supersedes an apparent external class.
1580 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1581 unescaped square brackets may also appear as part of class names. For
1582 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1583 more like Perl. (But see 8.20/1 above.)
1584
1585 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1586 was because it thought it was \N{name}, which is not supported).
1587
1588 39. Add minix to OS list not supporting the -S option in pcretest.
1589
1590 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1591 cannot analyze patterns in sufficient detail to catch mutual recursions
1592 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1593 subgroup is called recursively as a subpattern for a second time at the
1594 same position in the subject string. In previous releases this might have
1595 been caught by the recursion limit, or it might have run out of stack.
1596
1597 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1598 happen only once. PCRE was, however incorrectly giving a compile time error
1599 "recursive call could loop indefinitely" because it cannot analyze the
1600 pattern in sufficient detail. The compile time test no longer happens when
1601 PCRE is compiling a conditional subpattern, but actual runaway loops are
1602 now caught at runtime (see 40 above).
1603
1604 42. It seems that Perl allows any characters other than a closing parenthesis
1605 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1606 has been changed to be the same.
1607
1608 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1609 as not to get warnings when autogen.sh is called. Also changed
1610 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1611
1612 44. To help people who use pcregrep to scan files containing exceedingly long
1613 lines, the following changes have been made:
1614
1615 (a) The default value of the buffer size parameter has been increased from
1616 8K to 20K. (The actual buffer used is three times this size.)
1617
1618 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1619 PCRE is built.
1620
1621 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1622 to be set at run time.
1623
1624 (d) Numerical values in pcregrep options can be followed by K or M, for
1625 example --buffer-size=50K.
1626
1627 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1628 given and the return code is set to 2.
1629
1630 45. Add a pointer to the latest mark to the callout data block.
1631
1632 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1633 partial match of an empty string instead of no match. This was specific to
1634 the use of ".".
1635
1636 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1637 complete match instead of a partial match. This bug was dependent on both
1638 the PCRE_UTF8 and PCRE_DOTALL options being set.
1639
1640 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1641 starting byte set, because \b was not being ignored.
1642
1643
1644 Version 8.12 15-Jan-2011
1645 ------------------------
1646
1647 1. Fixed some typos in the markup of the man pages, and wrote a script that
1648 checks for such things as part of the documentation building process.
1649
1650 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1651 --match-limit and --recursion-limit options (added for 8.11). In
1652 particular, this made one of the standard tests fail. (The integer value
1653 went into the wrong half of a long int.)
1654
1655 3. If the --colour option was given to pcregrep with -v (invert match), it
1656 did strange things, either producing crazy output, or crashing. It should,
1657 of course, ignore a request for colour when reporting lines that do not
1658 match.
1659
1660 4. Another pcregrep bug caused similar problems if --colour was specified with
1661 -M (multiline) and the pattern match finished with a line ending.
1662
1663 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1664 matched in multiline mode, the following line was shown as part of the
1665 match. This seems wrong, so I have changed it.
1666
1667 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1668 the check for further matches in the same line (so they could be coloured)
1669 to overrun the end of the current line. If another match was found, it was
1670 incorrectly shown (and then shown again when found in the next line).
1671
1672 7. If pcregrep was compiled under Windows, there was a reference to the
1673 function pcregrep_exit() before it was defined. I am assuming this was
1674 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1675 reported by a user. I've moved the definition above the reference.
1676
1677
1678 Version 8.11 10-Dec-2010
1679 ------------------------
1680
1681 1. (*THEN) was not working properly if there were untried alternatives prior
1682 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1683 backtracked to try for "b" instead of moving to the next alternative branch
1684 at the same level (in this case, to look for "c"). The Perl documentation
1685 is clear that when (*THEN) is backtracked onto, it goes to the "next
1686 alternative in the innermost enclosing group".
1687
1688 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1689 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1690 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1691 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1692 (*THEN).
1693
1694 3. If \s appeared in a character class, it removed the VT character from
1695 the class, even if it had been included by some previous item, for example
1696 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1697 of \s, but is part of the POSIX "space" class.)
1698
1699 4. A partial match never returns an empty string (because you can always
1700 match an empty string at the end of the subject); however the checking for
1701 an empty string was starting at the "start of match" point. This has been
1702 changed to the "earliest inspected character" point, because the returned
1703 data for a partial match starts at this character. This means that, for
1704 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1705 (previously it gave "no match").
1706
1707 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1708 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1709 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1710 has an implication that the given string is incomplete (because a partial
1711 match is preferred over a full match). For this reason, these items now
1712 give a partial match in this situation. [Aside: previously, the one case
1713 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1714 match rather than a full match, which was wrong by the old rules, but is
1715 now correct.]
1716
1717 6. There was a bug in the handling of #-introduced comments, recognized when
1718 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1719 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1720 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1721 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1722 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1723 places in pcre_compile().
1724
1725 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1726 comments when looking ahead for named forward references to subpatterns,
1727 the only newline sequence it recognized was NL. It now handles newlines
1728 according to the set newline convention.
1729
1730 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1731 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1732 cater for a lack of strerror(). These oversights have been fixed.
1733
1734 9. Added --match-limit and --recursion-limit to pcregrep.
1735
1736 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1737
1738 11. When the -o option was used, pcregrep was setting a return code of 1, even
1739 when matches were found, and --line-buffered was not being honoured.
1740
1741 12. Added an optional parentheses number to the -o and --only-matching options
1742 of pcregrep.
1743
1744 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1745 can match an empty string. The code to do it in pcretest and pcredemo
1746 needed fixing:
1747
1748 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1749 only one byte after an empty string match just before CRLF (this case
1750 just got forgotten; "any" and "anycrlf" were OK).
1751
1752 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1753 mode when an empty string match preceded an ASCII character followed by
1754 a non-ASCII character. (The code for advancing by one character rather
1755 than one byte was nonsense.)
1756
1757 (c) The pcredemo.c sample program did not have any code at all to handle
1758 the cases when CRLF is a valid newline sequence.
1759
1760 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1761 as a starting offset was within the subject string. There is now a new
1762 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1763 negative or greater than the length of the string. In order to test this,
1764 pcretest is extended to allow the setting of negative starting offsets.
1765
1766 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1767 starting offset points to the beginning of a UTF-8 character was
1768 unnecessarily clumsy. I tidied it up.
1769
1770 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1771 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1772
1773 17. Nobody had reported that the --include_dir option, which was added in
1774 release 7.7 should have been called --include-dir (hyphen, not underscore)
1775 for compatibility with GNU grep. I have changed it to --include-dir, but
1776 left --include_dir as an undocumented synonym, and the same for
1777 --exclude-dir, though that is not available in GNU grep, at least as of
1778 release 2.5.4.
1779
1780 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1781 characters from a string of bytes) have been redefined so as not to use
1782 loops, in order to improve performance in some environments. At the same
1783 time, I abstracted some of the common code into auxiliary macros to save
1784 repetition (this should not affect the compiled code).
1785
1786 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1787 compile-time error is now given if \c is not followed by an ASCII
1788 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1789 different, and any byte value is allowed.)
1790
1791 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1792 START_OPTIMIZE option, which is now allowed at compile time - but just
1793 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1794 to pcregrep and other applications that have no direct access to PCRE
1795 options. The new /Y option in pcretest sets this option when calling
1796 pcre_compile().
1797
1798 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1799 back references. Groups containing recursive back references were forced to
1800 be atomic by that change, but in the case of named groups, the amount of
1801 memory required was incorrectly computed, leading to "Failed: internal
1802 error: code overflow". This has been fixed.
1803
1804 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1805 pcretest.c, to avoid build problems in some Borland environments.
1806
1807
1808 Version 8.10 25-Jun-2010
1809 ------------------------
1810
1811 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1812 THEN.
1813
1814 2. (*ACCEPT) was not working when inside an atomic group.
1815
1816 3. Inside a character class, \B is treated as a literal by default, but
1817 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1818 causes the error). The code is unchanged, but I tidied the documentation.
1819
1820 4. Inside a character class, PCRE always treated \R and \X as literals,
1821 whereas Perl faults them if its -w option is set. I have changed PCRE so
1822 that it faults them when PCRE_EXTRA is set.
1823
1824 5. Added support for \N, which always matches any character other than
1825 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1826
1827 6. When compiling pcregrep with newer versions of gcc which may have
1828 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1829 declared with attribute warn_unused_result" were given. Just casting the
1830 result to (void) does not stop the warnings; a more elaborate fudge is
1831 needed. I've used a macro to implement this.
1832
1833 7. Minor change to pcretest.c to avoid a compiler warning.
1834
1835 8. Added four artifical Unicode properties to help with an option to make
1836 \s etc use properties (see next item). The new properties are: Xan
1837 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1838
1839 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1840 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1841 this option. Modified pcretest to add /W to test this facility. Added
1842 REG_UCP to make it available via the POSIX interface.
1843
1844 10. Added --line-buffered to pcregrep.
1845
1846 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1847 studied, and the match started with a letter with a code point greater than
1848 127 whose first byte was different to the first byte of the other case of
1849 the letter, the other case of this starting letter was not recognized
1850 (#976).
1851
1852 12. If a pattern that was studied started with a repeated Unicode property
1853 test, for example, \p{Nd}+, there was the theoretical possibility of
1854 setting up an incorrect bitmap of starting bytes, but fortunately it could
1855 not have actually happened in practice until change 8 above was made (it
1856 added property types that matched character-matching opcodes).
1857
1858 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1859 possible starting bytes for non-anchored patterns.
1860
1861 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1862 \R, and also a number of cases that involve Unicode properties, both
1863 explicit and implicit when PCRE_UCP is set.
1864
1865 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1866 input, it could crash or give wrong results if characters with values
1867 greater than 0xc0 were present in the subject string. (Detail: it assumed
1868 UTF-8 input when processing these items.)
1869
1870 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1871 size_t is 64-bit (#991).
1872
1873 17. Added a check for running out of memory when PCRE is compiled with
1874 --disable-stack-for-recursion (#990).
1875
1876 18. If the last data line in a file for pcretest does not have a newline on
1877 the end, a newline was missing in the output.
1878
1879 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1880 less than 128) in its various bitmaps. However, there is a facility for
1881 generating tables according to the current locale when PCRE is compiled. It
1882 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1883 characters, are recognized by isspace() and therefore were getting set in
1884 these tables, and indeed these tables seem to approximate to ISO 8859. This
1885 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1886 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1887 which of course cannot start UTF-8 characters. I have changed the code so
1888 that only real ASCII characters (less than 128) and the correct starting
1889 bytes for UTF-8 encodings are set for characters greater than 127 when in
1890 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1891 altogether.)
1892
1893 20. Added the /T option to pcretest so as to be able to run tests with non-
1894 standard character tables, thus making it possible to include the tests
1895 used for 19 above in the standard set of tests.
1896
1897 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1898 reference to a subpattern the other side of a comment that contains an
1899 opening parenthesis caused either an internal compiling error, or a
1900 reference to the wrong subpattern.
1901
1902
1903 Version 8.02 19-Mar-2010
1904 ------------------------
1905
1906 1. The Unicode data tables have been updated to Unicode 5.2.0.
1907
1908 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1909 configured.
1910
1911 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1912 original author of that file, following a query about its status.
1913
1914 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1915 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1916
1917 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1918 quantifier applied to a forward-referencing subroutine call, could compile
1919 incorrect code or give the error "internal error: previously-checked
1920 referenced subpattern not found".
1921
1922 6. Both MS Visual Studio and Symbian OS have problems with initializing
1923 variables to point to external functions. For these systems, therefore,
1924 pcre_malloc etc. are now initialized to local functions that call the
1925 relevant global functions.
1926
1927 7. There were two entries missing in the vectors called coptable and poptable
1928 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1929 I've fixed the data, and added a kludgy way of testing at compile time that
1930 the lengths are correct (equal to the number of opcodes).
1931
1932 8. Following on from 7, I added a similar kludge to check the length of the
1933 eint vector in pcreposix.c.
1934
1935 9. Error texts for pcre_compile() are held as one long string to avoid too
1936 much relocation at load time. To find a text, the string is searched,
1937 counting zeros. There was no check for running off the end of the string,
1938 which could happen if a new error number was added without updating the
1939 string.
1940
1941 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1942
1943 11. \K was not working if it appeared in an atomic group or in a group that
1944 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1945 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1946 the assertion is positive, but not if it is negative.
1947
1948 12. Change 11 fortuitously reduced the size of the stack frame used in the
1949 "match()" function of pcre_exec.c by one pointer. Forthcoming
1950 implementation of support for (*MARK) will need an extra pointer on the
1951 stack; I have reserved it now, so that the stack frame size does not
1952 decrease.
1953
1954 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1955 item in branch that calls a recursion is a subroutine call - as in the
1956 second branch in the above example - was incorrectly given the compile-
1957 time error "recursive call could loop indefinitely" because pcre_compile()
1958 was not correctly checking the subroutine for matching a non-empty string.
1959
1960 14. The checks for overrunning compiling workspace could trigger after an
1961 overrun had occurred. This is a "should never occur" error, but it can be
1962 triggered by pathological patterns such as hundreds of nested parentheses.
1963 The checks now trigger 100 bytes before the end of the workspace.
1964
1965 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1966
1967
1968 Version 8.01 19-Jan-2010
1969 ------------------------
1970
1971 1. If a pattern contained a conditional subpattern with only one branch (in
1972 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1973 computed the wrong minimum data length (which is of course zero for such
1974 subpatterns). This could cause incorrect "no match" results.
1975
1976 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1977 the pattern is reset in the first branch, pcre_compile() failed with
1978 "internal error: code overflow at offset...". This happened only when
1979 the reset was to the original external option setting. (An optimization
1980 abstracts leading options settings into an external setting, which was the
1981 cause of this.)
1982
1983 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1984 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1985 assertion pattern did not match (meaning that the assertion was true), it
1986 was incorrectly treated as false if the SKIP had been reached during the
1987 matching. This also applied to assertions used as conditions.
1988
1989 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1990 assertion subpattern, including such a pattern used as a condition,
1991 unpredictable results occurred, instead of the error return
1992 PCRE_ERROR_DFA_UITEM.
1993
1994 5. The C++ GlobalReplace function was not working like Perl for the special
1995 situation when an empty string is matched. It now does the fancy magic
1996 stuff that is necessary.
1997
1998 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1999 removed. (These were left over from very, very early versions of PCRE.)
2000
2001 7. Some cosmetic changes to the code to make life easier when compiling it
2002 as part of something else:
2003
2004 (a) Change DEBUG to PCRE_DEBUG.
2005
2006 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2007 called "current" as "current_branch", to prevent a collision with the
2008 Linux macro when compiled as a kernel module.
2009
2010 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2011 prevent a collision with the Linux macro when compiled as a kernel
2012 module.
2013
2014 8. In pcre_compile() there are some checks for integer overflows that used to
2015 cast potentially large values to (double). This has been changed to that
2016 when building, a check for int64_t is made, and if it is found, it is used
2017 instead, thus avoiding the use of floating point arithmetic. (There is no
2018 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2019 double.
2020
2021 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2022 2005 (difference between two addresses compared to an unsigned value).
2023
2024 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2025 custom one, because of the following reported problem in Windows:
2026
2027 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2028 under Win32.
2029 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2030 therefore missing the function definition.
2031 - The compiler thus generates a "C" signature for the test function.
2032 - The linker fails to find the "C" function.
2033 - PCRE fails to configure if asked to do so against libbz2.
2034
2035 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2036 messages were output:
2037
2038 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2039 rerunning libtoolize, to keep the correct libtool macros in-tree.
2040 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2041
2042 I have done both of these things.
2043
2044 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2045 most of the time, it *can* run out if it is given a pattern that contains a
2046 runaway infinite recursion. I updated the discussion in the pcrestack man
2047 page.
2048
2049 13. Now that we have gone to the x.xx style of version numbers, the minor
2050 version may start with zero. Using 08 or 09 is a bad idea because users
2051 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2052 interpreted as invalid octal numbers. I've updated the previous comment in
2053 configure.ac, and also added a check that gives an error if 08 or 09 are
2054 used.
2055
2056 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2057 causing partial matching to fail when the end of the subject matched \W
2058 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2059
2060 15. There were some discrepancies between the declarations in pcre_internal.h
2061 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2062 their definitions. The declarations used "const uschar *" and the
2063 definitions used USPTR. Even though USPTR is normally defined as "const
2064 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2065 reported that: "This difference in casting confuses some C++ compilers, for
2066 example, SunCC recognizes above declarations as different functions and
2067 generates broken code for hbpcre." I have changed the declarations to use
2068 USPTR.
2069
2070 16. GNU libtool is named differently on some systems. The autogen.sh script now
2071 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2072 (FreeBSD).
2073
2074 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2075 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2076 comment: "Figure out how to create a longlong from a string: strtoll and
2077 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2078 instance, but it only takes 2 args instead of 3!"
2079
2080 18. A subtle bug concerned with back references has been fixed by a change of
2081 specification, with a corresponding code fix. A pattern such as
2082 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2083 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2084 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2085 same bug. Such groups have to be quantified to be useful, or contained
2086 inside another quantified group. (If there's no repetition, the reference
2087 can never match.) The problem arises because, having left the group and
2088 moved on to the rest of the pattern, a later failure that backtracks into
2089 the group uses the captured value from the final iteration of the group
2090 rather than the correct earlier one. I have fixed this in PCRE by forcing
2091 any group that contains a reference to itself to be an atomic group; that
2092 is, there cannot be any backtracking into it once it has completed. This is
2093 similar to recursive and subroutine calls.
2094
2095
2096 Version 8.00 19-Oct-09
2097 ----------------------
2098
2099 1. The table for translating pcre_compile() error codes into POSIX error codes
2100 was out-of-date, and there was no check on the pcre_compile() error code
2101 being within the table. This could lead to an OK return being given in
2102 error.
2103
2104 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2105 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2106 in a Windows environment.
2107
2108 3. The pcregrep --count option prints the count for each file even when it is
2109 zero, as does GNU grep. However, pcregrep was also printing all files when
2110 --files-with-matches was added. Now, when both options are given, it prints
2111 counts only for those files that have at least one match. (GNU grep just
2112 prints the file name in this circumstance, but including the count seems
2113 more useful - otherwise, why use --count?) Also ensured that the
2114 combination -clh just lists non-zero counts, with no names.
2115
2116 4. The long form of the pcregrep -F option was incorrectly implemented as
2117 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2118 but it seems right to fix it, and I didn't think it was worth preserving
2119 the old behaviour.
2120
2121 5. The command line items --regex=pattern and --regexp=pattern were not
2122 recognized by pcregrep, which required --regex pattern or --regexp pattern
2123 (with a space rather than an '='). The man page documented the '=' forms,
2124 which are compatible with GNU grep; these now work.
2125
2126 6. No libpcreposix.pc file was created for pkg-config; there was just
2127 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2128
2129 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2130 when UCP support is not needed, by modifying the Python script that
2131 generates it from Unicode data files. This should not matter if the module
2132 is correctly used as a library, but I received one complaint about 50K of
2133 unwanted data. My guess is that the person linked everything into his
2134 program rather than using a library. Anyway, it does no harm.
2135
2136 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2137 was a minimum greater than 1 for a wide character in a possessive
2138 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2139 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2140 character. Chaos in the form of incorrect output or a compiling loop could
2141 result.
2142
2143 9. The restrictions on what a pattern can contain when partial matching is
2144 requested for pcre_exec() have been removed. All patterns can now be
2145 partially matched by this function. In addition, if there are at least two
2146 slots in the offset vector, the offset of the earliest inspected character
2147 for the match and the offset of the end of the subject are set in them when
2148 PCRE_ERROR_PARTIAL is returned.
2149
2150 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2151 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2152 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2153 and may be more useful for multi-segment matching.
2154
2155 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2156 used to be given if ever the end of the subject was reached; now it is
2157 given only if matching could not proceed because another character was
2158 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2159 string "Z", which now yields "no match" instead of "partial match". In the
2160 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2161 final character ended with (*FAIL).
2162
2163 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2164 if the pattern had a "must contain" character that was already found in the
2165 earlier partial match, unless partial matching was again requested. For
2166 example, with the pattern /dog.(body)?/, the "must contain" character is
2167 "g". If the first part-match was for the string "dog", restarting with
2168 "sbody" failed. This bug has been fixed.
2169
2170 13. The string returned by pcre_dfa_exec() after a partial match has been
2171 changed so that it starts at the first inspected character rather than the
2172 first character of the match. This makes a difference only if the pattern
2173 starts with a lookbehind assertion or \b or \B (\K is not supported by
2174 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2175 matching functions compatible, and I think it's the right thing to do.
2176
2177 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2178 so that the demonstration program is easily available in environments where
2179 PCRE has not been installed from source.
2180
2181 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2182 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2183 library.
2184
2185 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2186 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2187 is not the first non-POSIX option to be added. Clearly some people find
2188 these options useful.
2189
2190 17. If a caller to the POSIX matching function regexec() passes a non-zero
2191 value for nmatch with a NULL value for pmatch, the value of
2192 nmatch is forced to zero.
2193
2194 18. RunGrepTest did not have a test for the availability of the -u option of
2195 the diff command, as RunTest does. It now checks in the same way as
2196 RunTest, and also checks for the -b option.
2197
2198 19. If an odd number of negated classes containing just a single character
2199 interposed, within parentheses, between a forward reference to a named
2200 subpattern and the definition of the subpattern, compilation crashed with
2201 an internal error, complaining that it could not find the referenced
2202 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2203 [The bug was that it was starting one character too far in when skipping
2204 over the character class, thus treating the ] as data rather than
2205 terminating the class. This meant it could skip too much.]
2206
2207 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2208 /g option in pcretest when the pattern contains \K, which makes it possible
2209 to have an empty string match not at the start, even when the pattern is
2210 anchored. Updated pcretest and pcredemo to use this option.
2211
2212 21. If the maximum number of capturing subpatterns in a recursion was greater
2213 than the maximum at the outer level, the higher number was returned, but
2214 with unset values at the outer level. The correct (outer level) value is
2215 now given.
2216
2217 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2218 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2219 make it do so. The string so far is captured, making this feature
2220 compatible with Perl.
2221
2222 23. The tests have been re-organized, adding tests 11 and 12, to make it
2223 possible to check the Perl 5.10 features against Perl 5.10.
2224
2225 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2226 pattern matches a fixed length string. PCRE did not allow this; now it
2227 does. Neither allows recursion.
2228
2229 25. I finally figured out how to implement a request to provide the minimum
2230 length of subject string that was needed in order to match a given pattern.
2231 (It was back references and recursion that I had previously got hung up
2232 on.) This code has now been added to pcre_study(); it finds a lower bound
2233 to the length of subject needed. It is not necessarily the greatest lower
2234 bound, but using it to avoid searching strings that are too short does give
2235 some useful speed-ups. The value is available to calling programs via
2236 pcre_fullinfo().
2237
2238 26. While implementing 25, I discovered to my embarrassment that pcretest had
2239 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2240 study optimizations had never been tested with that matching function.
2241 Oops. What is worse, even when it was passed study data, there was a bug in
2242 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2243 were also very few tests of studied patterns with pcre_dfa_exec().
2244
2245 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2246 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2247 on the other side of the coin, they are no longer allowed to have different
2248 names, because these cannot be distinguished in PCRE, and this has caused
2249 confusion. (This is a difference from Perl.)
2250
2251 28. When duplicate subpattern names are present (necessarily with different
2252 numbers, as required by 27 above), and a test is made by name in a
2253 conditional pattern, either for a subpattern having been matched, or for
2254 recursion in such a pattern, all the associated numbered subpatterns are
2255 tested, and the overall condition is true if the condition is true for any
2256 one of them. This is the way Perl works, and is also more like the way
2257 testing by number works.
2258
2259
2260 Version 7.9 11-Apr-09
2261 ---------------------
2262
2263 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2264 (pcretest), all targets were linked against these libraries. This included
2265 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2266 libraries. This caused unwanted dependencies to be created. This problem
2267 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2268 pcretest is linked with readline.
2269
2270 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2271 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2272 moved outside it again, because FALSE and TRUE are already defined in AIX,
2273 but BOOL is not.
2274
2275 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2276 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2277
2278 4. The pcregrep documentation said spaces were inserted as well as colons (or
2279 hyphens) following file names and line numbers when outputting matching
2280 lines. This is not true; no spaces are inserted. I have also clarified the
2281 wording for the --colour (or --color) option.
2282
2283 5. In pcregrep, when --colour was used with -o, the list of matching strings
2284 was not coloured; this is different to GNU grep, so I have changed it to be
2285 the same.
2286
2287 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2288 each matching line was coloured. Now it goes on to look for further matches
2289 of any of the test patterns, which is the same behaviour as GNU grep.
2290
2291 7. A pattern that could match an empty string could cause pcregrep to loop; it
2292 doesn't make sense to accept an empty string match in pcregrep, so I have
2293 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2294 seems to be how GNU grep behaves. [But see later change 40 for release
2295 8.33.]
2296
2297 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2298 start or after a newline", because the conditional assertion was not being
2299 correctly handled. The rule now is that both the assertion and what follows
2300 in the first alternative must satisfy the test.
2301
2302 9. If auto-callout was enabled in a pattern with a conditional group whose
2303 condition was an assertion, PCRE could crash during matching, both with
2304 pcre_exec() and pcre_dfa_exec().
2305
2306 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2307 used for matching.
2308
2309 11. Unicode property support in character classes was not working for
2310 characters (bytes) greater than 127 when not in UTF-8 mode.
2311
2312 12. Added the -M command line option to pcretest.
2313
2314 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2315
2316 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2317
2318 16. Added comments and documentation about mis-use of no_arg in the C++
2319 wrapper.
2320
2321 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2322 from Martin Jerabek that uses macro names for all relevant character and
2323 string constants.
2324
2325 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2326 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2327 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2328 these, but not everybody uses configure.
2329
2330 19. A conditional group that had only one branch was not being correctly
2331 recognized as an item that could match an empty string. This meant that an
2332 enclosing group might also not be so recognized, causing infinite looping
2333 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2334 with the subject "ab", where knowledge that the repeated group can match
2335 nothing is needed in order to break the loop.
2336
2337 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2338 exec(), but without supplying a callout function, matching went wrong.
2339
2340 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2341 leak if the size of the offset vector was greater than 30. When the vector
2342 is smaller, the saved offsets during recursion go onto a local stack
2343 vector, but for larger vectors malloc() is used. It was failing to free
2344 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2345 error, in fact).
2346
2347 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2348 heapframe that is used only when UTF-8 support is enabled. This caused no
2349 problem, but was untidy.
2350
2351 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2352 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2353 included within another project.
2354
2355 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2356 slightly modified by me:
2357
2358 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2359 not building pcregrep.
2360
2361 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2362 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2363
2364 25. Forward references, both numeric and by name, in patterns that made use of
2365 duplicate group numbers, could behave incorrectly or give incorrect errors,
2366 because when scanning forward to find the reference group, PCRE was not
2367 taking into account the duplicate group numbers. A pattern such as
2368 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2369
2370 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2371 the feature of a custom pointer more persuasive (as requested by a user).
2372
2373 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2374 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2375 pre-defined.
2376
2377 28. Added support for (*UTF8) at the start of a pattern.
2378
2379 29. Arrange for flags added by the "release type" setting in CMake to be shown
2380 in the configuration summary.
2381
2382
2383 Version 7.8 05-Sep-08
2384 ---------------------
2385
2386 1. Replaced UCP searching code with optimized version as implemented for Ad
2387 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2388 stage table and inline lookup instead of a function, giving speed ups of 2
2389 to 5 times on some simple patterns that I tested. Permission was given to
2390 distribute the MultiStage2.py script that generates the tables (it's not in
2391 the tarball, but is in the Subversion repository).
2392
2393 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2394 scripts.
2395
2396 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2397 a group with a zero qualifier. The result of the study could be incorrect,
2398 or the function might crash, depending on the pattern.
2399
2400 4. Caseless matching was not working for non-ASCII characters in back
2401 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2402 It now works when Unicode Property Support is available.
2403
2404 5. In pcretest, an escape such as \x{de} in the data was always generating
2405 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2406 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2407 truncation.
2408
2409 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2410
2411 7. Added two (int) casts to pcregrep when printing the difference of two
2412 pointers, in case they are 64-bit values.
2413
2414 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2415 test 2 if it fails.
2416
2417 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2418 and a #define of that name to empty if it is not externally set. This is to
2419 allow users of MSVC to set it if necessary.
2420
2421 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2422 the convenience functions in the pcre_get.c source file.
2423
2424 11. An option change at the start of a pattern that had top-level alternatives
2425 could cause overwriting and/or a crash. This command provoked a crash in
2426 some environments:
2427
2428 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2429
2430 This potential security problem was recorded as CVE-2008-2371.
2431
2432 12. For a pattern where the match had to start at the beginning or immediately
2433 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2434 pcre_dfa_exec() could read past the end of the passed subject if there was
2435 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2436 pcretest so that it places the subject at the end of its malloc-ed buffer.
2437
2438 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2439 exec() might read past the end of the data buffer in UTF-8 mode.
2440
2441 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2442 the data contained the byte 0x85 as part of a UTF-8 character within its
2443 first line. This applied both to normal and DFA matching.
2444
2445 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2446 /^[^d]*?$/8 failed to match "abc".
2447
2448 16. Added a missing copyright notice to pcrecpp_internal.h.
2449
2450 17. Make it more clear in the documentation that values returned from
2451 pcre_exec() in ovector are byte offsets, not character counts.
2452
2453 18. Tidied a few places to stop certain compilers from issuing warnings.
2454
2455 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2456 supplied by Stefan Weber. I made a further small update for 7.8 because
2457 there is a change of source arrangements: the pcre_searchfuncs.c module is
2458 replaced by pcre_ucd.c.
2459
2460
2461 Version 7.7 07-May-08
2462 ---------------------
2463
2464 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2465 a string to a long long, pretend we don't even have a long long." This is
2466 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2467
2468 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2469 pre-7.6 versions, which defined a global no_arg variable instead of putting
2470 it in the RE class. (See also #8 below.)
2471
2472 3. Remove a line of dead code, identified by coverity and reported by Nuno
2473 Lopes.
2474
2475 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2476
2477 (1) The include/exclude patterns were being applied to the whole pathnames
2478 of files, instead of just to the final components.
2479
2480 (2) If there was more than one level of directory, the subdirectories were
2481 skipped unless they satisfied the include/exclude conditions. This is
2482 inconsistent with GNU grep (and could even be seen as contrary to the
2483 pcregrep specification - which I improved to make it absolutely clear).
2484 The action now is always to scan all levels of directory, and just
2485 apply the include/exclude patterns to regular files.
2486
2487 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2488 --exclude_dir in the tests to avoid scanning .svn directories.
2489
2490 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2491 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2492 doesn't support NULs in patterns.
2493
2494 7. Added some missing "const"s to declarations of static tables in
2495 pcre_compile.c and pcre_dfa_exec.c.
2496
2497 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2498 caused by fix #2 above. (Subsequently also a second patch to fix the
2499 first patch. And a third patch - this was a messy problem.)
2500
2501 9. Applied Craig's patch to remove the use of push_back().
2502
2503 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2504 matching function regexec().
2505
2506 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2507 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2508 references. PCRE supports relative numbers with this syntax (I don't think
2509 Oniguruma does).
2510
2511 12. Previously, a group with a zero repeat such as (...){0} was completely
2512 omitted from the compiled regex. However, this means that if the group
2513 was called as a subroutine from elsewhere in the pattern, things went wrong
2514 (an internal error was given). Such groups are now left in the compiled
2515 pattern, with a new opcode that causes them to be skipped at execution
2516 time.
2517
2518 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2519 to the way PCRE behaves:
2520
2521 (a) A lone ] character is dis-allowed (Perl treats it as data).
2522
2523 (b) A back reference to an unmatched subpattern matches an empty string
2524 (Perl fails the current match path).
2525
2526 (c) A data ] in a character class must be notated as \] because if the
2527 first data character in a class is ], it defines an empty class. (In
2528 Perl it is not possible to have an empty class.) The empty class []
2529 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2530 The negative empty class [^] matches any one character, independently
2531 of the DOTALL setting.
2532
2533 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2534 non-existent subpattern following a character class starting with ']' and
2535 containing () gave an internal compiling error instead of "reference to
2536 non-existent subpattern". Fortunately, when the pattern did exist, the
2537 compiled code was correct. (When scanning forwards to check for the
2538 existence of the subpattern, it was treating the data ']' as terminating
2539 the class, so got the count wrong. When actually compiling, the reference
2540 was subsequently set up correctly.)
2541
2542 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2543 it was being rejected as not supported by pcre_dfa_exec(), even though
2544 other assertions are supported. I have made pcre_dfa_exec() support
2545 (*FAIL).
2546
2547 16. The implementation of 13c above involved the invention of a new opcode,
2548 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2549 cannot be changed at match time, I realized I could make a small
2550 improvement to matching performance by compiling OP_ALLANY instead of
2551 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2552 on the OP_ANY path.
2553
2554 17. Compiling pcretest on Windows with readline support failed without the
2555 following two fixes: (1) Make the unistd.h include conditional on
2556 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2557
2558 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2559 ncurses library to be included for pcretest when ReadLine support is
2560 requested, but also to allow for it to be overridden. This patch came from
2561 Daniel Bergström.
2562
2563 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2564 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2565 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2566 spotting this.
2567
2568
2569 Version 7.6 28-Jan-08
2570 ---------------------
2571
2572 1. A character class containing a very large number of characters with
2573 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2574 overflow.
2575
2576 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2577 HAVE_LONG_LONG is not defined.
2578
2579 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2580 bring them up to date and include new features. This patch includes:
2581
2582 - Fixed PH's badly added libz and libbz2 support.
2583 - Fixed a problem with static linking.
2584 - Added pcredemo. [But later removed - see 7 below.]
2585 - Fixed dftables problem and added an option.
2586 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2587 HAVE_LONG_LONG.
2588 - Added readline support for pcretest.
2589 - Added an listing of the option settings after cmake has run.
2590
2591 4. A user submitted a patch to Makefile that makes it easy to create
2592 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2593 Makefile.am that cause it to include this special target, without
2594 affecting anything else. Note that the same mingw target plus all
2595 the other distribution libraries and programs are now supported
2596 when configuring with CMake (see 6 below) instead of with
2597 Configure/Make.
2598
2599 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2600 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2601 exported in the Windows port". It has not yet been confirmed that the patch
2602 solves the problem, but it does no harm.
2603
2604 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2605 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2606 with CMake, and also correct the comment about stack recursion.
2607
2608 7. Remove the automatic building of pcredemo from the ./configure system and
2609 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2610 of a program that users should build themselves after PCRE is installed, so
2611 building it automatically is not really right. What is more, it gave
2612 trouble in some build environments.
2613
2614 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2615
2616
2617 Version 7.5 10-Jan-08
2618 ---------------------
2619
2620 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2621 values in parens when parsing an RE using the C++ wrapper."
2622
2623 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2624 Characters greater than 255 were excluded from the class instead of being
2625 included.
2626
2627 3. The same bug as (2) above applied to negated POSIX classes such as
2628 [:^space:].
2629
2630 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2631 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2632 I have changed it.
2633
2634 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2635 first named subpattern) and a construct such as (?&a) would reference the
2636 first named subpattern whose name started with "a" (in other words, the
2637 length check was missing). Both these problems are fixed. "Subpattern name
2638 expected" is now given for (?&) (a zero-length name), and this patch also
2639 makes it give the same error for \k'' (previously it complained that that
2640 was a reference to a non-existent subpattern).
2641
2642 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2643 this is right because (?- can be followed by option settings as well as by
2644 digits. I have, however, made the messages clearer.
2645
2646 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2647 than the number used in the conditional) now cause a compile-time error.
2648 This is actually not compatible with Perl, which accepts such patterns, but
2649 treats the conditional as always being FALSE (as PCRE used to), but it
2650 seems to me that giving a diagnostic is better.
2651
2652 8. Change "alphameric" to the more common word "alphanumeric" in comments
2653 and messages.
2654
2655 9. Fix two occurrences of "backslash" in comments that should have been
2656 "backspace".
2657
2658 10. Remove two redundant lines of code that can never be obeyed (their function
2659 was moved elsewhere).
2660
2661 11. The program that makes PCRE's Unicode character property table had a bug
2662 which caused it to generate incorrect table entries for sequences of
2663 characters that have the same character type, but are in different scripts.
2664 It amalgamated them into a single range, with the script of the first of
2665 them. In other words, some characters were in the wrong script. There were
2666 thirteen such cases, affecting characters in the following ranges:
2667
2668 U+002b0 - U+002c1
2669 U+0060c - U+0060d
2670 U+0061e - U+00612
2671 U+0064b - U+0065e
2672 U+0074d - U+0076d
2673 U+01800 - U+01805
2674 U+01d00 - U+01d77
2675 U+01d9b - U+01dbf
2676 U+0200b - U+0200f
2677 U+030fc - U+030fe
2678 U+03260 - U+0327f
2679 U+0fb46 - U+0fbb1
2680 U+10450 - U+1049d
2681
2682 12. The -o option (show only the matching part of a line) for pcregrep was not
2683 compatible with GNU grep in that, if there was more than one match in a
2684 line, it showed only the first of them. It now behaves in the same way as
2685 GNU grep.
2686
2687 13. If the -o and -v options were combined for pcregrep, it printed a blank
2688 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2689 does the same. The return code can be used to tell if there were any
2690 non-matching lines.
2691
2692 14. Added --file-offsets and --line-offsets to pcregrep.
2693
2694 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2695 infinitely looping recursion. The bug was that positive lookaheads were not
2696 being skipped when checking for a possible empty match (negative lookaheads
2697 and both kinds of lookbehind were skipped).
2698
2699 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2700 inclusion of <windows.h> to before rather than after the definition of
2701 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2702
2703 17. Specifying a possessive quantifier with a specific limit for a Unicode
2704 character property caused pcre_compile() to compile bad code, which led at
2705 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2706 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2707 caused the error; without that there was no problem.
2708
2709 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2710
2711 19. Added --enable-pcretest-libreadline.
2712
2713 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2714 RE::GlobalReplace(). As a result, the number of replacements returned was
2715 double what it should be. I removed one of the increments, but Craig sent a
2716 later patch that removed the other one (the right fix) and added unit tests
2717 that check the return values (which was not done before).
2718
2719 21. Several CMake things:
2720
2721 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2722 the names libpcre and libpcreposix, not just pcre and pcreposix.
2723
2724 (2) The above change means that pcretest and pcregrep are now correctly
2725 linked with the newly-built libraries, not previously installed ones.
2726
2727 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2728
2729 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2730 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2731 UTF-8 newline character). The key issue is that the pattern starts .*;
2732 this means that the match must be either at the beginning, or after a
2733 newline. The bug was in the code for advancing after a failed match and
2734 checking that the new position followed a newline. It was not taking
2735 account of UTF-8 characters correctly.
2736
2737 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2738 character classes. PCRE was not treating the sequence [:...:] as a
2739 character class unless the ... were all letters. Perl, however, seems to
2740 allow any characters between [: and :], though of course it rejects as
2741 unknown any "names" that contain non-letters, because all the known class
2742 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2743 for example, whereas PCRE did not - it did not recognize a POSIX character
2744 class. This seemed a bit dangerous, so the code has been changed to be
2745 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2746 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2747 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2748 Perl does, and where it didn't before.
2749
2750 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2751 Windows environments %n is disabled by default.
2752
2753
2754 Version 7.4 21-Sep-07
2755 ---------------------
2756
2757 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2758 means that a class such as [\s] counted as "explicit reference to CR or
2759 LF". That isn't really right - the whole point of the change was to try to
2760 help when there was an actual mention of one of the two characters. So now
2761 the change happens only if \r or \n (or a literal CR or LF) character is
2762 encountered.
2763
2764 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2765 of both had grown to the point where there were only 3 bits left.
2766 Fortunately, there was spare space in the data structure, and so I have
2767 moved the internal flags into a new 16-bit field to free up more option
2768 bits.
2769
2770 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2771 but did not set the internal JCHANGED flag - either of these is enough to
2772 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2773 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2774 start sets both bits.
2775
2776 4. Added options (at build time, compile time, exec time) to change \R from
2777 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2778
2779 5. doc/pcresyntax.html was missing from the distribution.
2780
2781 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2782 compatibility, even though it is no longer used.
2783
2784 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2785 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2786 windows.h file is present (where different names are used). [This was
2787 reversed later after testing - see 16 below.]
2788
2789 8. Changed all #include <config.h> to #include "config.h". There were also
2790 some further <pcre.h> cases that I changed to "pcre.h".
2791
2792 9. When pcregrep was used with the --colour option, it missed the line ending
2793 sequence off the lines that it output.
2794
2795 10. It was pointed out to me that arrays of string pointers cause lots of
2796 relocations when a shared library is dynamically loaded. A technique of
2797 using a single long string with a table of offsets can drastically reduce
2798 these. I have refactored PCRE in four places to do this. The result is
2799 dramatic:
2800
2801 Originally: 290
2802 After changing UCP table: 187
2803 After changing error message table: 43
2804 After changing table of "verbs" 36
2805 After changing table of Posix names 22
2806
2807 Thanks to the folks working on Gregex for glib for this insight.
2808
2809 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2810 unicode-properties was also set.
2811
2812 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2813
2814 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2815 checked only for CRLF.
2816
2817 14. Added casts to pcretest.c to avoid compiler warnings.
2818
2819 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2820
2821 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2822 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2823 entirely. This removes changes made in 7 above.
2824
2825 17. The CMake files have been updated, and there is now more information about
2826 building with CMake in the NON-UNIX-USE document.
2827
2828
2829 Version 7.3 28-Aug-07
2830 ---------------------
2831
2832 1. In the rejigging of the build system that eventually resulted in 7.1, the
2833 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2834 brackets there is not right, since it causes compilers to look for an
2835 installed pcre.h, not the version that is in the source that is being
2836 compiled (which of course may be different). I have changed it back to:
2837
2838 #include "pcre.h"
2839
2840 I have a vague recollection that the change was concerned with compiling in
2841 different directories, but in the new build system, that is taken care of
2842 by the VPATH setting the Makefile.
2843
2844 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2845 when the subject happened to end in the byte 0x85 (e.g. if the last
2846 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2847 characters but of course it shouldn't be taken as a newline when it is part
2848 of another character. The bug was that, for an unlimited repeat of . in
2849 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2850 characters when looking for a newline.
2851
2852 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2853
2854 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2855 in debug output.
2856
2857 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2858 long printing in the pcrecpp unittest when running under MinGW.
2859
2860 6. ESC_K was left out of the EBCDIC table.
2861
2862 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2863 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2864 limit also applies to "virtual nesting" when a pattern is recursive, and in
2865 this case 1000 isn't so big. I have been able to remove this limit at the
2866 expense of backing off one optimization in certain circumstances. Normally,
2867 when pcre_exec() would call its internal match() function recursively and
2868 immediately return the result unconditionally, it uses a "tail recursion"
2869 feature to save stack. However, when a subpattern that can match an empty
2870 string has an unlimited repetition quantifier, it no longer makes this
2871 optimization. That gives it a stack frame in which to save the data for
2872 checking that an empty string has been matched. Previously this was taken
2873 from the 1000-entry workspace that had been reserved. So now there is no
2874 explicit limit, but more stack is used.
2875
2876 8. Applied Daniel's patches to solve problems with the import/export magic
2877 syntax that is required for Windows, and which was going wrong for the
2878 pcreposix and pcrecpp parts of the library. These were overlooked when this
2879 problem was solved for the main library.
2880
2881 9. There were some crude static tests to avoid integer overflow when computing
2882 the size of patterns that contain repeated groups with explicit upper
2883 limits. As the maximum quantifier is 65535, the maximum group length was
2884 set at 30,000 so that the product of these two numbers did not overflow a
2885 32-bit integer. However, it turns out that people want to use groups that
2886 are longer than 30,000 bytes (though not repeat them that many times).
2887 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2888 made it possible to implement the integer overflow checks in a much more
2889 dynamic way, which I have now done. The artificial limitation on group
2890 length has been removed - we now have only the limit on the total length of
2891 the compiled pattern, which depends on the LINK_SIZE setting.
2892
2893 10. Fixed a bug in the documentation for get/copy named substring when
2894 duplicate names are permitted. If none of the named substrings are set, the
2895 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2896 empty string.
2897
2898 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2899 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2900 because the ] is interpreted as the first data character and the
2901 terminating ] is not found. PCRE has been made compatible with Perl in this
2902 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2903 cause memory overwriting.
2904
2905 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2906 string has been matched (to stop an infinite loop). It was not recognizing
2907 a conditional subpattern that could match an empty string if that
2908 subpattern was within another subpattern. For example, it looped when
2909 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2910 condition was not nested. This bug has been fixed.
2911
2912 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2913 past the start of the subject in the presence of bytes with the top bit
2914 set, for example "\x8aBCD".
2915
2916 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2917 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2918
2919 14. Optimized (?!) to (*FAIL).
2920
2921 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2922 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2923 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2924 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2925 does: it's just the validity check that is more restrictive.
2926
2927 16. Inserted checks for integer overflows during escape sequence (backslash)
2928 processing, and also fixed erroneous offset values for syntax errors during
2929 backslash processing.
2930
2931 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2932 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2933
2934 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2935 caused an overrun.
2936
2937 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2938 something other than just ASCII characters) inside a group that had an
2939 unlimited repeat caused a loop at compile time (while checking to see
2940 whether the group could match an empty string).
2941
2942 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2943 [\P{Any}] did so. (Error in the code for printing property names.)
2944
2945 21. An orphan \E inside a character class could cause a crash.
2946
2947 22. A repeated capturing bracket such as (A)? could cause a wild memory
2948 reference during compilation.
2949
2950 23. There are several functions in pcre_compile() that scan along a compiled
2951 expression for various reasons (e.g. to see if it's fixed length for look
2952 behind). There were bugs in these functions when a repeated \p or \P was
2953 present in the pattern. These operators have additional parameters compared
2954 with \d, etc, and these were not being taken into account when moving along
2955 the compiled data. Specifically:
2956
2957 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2958 length.
2959
2960 (b) An item such as \pL+ within a repeated group could cause crashes or
2961 loops.
2962
2963 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2964 "reference to non-existent subpattern" error.
2965
2966 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2967
2968 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2969 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2970
2971 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2972
2973 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2974 character were causing crashes (broken optimization).
2975
2976 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2977 \p or \P) caused a compile-time loop.
2978
2979 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2980 break. For example, the unstudied pattern [\r\n]A does not match the string
2981 "\r\nA" because change 7.0/46 below moves the current point on by two
2982 characters after failing to match at the start. However, the pattern \nA
2983 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2984 the same is true. There doesn't seem any very clean way out of this, but
2985 what I have chosen to do makes the common cases work: PCRE now takes note
2986 of whether there can be an explicit match for \r or \n anywhere in the
2987 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2988 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2989 pattern has explicit CR or LF references.
2990
2991 29. Added (*CR) etc for changing newline setting at start of pattern.
2992
2993
2994 Version 7.2 19-Jun-07
2995 ---------------------
2996
2997 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2998 which is apparently normally available under Windows.
2999
3000 2. Re-jig the pcregrep tests with different newline settings in an attempt
3001 to make them independent of the local environment's newline setting.
3002
3003 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3004
3005 4. Some of the "internals" tests were previously cut out when the link size
3006 was not 2, because the output contained actual offsets. The recent new
3007 "Z" feature of pcretest means that these can be cut out, making the tests
3008 usable with all link sizes.
3009
3010 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3011 stack recursion. This gives a massive performance boost under BSD, but just
3012 a small improvement under Linux. However, it saves one field in the frame
3013 in all cases.
3014
3015 6. Added more features from the forthcoming Perl 5.10:
3016
3017 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3018 recursion call. It refers to the nth most recently opened parentheses.
3019
3020 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3021 to be opened parentheses.
3022
3023 (c) Conditions that refer to capturing parentheses can be specified
3024 relatively, for example, (?(-2)... or (?(+3)...
3025
3026 (d) \K resets the start of the current match so that everything before
3027 is not part of it.
3028
3029 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3030
3031 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3032 reference syntax.
3033
3034 (g) (?| introduces a group in which the numbering of parentheses in each
3035 alternative starts with the same number.
3036
3037 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3038
3039 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3040 PCRE_INFO_JCHANGED.
3041
3042 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3043 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3044 for detecting groups that can match an empty string.
3045
3046 9. A pattern with a very large number of alternatives (more than several
3047 hundred) was running out of internal workspace during the pre-compile
3048 phase, where pcre_compile() figures out how much memory will be needed. A
3049 bit of new cunning has reduced the workspace needed for groups with
3050 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3051 workspace instead of running out of the 4096 that are available.
3052
3053 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3054
3055 11. Applied patch from Google to remove an optimization that didn't quite work.
3056 The report of the bug said:
3057
3058 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3059 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3060 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3061
3062 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3063 it matched the wrong number of bytes.
3064
3065
3066 Version 7.1 24-Apr-07
3067 ---------------------
3068
3069 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3070 that is more "standard", making use of automake and other Autotools. There
3071 is some re-arrangement of the files and adjustment of comments consequent
3072 on this.
3073
3074 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3075 for recursive directory scanning broke on some systems because the files
3076 are not scanned in any specific order and on different systems the order
3077 was different. A call to "sort" has been inserted into RunGrepTest for the
3078 approprate test as a short-term fix. In the longer term there may be an
3079 alternative.
3080
3081 3. I had an email from Eric Raymond about problems translating some of PCRE's
3082 man pages to HTML (despite the fact that I distribute HTML pages, some
3083 people do their own conversions for various reasons). The problems
3084 concerned the use of low-level troff macros .br and .in. I have therefore
3085 removed all such uses from the man pages (some were redundant, some could
3086 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3087 HTML has been updated to handle .nf/.fi and to complain if it encounters
3088 .br or .in.
3089
3090 4. Updated comments in configure.ac that get placed in config.h.in and also
3091 arranged for config.h to be included in the distribution, with the name
3092 config.h.generic, for the benefit of those who have to compile without
3093 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3094
3095 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3096 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3097 makevp.bat for the current PCRE, using the additional files
3098 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3099
3100 6. A Windows user reported a minor discrepancy with test 2, which turned out
3101 to be caused by a trailing space on an input line that had got lost in his
3102 copy. The trailing space was an accident, so I've just removed it.
3103
3104 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3105 that is needed.
3106
3107 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3108 as "const" (a) because they are and (b) because it helps the PHP
3109 maintainers who have recently made a script to detect big data structures
3110 in the php code that should be moved to the .rodata section. I remembered
3111 to update Builducptable as well, so it won't revert if ucptable.h is ever
3112 re-created.
3113
3114 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3115 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3116 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3117 support is not required. This saves 1.5-2K of code, which is important in
3118 some applications.
3119
3120 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3121 so as not to refer to the tables, even though these functions will never be
3122 called when UTF-8 support is disabled. Otherwise there are problems with a
3123 shared library.
3124
3125 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3126
3127 (a) It was defining its arguments as char * instead of void *.
3128
3129 (b) It was assuming that all moves were upwards in memory; this was true
3130 a long time ago when I wrote it, but is no longer the case.
3131
3132 The emulated memove() is provided for those environments that have neither
3133 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3134 is clearly not the case, as these two bugs were recently reported.
3135
3136 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3137 and Detrail to create the HTML documentation, the .txt form of the man
3138 pages, and it removes trailing spaces from listed files. It also creates
3139 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3140 case, it wraps all the #defines with #ifndefs. This script should be run
3141 before "make dist".
3142
3143 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3144 with Unicode property support.
3145
3146 (a) For a maximizing quantifier, if the two different cases of the
3147 character were of different lengths in their UTF-8 codings (there are
3148 some cases like this - I found 11), and the matching function had to
3149 back up over a mixture of the two cases, it incorrectly assumed they
3150 were both the same length.
3151
3152 (b) When PCRE was configured to use the heap rather than the stack for
3153 recursion during matching, it was not correctly preserving the data for
3154 the other case of a UTF-8 character when checking ahead for a match
3155 while processing a minimizing repeat. If the check also involved
3156 matching a wide character, but failed, corruption could cause an
3157 erroneous result when trying to check for a repeat of the original
3158 character.
3159
3160 13. Some tidying changes to the testing mechanism:
3161
3162 (a) The RunTest script now detects the internal link size and whether there
3163 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3164 values substituted by "configure". (The RunGrepTest script already did
3165 this for UTF-8.) The configure.ac script no longer substitutes the
3166 relevant variables.
3167
3168 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3169 with length and offset values. This means that the output is different
3170 for different internal link sizes. Test 2 is skipped for link sizes
3171 other than 2 because of this, bypassing the problem. Unfortunately,
3172 there was also a test in test 3 (the locale tests) that used /B and
3173 failed for link sizes other than 2. Rather than cut the whole test out,
3174 I have added a new /Z option to pcretest that replaces the length and
3175 offset values with spaces. This is now used to make test 3 independent
3176 of link size. (Test 2 will be tidied up later.)
3177
3178 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3179 segmentation fault instead of returning the appropriate error message.
3180
3181 15. In multiline mode when the newline sequence was set to "any", the pattern
3182 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3183 This doesn't seem right; it now treats the CRLF combination as the line
3184 ending, and so does not match in that case. It's only a pattern such as ^$
3185 that would hit this one: something like ^ABC$ would have failed after \r
3186 and then tried again after \r\n.
3187
3188 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3189 in an attempt to make files that differ only in their line terminators
3190 compare equal. This works on Linux.
3191
3192 17. Under certain error circumstances pcregrep might try to free random memory
3193 as it exited. This is now fixed, thanks to valgrind.
3194
3195 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3196 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3197 was because its rules for how to advance for /g after matching an empty
3198 string at the end of a line did not allow for this case. They now check for
3199 it specially.
3200
3201 20. pcretest is supposed to handle patterns and data of any length, by
3202 extending its buffers when necessary. It was getting this wrong when the
3203 buffer for a data line had to be extended.
3204
3205 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3206 CRLF as a newline sequence.
3207
3208 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3209 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3210 I have nevertheless tidied it up.
3211
3212 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3213
3214 24. Added a man page for pcre-config.
3215
3216
3217 Version 7.0 19-Dec-06
3218 ---------------------
3219
3220 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3221 moving to gcc 4.1.1.
3222
3223 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3224 sys/time.h, which is documented as needed for this function. It doesn't
3225 seem to matter on Linux, but it showed up on some releases of OS X.
3226
3227 3. It seems that there are systems where bytes whose values are greater than
3228 127 match isprint() in the "C" locale. The "C" locale should be the
3229 default when a C program starts up. In most systems, only ASCII printing
3230 characters match isprint(). This difference caused the output from pcretest
3231 to vary, making some of the tests fail. I have changed pcretest so that:
3232
3233 (a) When it is outputting text in the compiled version of a pattern, bytes
3234 other than 32-126 are always shown as hex escapes.
3235
3236 (b) When it is outputting text that is a matched part of a subject string,
3237 it does the same, unless a different locale has been set for the match
3238 (using the /L modifier). In this case, it uses isprint() to decide.
3239
3240 4. Fixed a major bug that caused incorrect computation of the amount of memory
3241 required for a compiled pattern when options that changed within the
3242 pattern affected the logic of the preliminary scan that determines the
3243 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3244 that the computed length was too small. The symptoms of this bug were
3245 either the PCRE error "internal error: code overflow" from pcre_compile(),
3246 or a glibc crash with a message such as "pcretest: free(): invalid next
3247 size (fast)". Examples of patterns that provoked this bug (shown in
3248 pcretest format) are:
3249
3250 /(?-x: )/x
3251 /(?x)(?-x: \s*#\s*)/
3252 /((?i)[\x{c0}])/8
3253 /(?i:[\x{c0}])/8
3254
3255 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3256 is now done differently.
3257
3258 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3259 wrapper classes; (b) implement a new function in the C++ scanner that is
3260 more efficient than the old way of doing things because it avoids levels of
3261 recursion in the regex matching; (c) add a paragraph to the documentation
3262 for the FullMatch() function.
3263
3264 6. The escape sequence \n was being treated as whatever was defined as
3265 "newline". Not only was this contrary to the documentation, which states
3266 that \n is character 10 (hex 0A), but it also went horribly wrong when
3267 "newline" was defined as CRLF. This has been fixed.
3268
3269 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3270 was being set to -1 for the "end of line" case (supposedly a value that no
3271 character can have). Though this value is never used (the check for end of
3272 line is "zero bytes in current character"), it caused compiler complaints.
3273 I've changed it to 0xffffffff.
3274
3275 8. In pcre_version.c, the version string was being built by a sequence of
3276 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3277 string (as it is for production releases) called a macro with an empty
3278 argument. The C standard says the result of this is undefined. The gcc
3279 compiler treats it as an empty string (which was what was wanted) but it is
3280 reported that Visual C gives an error. The source has been hacked around to
3281 avoid this problem.
3282
3283 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3284 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3285 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3286 of them did).
3287
3288 10. Originally, pcretest opened its input and output without "b"; then I was
3289 told that "b" was needed in some environments, so it was added for release
3290 5.0 to both the input and output. (It makes no difference on Unix-like
3291 systems.) Later I was told that it is wrong for the input on Windows. I've
3292 now abstracted the modes into two macros, to make it easier to fiddle with
3293 them, and removed "b" from the input mode under Windows.
3294
3295 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3296
3297 12. Added -help and --help to pcretest as an official way of being reminded
3298 of the options.
3299
3300 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3301 and pcrecpp.cc because they annoy compilers at high warning levels.
3302
3303 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3304
3305 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3306 scripts are not C programs :-) and which was not noticed because it works
3307 on Linux.
3308
3309 16. pcretest is supposed to handle any length of pattern and data line (as one
3310 line or as a continued sequence of lines) by extending its input buffer if
3311 necessary. This feature was broken for very long pattern lines, leading to
3312 a string of junk being passed to pcre_compile() if the pattern was longer
3313 than about 50K.
3314
3315 17. I have done a major re-factoring of the way pcre_compile() computes the
3316 amount of memory needed for a compiled pattern. Previously, there was code
3317 that made a preliminary scan of the pattern in order to do this. That was
3318 OK when PCRE was new, but as the facilities have expanded, it has become
3319 harder and harder to keep it in step with the real compile phase, and there
3320 have been a number of bugs (see for example, 4 above). I have now found a
3321 cunning way of running the real compile function in a "fake" mode that
3322 enables it to compute how much memory it would need, while actually only
3323 ever using a few hundred bytes of working memory and without too many
3324 tests of the mode. This should make future maintenance and development
3325 easier. A side effect of this work is that the limit of 200 on the nesting
3326 depth of parentheses has been removed (though this was never a serious
3327 limitation, I suspect). However, there is a downside: pcre_compile() now
3328 runs more slowly than before (30% or more, depending on the pattern). I
3329 hope this isn't a big issue. There is no effect on runtime performance.
3330
3331 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3332 newline (only possible for the last line of a file) and it was a
3333 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3334
3335 19. Added additional timing features to pcretest. (1) The -tm option now times
3336 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3337 separate command line item, by a number that specifies the number of
3338 repeats to use when timing. The default is 50000; this gives better
3339 precision, but takes uncomfortably long for very large patterns.
3340
3341 20. Extended pcre_study() to be more clever in cases where a branch of a
3342 subpattern has no definite first character. For example, (a*|b*)[cd] would
3343 previously give no result from pcre_study(). Now it recognizes that the
3344 first character must be a, b, c, or d.
3345
3346 21. There was an incorrect error "recursive call could loop indefinitely" if
3347 a subpattern (or the entire pattern) that was being tested for matching an
3348 empty string contained only one non-empty item after a nested subpattern.
3349 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3350 incorrectly, because the \d was being skipped in the check.
3351
3352 22. The pcretest program now has a new pattern option /B and a command line
3353 option -b, which is equivalent to adding /B to every pattern. This causes
3354 it to show the compiled bytecode, without the additional information that
3355 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3356 is the same as /B/I).
3357
3358 23. A new optimization is now able automatically to treat some sequences such
3359 as a*b as a*+b. More specifically, if something simple (such as a character
3360 or a simple class like \d) has an unlimited quantifier, and is followed by
3361 something that cannot possibly match the quantified thing, the quantifier
3362 is automatically "possessified".
3363
3364 24. A recursive reference to a subpattern whose number was greater than 39
3365 went wrong under certain circumstances in UTF-8 mode. This bug could also
3366 have affected the operation of pcre_study().
3367
3368 25. Realized that a little bit of performance could be had by replacing
3369 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3370
3371 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3372
3373 27. Possessive quantifiers such as a++ were previously implemented by turning
3374 them into atomic groups such as ($>a+). Now they have their own opcodes,
3375 which improves performance. This includes the automatically created ones
3376 from 23 above.
3377
3378 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3379 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3380 the first matched character to be a colon. This applied both to named and
3381 numbered groups.
3382
3383 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3384
3385 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3386 building PCRE on Windows easier, so I have included it in the distribution.
3387
3388 31. There is now a check in pcretest against a ridiculously large number being
3389 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3390 loop, the loop is abandoned.
3391
3392 32. Forward references to subpatterns in conditions such as (?(2)...) where
3393 subpattern 2 is defined later cause pcre_compile() to search forwards in
3394 the pattern for the relevant set of parentheses. This search went wrong
3395 when there were unescaped parentheses in a character class, parentheses
3396 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3397
3398 33. "Subroutine" calls and backreferences were previously restricted to
3399 referencing subpatterns earlier in the regex. This restriction has now
3400 been removed.
3401
3402 34. Added a number of extra features that are going to be in Perl 5.10. On the
3403 whole, these are just syntactic alternatives for features that PCRE had
3404 previously implemented using the Python syntax or my own invention. The
3405 other formats are all retained for compatibility.
3406
3407 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3408 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3409 also .NET compatible.
3410
3411 (b) A recursion or subroutine call to a named group can now be defined as
3412 (?&name) as well as (?P>name).
3413
3414 (c) A backreference to a named group can now be defined as \k<name> or
3415 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3416 5.10, are also .NET compatible.
3417
3418 (d) A conditional reference to a named group can now use the syntax
3419 (?(<name>) or (?('name') as well as (?(name).
3420
3421 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3422 groups (named and numbered) that are never evaluated inline, but can be
3423 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3424 is always false. There may be only one alternative in such a group.
3425
3426 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3427 as the simple (?(R). The condition is true only if the most recent
3428 recursion is that of the given number or name. It does not search out
3429 through the entire recursion stack.
3430
3431 (g) The escape \gN or \g{N} has been added, where N is a positive or
3432 negative number, specifying an absolute or relative reference.
3433
3434 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3435 some "unreachable code" warnings.
3436
3437 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3438 things, this adds five new scripts.
3439
3440 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3441 There were also incompatibilities regarding the handling of \Q..\E inside
3442 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3443 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3444
3445 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3446 matches an empty string, and forcibly breaks the loop. There were bugs in
3447 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3448 against aaaa the result was just "a" rather than "aaaa", for example. Two
3449 separate and independent bugs (that affected different cases) have been
3450 fixed.
3451
3452 39. Refactored the code to abolish the use of different opcodes for small
3453 capturing bracket numbers. This is a tidy that I avoided doing when I
3454 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3455 The new approach is not only tidier, it makes it possible to reduce the
3456 memory needed to fix the previous bug (38).
3457
3458 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3459 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3460 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3461 mode.
3462
3463 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3464 report.
3465
3466 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3467 copy construction and assignment in the C++ wrapper.
3468
3469 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3470 couple of bugs that could have given wrong results in the "--newline=crlf"
3471 case.
3472
3473 44. Added a number of casts and did some reorganization of signed/unsigned int
3474 variables following suggestions from Dair Grant. Also renamed the variable
3475 "this" as "item" because it is a C++ keyword.
3476
3477 45. Arranged for dftables to add
3478
3479 #include "pcre_internal.h"
3480
3481 to pcre_chartables.c because without it, gcc 4.x may remove the array
3482 definition from the final binary if PCRE is built into a static library and
3483 dead code stripping is activated.
3484
3485 46. For an unanchored pattern, if a match attempt fails at the start of a
3486 newline sequence, and the newline setting is CRLF or ANY, and the next two
3487 characters are CRLF, advance by two characters instead of one.
3488
3489
3490 Version 6.7 04-Jul-06
3491 ---------------------
3492
3493 1. In order to handle tests when input lines are enormously long, pcretest has
3494 been re-factored so that it automatically extends its buffers when
3495 necessary. The code is crude, but this _is_ just a test program. The
3496 default size has been increased from 32K to 50K.
3497
3498 2. The code in pcre_study() was using the value of the re argument before
3499 testing it for NULL. (Of course, in any sensible call of the function, it
3500 won't be NULL.)
3501
3502 3. The memmove() emulation function in pcre_internal.h, which is used on
3503 systems that lack both memmove() and bcopy() - that is, hardly ever -
3504 was missing a "static" storage class specifier.
3505
3506 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3507 containing an extended class (one that cannot be represented by a bitmap
3508 because it contains high-valued characters or Unicode property items, e.g.
3509 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3510 pattern, but PCRE should not loop if you do not (it no longer does).
3511 [Detail: two cases were found: (a) a repeated subpattern containing an
3512 extended class; (b) a recursive reference to a subpattern that followed a
3513 previous extended class. It wasn't skipping over the extended class
3514 correctly when UTF-8 mode was not set.]
3515
3516 5. A negated single-character class was not being recognized as fixed-length
3517 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3518 compile error "lookbehind assertion is not fixed length".
3519
3520 6. The RunPerlTest auxiliary script was showing an unexpected difference
3521 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3522 write a Perl script that can interpret lines of an input file either as
3523 byte characters or as UTF-8, which is what "perltest" was being required to
3524 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3525 can't do is switch easily at run time between having the "use utf8;" pragma
3526 or not. In the end, I fudged it by using the RunPerlTest script to insert
3527 "use utf8;" explicitly for the UTF-8 tests.
3528
3529 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3530 the end of the subject string, contrary to the documentation and to what
3531 Perl does. This was true of both matching functions. Now it matches only at
3532 the start of the subject and immediately after *internal* newlines.
3533
3534 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3535 a pointer to an int instead of a pointer to an unsigned long int. This
3536 caused problems on 64-bit systems.
3537
3538 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3539 instance of the 'standard' template library not being so standard".
3540
3541 10. There was no check on the number of named subpatterns nor the maximum
3542 length of a subpattern name. The product of these values is used to compute
3543 the size of the memory block for a compiled pattern. By supplying a very
3544 long subpattern name and a large number of named subpatterns, the size
3545 computation could be caused to overflow. This is now prevented by limiting
3546 the length of names to 32 characters, and the number of named subpatterns
3547 to 10,000.
3548
3549 11. Subpatterns that are repeated with specific counts have to be replicated in
3550 the compiled pattern. The size of memory for this was computed from the
3551 length of the subpattern and the repeat count. The latter is limited to
3552 65535, but there was no limit on the former, meaning that integer overflow
3553 could in principle occur. The compiled length of a repeated subpattern is
3554 now limited to 30,000 bytes in order to prevent this.
3555
3556 12. Added the optional facility to have named substrings with the same name.
3557
3558 13. Added the ability to use a named substring as a condition, using the
3559 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3560 are numbers (not recommended). Forward references are permitted.
3561
3562 14. Added forward references in named backreferences (if you see what I mean).
3563
3564 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3565 pattern could run off the end of the subject. For example, the pattern
3566 "(?s)(.{1,5})"8 did this with the subject "ab".
3567
3568 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3569 PCRE_CASELESS was set when matching characters that were quantified with ?
3570 or *.
3571
3572 17. A character class other than a single negated character that had a minimum
3573 but no maximum quantifier - for example [ab]{6,} - was not handled
3574 correctly by pce_dfa_exec(). It would match only one character.
3575
3576 18. A valid (though odd) pattern that looked like a POSIX character
3577 class but used an invalid character after [ (for example [[,abc,]]) caused
3578 pcre_compile() to give the error "Failed: internal error: code overflow" or
3579 in some cases to crash with a glibc free() error. This could even happen if
3580 the pattern terminated after [[ but there just happened to be a sequence of
3581 letters, a binary zero, and a closing ] in the memory that followed.
3582
3583 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3584 over the years. Originally (before any Unicode support), just the bottom 8
3585 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3586 output from "man perlunicode" includes this:
3587
3588 The regular expression compiler produces polymorphic opcodes. That
3589 is, the pattern adapts to the data and automatically switches to
3590 the Unicode character scheme when presented with Unicode data--or
3591 instead uses a traditional byte scheme when presented with byte
3592 data.
3593
3594 Sadly, a wide octal escape does not cause a switch, and in a string with
3595 no other multibyte characters, these octal escapes are treated as before.
3596 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3597 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3598 Unicode string.
3599
3600 I have not perpetrated such confusion in PCRE. Up till now, it took just
3601 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3602 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3603 translate to the appropriate multibyte character.
3604
3605 29. Applied some refactoring to reduce the number of warnings from Microsoft
3606 and Borland compilers. This has included removing the fudge introduced
3607 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3608 a warning about an unused variable.
3609
3610 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3611 characters since release 4.0, because Perl (from release 5.004) does not.
3612 [Or at least, is documented not to: some releases seem to be in conflict
3613 with the documentation.] However, when a pattern was studied with
3614 pcre_study() and all its branches started with \s, PCRE still included VT
3615 as a possible starting character. Of course, this did no harm; it just
3616 caused an unnecessary match attempt.
3617
3618 22. Removed a now-redundant internal flag bit that recorded the fact that case
3619 dependency changed within the pattern. This was once needed for "required
3620 byte" processing, but is no longer used. This recovers a now-scarce options
3621 bit. Also moved the least significant internal flag bit to the most-
3622 significant bit of the word, which was not previously used (hangover from
3623 the days when it was an int rather than a uint) to free up another bit for
3624 the future.
3625
3626 23. Added support for CRLF line endings as well as CR and LF. As well as the
3627 default being selectable at build time, it can now be changed at runtime
3628 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3629 specify that it is scanning data with non-default line endings.
3630
3631 24. Changed the definition of CXXLINK to make it agree with the definition of
3632 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3633
3634 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3635 recursions. This makes a big different to stack usage for some patterns.
3636
3637 26. If a subpattern containing a named recursion or subroutine reference such
3638 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3639 the space required for the compiled pattern went wrong and gave too small a
3640 value. Depending on the environment, this could lead to "Failed: internal
3641 error: code overflow at offset 49" or "glibc detected double free or
3642 corruption" errors.
3643
3644 27. Applied patches from Google (a) to support the new newline modes and (b) to
3645 advance over multibyte UTF-8 characters in GlobalReplace.
3646
3647 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3648 difference for some implementation of PCRE in some Windows version.
3649
3650 29. Added some extra testing facilities to pcretest:
3651
3652 \q<number> in a data line sets the "match limit" value
3653 \Q<number> in a data line sets the "match recursion limt" value
3654 -S <number> sets the stack size, where <number> is in megabytes
3655
3656 The -S option isn't available for Windows.
3657
3658
3659 Version 6.6 06-Feb-06
3660 ---------------------
3661
3662 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3663 in pcreposix.h. I have copied the definition from pcre.h.
3664
3665 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3666 because pcre.h is no longer a built file.
3667
3668 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3669 not normally included in the compiled code.
3670
3671
3672 Version 6.5 01-Feb-06
3673 ---------------------
3674
3675 1. When using the partial match feature with pcre_dfa_exec(), it was not
3676 anchoring the second and subsequent partial matches at the new starting
3677 point. This could lead to incorrect results. For example, with the pattern
3678 /1234/, partially matching against "123" and then "a4" gave a match.
3679
3680 2. Changes to pcregrep:
3681
3682 (a) All non-match returns from pcre_exec() were being treated as failures
3683 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3684 error message is output. Some extra information is given for the
3685 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3686 probably the only errors that are likely to be caused by users (by
3687 specifying a regex that has nested indefinite repeats, for instance).
3688 If there are more than 20 of these errors, pcregrep is abandoned.
3689
3690 (b) A binary zero was treated as data while matching, but terminated the
3691 output line if it was written out. This has been fixed: binary zeroes
3692 are now no different to any other data bytes.
3693
3694 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3695 used to set a locale for matching. The --locale=xxxx long option has
3696 been added (no short equivalent) to specify a locale explicitly on the
3697 pcregrep command, overriding the environment variables.
3698
3699 (d) When -B was used with -n, some line numbers in the output were one less
3700 than they should have been.
3701
3702 (e) Added the -o (--only-matching) option.
3703
3704 (f) If -A or -C was used with -c (count only), some lines of context were
3705 accidentally printed for the final match.
3706
3707 (g) Added the -H (--with-filename) option.
3708
3709 (h) The combination of options -rh failed to suppress file names for files
3710 that were found from directory arguments.
3711
3712 (i) Added the -D (--devices) and -d (--directories) options.
3713
3714 (j) Added the -F (--fixed-strings) option.
3715
3716 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3717
3718 (l) Added the --colo(u)r option.
3719
3720 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3721 is not present by default.
3722
3723 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3724 items such as (?R) or (?1), when the recursion could match a number of
3725 alternatives. If it matched one of the alternatives, but subsequently,
3726 outside the recursion, there was a failure, the code tried to back up into
3727 the recursion. However, because of the way PCRE is implemented, this is not
3728 possible, and the result was an incorrect result from the match.
3729
3730 In order to prevent this happening, the specification of recursion has
3731 been changed so that all such subpatterns are automatically treated as
3732 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3733
3734 4. I had overlooked the fact that, in some locales, there are characters for
3735 which isalpha() is true but neither isupper() nor islower() are true. In
3736 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3737 and ordfeminine) are like this. This affected the treatment of \w and \W
3738 when they appeared in character classes, but not when they appeared outside
3739 a character class. The bit map for "word" characters is now created
3740 separately from the results of isalnum() instead of just taking it from the
3741 upper, lower, and digit maps. (Plus the underscore character, of course.)
3742
3743 5. The above bug also affected the handling of POSIX character classes such as
3744 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3745 permanent tables. Instead, the bit maps for such a class were previously
3746 created as the appropriate unions of the upper, lower, and digit bitmaps.
3747 Now they are created by subtraction from the [[:word:]] class, which has
3748 its own bitmap.
3749
3750 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3751 It is created by subtracting the vertical space characters (\x09, \x0a,
3752 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3753 subtraction was done in the overall bitmap for a character class, meaning
3754 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3755 be recognized. This bug has been fixed.
3756
3757 7. Patches from the folks at Google:
3758
3759 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3760 real life, but is still worth protecting against".
3761
3762 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3763 regular expressions".
3764
3765 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3766 have it.
3767
3768 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3769 "configure" and the latter not, in order to fix a problem somebody had
3770 with compiling the Arg class on HP-UX.
3771
3772 (e) Improve the error-handling of the C++ wrapper a little bit.
3773
3774 (f) New tests for checking recursion limiting.
3775
3776 8. The pcre_memmove() function, which is used only if the environment does not
3777 have a standard memmove() function (and is therefore rarely compiled),
3778 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3779 returning a result (though PCRE never actually uses the result).
3780
3781 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3782 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3783 returned instead of calling malloc() with an overflowing number that would
3784 most likely cause subsequent chaos.
3785
3786 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3787
3788 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3789 with this option is matched, the nmatch and pmatch options of regexec() are
3790 ignored.
3791
3792 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3793 provided in case anyone wants to the the POSIX interface with UTF-8
3794 strings.
3795
3796 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3797 C++ linking (needed for some HP-UX environments).
3798
3799 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3800 (unused parameter) and in the pcre_printint() function (omitted "default"
3801 switch label when the default is to do nothing).
3802
3803 15. Added some code to make it possible, when PCRE is compiled as a C++
3804 library, to replace subject pointers for pcre_exec() with a smart pointer
3805 class, thus making it possible to process discontinuous strings.
3806
3807 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3808 much the same function. They were added by different people who were trying
3809 to make PCRE easy to compile on non-Unix systems. It has been suggested
3810 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3811 for compiling on Windows systems. I have therefore replaced it with
3812 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3813 defaults to "extern" for C or "extern C" for C++, which works fine on
3814 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3815 SCOPE with something explicit in config.h. In addition:
3816
3817 (a) pcreposix.h still had just "extern" instead of either of these macros;
3818 I have replaced it with PCRE_DATA_SCOPE.
3819
3820 (b) Functions such as _pcre_xclass(), which are internal to the library,
3821 but external in the C sense, all had PCRE_EXPORT in their definitions.
3822 This is apparently wrong for the Windows case, so I have removed it.
3823 (It makes no difference on Unix-like systems.)
3824
3825 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3826 of recursive calls to match(). This is different to MATCH_LIMIT because
3827 that limits the total number of calls to match(), not all of which increase
3828 the depth of recursion. Limiting the recursion depth limits the amount of
3829 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3830 when PCRE is compiled, and changed at run time. A patch from Google adds
3831 this functionality to the C++ interface.
3832
3833 18. Changes to the handling of Unicode character properties:
3834
3835 (a) Updated the table to Unicode 4.1.0.
3836
3837 (b) Recognize characters that are not in the table as "Cn" (undefined).
3838
3839 (c) I revised the way the table is implemented to a much improved format
3840 which includes recognition of ranges. It now supports the ranges that
3841 are defined in UnicodeData.txt, and it also amalgamates other
3842 characters into ranges. This has reduced the number of entries in the
3843 table from around 16,000 to around 3,000, thus reducing its size
3844 considerably. I realized I did not need to use a tree structure after
3845 all - a binary chop search is just as efficient. Having reduced the
3846 number of entries, I extended their size from 6 bytes to 8 bytes to
3847 allow for more data.
3848
3849 (d) Added support for Unicode script names via properties such as \p{Han}.
3850
3851 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3852 matching that character.
3853
3854 20. When matching a repeated Unicode property with a minimum greater than zero,
3855 (for example \pL{2,}), PCRE could look past the end of the subject if it
3856 reached it while seeking the minimum number of characters. This could
3857 happen only if some of the characters were more than one byte long, because
3858 there is a check for at least the minimum number of bytes.
3859
3860 21. Refactored the implementation of \p and \P so as to be more general, to
3861 allow for more different types of property in future. This has changed the
3862 compiled form incompatibly. Anybody with saved compiled patterns that use
3863 \p or \P will have to recompile them.
3864
3865 22. Added "Any" and "L&" to the supported property types.
3866
3867 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3868 but give a compile time error if the value is greater than 0xff.
3869
3870 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3871 accidentally not being installed or uninstalled.
3872
3873 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3874 made were to insert the current release number. This seemed silly, because
3875 it made things harder for people building PCRE on systems that don't run
3876 "configure". I have turned pcre.h into a distributed file, no longer built
3877 by "configure", with the version identification directly included. There is
3878 no longer a pcre.h.in file.
3879
3880 However, this change necessitated a change to the pcre-config script as
3881 well. It is built from pcre-config.in, and one of the substitutions was the
3882 release number. I have updated configure.ac so that ./configure now finds
3883 the release number by grepping pcre.h.
3884
3885 26. Added the ability to run the tests under valgrind.
3886
3887
3888 Version 6.4 05-Sep-05
3889 ---------------------
3890
3891 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3892 "--" to be printed when multiple files were scanned, even when none of the
3893 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3894 consider it to be a bug, and have restored the previous behaviour.
3895
3896 2. A couple of code tidies to get rid of compiler warnings.
3897
3898 3. The pcretest program used to cheat by referring to symbols in the library
3899 whose names begin with _pcre_. These are internal symbols that are not
3900 really supposed to be visible externally, and in some environments it is
3901 possible to suppress them. The cheating is now confined to including
3902 certain files from the library's source, which is a bit cleaner.
3903
3904 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3905 file's purpose clearer.
3906
3907 5. Reorganized pcre_ucp_findchar().
3908
3909
3910 Version 6.3 15-Aug-05
3911 ---------------------
3912
3913 1. The file libpcre.pc.in did not have general read permission in the tarball.
3914
3915 2. There were some problems when building without C++ support:
3916
3917 (a) If C++ support was not built, "make install" and "make test" still
3918 tried to test it.
3919
3920 (b) There were problems when the value of CXX was explicitly set. Some
3921 changes have been made to try to fix these, and ...
3922
3923 (c) --disable-cpp can now be used to explicitly disable C++ support.
3924
3925 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3926 backslash in a target when C++ was disabled. This confuses some
3927 versions of "make", apparently. Using an intermediate variable solves
3928 this. (Same for CPP_LOBJ.)
3929
3930 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3931 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3932 necessary on certain architectures.
3933
3934 4. Added a setting of -export-symbols-regex to the link command to remove
3935 those symbols that are exported in the C sense, but actually are local
3936 within the library, and not documented. Their names all begin with
3937 "_pcre_". This is not a perfect job, because (a) we have to except some
3938 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3939 available (and never for static libraries). I have made a note to try to
3940 find a way round (a) in the future.
3941
3942
3943 Version 6.2 01-Aug-05
3944 ---------------------
3945
3946 1. There was no test for integer overflow of quantifier values. A construction
3947 such as {1111111111111111} would give undefined results. What is worse, if
3948 a minimum quantifier for a parenthesized subpattern overflowed and became
3949 negative, the calculation of the memory size went wrong. This could have
3950 led to memory overwriting.
3951
3952 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3953
3954 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3955 operating environments where this matters.
3956
3957 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3958 PCRE options from within the C++ wrapper.
3959
3960 5. Named capturing subpatterns were not being correctly counted when a pattern
3961 was compiled. This caused two problems: (a) If there were more than 100
3962 such subpatterns, the calculation of the memory needed for the whole
3963 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3964 back references of the form \12, where the number was greater than 9, were
3965 not recognized as back references, even though there were sufficient
3966 previous subpatterns.
3967
3968 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3969 versions of gcc, e.g. 2.95.4.
3970
3971
3972 Version 6.1 21-Jun-05
3973 ---------------------
3974
3975 1. There was one reference to the variable "posix" in pcretest.c that was not
3976 surrounded by "#if !defined NOPOSIX".
3977
3978 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3979 the cross-check on the old pcre_info() function, for the benefit of the
3980 cut-down version of PCRE that is currently imported into Exim.
3981
3982 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3983 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3984 patterns that start (?i) but I don't think that matters. The use of (?i) is
3985 just an example; this all applies to the other options as well.
3986
3987 4. Since libtool seems to echo the compile commands it is issuing, the output
3988 from "make" can be reduced a bit by putting "@" in front of each libtool
3989 compile command.
3990
3991 5. Patch from the folks at Google for configure.in to be a bit more thorough
3992 in checking for a suitable C++ installation before trying to compile the
3993 C++ stuff. This should fix a reported problem when a compiler was present,
3994 but no suitable headers.
3995
3996 6. The man pages all had just "PCRE" as their title. I have changed them to
3997 be the relevant file name. I have also arranged that these names are
3998 retained in the file doc/pcre.txt, which is a concatenation in text format
3999 of all the man pages except the little individual ones for each function.
4000
4001 7. The NON-UNIX-USE file had not been updated for the different set of source
4002 files that come with release 6. I also added a few comments about the C++
4003 wrapper.
4004
4005
4006 Version 6.0 07-Jun-05
4007 ---------------------
4008
4009 1. Some minor internal re-organization to help with my DFA experiments.
4010
4011 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4012 didn't matter for the library itself when fully configured, but did matter
4013 when compiling without UCP support, or within Exim, where the ucp files are
4014 not imported.
4015
4016 3. Refactoring of the library code to split up the various functions into
4017 different source modules. The addition of the new DFA matching code (see
4018 below) to a single monolithic source would have made it really too
4019 unwieldy, quite apart from causing all the code to be include in a
4020 statically linked application, when only some functions are used. This is
4021 relevant even without the DFA addition now that patterns can be compiled in
4022 one application and matched in another.
4023
4024 The downside of splitting up is that there have to be some external
4025 functions and data tables that are used internally in different modules of
4026 the library but which are not part of the API. These have all had their
4027 names changed to start with "_pcre_" so that they are unlikely to clash
4028 with other external names.
4029
4030 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4031 a different (DFA) algorithm. Although it is slower than the original
4032 function, it does have some advantages for certain types of matching
4033 problem.
4034
4035 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4036 including restarting after a partial match.
4037
4038 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4039 defined when compiling for Windows was sent to me. I have put it into the
4040 code, though I have no means of testing or verifying it.
4041
4042 7. Added the pcre_refcount() auxiliary function.
4043
4044 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4045 match before or at the first newline in the subject string. In pcretest,
4046 the /f option on a pattern can be used to set this.
4047
4048 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4049 would behave wrongly. This has been present in PCRE since release 4.0.
4050
4051 10. A number of changes to the pcregrep command:
4052
4053 (a) Refactored how -x works; insert ^(...)$ instead of setting
4054 PCRE_ANCHORED and checking the length, in preparation for adding
4055 something similar for -w.
4056
4057 (b) Added the -w (match as a word) option.
4058
4059 (c) Refactored the way lines are read and buffered so as to have more
4060 than one at a time available.
4061
4062 (d) Implemented a pcregrep test script.
4063
4064 (e) Added the -M (multiline match) option. This allows patterns to match
4065 over several lines of the subject. The buffering ensures that at least
4066 8K, or the rest of the document (whichever is the shorter) is available
4067 for matching (and similarly the previous 8K for lookbehind assertions).
4068
4069 (f) Changed the --help output so that it now says
4070
4071 -w, --word-regex(p)
4072
4073 instead of two lines, one with "regex" and the other with "regexp"
4074 because that confused at least one person since the short forms are the
4075 same. (This required a bit of code, as the output is generated
4076 automatically from a table. It wasn't just a text change.)
4077
4078 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4079 option but starts with a hyphen. Could be a pattern or a path name
4080 starting with a hyphen, for instance.
4081
4082 (h) "-" can be given as a file name to represent stdin.
4083
4084 (i) When file names are being printed, "(standard input)" is used for
4085 the standard input, for compatibility with GNU grep. Previously
4086 "<stdin>" was used.
4087
4088 (j) The option --label=xxx can be used to supply a name to be used for
4089 stdin when file names are being printed. There is no short form.
4090
4091 (k) Re-factored the options decoding logic because we are going to add
4092 two more options that take data. Such options can now be given in four
4093 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4094
4095 (l) Added the -A, -B, and -C options for requesting that lines of context
4096 around matches be printed.
4097
4098 (m) Added the -L option to print the names of files that do not contain
4099 any matching lines, that is, the complement of -l.
4100
4101 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4102 continue to scan other files.
4103
4104 (o) The -s option was incorrectly implemented. For compatibility with other
4105 greps, it now suppresses the error message for a non-existent or non-
4106 accessible file (but not the return code). There is a new option called
4107 -q that suppresses the output of matching lines, which was what -s was
4108 previously doing.
4109
4110 (p) Added --include and --exclude options to specify files for inclusion
4111 and exclusion when recursing.
4112
4113 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4114 Hopefully, it now does.
4115
4116 12. Missing cast in pcre_study().
4117
4118 13. Added an "uninstall" target to the makefile.
4119
4120 14. Replaced "extern" in the function prototypes in Makefile.in with
4121 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4122 world, but is set differently for Windows.
4123
4124 15. Added a second compiling function called pcre_compile2(). The only
4125 difference is that it has an extra argument, which is a pointer to an
4126 integer error code. When there is a compile-time failure, this is set
4127 non-zero, in addition to the error test pointer being set to point to an
4128 error message. The new argument may be NULL if no error number is required
4129 (but then you may as well call pcre_compile(), which is now just a
4130 wrapper). This facility is provided because some applications need a
4131 numeric error indication, but it has also enabled me to tidy up the way
4132 compile-time errors are handled in the POSIX wrapper.
4133
4134 16. Added VPATH=.libs to the makefile; this should help when building with one
4135 prefix path and installing with another. (Or so I'm told by someone who
4136 knows more about this stuff than I do.)
4137
4138 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4139 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4140 match everything, including newlines. This is not POSIX-compatible, but
4141 somebody wanted the feature. From pcretest it can be activated by using
4142 both the P and the s flags.
4143
4144 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4145
4146 19. libpcre.pc was being incorrectly installed as executable.
4147
4148 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4149 it now also looks for '\r' so that it will work unmodified on Windows.
4150
4151 21. Added Google's contributed C++ wrapper to the distribution.
4152
4153 22. Added some untidy missing memory free() calls in pcretest, to keep
4154 Electric Fence happy when testing.
4155
4156
4157
4158 Version 5.0 13-Sep-04
4159 ---------------------
4160
4161 1. Internal change: literal characters are no longer packed up into items
4162 containing multiple characters in a single byte-string. Each character
4163 is now matched using a separate opcode. However, there may be more than one
4164 byte in the character in UTF-8 mode.
4165
4166 2. The pcre_callout_block structure has two new fields: pattern_position and
4167 next_item_length. These contain the offset in the pattern to the next match
4168 item, and its length, respectively.
4169
4170 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4171 insertion of callouts before each pattern item. Added the /C option to
4172 pcretest to make use of this.
4173
4174 4. On the advice of a Windows user, the lines
4175
4176 #if defined(_WIN32) || defined(WIN32)
4177 _setmode( _fileno( stdout ), 0x8000 );
4178 #endif /* defined(_WIN32) || defined(WIN32) */
4179
4180 have been added to the source of pcretest. This apparently does useful
4181 magic in relation to line terminators.
4182
4183 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4184 for the benefit of those environments where the "b" makes a difference.
4185
4186 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4187 to know about it. I have put a hack into configure.in that adds in code
4188 to set GCC=yes if CC=icc. This seems to end up at a point in the
4189 generated configure script that is early enough to affect the setting of
4190 compiler options, which is what is needed, but I have no means of testing
4191 whether it really works. (The user who reported this had patched the
4192 generated configure script, which of course I cannot do.)
4193
4194 LATER: After change 22 below (new libtool files), the configure script
4195 seems to know about icc (and also ecc). Therefore, I have commented out
4196 this hack in configure.in.
4197
4198 7. Added support for pkg-config (2 patches were sent in).
4199
4200 8. Negated POSIX character classes that used a combination of internal tables
4201 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4202 [[:^ascii]]. Typically, they would match almost any characters. The other
4203 POSIX classes were not broken in this way.
4204
4205 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4206 to find the match, as PCRE was deluded into thinking that the match had to
4207 start at the start point or following a newline. The same bug applied to
4208 patterns with negative forward assertions or any backward assertions
4209 preceding ".*" at the start, unless the pattern required a fixed first
4210 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4211
4212 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4213 starting at the last subject character, bytes beyond the end of the subject
4214 string were read.
4215
4216 11. Renamed the variable "class" as "classbits" to make life easier for C++
4217 users. (Previously there was a macro definition, but it apparently wasn't
4218 enough.)
4219
4220 12. Added the new field "tables" to the extra data so that tables can be passed
4221 in at exec time, or the internal tables can be re-selected. This allows
4222 a compiled regex to be saved and re-used at a later time by a different
4223 program that might have everything at different addresses.
4224
4225 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4226 -R library as well as a -L library.
4227
4228 14. The debugging options of pcretest (-d on the command line or D on a
4229 pattern) showed incorrect output for anything following an extended class
4230 that contained multibyte characters and which was followed by a quantifier.
4231
4232 15. Added optional support for general category Unicode character properties
4233 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4234 support. It adds about 90K to the size of the library. The meanings of the
4235 inbuilt class escapes such as \d and \s have NOT been changed.
4236
4237 16. Updated pcredemo.c to include calls to free() to release the memory for the
4238 compiled pattern.
4239
4240 17. The generated file chartables.c was being created in the source directory
4241 instead of in the building directory. This caused the build to fail if the
4242 source directory was different from the building directory, and was
4243 read-only.
4244
4245 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4246 file. No doubt somebody will tell me if they don't make sense... Also added
4247 Dan Mooney's comments about building on OpenVMS.
4248
4249 19. Added support for partial matching via the PCRE_PARTIAL option for
4250 pcre_exec() and the \P data escape in pcretest.
4251
4252 20. Extended pcretest with 3 new pattern features:
4253
4254 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4255 write the compiled pattern to the file whose name is "rest-of-line".
4256 This is a straight binary dump of the data, with the saved pointer to
4257 the character tables forced to be NULL. The study data, if any, is
4258 written too. After writing, pcretest reads a new pattern.
4259
4260 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4261 compiled pattern from the given file. There must not be any
4262 occurrences of "<" in the file name (pretty unlikely); if there are,
4263 pcretest will instead treat the initial "<" as a pattern delimiter.
4264 After reading in the pattern, pcretest goes on to read data lines as
4265 usual.
4266
4267 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4268 and 16-bit fields in a compiled pattern, to simulate a pattern that
4269 was compiled on a host of opposite endianness.
4270
4271 21. The pcre-exec() function can now cope with patterns that were compiled on
4272 hosts of opposite endianness, with this restriction:
4273
4274 As for any compiled expression that is saved and used later, the tables
4275 pointer field cannot be preserved; the extra_data field in the arguments
4276 to pcre_exec() should be used to pass in a tables address if a value
4277 other than the default internal tables were used at compile time.
4278
4279 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4280 now diagnosed as an error. Previously, most of the time, a negative number
4281 would have been treated as zero, but if in addition "ovector" was passed as
4282 NULL, a crash could occur.
4283
4284 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4285 new versions from the libtool 1.5 distribution (the last one is a copy of
4286 a file called libtool.m4). This seems to have fixed the need to patch
4287 "configure" to support Darwin 1.3 (which I used to do). However, I still
4288 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4289 workstation).
4290
4291 24. Changed the PCRE licence to be the more standard "BSD" licence.
4292
4293
4294 Version 4.5 01-Dec-03
4295 ---------------------
4296
4297 1. There has been some re-arrangement of the code for the match() function so
4298 that it can be compiled in a version that does not call itself recursively.
4299 Instead, it keeps those local variables that need separate instances for
4300 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4301 needs to "recurse". Keeping track of where control must go is done by means
4302 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4303 hide most of the details from the main code, and operates only if
4304 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4305 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4306 operating.
4307
4308 To make it easier for callers to provide specially tailored get/free
4309 functions for this usage, two new functions, pcre_stack_malloc, and
4310 pcre_stack_free, are used. They are always called in strict stacking order,
4311 and the size of block requested is always the same.
4312
4313 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4314 PCRE has been compiled to use the stack or the heap for recursion. The
4315 -C option of pcretest uses this to show which version is compiled.
4316
4317 A new data escape \S, is added to pcretest; it causes the amounts of store
4318 obtained and freed by both kinds of malloc/free at match time to be added
4319 to the output.
4320
4321 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4322 what's available on my current Linux desktop machine.
4323
4324 3. When matching a UTF-8 string, the test for a valid string at the start has
4325 been extended. If start_offset is not zero, PCRE now checks that it points
4326 to a byte that is the start of a UTF-8 character. If not, it returns
4327 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4328 this is necessary because there may be backward assertions in the pattern.
4329 When matching the same subject several times, it may save resources to use
4330 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4331
4332 4. The code for checking the validity of UTF-8 strings has been tightened so
4333 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4334 containing "overlong sequences".
4335
4336 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4337 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4338 should have been "&", but it just so happened that all the cases this let
4339 through by mistake were picked up later in the function.
4340
4341 6. I had used a variable called "isblank" - this is a C99 function, causing
4342 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4343
4344 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4345 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4346 so the version is shown; (c) stop "make test" echoing "./RunTest".
4347
4348 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4349
4350 9. The prototype for memmove() for systems that don't have it was using
4351 size_t, but the inclusion of the header that defines size_t was later. I've
4352 moved the #includes for the C headers earlier to avoid this.
4353
4354 10. Added some adjustments to the code to make it easier to compiler on certain
4355 special systems:
4356
4357 (a) Some "const" qualifiers were missing.
4358 (b) Added the macro EXPORT before all exported functions; by default this
4359 is defined to be empty.
4360 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4361 that it reads its output file name as an argument instead of writing
4362 to the standard output and assuming this can be redirected.
4363
4364 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4365 class containing characters with values greater than 255, PCRE compilation
4366 went into a loop.
4367
4368 12. A recursive reference to a subpattern that was within another subpattern
4369 that had a minimum quantifier of zero caused PCRE to crash. For example,
4370 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4371 recursion. If the recursively-called subpattern itself had a zero repeat,
4372 that was OK.
4373
4374 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4375 buffer into which it was copied (for escape processing) was still set at
4376 1024, so long lines caused crashes.
4377
4378 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4379 "internal error: code overflow...". This applied to any character class
4380 that was followed by a possessive quantifier.
4381
4382 15. Modified the Makefile to add libpcre.la as a prerequisite for
4383 libpcreposix.la because I was told this is needed for a parallel build to
4384 work.
4385
4386 16. If a pattern that contained .* following optional items at the start was
4387 studied, the wrong optimizing data was generated, leading to matching
4388 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4389 matching string must start with a or b or c. The correct conclusion for
4390 this pattern is that a match can start with any character.
4391
4392
4393 Version 4.4 13-Aug-03
4394 ---------------------
4395
4396 1. In UTF-8 mode, a character class containing characters with values between
4397 127 and 255 was not handled correctly if the compiled pattern was studied.
4398 In fixing this, I have also improved the studying algorithm for such
4399 classes (slightly).
4400
4401 2. Three internal functions had redundant arguments passed to them. Removal
4402 might give a very teeny performance improvement.
4403
4404 3. Documentation bug: the value of the capture_top field in a callout is *one
4405 more than* the number of the hightest numbered captured substring.
4406
4407 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4408 in incorrectly linking with a previously installed version. They now link
4409 explicitly with libpcre.la.
4410
4411 5. configure.in no longer needs to recognize Cygwin specially.
4412
4413 6. A problem in pcre.in for Windows platforms is fixed.
4414
4415 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4416 pcretest, it used to include the size of the study block as part of its
4417 output. Unfortunately, the structure contains a field that has a different
4418 size on different hardware architectures. This meant that the tests that
4419 showed this size failed. As the block is currently always of a fixed size,
4420 this information isn't actually particularly useful in pcretest output, so
4421 I have just removed it.
4422
4423 8. Three pre-processor statements accidentally did not start in column 1.
4424 Sadly, there are *still* compilers around that complain, even though
4425 standard C has not required this for well over a decade. Sigh.
4426
4427 9. In pcretest, the code for checking callouts passed small integers in the
4428 callout_data field, which is a void * field. However, some picky compilers
4429 complained about the casts involved for this on 64-bit systems. Now
4430 pcretest passes the address of the small integer instead, which should get
4431 rid of the warnings.
4432
4433 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4434 both compile and run time, and gives an error if an invalid UTF-8 sequence
4435 is found. There is a option for disabling this check in cases where the
4436 string is known to be correct and/or the maximum performance is wanted.
4437
4438 11. In response to a bug report, I changed one line in Makefile.in from
4439
4440 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4441 to
4442 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4443
4444 to look similar to other lines, but I have no way of telling whether this
4445 is the right thing to do, as I do not use Windows. No doubt I'll get told
4446 if it's wrong...
4447
4448
4449 Version 4.3 21-May-03
4450 ---------------------
4451
4452 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4453 Makefile.
4454
4455 2. Some refactoring to improve the quality of the code:
4456
4457 (i) The utf8_table... variables are now declared "const".
4458
4459 (ii) The code for \cx, which used the "case flipping" table to upper case
4460 lower case letters, now just substracts 32. This is ASCII-specific,
4461 but the whole concept of \cx is ASCII-specific, so it seems
4462 reasonable.
4463
4464 (iii) PCRE was using its character types table to recognize decimal and
4465 hexadecimal digits in the pattern. This is silly, because it handles
4466 only 0-9, a-f, and A-F, but the character types table is locale-
4467 specific, which means strange things might happen. A private
4468 table is now used for this - though it costs 256 bytes, a table is
4469 much faster than multiple explicit tests. Of course, the standard
4470 character types table is still used for matching digits in subject
4471 strings against \d.
4472
4473 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4474 ending in _t are). So I've renamed it as ESC_tee.
4475
4476 3. The first argument for regexec() in the POSIX wrapper should have been
4477 defined as "const".
4478
4479 4. Changed pcretest to use malloc() for its buffers so that they can be
4480 Electric Fenced for debugging.
4481
4482 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4483 to read one or more bytes before the start of the subject string. Often this
4484 had no effect on PCRE's behaviour, but in some circumstances it could
4485 provoke a segmentation fault.
4486
4487 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4488 to try to read one or more bytes before the start of the subject string.
4489
4490 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4491 UTF-8 support could misbehave in various ways if the subject string
4492 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4493 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4494 back over UTF-8 characters.)
4495
4496
4497 Version 4.2 14-Apr-03
4498 ---------------------
4499
4500 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4501
4502 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4503 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4504 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4505 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4506 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4507 and BUILD_EXEEXT
4508 Note: automatic setting of the BUILD variables is not yet working
4509 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4510 compile-time but not at link-time
4511 [LINK]: use for linking executables only
4512 make different versions for Windows and non-Windows
4513 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4514 libraries
4515 [LINK_FOR_BUILD]: new variable
4516 [OBJEXT]: use throughout
4517 [EXEEXT]: use throughout
4518 <winshared>: new target
4519 <wininstall>: new target
4520 <dftables.o>: use native compiler
4521 <dftables>: use native linker
4522 <install>: handle Windows platform correctly
4523 <clean>: ditto
4524 <check>: ditto
4525 copy DLL to top builddir before testing
4526
4527 As part of these changes, -no-undefined was removed again. This was reported
4528 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4529 in any case.
4530
4531 3. Some tidies to get rid of compiler warnings:
4532
4533 . In the match_data structure, match_limit was an unsigned long int, whereas
4534 match_call_count was an int. I've made them both unsigned long ints.
4535
4536 . In pcretest the fact that a const uschar * doesn't automatically cast to
4537 a void * provoked a warning.
4538
4539 . Turning on some more compiler warnings threw up some "shadow" variables
4540 and a few more missing casts.
4541
4542 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4543 option, a class that contained a single character with a value between 128
4544 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4545
4546 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4547 option, a class that contained several characters, but with at least one
4548 whose value was between 128 and 255 caused PCRE to crash.
4549
4550
4551 Version 4.1 12-Mar-03
4552 ---------------------
4553
4554 1. Compiling with gcc -pedantic found a couple of places where casts were
4555 needed, and a string in dftables.c that was longer than standard compilers are
4556 required to support.
4557
4558 2. Compiling with Sun's compiler found a few more places where the code could
4559 be tidied up in order to avoid warnings.
4560
4561 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4562 first of these names is deprecated in the latest Autoconf in favour of the name
4563 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4564 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4565 analogy I have changed it to CFLAGS_FOR_BUILD.
4566
4567 4. Added -no-undefined to the linking command in the Makefile, because this is
4568 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4569 linking step for the pcreposix library.
4570
4571 5. PCRE was failing to diagnose the case of two named groups with the same
4572 name.
4573
4574 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4575 literal character that is needed in the subject for a match, and scans along to
4576 ensure that it is present before embarking on the full matching process. This
4577 saves time in cases of nested unlimited repeats that are never going to match.
4578 Problem: the scan can take a lot of time if the subject is very long (e.g.
4579 megabytes), thus penalizing straightforward matches. It is now done only if the
4580 amount of subject to be scanned is less than 1000 bytes.
4581
4582 7. A lesser problem with the same optimization is that it was recording the
4583 first character of an anchored pattern as "needed", thus provoking a search
4584 right along the subject, even when the first match of the pattern was going to
4585 fail. The "needed" character is now not set for anchored patterns, unless it
4586 follows something in the pattern that is of non-fixed length. Thus, it still
4587 fulfils its original purpose of finding quick non-matches in cases of nested
4588 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4589
4590
4591 Version 4.0 17-Feb-03
4592 ---------------------
4593
4594 1. If a comment in an extended regex that started immediately after a meta-item
4595 extended to the end of string, PCRE compiled incorrect data. This could lead to
4596 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4597
4598 2. Moved to autoconf 2.53 and libtool 1.4.2.
4599
4600 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4601 the special perltest8 script is no longer needed - all the tests can be run
4602 from a single perltest script.
4603
4604 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4605 by \s. It has now been removed in PCRE. This means it isn't recognized as
4606 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4607 class [:space:] *does* include VT, thereby creating a mess.
4608
4609 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4610 space and tab.
4611
4612 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4613 its new features into the main test script, reducing the number of scripts.
4614
4615 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4616 were backward compatible, and made the (?i) apply to the whole pattern, as if
4617 /i were given. Now it behaves more logically, and applies the option setting
4618 only to what follows. PCRE has been changed to follow suit. However, if it
4619 finds options settings right at the start of the pattern, it extracts them into
4620 the global options, as before. Thus, they show up in the info data.
4621
4622 8. Added support for the \Q...\E escape sequence. Characters in between are
4623 treated as literals. This is slightly different from Perl in that $ and @ are
4624 also handled as literals inside the quotes. In Perl, they will cause variable
4625 interpolation. Note the following examples:
4626
4627 Pattern PCRE matches Perl matches
4628
4629 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4630 \Qabc\$xyz\E abc\$xyz abc\$xyz
4631 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4632
4633 For compatibility with Perl, \Q...\E sequences are recognized inside character
4634 classes as well as outside them.
4635
4636 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4637 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4638 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4639 signed/unsigned warnings.
4640
4641 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4642 option for pcretest, so I've replaced it by a simple function that does just
4643 that job.
4644
4645 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4646 "pcregrep -".
4647
4648 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4649 Java package. This provides some syntactic sugar for simple cases of what my
4650 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4651 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4652 item, you can use this simplified notation. Note that only makes sense with
4653 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4654 greediness, whatever the setting of the PCRE_UNGREEDY option.
4655
4656 13. A change of greediness default within a pattern was not taking effect at
4657 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4658 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4659 was abstracted outside.
4660
4661 14. PCRE now supports the \G assertion. It is true when the current matching
4662 position is at the start point of the match. This differs from \A when the
4663 starting offset is non-zero. Used with the /g option of pcretest (or similar
4664 code), it works in the same way as it does for Perl's /g option. If all
4665 alternatives of a regex begin with \G, the expression is anchored to the start
4666 match position, and the "anchored" flag is set in the compiled expression.
4667
4668 15. Some bugs concerning the handling of certain option changes within patterns
4669 have been fixed. These applied to options other than (?ims). For example,
4670 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4671 been the other way round. Some of this was related to change 7 above.
4672
4673 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4674 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4675 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4676 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4677
4678 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4679 mode. Unlike ".", it always matches newline, whatever the setting of
4680 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4681 assertions. Perl allows it, but it doesn't (in general) work because it can't
4682 calculate the length of the lookbehind. At least, that's the case for Perl
4683 5.8.0 - I've been told they are going to document that it doesn't work in
4684 future.
4685
4686 18. Added an error diagnosis for escapes that PCRE does not support: these are
4687 \L, \l, \N, \P, \p, \U, \u, and \X.
4688
4689 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4690 reading past the end of the pattern in cases such as /[abcd/.
4691
4692 20. PCRE was getting more memory than necessary for patterns with classes that
4693 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4694
4695 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4696 compiling PCRE for use with Virtual Pascal.
4697
4698 22. Small fix to the Makefile to make it work properly if the build is done
4699 outside the source tree.
4700
4701 23. Added a new extension: a condition to go with recursion. If a conditional
4702 subpattern starts with (?(R) the "true" branch is used if recursion has
4703 happened, whereas the "false" branch is used only at the top level.
4704
4705 24. When there was a very long string of literal characters (over 255 bytes
4706 without UTF support, over 250 bytes with UTF support), the computation of how
4707 much memory was required could be incorrect, leading to segfaults or other
4708 strange effects.
4709
4710 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4711 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4712 there was a subsequent back reference to those brackets. This meant that, for
4713 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4714 possible to check for precisely this case. All we can do is abandon the
4715 optimization if .* occurs inside capturing brackets when there are any back
4716 references whatsoever. (See below for a better fix that came later.)
4717
4718 26. The handling of the optimization for finding the first character of a
4719 non-anchored pattern, and for finding a character that is required later in the
4720 match were failing in some cases. This didn't break the matching; it just
4721 failed to optimize when it could. The way this is done has been re-implemented.
4722
4723 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4724
4725 28. Added a new feature that provides some of the functionality that Perl
4726 provides with (?{...}). The facility is termed a "callout". The way it is done
4727 in PCRE is for the caller to provide an optional function, by setting
4728 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4729 global variable. By default it is unset, which disables all calling out. To get
4730 the function called, the regex must include (?C) at appropriate points. This
4731 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4732 This provides a means of identifying different callout points. When PCRE
4733 reaches such a point in the regex, if pcre_callout has been set, the external
4734 function is called. It is provided with data in a structure called
4735 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4736 matching continues; if it returns a non-zero value, the match at the current
4737 point fails. However, backtracking will occur if possible. [This was changed
4738 later and other features added - see item 49 below.]
4739
4740 29. pcretest is upgraded to test the callout functionality. It provides a
4741 callout function that displays information. By default, it shows the start of
4742 the match and the current position in the text. There are some new data escapes
4743 to vary what happens:
4744
4745 \C+ in addition, show current contents of captured substrings
4746 \C- do not supply a callout function
4747 \C!n return 1 when callout number n is reached
4748 \C!n!m return 1 when callout number n is reached for the mth time
4749
4750 30. If pcregrep was called with the -l option and just a single file name, it
4751 output "<stdin>" if a match was found, instead of the file name.
4752
4753 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4754 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4755 pcre_exec(). This saves a malloc/free per call. The default value of
4756 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4757 when configuring.
4758
4759 32. The default maximum size of a compiled pattern is 64K. There have been a
4760 few cases of people hitting this limit. The code now uses macros to handle the
4761 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4762 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4763 configuring. Tests 2 and 5 work only with 2-byte links because they output
4764 debugging information about compiled patterns.
4765
4766 33. Internal code re-arrangements:
4767
4768 (a) Moved the debugging function for printing out a compiled regex into
4769 its own source file (printint.c) and used #include to pull it into
4770 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4771 separate copies.
4772
4773 (b) Defined the list of op-code names for debugging as a macro in
4774 internal.h so that it is next to the definition of the opcodes.
4775
4776 (c) Defined a table of op-code lengths for simpler skipping along compiled
4777 code. This is again a macro in internal.h so that it is next to the
4778 definition of the opcodes.
4779
4780 34. Added support for recursive calls to individual subpatterns, along the
4781 lines of Robin Houston's patch (but implemented somewhat differently).
4782
4783 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4784 allow it to read and process whole directories in Win32. This code was
4785 contributed by Lionel Fourquaux; it has not been tested by me.
4786
4787 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4788 used to name a group. Names consist of alphanumerics and underscores, and must
4789 be unique. Back references use the syntax (?P=name) and recursive calls use
4790 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4791 numbers. The function pcre_fullinfo() can be used after compilation to extract
4792 a name/number map. There are three relevant calls:
4793
4794 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4795 PCRE_INFO_NAMECOUNT yields the number of entries
4796 PCRE_INFO_NAMETABLE yields a pointer to the map.
4797
4798 The map is a vector of fixed-size entries. The size of each entry depends on
4799 the length of the longest name used. The first two bytes of each entry are the
4800 group number, most significant byte first. There follows the corresponding
4801 name, zero terminated. The names are in alphabetical order.
4802
4803 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4804 case instead of 255. Making it the same both with and without UTF-8 support
4805 means that the same test output works with both.
4806
4807 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4808 calling malloc() with a zero argument.
4809
4810 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4811 optimization. I've improved things by keeping a bitmap of backreferences with
4812 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4813 fact referenced, the optimization can be applied. It is unlikely that a
4814 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4815 the match to follow \n) will appear inside brackets with a number greater than
4816 31, but if it does, any back reference > 31 suppresses the optimization.
4817
4818 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4819 of disabling numbered capturing parentheses. Any opening parenthesis that is
4820 not followed by ? behaves as if it were followed by ?: but named parentheses
4821 can still be used for capturing (and they will acquire numbers in the usual
4822 way).
4823
4824 41. Redesigned the return codes from the match() function into yes/no/error so
4825 that errors can be passed back from deep inside the nested calls. A malloc
4826 failure while inside a recursive subpattern call now causes the
4827 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4828
4829 42. It is now possible to set a limit on the number of times the match()
4830 function is called in a call to pcre_exec(). This facility makes it possible to
4831 limit the amount of recursion and backtracking, though not in a directly
4832 obvious way, because the match() function is used in a number of different
4833 circumstances. The count starts from zero for each position in the subject
4834 string (for non-anchored patterns). The default limit is, for compatibility, a
4835 large number, namely 10 000 000. You can change this in two ways:
4836
4837 (a) When configuring PCRE before making, you can use --with-match-limit=n
4838 to set a default value for the compiled library.
4839
4840 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4841 a different value is set. See 45 below.
4842
4843 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4844
4845 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4846 of things that can be changed at compile time. The first argument specifies
4847 what is wanted and the second points to where the information is to be placed.
4848 The current list of available information is:
4849
4850 PCRE_CONFIG_UTF8
4851
4852 The output is an integer that is set to one if UTF-8 support is available;
4853 otherwise it is set to zero.
4854
4855 PCRE_CONFIG_NEWLINE
4856
4857 The output is an integer that it set to the value of the code that is used for
4858 newline. It is either LF (10) or CR (13).
4859
4860 PCRE_CONFIG_LINK_SIZE
4861
4862 The output is an integer that contains the number of bytes used for internal
4863 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4864
4865 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4866
4867 The output is an integer that contains the threshold above which the POSIX
4868 interface uses malloc() for output vectors. See item 31 above.
4869
4870 PCRE_CONFIG_MATCH_LIMIT
4871
4872 The output is an unsigned integer that contains the default limit of the number
4873 of match() calls in a pcre_exec() execution. See 42 above.
4874
4875 44. pcretest has been upgraded by the addition of the -C option. This causes it
4876 to extract all the available output from the new pcre_config() function, and to
4877 output it. The program then exits immediately.
4878
4879 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4880 order to support additional features. One way would have been to define
4881 pcre_exec2() (for example) with extra arguments, but this would not have been
4882 extensible, and would also have required all calls to the original function to
4883 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4884 is used for passing in "extra" data from pcre_study().
4885
4886 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4887 contains the following fields:
4888
4889 flags a bitmap indicating which of the following fields are set
4890 study_data opaque data from pcre_study()
4891 match_limit a way of specifying a limit on match() calls for a specific
4892 call to pcre_exec()
4893 callout_data data for callouts (see 49 below)
4894
4895 The flag bits are also defined in pcre.h, and are
4896
4897 PCRE_EXTRA_STUDY_DATA
4898 PCRE_EXTRA_MATCH_LIMIT
4899 PCRE_EXTRA_CALLOUT_DATA
4900
4901 The pcre_study() function now returns one of these new pcre_extra blocks, with
4902 the actual study data pointed to by the study_data field, and the
4903 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4904 before. That is, this change is entirely upwards-compatible and requires no
4905 change to existing code.
4906
4907 If you want to pass in additional data to pcre_exec(), you can either place it
4908 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4909 block.
4910
4911 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4912 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4913 times with different match limits, until it finds the minimum value needed for
4914 pcre_exec() to complete. The value is then output. This can be instructive; for
4915 most simple matches the number is quite small, but for pathological cases it
4916 gets very large very quickly.
4917
4918 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4919 returns the size of the data block pointed to by the study_data field in a
4920 pcre_extra block, that is, the value that was passed as the argument to
4921 pcre_malloc() when PCRE was getting memory in which to place the information
4922 created by pcre_study(). The fourth argument should point to a size_t variable.
4923 pcretest has been extended so that this information is shown after a successful
4924 pcre_study() call when information about the compiled regex is being displayed.
4925
4926 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4927 because what follows is always an absolute path. (Later: it turns out that this
4928 is more than cosmetic for MinGW, because it doesn't like empty path
4929 components.)
4930
4931 49. Some changes have been made to the callout feature (see 28 above):
4932
4933 (i) A callout function now has three choices for what it returns:
4934
4935 0 => success, carry on matching
4936 > 0 => failure at this point, but backtrack if possible
4937 < 0 => serious error, return this value from pcre_exec()
4938
4939 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4940 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4941 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4942 use by callout functions. It will never be used by PCRE itself.
4943
4944 (ii) The pcre_extra structure (see 45 above) has a void * field called
4945 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4946 pcre_callout_block structure has a field of the same name. The contents of
4947 the field passed in the pcre_extra structure are passed to the callout
4948 function in the corresponding field in the callout block. This makes it
4949 easier to use the same callout-containing regex from multiple threads. For
4950 testing, the pcretest program has a new data escape
4951
4952 \C*n pass the number n (may be negative) as callout_data
4953
4954 If the callout function in pcretest receives a non-zero value as
4955 callout_data, it returns that value.
4956
4957 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4958 there were some redundant $(CFLAGS) in commands that are now specified as
4959 $(LINK), which already includes $(CFLAGS).
4960
4961 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4962 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4963 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4964 one-byte characters throughout. Note that case-insensitive matching applies
4965 only to characters whose values are less than 256. PCRE doesn't support the
4966 notion of cases for higher-valued characters.
4967
4968 (i) A character class whose characters are all within 0-255 is handled as
4969 a bit map, and the map is inverted for negative classes. Previously, a
4970 character > 255 always failed to match such a class; however it should
4971 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4972
4973 (ii) A negated character class with a single character < 255 is coded as
4974 "not this character" (OP_NOT). This wasn't working properly when the test
4975 character was multibyte, either singly or repeated.
4976
4977 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4978 mode, for example: \x{100}{2,3}.
4979
4980 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4981 singly or repeated) now correctly test multibyte characters. However,
4982 PCRE doesn't recognize any characters with values greater than 255 as
4983 digits, spaces, or word characters. Such characters always match \D, \S,
4984 and \W, and never match \d, \s, or \w.
4985
4986 (v) Classes may now contain characters and character ranges with values
4987 greater than 255. For example: [ab\x{100}-\x{400}].
4988
4989 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4990 PCRE in UTF-8 mode.
4991
4992 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4993 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4994 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4995 value.)
4996
4997 53. The single man page has become too large. I have therefore split it up into
4998 a number of separate man pages. These also give rise to individual HTML pages;
4999 these are now put in a separate directory, and there is an index.html page that
5000 lists them all. Some hyperlinking between the pages has been installed.
5001
5002 54. Added convenience functions for handling named capturing parentheses.
5003
5004 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5005 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5006 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5007 are faulted.
5008
5009 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5010 calling configure. These values are used when compiling the dftables.c program
5011 which is run to generate the source of the default character tables. They
5012 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5013 you will need to set these values.
5014
5015 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5016
5017
5018 Version 3.9 02-Jan-02
5019 ---------------------
5020
5021 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5022
5023 2. If --disable-static was given, the building process failed when trying to
5024 build pcretest and pcregrep. (For some reason it was using libtool to compile
5025 them, which is not right, as they aren't part of the library.)
5026
5027
5028 Version 3.8 18-Dec-01
5029 ---------------------
5030
5031 1. The experimental UTF-8 code was completely screwed up. It was packing the
5032 bytes in the wrong order. How dumb can you get?
5033
5034
5035 Version 3.7 29-Oct-01
5036 ---------------------
5037
5038 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5039 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5040 this didn't happen under Solaris 8, where I normally test things.
5041
5042 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5043 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5044 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5045 this fix an hour or so after the initial 3.7 release.)
5046
5047
5048 Version 3.6 23-Oct-01
5049 ---------------------
5050
5051 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5052 offsets passed as NULL with zero offset count.
5053
5054 2. The config.guess and config.sub files had not been updated when I moved to
5055 the latest autoconf.
5056
5057
5058 Version 3.5 15-Aug-01
5059 ---------------------
5060
5061 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5062 had been forgotten.
5063
5064 2. By using declared but undefined structures, we can avoid using "void"
5065 definitions in pcre.h while keeping the internal definitions of the structures
5066 private.
5067
5068 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5069 user point of view, this means that both static and shared libraries are built
5070 by default, but this can be individually controlled. More of the work of
5071 handling this static/shared cases is now inside libtool instead of PCRE's make
5072 file.
5073
5074 4. The pcretest utility is now installed along with pcregrep because it is
5075 useful for users (to test regexs) and by doing this, it automatically gets
5076 relinked by libtool. The documentation has been turned into a man page, so
5077 there are now .1, .txt, and .html versions in /doc.
5078
5079 5. Upgrades to pcregrep:
5080 (i) Added long-form option names like gnu grep.
5081 (ii) Added --help to list all options with an explanatory phrase.
5082 (iii) Added -r, --recursive to recurse into sub-directories.
5083 (iv) Added -f, --file to read patterns from a file.
5084
5085 6. pcre_exec() was referring to its "code" argument before testing that
5086 argument for NULL (and giving an error if it was NULL).
5087
5088 7. Upgraded Makefile.in to allow for compiling in a different directory from
5089 the source directory.
5090
5091 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5092 options bits, the pointer it was passed was to an int instead of to an unsigned
5093 long int. This mattered only on 64-bit systems.
5094
5095 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5096 generated) instead of pcre.in, which it its source. Also made the same change
5097 in several of the .c files.
5098
5099 10. A new release of gcc defines printf() as a macro, which broke pcretest
5100 because it had an ifdef in the middle of a string argument for printf(). Fixed
5101 by using separate calls to printf().
5102
5103 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5104 script, to force use of CR or LF instead of \n in the source. On non-Unix
5105 systems, the value can be set in config.h.
5106
5107 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5108 absolute limit. Changed the text of the error message to make this clear, and
5109 likewise updated the man page.
5110
5111 13. The limit of 99 on the number of capturing subpatterns has been removed.
5112 The new limit is 65535, which I hope will not be a "real" limit.
5113
5114
5115 Version 3.4 22-Aug-00
5116 ---------------------
5117
5118 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5119
5120 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5121
5122
5123 Version 3.3 01-Aug-00
5124 ---------------------
5125
5126 1. If an octal character was given, but the value was greater than \377, it
5127 was not getting masked to the least significant bits, as documented. This could
5128 lead to crashes in some systems.
5129
5130 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5131 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5132
5133 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5134 These just pass their arguments on to (pcre_free)(), but they are provided
5135 because some uses of PCRE bind it to non-C systems that can call its functions,
5136 but cannot call free() or pcre_free() directly.
5137
5138 4. Add "make test" as a synonym for "make check". Corrected some comments in
5139 the Makefile.
5140
5141 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5142 Makefile.
5143
5144 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5145 command called pgrep for grepping around the active processes.
5146
5147 7. Added the beginnings of support for UTF-8 character strings.
5148
5149 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5150 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5151 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5152 out for the ar command.)
5153
5154
5155 Version 3.2 12-May-00
5156 ---------------------
5157
5158 This is purely a bug fixing release.
5159
5160 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5161 of ZA. This was just one example of several cases that could provoke this bug,
5162 which was introduced by change 9 of version 2.00. The code for breaking
5163 infinite loops after an iteration that matches an empty string was't working
5164 correctly.
5165
5166 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5167 when matched against abbab (for example). After matching an empty string, it
5168 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5169 caused it to match further down the string than it should.
5170
5171 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5172 was there because it doesn't seem to be needed, and it causes trouble on some
5173 systems, as it is not a Standard C header. It has been removed.
5174
5175 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5176 were reported on the Macintosh. The changes were from
5177
5178 while ((c = *(++ptr)) != 0 && c != '\n');
5179 to
5180 while ((c = *(++ptr)) != 0 && c != '\n') ;
5181
5182 Totally extraordinary, but if that's what it takes...
5183
5184 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5185 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5186 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5187 assumes the way PCRE uses memmove() (always moving upwards).
5188
5189 6. PCRE is being used in one environment where strchr() is not available. There
5190 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5191 faster code anyway.
5192
5193
5194 Version 3.1 09-Feb-00
5195 ---------------------
5196
5197 The only change in this release is the fixing of some bugs in Makefile.in for
5198 the "install" target:
5199
5200 (1) It was failing to install pcreposix.h.
5201
5202 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5203
5204
5205 Version 3.0 01-Feb-00
5206 ---------------------
5207
5208 1. Add support for the /+ modifier to perltest (to output $` like it does in
5209 pcretest).
5210
5211 2. Add support for the /g modifier to perltest.
5212
5213 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5214 matches null strings.
5215
5216 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5217 pattern. Perl treats empty patterns specially - it reuses the most recent
5218 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5219 effect.
5220
5221 5. The POSIX interface was broken in that it was just handing over the POSIX
5222 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5223 required a bigger vector, with some working space on the end. This means that
5224 the POSIX wrapper now has to get and free some memory, and copy the results.
5225
5226 6. Added some simple autoconf support, placing the test data and the
5227 documentation in separate directories, re-organizing some of the
5228 information files, and making it build pcre-config (a GNU standard). Also added
5229 libtool support for building PCRE as a shared library, which is now the
5230 default.
5231
5232 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5233 09 are not valid octal constants. Single digits will be used for minor values
5234 less than 10.
5235
5236 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5237 existing programs that set these in the POSIX interface can use PCRE without
5238 modification.
5239
5240 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5241 return all that pcre_info() returns, plus additional data. The pcre_info()
5242 function is retained for compatibility, but is considered to be obsolete.
5243
5244 10. Added experimental recursion feature (?R) to handle one common case that
5245 Perl 5.6 will be able to do with (?p{...}).
5246
5247 11. Added support for POSIX character classes like [:alpha:], which Perl is
5248 adopting.
5249
5250
5251 Version 2.08 31-Aug-99
5252 ----------------------
5253
5254 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5255 trying to match at the startoffset position, but instead was moving forward to
5256 the next newline as if a previous match had failed.
5257
5258 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5259 and could get into a loop if a null string was matched other than at the start
5260 of the subject.
5261
5262 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5263 be distinguished at compile time, and for completeness also added PCRE_DATE.
5264
5265 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5266 in GnuWin32 environments.
5267
5268
5269 Version 2.07 29-Jul-99
5270 ----------------------
5271
5272 1. The documentation is now supplied in plain text form and HTML as well as in
5273 the form of man page sources.
5274
5275 2. C++ compilers don't like assigning (void *) values to other pointer types.
5276 In particular this affects malloc(). Although there is no problem in Standard
5277 C, I've put in casts to keep C++ compilers happy.
5278
5279 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5280 should be (const char *).
5281
5282 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5283 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5284 However, I haven't made this a standard facility. The documentation doesn't
5285 mention it, and the Makefile doesn't support it.
5286
5287 5. The Makefile now contains an "install" target, with editable destinations at
5288 the top of the file. The pcretest program is not installed.
5289
5290 6. pgrep -V now gives the PCRE version number and date.
5291
5292 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5293 causing the entire string to be ignored, instead of just the last character.
5294
5295 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5296 non-matching string, it can take a very, very long time, even for strings of
5297 quite modest length, because of the nested recursion. PCRE now does better in
5298 some of these cases. It does this by remembering the last required literal
5299 character in the pattern, and pre-searching the subject to ensure it is present
5300 before running the real match. In other words, it applies a heuristic to detect
5301 some types of certain failure quickly, and in the above example, if presented
5302 with a string that has no trailing " it gives "no match" very quickly.
5303
5304 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5305 other alternatives are tried instead.
5306
5307
5308 Version 2.06 09-Jun-99
5309 ----------------------
5310
5311 1. Change pcretest's output for amount of store used to show just the code
5312 space, because the remainder (the data block) varies in size between 32-bit and
5313 64-bit systems.
5314
5315 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5316 start matching at. This allows lookbehinds to work when searching for multiple
5317 occurrences in a string.
5318
5319 3. Added additional options to pcretest for testing multiple occurrences:
5320
5321 /+ outputs the rest of the string that follows a match
5322 /g loops for multiple occurrences, using the new startoffset argument
5323 /G loops for multiple occurrences by passing an incremented pointer
5324
5325 4. PCRE wasn't doing the "first character" optimization for patterns starting
5326 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5327 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5328 the letter 'x'. On long subject strings, this gives a significant speed-up.
5329
5330
5331 Version 2.05 21-Apr-99
5332 ----------------------
5333
5334 1. Changed the type of magic_number from int to long int so that it works
5335 properly on 16-bit systems.
5336
5337 2. Fixed a bug which caused patterns starting with .* not to work correctly
5338 when the subject string contained newline characters. PCRE was assuming
5339 anchoring for such patterns in all cases, which is not correct because .* will
5340 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5341 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5342 must be retried after every newline in the subject.
5343
5344
5345 Version 2.04 18-Feb-99
5346 ----------------------
5347
5348 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5349 computation of the store needed to hold the pattern was incorrect (too large).
5350 If such patterns were nested a few deep, this could multiply and become a real
5351 problem.
5352
5353 2. Added /M option to pcretest to show the memory requirement of a specific
5354 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5355
5356 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5357 compiled in such a way that the backtracking after subsequent failure was
5358 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5359 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5360
5361
5362 Version 2.03 02-Feb-99
5363 ----------------------
5364
5365 1. Fixed typo and small mistake in man page.
5366
5367 2. Added 4th condition (GPL supersedes if conflict) and created separate
5368 LICENCE file containing the conditions.
5369
5370 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5371 Perl, that is the internal \ allows the delimiter to be included in the
5372 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5373 the final delimiter, add \ to the end of the pattern (to test the error).
5374
5375 4. Added the convenience functions for extracting substrings after a successful
5376 match. Updated pcretest to make it able to test these functions.
5377
5378
5379 Version 2.02 14-Jan-99
5380 ----------------------
5381
5382 1. Initialized the working variables associated with each extraction so that
5383 their saving and restoring doesn't refer to uninitialized store.
5384
5385 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5386 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5387 fix the problem.
5388
5389 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5390 calls, and wasn't printing the correct value for compiling calls. Increased the
5391 default value of LOOPREPEAT, and the number of significant figures in the
5392 times.
5393
5394 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5395
5396 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5397 a building problem on Windows NT with a FAT file system.
5398
5399
5400 Version 2.01 21-Oct-98
5401 ----------------------
5402
5403 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5404 to character tables built by pcre_maketables() in the current locale. If NULL
5405 is passed, the default tables are used.
5406
5407
5408 Version 2.00 24-Sep-98
5409 ----------------------
5410
5411 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5412 it any more.
5413
5414 2. Allow quantification of (?>) groups, and make it work correctly.
5415
5416 3. The first character computation wasn't working for (?>) groups.
5417
5418 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5419 end of the subject) and add 5.005's \z, which really does match only at the
5420 very end of the subject.
5421
5422 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5423
5424 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5425 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5426 localized options. All options to pcre_study() were also removed.
5427
5428 7. Add other new features from 5.005:
5429
5430 $(?<= positive lookbehind
5431 $(?<! negative lookbehind
5432 (?imsx-imsx) added the unsetting capability
5433 such a setting is global if at outer level; local otherwise
5434 (?imsx-imsx:) non-capturing groups with option setting
5435 (?(cond)re|re) conditional pattern matching
5436
5437 A backreference to itself in a repeated group matches the previous
5438 captured string.
5439
5440 8. General tidying up of studying (both automatic and via "study")
5441 consequential on the addition of new assertions.
5442
5443 9. As in 5.005, unlimited repeated groups that could match an empty substring
5444 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5445 runtime if any iteration does actually match an empty substring.
5446
5447 10. Include the RunTest script in the distribution.
5448
5449 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5450 discrepancies, some of which were old and were also with respect to 5.004. They
5451 have now been fixed.
5452
5453
5454 Version 1.09 28-Apr-98
5455 ----------------------
5456
5457 1. A negated single character class followed by a quantifier with a minimum
5458 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5459 program crashes, or just wrong answers. This did not apply to negated classes
5460 containing more than one character, or to minima other than one.
5461
5462
5463 Version 1.08 27-Mar-98
5464 ----------------------
5465
5466 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5467
5468 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5469 latter must appear before anything that relies on it in the pattern.
5470
5471
5472 Version 1.07 16-Feb-98
5473 ----------------------
5474
5475 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5476 repeat of a potentially empty string).
5477
5478
5479 Version 1.06 23-Jan-98
5480 ----------------------
5481
5482 1. Added Markus Oberhumer's little patches for C++.
5483
5484 2. Literal strings longer than 255 characters were broken.
5485
5486
5487 Version 1.05 23-Dec-97
5488 ----------------------
5489
5490 1. Negated character classes containing more than one character were failing if
5491 PCRE_CASELESS was set at run time.
5492
5493
5494 Version 1.04 19-Dec-97
5495 ----------------------
5496
5497 1. Corrected the man page, where some "const" qualifiers had been omitted.
5498
5499 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5500 input syntax.
5501
5502 3. Fixed memory leak which occurred when a regex with back references was
5503 matched with an offsets vector that wasn't big enough. The temporary memory
5504 that is used in this case wasn't being freed if the match failed.
5505
5506 4. Tidied pcretest to ensure it frees memory that it gets.
5507
5508 5. Temporary memory was being obtained in the case where the passed offsets
5509 vector was exactly big enough.
5510
5511 6. Corrected definition of offsetof() from change 5 below.
5512
5513 7. I had screwed up change 6 below and broken the rules for the use of
5514 setjmp(). Now fixed.
5515
5516
5517 Version 1.03 18-Dec-97
5518 ----------------------
5519
5520 1. A erroneous regex with a missing opening parenthesis was correctly
5521 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5522 on some systems.
5523
5524 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5525 it was reported that one broken compiler failed on the former because "code" is
5526 also an independent variable.
5527
5528 3. The erroneous regex a[]b caused an array overrun reference.
5529
5530 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5531 fail on data ending with that character. (It was going on too far, and checking
5532 the next character, typically a binary zero.) This was specific to the
5533 optimized code for single-character negative classes.
5534
5535 5. Added a contributed patch from the TIN world which does the following:
5536
5537 + Add an undef for memmove, in case the the system defines a macro for it.
5538
5539 + Add a definition of offsetof(), in case there isn't one. (I don't know
5540 the reason behind this - offsetof() is part of the ANSI standard - but
5541 it does no harm).
5542
5543 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5544 most of the places where whitespace preceded '#'. I have given up and
5545 allowed the remaining 2 cases to be at the margin.
5546
5547 + Rename some variables in pcre to eliminate shadowing. This seems very
5548 pedantic, but does no harm, of course.
5549
5550 6. Moved the call to setjmp() into its own function, to get rid of warnings
5551 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5552
5553 7. Constructs such as \d{8,} were compiling into the equivalent of
5554 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5555 outcome, but in this particular case used more store than had been allocated,
5556 which caused the bug to be discovered because it threw up an internal error.
5557
5558 8. The debugging code in both pcre and pcretest for outputting the compiled
5559 form of a regex was going wrong in the case of back references followed by
5560 curly-bracketed repeats.
5561
5562
5563 Version 1.02 12-Dec-97
5564 ----------------------
5565
5566 1. Typos in pcre.3 and comments in the source fixed.
5567
5568 2. Applied a contributed patch to get rid of places where it used to remove
5569 'const' from variables, and fixed some signed/unsigned and uninitialized
5570 variable warnings.
5571
5572 3. Added the "runtest" target to Makefile.
5573
5574 4. Set default compiler flag to -O2 rather than just -O.
5575
5576
5577 Version 1.01 19-Nov-97
5578 ----------------------
5579
5580 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5581 like /([ab]*)*/, that is, for classes with more than one character in them.
5582
5583 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5584 as /((?>a*))*/ (a PCRE_EXTRA facility).
5585
5586
5587 Version 1.00 18-Nov-97
5588 ----------------------
5589
5590 1. Added compile-time macros to support systems such as SunOS4 which don't have
5591 memmove() or strerror() but have other things that can be used instead.
5592
5593 2. Arranged that "make clean" removes the executables.
5594
5595
5596 Version 0.99 27-Oct-97
5597 ----------------------
5598
5599 1. Fixed bug in code for optimizing classes with only one character. It was
5600 initializing a 32-byte map regardless, which could cause it to run off the end
5601 of the memory it had got.
5602
5603 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5604
5605
5606 Version 0.98 22-Oct-97
5607 ----------------------
5608
5609 1. Fixed bug in code for handling temporary memory usage when there are more
5610 back references than supplied space in the ovector. This could cause segfaults.
5611
5612
5613 Version 0.97 21-Oct-97
5614 ----------------------
5615
5616 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5617
5618 2. Optimized negated single characters not to use a bit map.
5619
5620 3. Brought error texts together as macro definitions; clarified some of them;
5621 fixed one that was wrong - it said "range out of order" when it meant "invalid
5622 escape sequence".
5623
5624 4. Changed some char * arguments to const char *.
5625
5626 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5627
5628 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5629 pcretest.
5630
5631
5632 Version 0.96 16-Oct-97
5633 ----------------------
5634
5635 1. Added a simple "pgrep" utility to the distribution.
5636
5637 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5638 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5639 where "ddd" means "one or more decimal digits".
5640
5641 3. Fixed serious bug. If a pattern had a back reference, but the call to
5642 pcre_exec() didn't supply a large enough ovector to record the related
5643 identifying subpattern, the match always failed. PCRE now remembers the number
5644 of the largest back reference, and gets some temporary memory in which to save
5645 the offsets during matching if necessary, in order to ensure that
5646 backreferences always work.
5647
5648 4. Increased the compatibility with Perl in a number of ways:
5649
5650 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5651 to request this handling. The option can be set at compile or exec time.
5652
5653 (b) $ matches before a terminating newline by default; an option
5654 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5655 mode). The option can be set at compile or exec time.
5656
5657 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5658 the same as Perl's. If the decimal number it represents is less than 10
5659 or there aren't that many previous left capturing parentheses, an octal
5660 escape is read. Inside a character class, it's always an octal escape,
5661 even if it is a single digit.
5662
5663 (d) An escaped but undefined alphabetic character is taken as a literal,
5664 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5665 escapes.
5666
5667 (e) {0} is now permitted. (The previous item is removed from the compiled
5668 pattern).
5669
5670 5. Changed all the names of code files so that the basic parts are no longer
5671 than 10 characters, and abolished the teeny "globals.c" file.
5672
5673 6. Changed the handling of character classes; they are now done with a 32-byte
5674 bit map always.
5675
5676 7. Added the -d and /D options to pcretest to make it possible to look at the
5677 internals of compilation without having to recompile pcre.
5678
5679
5680 Version 0.95 23-Sep-97
5681 ----------------------
5682
5683 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5684 \x20 at the start of a run of normal characters. These were being treated as
5685 real characters, instead of the source characters being re-checked.
5686
5687
5688 Version 0.94 18-Sep-97
5689 ----------------------
5690
5691 1. The functions are now thread-safe, with the caveat that the global variables
5692 containing pointers to malloc() and free() or alternative functions are the
5693 same for all threads.
5694
5695 2. Get pcre_study() to generate a bitmap of initial characters for non-
5696 anchored patterns when this is possible, and use it if passed to pcre_exec().
5697
5698
5699 Version 0.93 15-Sep-97
5700 ----------------------
5701
5702 1. /(b)|(:+)/ was computing an incorrect first character.
5703
5704 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5705 but not actually doing anything yet.
5706
5707 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5708 as Perl does (e.g. [-az] or [az-]).
5709
5710 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5711 all possible positions.
5712
5713 5. Split up into different modules to avoid including unneeded functions in a
5714 compiled binary. However, compile and exec are still in one module. The "study"
5715 function is split off.
5716
5717 6. The character tables are now in a separate module whose source is generated
5718 by an auxiliary program - but can then be edited by hand if required. There are
5719 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5720 toupper() in the code.
5721
5722 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5723 make them global. Abolish the function for setting them, as the caller can now
5724 set them directly.
5725
5726
5727 Version 0.92 11-Sep-97
5728 ----------------------
5729
5730 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5731 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5732
5733 2. Caseless matching was not working in character classes if the characters in
5734 the pattern were in upper case.
5735
5736 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5737
5738 4. Make PCRE_ANCHORED public and accept as a compile option.
5739
5740 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5741 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5742 pass them.
5743
5744 6. Give an error if bad option bits passed at compile or run time.
5745
5746 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5747 pcretest to cause it to pass that flag.
5748
5749 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5750 options, and the first character, if set.
5751
5752 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5753
5754
5755 Version 0.91 10-Sep-97
5756 ----------------------
5757
5758 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5759 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5760
5761 2. PCRE was looping on encountering an indefinitely repeated back reference to
5762 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5763 Perl does - treats the match as successful.
5764
5765 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5