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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1532 - (show annotations)
Fri Mar 6 11:59:39 2015 UTC (4 years, 7 months ago) by ph10
File size: 273726 byte(s)
Error occurred while calculating annotation data.
Fix bad compile of patterns like /[A-`]/i8 where the range contains 
characters with multiple other cases and the ranges adjoin.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.37 xx-xxx-2015
5 ------------------------
6
7 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
8 for those parentheses to be closed with whatever has been captured so far.
9 However, it was failing to mark any other groups between the hightest
10 capture so far and the currrent group as "unset". Thus, the ovector for
11 those groups contained whatever was previously there. An example is the
12 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
13
14 2. If an assertion condition was quantified with a minimum of zero (an odd
15 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
16
17 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
18 unrecognized modifier, a crash could occur.
19
20 4. An attempt to do global matching in pcretest with a zero-length ovector
21 caused a crash.
22
23 5. Fixed a memory leak during matching that could occur for a subpattern
24 subroutine call (recursive or otherwise) if the number of captured groups
25 that had to be saved was greater than ten.
26
27 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
28 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
29 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
30
31 7. A UTF pattern containing a "not" match of a non-ASCII character and a
32 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
33
34 8. When a pattern is compiled, it remembers the highest back reference so that
35 when matching, if the ovector is too small, extra memory can be obtained to
36 use instead. A conditional subpattern whose condition is a check on a
37 capture having happened, such as, for example in the pattern
38 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
39 setting the highest backreference number. This mattered only if pcre_exec()
40 was called with an ovector that was too small to hold the capture, and there
41 was no other kind of back reference (a situation which is probably quite
42 rare). The effect of the bug was that the condition was always treated as
43 FALSE when the capture could not be consulted, leading to a incorrect
44 behaviour by pcre2_match(). This bug has been fixed.
45
46 9. A reference to a duplicated named group (either a back reference or a test
47 for being set in a conditional) that occurred in a part of the pattern where
48 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
49 to be incorrectly calculated, leading to overwriting.
50
51 10. A mutually recursive set of back references such as (\2)(\1) caused a
52 segfault at study time (while trying to find the minimum matching length).
53 The infinite loop is now broken (with the minimum length unset, that is,
54 zero).
55
56 11. If an assertion that was used as a condition was quantified with a minimum
57 of zero, matching went wrong. In particular, if the whole group had
58 unlimited repetition and could match an empty string, a segfault was
59 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
60 assertions to be quantified, but not if they are being used as conditions,
61 so the above pattern is faulted by Perl. PCRE has now been changed so that
62 it also rejects such patterns.
63
64 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
65 failed to allow the zero-repeat case if pcre2_exec() was called with an
66 ovector too small to capture the group.
67
68 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
69 Red Hat Product Security:
70
71 (a) A crash if /K and /F were both set with the option to save the compiled
72 pattern.
73
74 (b) Another crash if the option to print captured substrings in a callout
75 was combined with setting a null ovector, for example \O\C+ as a subject
76 string.
77
78 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
79 forward reference repeated a large (but limited) number of times within a
80 repeated outer group that has a zero minimum quantifier, caused incorrect
81 code to be compiled, leading to the error "internal error:
82 previously-checked referenced subpattern not found" when an incorrect
83 memory address was read. This bug was reported as "heap overflow",
84 discovered by Kai Lu of Fortinet's FortiGuard Labs.
85
86 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
87 call within a group that also contained a recursive back reference caused
88 incorrect code to be compiled. This bug was reported as "heap overflow",
89 discovered by Kai Lu of Fortinet's FortiGuard Labs.
90
91 24. Computing the size of the JIT read-only data in advance has been a source
92 of various issues, and new ones are still appear unfortunately. To fix
93 existing and future issues, size computation is eliminated from the code,
94 and replaced by on-demand memory allocation.
95
96 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
97 adjacent to the end of the range, and the range contained characters with
98 more than one other case, caused incorrect behaviour when compiled in UTF
99 mode. In that example, the range a-j was left out of the class.
100
101
102 Version 8.36 26-September-2014
103 ------------------------------
104
105 1. Got rid of some compiler warnings in the C++ modules that were shown up by
106 -Wmissing-field-initializers and -Wunused-parameter.
107
108 2. The tests for quantifiers being too big (greater than 65535) were being
109 applied after reading the number, and stupidly assuming that integer
110 overflow would give a negative number. The tests are now applied as the
111 numbers are read.
112
113 3. Tidy code in pcre_exec.c where two branches that used to be different are
114 now the same.
115
116 4. The JIT compiler did not generate match limit checks for certain
117 bracketed expressions with quantifiers. This may lead to exponential
118 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
119 issue should be resolved now.
120
121 5. Fixed an issue, which occures when nested alternatives are optimized
122 with table jumps.
123
124 6. Inserted two casts and changed some ints to size_t in the light of some
125 reported 64-bit compiler warnings (Bugzilla 1477).
126
127 7. Fixed a bug concerned with zero-minimum possessive groups that could match
128 an empty string, which sometimes were behaving incorrectly in the
129 interpreter (though correctly in the JIT matcher). This pcretest input is
130 an example:
131
132 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
133 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
134
135 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
136 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
137 for an empty string was breaking the inner loop and carrying on at a lower
138 level, when possessive repeated groups should always return to a higher
139 level as they have no backtrack points in them. The empty string test now
140 occurs at the outer level.
141
142 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
143 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
144
145 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
146 doesn't).
147
148 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
149 to hold a pattern to be incorrectly computed (too small) when there were
150 named back references to duplicated names. This could cause "internal
151 error: code overflow" or "double free or corruption" or other memory
152 handling errors.
153
154 11. When named subpatterns had the same prefixes, back references could be
155 confused. For example, in this pattern:
156
157 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
158
159 the reference to 'Name' was incorrectly treated as a reference to a
160 duplicate name.
161
162 12. A pattern such as /^s?c/mi8 where the optional character has more than
163 one "other case" was incorrectly compiled such that it would only try to
164 match starting at "c".
165
166 13. When a pattern starting with \s was studied, VT was not included in the
167 list of possible starting characters; this should have been part of the
168 8.34/18 patch.
169
170 14. If a character class started [\Qx]... where x is any character, the class
171 was incorrectly terminated at the ].
172
173 15. If a pattern that started with a caseless match for a character with more
174 than one "other case" was studied, PCRE did not set up the starting code
175 unit bit map for the list of possible characters. Now it does. This is an
176 optimization improvement, not a bug fix.
177
178 16. The Unicode data tables have been updated to Unicode 7.0.0.
179
180 17. Fixed a number of memory leaks in pcregrep.
181
182 18. Avoid a compiler warning (from some compilers) for a function call with
183 a cast that removes "const" from an lvalue by using an intermediate
184 variable (to which the compiler does not object).
185
186 19. Incorrect code was compiled if a group that contained an internal recursive
187 back reference was optional (had quantifier with a minimum of zero). This
188 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
189 caused segmentation faults because of stack overflows at compile time.
190
191 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
192 group that is quantified with an indefinite repeat, caused a compile-time
193 loop which used up all the system stack and provoked a segmentation fault.
194 This was not the same bug as 19 above.
195
196 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
197 Patch by Mike Frysinger.
198
199
200 Version 8.35 04-April-2014
201 --------------------------
202
203 1. A new flag is set, when property checks are present in an XCLASS.
204 When this flag is not set, PCRE can perform certain optimizations
205 such as studying these XCLASS-es.
206
207 2. The auto-possessification of character sets were improved: a normal
208 and an extended character set can be compared now. Furthermore
209 the JIT compiler optimizes more character set checks.
210
211 3. Got rid of some compiler warnings for potentially uninitialized variables
212 that show up only when compiled with -O2.
213
214 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
215 of a match later then the end of the match. The pcretest program was not
216 handling the case sensibly - it was outputting from the start to the next
217 binary zero. It now reports this situation in a message, and outputs the
218 text from the end to the start.
219
220 5. Fast forward search is improved in JIT. Instead of the first three
221 characters, any three characters with fixed position can be searched.
222 Search order: first, last, middle.
223
224 6. Improve character range checks in JIT. Characters are read by an inprecise
225 function now, which returns with an unknown value if the character code is
226 above a certain threshold (e.g: 256). The only limitation is that the value
227 must be bigger than the threshold as well. This function is useful when
228 the characters above the threshold are handled in the same way.
229
230 7. The macros whose names start with RAWUCHAR are placeholders for a future
231 mode in which only the bottom 21 bits of 32-bit data items are used. To
232 make this more memorable for those maintaining the code, the names have
233 been changed to start with UCHAR21, and an extensive comment has been added
234 to their definition.
235
236 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
237 to the export list in Makefile.am (they were accidentally omitted from the
238 8.34 tarball).
239
240 9. The informational output from pcretest used the phrase "starting byte set"
241 which is inappropriate for the 16-bit and 32-bit libraries. As the output
242 for "first char" and "need char" really means "non-UTF-char", I've changed
243 "byte" to "char", and slightly reworded the output. The documentation about
244 these values has also been (I hope) clarified.
245
246 10. Another JIT related optimization: use table jumps for selecting the correct
247 backtracking path, when more than four alternatives are present inside a
248 bracket.
249
250 11. Empty match is not possible, when the minimum length is greater than zero,
251 and there is no \K in the pattern. JIT should avoid empty match checks in
252 such cases.
253
254 12. In a caseless character class with UCP support, when a character with more
255 than one alternative case was not the first character of a range, not all
256 the alternative cases were added to the class. For example, s and \x{17f}
257 are both alternative cases for S: the class [RST] was handled correctly,
258 but [R-T] was not.
259
260 13. The configure.ac file always checked for pthread support when JIT was
261 enabled. This is not used in Windows, so I have put this test inside a
262 check for the presence of windows.h (which was already tested for).
263
264 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
265 The algorithm provides a way to skip certain starting offsets, and usually
266 faster than linear prefix searches.
267
268 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
269 as for 'fr_FR' and 'french'. For some reason, however, it then used the
270 Windows-specific input and output files, which have 'french' screwed in.
271 So this could never have worked. One of the problems with locales is that
272 they aren't always the same. I have now updated RunTest so that it checks
273 the output of the locale test (test 3) against three different output
274 files, and it allows the test to pass if any one of them matches. With luck
275 this should make the test pass on some versions of Solaris where it was
276 failing. Because of the uncertainty, the script did not used to stop if
277 test 3 failed; it now does. If further versions of a French locale ever
278 come to light, they can now easily be added.
279
280 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
281 there was a message during ./configure, but it did not stop. This now
282 provokes an error. The invalid example in README has been corrected.
283 If a value less than the minimum is given, the minimum value has always
284 been used, but now a warning is given.
285
286 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
287 was a bug in the test system, which is now fixed. Also, the list of various
288 configurations that are tested for each release did not have one with both
289 16/32 bits and --enable-bar-anycrlf. It now does.
290
291 18. pcretest was missing "-C bsr" for displaying the \R default setting.
292
293 19. Little endian PowerPC systems are supported now by the JIT compiler.
294
295 20. The fast forward newline mechanism could enter to an infinite loop on
296 certain invalid UTF-8 input. Although we don't support these cases
297 this issue can be fixed by a performance optimization.
298
299 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
300 not take account if existing stack usage. There is now a new global
301 variable called pcre_stack_guard that can be set to point to an external
302 function to check stack availability. It is called at the start of
303 processing every parenthesized group.
304
305 22. A typo in the code meant that in ungreedy mode the max/min qualifier
306 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
307 not match "ab".
308
309 23. When UTF was disabled, the JIT program reported some incorrect compile
310 errors. These messages are silenced now.
311
312 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
313 compiler.
314
315 25. Change all the temporary files used in RunGrepTest to be different to those
316 used by RunTest so that the tests can be run simultaneously, for example by
317 "make -j check".
318
319
320 Version 8.34 15-December-2013
321 -----------------------------
322
323 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
324 executable memory. Patch inspired by Carsten Klein.
325
326 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
327 this macro is never tested and has no effect, because the work to support
328 coverage involves only compiling and linking options and special targets in
329 the Makefile. The comment in config.h implied that defining the macro would
330 enable coverage support, which is totally false. There was also support for
331 setting this macro in the CMake files (my fault, I just copied it from
332 configure). SUPPORT_GCOV has now been removed.
333
334 3. Make a small performance improvement in strlen16() and strlen32() in
335 pcretest.
336
337 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
338 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
339 these cases). There was also one in pcretest.c.
340
341 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
342
343 6. In UTF mode, the code for checking whether a group could match an empty
344 string (which is used for indefinitely repeated groups to allow for
345 breaking an infinite loop) was broken when the group contained a repeated
346 negated single-character class with a character that occupied more than one
347 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
348 UTF-8 mode). The effect was undefined: the group might or might not be
349 deemed as matching an empty string, or the program might have crashed.
350
351 7. The code for checking whether a group could match an empty string was not
352 recognizing that \h, \H, \v, \V, and \R must match a character.
353
354 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
355 an empty string. If it can, pcretest shows this in its information output.
356
357 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
358 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
359 matched by pcre_exec() without using JIT:
360
361 (a) If the rest of the pattern did not match after a maximal run of
362 grapheme clusters, the code for backing up to try with fewer of them
363 did not always back up over a full grapheme when characters that do not
364 have the modifier quality were involved, e.g. Hangul syllables.
365
366 (b) If the match point in a subject started with modifier character, and
367 there was no match, the code could incorrectly back up beyond the match
368 point, and potentially beyond the first character in the subject,
369 leading to a segfault or an incorrect match result.
370
371 10. A conditional group with an assertion condition could lead to PCRE
372 recording an incorrect first data item for a match if no other first data
373 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
374 first data item, and therefore matched "ca" after "c" instead of at the
375 start.
376
377 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
378 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
379
380 12. The source of pcregrep now includes z/OS-specific code so that it can be
381 compiled for z/OS as part of the special z/OS distribution.
382
383 13. Added the -T and -TM options to pcretest.
384
385 14. The code in pcre_compile.c for creating the table of named capturing groups
386 has been refactored. Instead of creating the table dynamically during the
387 actual compiling pass, the information is remembered during the pre-compile
388 pass (on the stack unless there are more than 20 named groups, in which
389 case malloc() is used) and the whole table is created before the actual
390 compile happens. This has simplified the code (it is now nearly 150 lines
391 shorter) and prepared the way for better handling of references to groups
392 with duplicate names.
393
394 15. A back reference to a named subpattern when there is more than one of the
395 same name now checks them in the order in which they appear in the pattern.
396 The first one that is set is used for the reference. Previously only the
397 first one was inspected. This change makes PCRE more compatible with Perl.
398
399 16. Unicode character properties were updated from Unicode 6.3.0.
400
401 17. The compile-time code for auto-possessification has been refactored, based
402 on a patch by Zoltan Herczeg. It now happens after instead of during
403 compilation. The code is cleaner, and more cases are handled. The option
404 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
405 options in pcretest are provided to set it. It can also be set by
406 (*NO_AUTO_POSSESS) at the start of a pattern.
407
408 18. The character VT has been added to the default ("C" locale) set of
409 characters that match \s and are generally treated as white space,
410 following this same change in Perl 5.18. There is now no difference between
411 "Perl space" and "POSIX space". Whether VT is treated as white space in
412 other locales depends on the locale.
413
414 19. The code for checking named groups as conditions, either for being set or
415 for being recursed, has been refactored (this is related to 14 and 15
416 above). Processing unduplicated named groups should now be as fast at
417 numerical groups, and processing duplicated groups should be faster than
418 before.
419
420 20. Two patches to the CMake build system, by Alexander Barkov:
421
422 (1) Replace the "source" command by "." in CMakeLists.txt because
423 "source" is a bash-ism.
424
425 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
426 without these the CMake build does not work on Solaris.
427
428 21. Perl has changed its handling of \8 and \9. If there is no previously
429 encountered capturing group of those numbers, they are treated as the
430 literal characters 8 and 9 instead of a binary zero followed by the
431 literals. PCRE now does the same.
432
433 22. Following Perl, added \o{} to specify codepoints in octal, making it
434 possible to specify values greater than 0777 and also making them
435 unambiguous.
436
437 23. Perl now gives an error for missing closing braces after \x{... instead of
438 treating the string as literal. PCRE now does the same.
439
440 24. RunTest used to grumble if an inappropriate test was selected explicitly,
441 but just skip it when running all tests. This make it awkward to run ranges
442 of tests when one of them was inappropriate. Now it just skips any
443 inappropriate tests, as it always did when running all tests.
444
445 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
446 character types such as \d or \w, too many callouts were inserted, and the
447 data that they returned was rubbish.
448
449 26. In UCP mode, \s was not matching two of the characters that Perl matches,
450 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
451 were matched by \h. The code has now been refactored so that the lists of
452 the horizontal and vertical whitespace characters used for \h and \v (which
453 are defined only in one place) are now also used for \s.
454
455 27. Add JIT support for the 64 bit TileGX architecture.
456 Patch by Jiong Wang (Tilera Corporation).
457
458 28. Possessive quantifiers for classes (both explicit and automatically
459 generated) now use special opcodes instead of wrapping in ONCE brackets.
460
461 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
462 (because it's meaningless), this was not happening when PCRE_CASELESS was
463 set. Not wrong, but inefficient.
464
465 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
466 properties for \w, \d, etc) is present in a test regex. Otherwise if the
467 test contains no characters greater than 255, Perl doesn't realise it
468 should be using Unicode semantics.
469
470 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
471 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
472 does in Unicode mode.
473
474 32. Added the "forbid" facility to pcretest so that putting tests into the
475 wrong test files can sometimes be quickly detected.
476
477 33. There is now a limit (default 250) on the depth of nesting of parentheses.
478 This limit is imposed to control the amount of system stack used at compile
479 time. It can be changed at build time by --with-parens-nest-limit=xxx or
480 the equivalent in CMake.
481
482 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
483 errors. Perl warns for these when in warning mode, but PCRE has no facility
484 for giving warnings.
485
486 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
487 However, this was not working for (?!) because it is optimized to (*FAIL),
488 for which PCRE does not allow quantifiers. The optimization is now disabled
489 when a quantifier follows (?!). I can't see any use for this, but it makes
490 things uniform.
491
492 36. Perl no longer allows group names to start with digits, so I have made this
493 change also in PCRE. It simplifies the code a bit.
494
495 37. In extended mode, Perl ignores spaces before a + that indicates a
496 possessive quantifier. PCRE allowed a space before the quantifier, but not
497 before the possessive +. It now does.
498
499 38. The use of \K (reset reported match start) within a repeated possessive
500 group such as (a\Kb)*+ was not working.
501
502 40. Document that the same character tables must be used at compile time and
503 run time, and that the facility to pass tables to pcre_exec() and
504 pcre_dfa_exec() is for use only with saved/restored patterns.
505
506 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
507 features for Builds with MSVC:
508
509 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
510 builds). The .rc files can be used to set FileDescription and many other
511 attributes.
512
513 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
514 This allows higher-level build scripts which want .pdb files to avoid
515 hard-coding the exact files needed."
516
517 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
518 mean "start of word" and "end of word", respectively, as a transition aid.
519
520 43. A minimizing repeat of a class containing codepoints greater than 255 in
521 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
522 compiled to use the heap for recursion.
523
524 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
525 is configured.
526
527
528 Version 8.33 28-May-2013
529 ------------------------
530
531 1. Added 'U' to some constants that are compared to unsigned integers, to
532 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
533 variables that are added to signed variables, to ensure the result is
534 signed and can be negated.
535
536 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
537 CMake config files.
538
539 3. Revise the creation of config.h.generic so that all boolean macros are
540 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
541 overriding via -D on the command line possible.
542
543 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
544 to unsigned int is reported to make a quite noticeable speed difference in
545 a specific Windows environment. Testing on Linux did also appear to show
546 some benefit (and it is clearly not harmful). Also fixed the definition of
547 Xop which should be unsigned.
548
549 5. Related to (4), changing the definition of the intermediate variable cc
550 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
551 improvements.
552
553 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
554 unnecessary spaces.
555
556 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
557 and later.
558
559 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
560
561 9. Optimizing fast_forward_start_bits in JIT.
562
563 10. Adding support for callouts in JIT, and fixing some issues revealed
564 during this work. Namely:
565
566 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
567
568 (b) Minimum length was not checked before the matching is started.
569
570 11. The value of capture_last that is passed to callouts was incorrect in some
571 cases when there was a capture on one path that was subsequently abandoned
572 after a backtrack. Also, the capture_last value is now reset after a
573 recursion, since all captures are also reset in this case.
574
575 12. The interpreter no longer returns the "too many substrings" error in the
576 case when an overflowing capture is in a branch that is subsequently
577 abandoned after a backtrack.
578
579 13. In the pathological case when an offset vector of size 2 is used, pcretest
580 now prints out the matched string after a yield of 0 or 1.
581
582 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
583 Only supported by the JIT compiler at the moment.
584
585 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
586
587 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
588 offset of the starting point of the matching process, provided the offsets
589 vector is large enough.
590
591 17. The \A escape now records a lookbehind value of 1, though its execution
592 does not actually inspect the previous character. This is to ensure that,
593 in partial multi-segment matching, at least one character from the old
594 segment is retained when a new segment is processed. Otherwise, if there
595 are no lookbehinds in the pattern, \A might match incorrectly at the start
596 of a new segment.
597
598 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
599
600 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
601 gives some modest performance improvement in 8-bit mode.
602
603 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
604 be expressed in certain programming languages using Universal Character
605 Names.
606
607 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
608 which points out that "non characters" are not "characters that may not
609 appear in Unicode strings" but rather "characters that are reserved for
610 internal use and have only local meaning".
611
612 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
613 there was a conditional group that depended on an assertion, if the
614 assertion was false, the callout that immediately followed the alternation
615 in the condition was skipped when pcre_exec() was used for matching.
616
617 23. Allow an explicit callout to be inserted before an assertion that is the
618 condition for a conditional group, for compatibility with automatic
619 callouts, which always insert a callout at this point.
620
621 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
622 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
623
624 25. (*PRUNE) is now supported by the JIT compiler.
625
626 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
627
628 27. Fix the case where there are two or more SKIPs with arguments that may be
629 ignored.
630
631 28. (*SKIP) is now supported by the JIT compiler.
632
633 29. (*THEN) is now supported by the JIT compiler.
634
635 30. Update RunTest with additional test selector options.
636
637 31. The way PCRE handles backtracking verbs has been changed in two ways.
638
639 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
640 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
641 backtracking. In some cases this makes it more Perl-compatible, but Perl's
642 rather obscure rules do not always do the same thing.
643
644 (2) Previously, backtracking verbs were confined within assertions. This is
645 no longer the case for positive assertions, except for (*ACCEPT). Again,
646 this sometimes improves Perl compatibility, and sometimes does not.
647
648 32. A number of tests that were in test 2 because Perl did things differently
649 have been moved to test 1, because either Perl or PCRE has changed, and
650 these tests are now compatible.
651
652 32. Backtracking control verbs are now handled in the same way in JIT and
653 interpreter.
654
655 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
656 contained a forward subroutine reference caused a compile error.
657
658 34. Auto-detect and optimize limited repetitions in JIT.
659
660 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
661 blocking (*UTF) etc.
662
663 36. In the interpreter, maximizing pattern repetitions for characters and
664 character types now use tail recursion, which reduces stack usage.
665
666 37. The value of the max lookbehind was not correctly preserved if a compiled
667 and saved regex was reloaded on a host of different endianness.
668
669 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
670 of the compiled pattern block, expand the flags field from 16 to 32 bits
671 because it was almost full.
672
673 39. Try madvise first before posix_madvise.
674
675 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
676 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
677 original change locked out all matches of empty strings. This has been
678 changed so that one match of an empty string per line is recognized.
679 Subsequent searches on the same line (for colouring or for --only-matching,
680 for example) do not recognize empty strings.
681
682 41. Applied a user patch to fix a number of spelling mistakes in comments.
683
684 42. Data lines longer than 65536 caused pcretest to crash.
685
686 43. Clarified the data type for length and startoffset arguments for pcre_exec
687 and pcre_dfa_exec in the function-specific man pages, where they were
688 explicitly stated to be in bytes, never having been updated. I also added
689 some clarification to the pcreapi man page.
690
691 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
692 a segmentation fault.
693
694
695 Version 8.32 30-November-2012
696 -----------------------------
697
698 1. Improved JIT compiler optimizations for first character search and single
699 character iterators.
700
701 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
702 Patch by Daniel Richard G.
703
704 3. Single character iterator optimizations in the JIT compiler.
705
706 4. Improved JIT compiler optimizations for character ranges.
707
708 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
709 Reported by Giuseppe D'Angelo.
710
711 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
712 of a line, was being set incorrectly in cases where .* appeared inside
713 atomic brackets at the start of a pattern, or where there was a subsequent
714 *PRUNE or *SKIP.
715
716 7. Improved instruction cache flush for POWER/PowerPC.
717 Patch by Daniel Richard G.
718
719 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
720 grep:
721
722 (a) There is now no limit to the number of patterns to be matched.
723
724 (b) An error is given if a pattern is too long.
725
726 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
727 are now supported.
728
729 (d) --exclude-from and --include-from (multiple use) have been added.
730
731 (e) Exclusions and inclusions now apply to all files and directories, not
732 just to those obtained from scanning a directory recursively.
733
734 (f) Multiple uses of -f and --file-list are now supported.
735
736 (g) In a Windows environment, the default for -d has been changed from
737 "read" (the GNU grep default) to "skip", because otherwise the presence
738 of a directory in the file list provokes an error.
739
740 (h) The documentation has been revised and clarified in places.
741
742 9. Improve the matching speed of capturing brackets.
743
744 10. Changed the meaning of \X so that it now matches a Unicode extended
745 grapheme cluster.
746
747 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
748 out POSIX threads when JIT support is configured.
749
750 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
751
752 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
753 structure could go wrong in environments where size_t is not the same size
754 as int.
755
756 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
757 set.
758
759 15. The EBCDIC support had decayed; later updates to the code had included
760 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
761 general tidy up of EBCDIC-related issues, and the documentation was also
762 not quite right. There is now a test that can be run on ASCII systems to
763 check some of the EBCDIC-related things (but is it not a full test).
764
765 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
766 in a small tidy to the code.
767
768 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
769
770 18. If the --only-matching (-o) option in pcregrep is specified multiple
771 times, each one causes appropriate output. For example, -o1 -o2 outputs the
772 substrings matched by the 1st and 2nd capturing parentheses. A separating
773 string can be specified by --om-separator (default empty).
774
775 19. Improving the first n character searches.
776
777 20. Turn case lists for horizontal and vertical white space into macros so that
778 they are defined only once.
779
780 21. This set of changes together give more compatible Unicode case-folding
781 behaviour for characters that have more than one other case when UCP
782 support is available.
783
784 (a) The Unicode property table now has offsets into a new table of sets of
785 three or more characters that are case-equivalent. The MultiStage2.py
786 script that generates these tables (the pcre_ucd.c file) now scans
787 CaseFolding.txt instead of UnicodeData.txt for character case
788 information.
789
790 (b) The code for adding characters or ranges of characters to a character
791 class has been abstracted into a generalized function that also handles
792 case-independence. In UTF-mode with UCP support, this uses the new data
793 to handle characters with more than one other case.
794
795 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
796 whose other case is greater than 256 are now correctly matched
797 caselessly. Previously, the high codepoint matched the low one, but not
798 vice versa.
799
800 (d) The processing of \h, \H, \v, and \ in character classes now makes use
801 of the new class addition function, using character lists defined as
802 macros alongside the case definitions of 20 above.
803
804 (e) Caseless back references now work with characters that have more than
805 one other case.
806
807 (f) General caseless matching of characters with more than one other case
808 is supported.
809
810 22. Unicode character properties were updated from Unicode 6.2.0
811
812 23. Improved CMake support under Windows. Patch by Daniel Richard G.
813
814 24. Add support for 32-bit character strings, and UTF-32
815
816 25. Major JIT compiler update (code refactoring and bugfixing).
817 Experimental Sparc 32 support is added.
818
819 26. Applied a modified version of Daniel Richard G's patch to create
820 pcre.h.generic and config.h.generic by "make" instead of in the
821 PrepareRelease script.
822
823 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
824 pcre_compile.c when checking for a zero character.
825
826 28. Introducing a native interface for JIT. Through this interface, the compiled
827 machine code can be directly executed. The purpose of this interface is to
828 provide fast pattern matching, so several sanity checks are not performed.
829 However, feature tests are still performed. The new interface provides
830 1.4x speedup compared to the old one.
831
832 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
833 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
834 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
835
836 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
837 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
838 the "old" RFC 2279). Instead, it ended up passing a negative length to
839 pcre_exec().
840
841 31. Add support for GCC's visibility feature to hide internal functions.
842
843 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
844 "unknown -C option" after outputting 0 or 1.
845
846 33. There is now support for generating a code coverage report for the test
847 suite in environments where gcc is the compiler and lcov is installed. This
848 is mainly for the benefit of the developers.
849
850 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
851 unaddressable using valgrind annotations, allowing valgrind to detect
852 invalid memory accesses. This is mainly for the benefit of the developers.
853
854 25. (*UTF) can now be used to start a pattern in any of the three libraries.
855
856 26. Give configure error if --enable-cpp but no C++ compiler found.
857
858
859 Version 8.31 06-July-2012
860 -------------------------
861
862 1. Fixing a wrong JIT test case and some compiler warnings.
863
864 2. Removed a bashism from the RunTest script.
865
866 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
867 to unsigned type, result still unsigned" that was given by an MS compiler
868 on encountering the code "-sizeof(xxx)".
869
870 4. Partial matching support is added to the JIT compiler.
871
872 5. Fixed several bugs concerned with partial matching of items that consist
873 of more than one character:
874
875 (a) /^(..)\1/ did not partially match "aba" because checking references was
876 done on an "all or nothing" basis. This also applied to repeated
877 references.
878
879 (b) \R did not give a hard partial match if \r was found at the end of the
880 subject.
881
882 (c) \X did not give a hard partial match after matching one or more
883 characters at the end of the subject.
884
885 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
886 a partial match for the string "\r".
887
888 (e) When newline was set to CRLF, the metacharacter "." did not recognize
889 a partial match for a CR character at the end of the subject string.
890
891 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
892 running pcretest, the text "(JIT)" added to the output whenever JIT is
893 actually used to run the match.
894
895 7. Individual JIT compile options can be set in pcretest by following -s+[+]
896 or /S+[+] with a digit between 1 and 7.
897
898 8. OP_NOT now supports any UTF character not just single-byte ones.
899
900 9. (*MARK) control verb is now supported by the JIT compiler.
901
902 10. The command "./RunTest list" lists the available tests without actually
903 running any of them. (Because I keep forgetting what they all are.)
904
905 11. Add PCRE_INFO_MAXLOOKBEHIND.
906
907 12. Applied a (slightly modified) user-supplied patch that improves performance
908 when the heap is used for recursion (compiled with --disable-stack-for-
909 recursion). Instead of malloc and free for each heap frame each time a
910 logical recursion happens, frames are retained on a chain and re-used where
911 possible. This sometimes gives as much as 30% improvement.
912
913 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
914 call.
915
916 14. As documented, (*COMMIT) is now confined to within a positive assertion.
917
918 15. It is now possible to link pcretest with libedit as an alternative to
919 libreadline.
920
921 16. (*COMMIT) control verb is now supported by the JIT compiler.
922
923 17. The Unicode data tables have been updated to Unicode 6.1.0.
924
925 18. Added --file-list option to pcregrep.
926
927 19. Added binary file support to pcregrep, including the -a, --binary-files,
928 -I, and --text options.
929
930 20. The madvise function is renamed for posix_madvise for QNX compatibility
931 reasons. Fixed by Giuseppe D'Angelo.
932
933 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
934
935 22. Changed the option for creating symbolic links for 16-bit man pages from
936 -s to -sf so that re-installing does not cause issues.
937
938 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
939
940 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
941 matches in certain environments (the workspace was not being correctly
942 retained). Also added to pcre_dfa_exec() a simple plausibility check on
943 some of the workspace data at the beginning of a restart.
944
945 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
946 was not doing so when it should - probably a typo introduced by SVN 528
947 (change 8.10/14).
948
949 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
950 \w+ when the character tables indicated that \x{c4} was a word character.
951 There were several related cases, all because the tests for doing a table
952 lookup were testing for characters less than 127 instead of 255.
953
954 27. If a pattern contains capturing parentheses that are not used in a match,
955 their slots in the ovector are set to -1. For those that are higher than
956 any matched groups, this happens at the end of processing. In the case when
957 there were back references that the ovector was too small to contain
958 (causing temporary malloc'd memory to be used during matching), and the
959 highest capturing number was not used, memory off the end of the ovector
960 was incorrectly being set to -1. (It was using the size of the temporary
961 memory instead of the true size.)
962
963 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
964 ovector size, it uses memory at the end of the block that it has got.
965
966 29. Check for an overlong MARK name and give an error at compile time. The
967 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
968
969 30. JIT compiler update.
970
971 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
972 Rill for the patch.
973
974 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
975
976 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
977
978 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
979 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
980 was enabled.
981
982 35. Improve JIT code generation for greedy plus quantifier.
983
984 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
985 1 to "aa" instead of to an empty string. The bug affected repeated groups
986 that could potentially match an empty string.
987
988 37. Optimizing single character iterators in JIT.
989
990 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
991 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
992 codepoints that are too big for the mode are faulted, and in a UTF mode,
993 disallowed codepoints are also faulted.
994
995 39. If PCRE was compiled with UTF support, in three places in the DFA
996 matcher there was code that should only have been obeyed in UTF mode, but
997 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
998 processing when bytes with values greater than 127 were present. In 16-bit
999 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1000 both cases the values are those that cannot be the first data item in a UTF
1001 character. The three items that might have provoked this were recursions,
1002 possessively repeated groups, and atomic groups.
1003
1004 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1005 and pcregrep, because some OS require shared objects to be explicitly
1006 passed to ld, causing the link step to fail if they are not.
1007
1008 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1009 mode, patterns that started with \h* or \R* might be incorrectly matched.
1010
1011
1012 Version 8.30 04-February-2012
1013 -----------------------------
1014
1015 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1016 name is defined in ctype.h.
1017
1018 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1019 only in quite long subpatterns.
1020
1021 3. Removed the function pcre_info(), which has been obsolete and deprecated
1022 since it was replaced by pcre_fullinfo() in February 2000.
1023
1024 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1025 match a (*MARK), and the match failed at the start of the subject, a
1026 reference to memory before the start of the subject could occur. This bug
1027 was introduced by fix 17 of release 8.21.
1028
1029 5. A reference to an unset group with zero minimum repetition was giving
1030 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1031 /(another)?(\1?)test/ matched against "hello world test". This bug was
1032 introduced in release 8.13.
1033
1034 6. Add support for 16-bit character strings (a large amount of work involving
1035 many changes and refactorings).
1036
1037 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1038 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1039 from a file.
1040
1041 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1042 rounding is not applied in this particular case).
1043
1044 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1045 if they appear, or are escaped, in patterns.
1046
1047 10. Get rid of a number of -Wunused-but-set-variable warnings.
1048
1049 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1050 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1051 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1052 also returns the mark "x". This bug applied to capturing parentheses,
1053 non-capturing parentheses, and atomic parentheses. It also applied to some
1054 assertions.
1055
1056 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1057 information out of configure.ac instead of relying on pcre.h.generic, which
1058 is not stored in the repository.
1059
1060 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1061 -lreadline.
1062
1063 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1064
1065 15. Applied Graycode's patch to put the top-level frame on the stack rather
1066 than the heap when not using the stack for recursion. This gives a
1067 performance improvement in many cases when recursion is not deep.
1068
1069 16. Experimental code added to "pcretest -C" to output the stack frame size.
1070
1071
1072 Version 8.21 12-Dec-2011
1073 ------------------------
1074
1075 1. Updating the JIT compiler.
1076
1077 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1078 are added as well.
1079
1080 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1081 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1082 calling _pcre_jit_exec. Some extra comments are added.
1083
1084 4. (*MARK) settings inside atomic groups that do not contain any capturing
1085 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1086 was introduced by change 18 for 8.20.
1087
1088 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1089 ECMA-262 standard.
1090
1091 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1092 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1093 This bug was probably introduced by change 9 of 8.13.
1094
1095 7. While fixing 6 above, I noticed that a number of other items were being
1096 incorrectly rejected as "not fixed length". This arose partly because newer
1097 opcodes had not been added to the fixed-length checking code. I have (a)
1098 corrected the bug and added tests for these items, and (b) arranged for an
1099 error to occur if an unknown opcode is encountered while checking for fixed
1100 length instead of just assuming "not fixed length". The items that were
1101 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1102 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1103 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1104
1105 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1106 being incorrectly compiled and would have given unpredicatble results.
1107
1108 9. A possessively repeated subpattern with minimum repeat count greater than
1109 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1110 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1111 the first (A) could occur when it should not.
1112
1113 10. Add a cast and remove a redundant test from the code.
1114
1115 11. JIT should use pcre_malloc/pcre_free for allocation.
1116
1117 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1118 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1119 is anything other than /usr, -L is still shown).
1120
1121 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1122
1123 14. Perl does not support \N without a following name in a [] class; PCRE now
1124 also gives an error.
1125
1126 15. If a forward reference was repeated with an upper limit of around 2000,
1127 it caused the error "internal error: overran compiling workspace". The
1128 maximum number of forward references (including repeats) was limited by the
1129 internal workspace, and dependent on the LINK_SIZE. The code has been
1130 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1131 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1132 of around 200,000 forward references. While doing this, I also speeded up
1133 the filling in of repeated forward references.
1134
1135 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1136 incorrectly expecting the subject to contain another "a" after the start.
1137
1138 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1139 in the match, the SKIP should be ignored. This was not happening; instead
1140 the SKIP was being treated as NOMATCH. For patterns such as
1141 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1142 tested.
1143
1144 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1145 now much more compatible with Perl, in particular in cases where the result
1146 is a non-match for a non-anchored pattern. For example, if
1147 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1148 "m", where previously it did not return a name. A side effect of this
1149 change is that for partial matches, the last encountered mark name is
1150 returned, as for non matches. A number of tests that were previously not
1151 Perl-compatible have been moved into the Perl-compatible test files. The
1152 refactoring has had the pleasing side effect of removing one argument from
1153 the match() function, thus reducing its stack requirements.
1154
1155 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1156 subsequent uses of /S (without +) incorrectly behaved like /S+.
1157
1158 21. Retrieve executable code size support for the JIT compiler and fixing
1159 some warnings.
1160
1161 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1162 not work when the shorter character appeared right at the end of the
1163 subject string.
1164
1165 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1166 systems.
1167
1168 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1169 output it when the /M option is used in pcretest.
1170
1171 25. The CheckMan script was not being included in the distribution. Also, added
1172 an explicit "perl" to run Perl scripts from the PrepareRelease script
1173 because this is reportedly needed in Windows.
1174
1175 26. If study data was being save in a file and studying had not found a set of
1176 "starts with" bytes for the pattern, the data written to the file (though
1177 never used) was taken from uninitialized memory and so caused valgrind to
1178 complain.
1179
1180 27. Updated RunTest.bat as provided by Sheri Pierce.
1181
1182 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1183
1184 29. Computation of memory usage for the table of capturing group names was
1185 giving an unnecessarily large value.
1186
1187
1188 Version 8.20 21-Oct-2011
1189 ------------------------
1190
1191 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1192 a POSIX class. After further experiments with Perl, which convinced me that
1193 Perl has bugs and confusions, a closing square bracket is no longer allowed
1194 in a POSIX name. This bug also affected patterns with classes that started
1195 with full stops.
1196
1197 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1198 captured substring, but while checking the failing first alternative,
1199 substring 1 is temporarily captured. If the output vector supplied to
1200 pcre_exec() was not big enough for this capture, the yield of the function
1201 was still zero ("insufficient space for captured substrings"). This cannot
1202 be totally fixed without adding another stack variable, which seems a lot
1203 of expense for a edge case. However, I have improved the situation in cases
1204 such as /(a)(b)x|abc/ matched against "abc", where the return code
1205 indicates that fewer than the maximum number of slots in the ovector have
1206 been set.
1207
1208 3. Related to (2) above: when there are more back references in a pattern than
1209 slots in the output vector, pcre_exec() uses temporary memory during
1210 matching, and copies in the captures as far as possible afterwards. It was
1211 using the entire output vector, but this conflicts with the specification
1212 that only 2/3 is used for passing back captured substrings. Now it uses
1213 only the first 2/3, for compatibility. This is, of course, another edge
1214 case.
1215
1216 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1217 main code base, and can be used by building with --enable-jit. When this is
1218 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1219 runtime --no-jit option is given.
1220
1221 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1222 ovector, the return from the function was zero, implying that there were
1223 other matches that did not fit. The correct "exactly full" value is now
1224 returned.
1225
1226 6. If a subpattern that was called recursively or as a subroutine contained
1227 (*PRUNE) or any other control that caused it to give a non-standard return,
1228 invalid errors such as "Error -26 (nested recursion at the same subject
1229 position)" or even infinite loops could occur.
1230
1231 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1232 computing the minimum length on reaching *ACCEPT, and so ended up with the
1233 wrong value of 1 rather than 0. Further investigation indicates that
1234 computing a minimum subject length in the presence of *ACCEPT is difficult
1235 (think back references, subroutine calls), and so I have changed the code
1236 so that no minimum is registered for a pattern that contains *ACCEPT.
1237
1238 8. If (*THEN) was present in the first (true) branch of a conditional group,
1239 it was not handled as intended. [But see 16 below.]
1240
1241 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1242 Sheri Pierce.
1243
1244 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1245 the first byte in a match must be "a".
1246
1247 11. Change 17 for 8.13 increased the recursion depth for patterns like
1248 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1249 pattern contains any instances of (*THEN). If it does not, the old
1250 optimizations are restored. It would be nice to do this on a per-group
1251 basis, but at the moment that is not feasible.
1252
1253 12. In some environments, the output of pcretest -C is CRLF terminated. This
1254 broke RunTest's code that checks for the link size. A single white space
1255 character after the value is now allowed for.
1256
1257 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1258 For "fr", it uses the Windows-specific input and output files.
1259
1260 14. If (*THEN) appeared in a group that was called recursively or as a
1261 subroutine, it did not work as intended. [But see next item.]
1262
1263 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1264 pattern fragments (but not containing any | characters). If A and B are
1265 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1266 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1267 D. In other words, Perl considers parentheses that do not contain any |
1268 characters to be part of a surrounding alternative, whereas PCRE was
1269 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1270 differently. PCRE now behaves in the same way as Perl, except in the case
1271 of subroutine/recursion calls such as (?1) which have in any case always
1272 been different (but PCRE had them first :-).
1273
1274 16. Related to 15 above: Perl does not treat the | in a conditional group as
1275 creating alternatives. Such a group is treated in the same way as an
1276 ordinary group without any | characters when processing (*THEN). PCRE has
1277 been changed to match Perl's behaviour.
1278
1279 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1280 RunGrepTest script failed.
1281
1282 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1283 inevitable for groups that contain captures, but it can lead to a lot of
1284 stack use in large patterns. The old behaviour has been restored for atomic
1285 groups that do not contain any capturing parentheses.
1286
1287 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1288 suppress the check for a minimum subject length at run time. (If it was
1289 given to pcre_exec() or pcre_dfa_exec() it did work.)
1290
1291 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1292 fail to work when decoding hex characters in data strings in EBCDIC
1293 environments.
1294
1295 21. It appears that in at least one Mac OS environment, the isxdigit() function
1296 is implemented as a macro that evaluates to its argument more than once,
1297 contravening the C 90 Standard (I haven't checked a later standard). There
1298 was an instance in pcretest which caused it to go wrong when processing
1299 \x{...} escapes in subject strings. The has been rewritten to avoid using
1300 things like p++ in the argument of isxdigit().
1301
1302
1303 Version 8.13 16-Aug-2011
1304 ------------------------
1305
1306 1. The Unicode data tables have been updated to Unicode 6.0.0.
1307
1308 2. Two minor typos in pcre_internal.h have been fixed.
1309
1310 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1311 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1312 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1313
1314 4. There were a number of related bugs in the code for matching backrefences
1315 caselessly in UTF-8 mode when codes for the characters concerned were
1316 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1317 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1318 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1319 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1320 2-byte code at the end of the subject (it thought there wasn't enough data
1321 left).
1322
1323 5. Comprehensive information about what went wrong is now returned by
1324 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1325 as the output vector has at least 2 elements. The offset of the start of
1326 the failing character and a reason code are placed in the vector.
1327
1328 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1329 now returned is for the first byte of the failing character, instead of the
1330 last byte inspected. This is an incompatible change, but I hope it is small
1331 enough not to be a problem. It makes the returned offset consistent with
1332 pcre_exec() and pcre_dfa_exec().
1333
1334 7. pcretest now gives a text phrase as well as the error number when
1335 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1336 failure, the offset and reason code are output.
1337
1338 8. When \R was used with a maximizing quantifier it failed to skip backwards
1339 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1340 back over a single character (\n). This seems wrong (because it treated the
1341 two characters as a single entity when going forwards), conflicts with the
1342 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1343 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1344 has been changed.
1345
1346 9. Some internal refactoring has changed the processing so that the handling
1347 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1348 time (the PCRE_DOTALL option was changed this way some time ago: version
1349 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1350 which was always a bit of a fudge. It also means that there is one less
1351 argument for the match() function, which reduces its stack requirements
1352 slightly. This change also fixes an incompatibility with Perl: the pattern
1353 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1354
1355 10. More internal refactoring has drastically reduced the number of recursive
1356 calls to match() for possessively repeated groups such as (abc)++ when
1357 using pcre_exec().
1358
1359 11. While implementing 10, a number of bugs in the handling of groups were
1360 discovered and fixed:
1361
1362 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1363 (a|)*(?1) gave a compile-time internal error.
1364 ((a|)+)+ did not notice that the outer group could match an empty string.
1365 (^a|^)+ was not marked as anchored.
1366 (.*a|.*)+ was not marked as matching at start or after a newline.
1367
1368 12. Yet more internal refactoring has removed another argument from the match()
1369 function. Special calls to this function are now indicated by setting a
1370 value in a variable in the "match data" data block.
1371
1372 13. Be more explicit in pcre_study() instead of relying on "default" for
1373 opcodes that mean there is no starting character; this means that when new
1374 ones are added and accidentally left out of pcre_study(), testing should
1375 pick them up.
1376
1377 14. The -s option of pcretest has been documented for ages as being an old
1378 synonym of -m (show memory usage). I have changed it to mean "force study
1379 for every regex", that is, assume /S for every regex. This is similar to -i
1380 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1381 using it. It makes it easier to run collections of tests with and without
1382 study enabled, and thereby test pcre_study() more easily. All the standard
1383 tests are now run with and without -s (but some patterns can be marked as
1384 "never study" - see 20 below).
1385
1386 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1387 restoration of the capturing data to the outer values was not happening
1388 correctly.
1389
1390 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1391 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1392 pattern had matched an empty string, and so incorrectly returned a no
1393 match.
1394
1395 17. There was optimizing code for the last branch of non-capturing parentheses,
1396 and also for the obeyed branch of a conditional subexpression, which used
1397 tail recursion to cut down on stack usage. Unfortunately, now that there is
1398 the possibility of (*THEN) occurring in these branches, tail recursion is
1399 no longer possible because the return has to be checked for (*THEN). These
1400 two optimizations have therefore been removed. [But see 8.20/11 above.]
1401
1402 18. If a pattern containing \R was studied, it was assumed that \R always
1403 matched two bytes, thus causing the minimum subject length to be
1404 incorrectly computed because \R can also match just one byte.
1405
1406 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1407 was incorrectly computed.
1408
1409 20. If /S is present twice on a test pattern in pcretest input, it now
1410 *disables* studying, thereby overriding the use of -s on the command line
1411 (see 14 above). This is necessary for one or two tests to keep the output
1412 identical in both cases.
1413
1414 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1415 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1416
1417 22. When an atomic group that contained a capturing parenthesis was
1418 successfully matched, but the branch in which it appeared failed, the
1419 capturing was not being forgotten if a higher numbered group was later
1420 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1421 group 1 to "a", when in fact it should be unset. This applied to multi-
1422 branched capturing and non-capturing groups, repeated or not, and also to
1423 positive assertions (capturing in negative assertions does not happen
1424 in PCRE) and also to nested atomic groups.
1425
1426 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1427 subject after a captured substring, to make it easier to tell which of a
1428 number of identical substrings has been captured.
1429
1430 24. The way atomic groups are processed by pcre_exec() has been changed so that
1431 if they are repeated, backtracking one repetition now resets captured
1432 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1433 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1434 "aaa". Previously, it would have been "a". As part of this code
1435 refactoring, the way recursive calls are handled has also been changed.
1436
1437 25. If an assertion condition captured any substrings, they were not passed
1438 back unless some other capturing happened later. For example, if
1439 (?(?=(a))a) was matched against "a", no capturing was returned.
1440
1441 26. When studying a pattern that contained subroutine calls or assertions,
1442 the code for finding the minimum length of a possible match was handling
1443 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1444 group 1 called group 2 while simultaneously a separate group 2 called group
1445 1). A stack overflow occurred in this case. I have fixed this by limiting
1446 the recursion depth to 10.
1447
1448 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1449 Fortmann. This supports explicit test numbers on the command line, and has
1450 argument validation and error reporting.
1451
1452 28. An instance of \X with an unlimited repeat could fail if at any point the
1453 first character it looked at was a mark character.
1454
1455 29. Some minor code refactoring concerning Unicode properties and scripts
1456 should reduce the stack requirement of match() slightly.
1457
1458 30. Added the '=' option to pcretest to check the setting of unused capturing
1459 slots at the end of the pattern, which are documented as being -1, but are
1460 not included in the return count.
1461
1462 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1463 compiled something random. Now it gives a compile-time error (as does
1464 Perl).
1465
1466 32. A *MARK encountered during the processing of a positive assertion is now
1467 recorded and passed back (compatible with Perl).
1468
1469 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1470 had alternative anchored branches, the search for a second match in a line
1471 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1472 matched the line "0102" twice. The same bug affected patterns that started
1473 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1474 twice.
1475
1476 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1477 does, and because of capturing effects, quantifying parenthesized
1478 assertions may at times be useful. Quantifiers are now allowed for
1479 parenthesized assertions.
1480
1481 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1482
1483 36. \g was being checked for fancy things in a character class, when it should
1484 just be a literal "g".
1485
1486 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1487 appearance of a nested POSIX class supersedes an apparent external class.
1488 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1489 unescaped square brackets may also appear as part of class names. For
1490 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1491 more like Perl. (But see 8.20/1 above.)
1492
1493 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1494 was because it thought it was \N{name}, which is not supported).
1495
1496 39. Add minix to OS list not supporting the -S option in pcretest.
1497
1498 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1499 cannot analyze patterns in sufficient detail to catch mutual recursions
1500 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1501 subgroup is called recursively as a subpattern for a second time at the
1502 same position in the subject string. In previous releases this might have
1503 been caught by the recursion limit, or it might have run out of stack.
1504
1505 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1506 happen only once. PCRE was, however incorrectly giving a compile time error
1507 "recursive call could loop indefinitely" because it cannot analyze the
1508 pattern in sufficient detail. The compile time test no longer happens when
1509 PCRE is compiling a conditional subpattern, but actual runaway loops are
1510 now caught at runtime (see 40 above).
1511
1512 42. It seems that Perl allows any characters other than a closing parenthesis
1513 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1514 has been changed to be the same.
1515
1516 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1517 as not to get warnings when autogen.sh is called. Also changed
1518 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1519
1520 44. To help people who use pcregrep to scan files containing exceedingly long
1521 lines, the following changes have been made:
1522
1523 (a) The default value of the buffer size parameter has been increased from
1524 8K to 20K. (The actual buffer used is three times this size.)
1525
1526 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1527 PCRE is built.
1528
1529 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1530 to be set at run time.
1531
1532 (d) Numerical values in pcregrep options can be followed by K or M, for
1533 example --buffer-size=50K.
1534
1535 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1536 given and the return code is set to 2.
1537
1538 45. Add a pointer to the latest mark to the callout data block.
1539
1540 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1541 partial match of an empty string instead of no match. This was specific to
1542 the use of ".".
1543
1544 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1545 complete match instead of a partial match. This bug was dependent on both
1546 the PCRE_UTF8 and PCRE_DOTALL options being set.
1547
1548 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1549 starting byte set, because \b was not being ignored.
1550
1551
1552 Version 8.12 15-Jan-2011
1553 ------------------------
1554
1555 1. Fixed some typos in the markup of the man pages, and wrote a script that
1556 checks for such things as part of the documentation building process.
1557
1558 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1559 --match-limit and --recursion-limit options (added for 8.11). In
1560 particular, this made one of the standard tests fail. (The integer value
1561 went into the wrong half of a long int.)
1562
1563 3. If the --colour option was given to pcregrep with -v (invert match), it
1564 did strange things, either producing crazy output, or crashing. It should,
1565 of course, ignore a request for colour when reporting lines that do not
1566 match.
1567
1568 4. Another pcregrep bug caused similar problems if --colour was specified with
1569 -M (multiline) and the pattern match finished with a line ending.
1570
1571 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1572 matched in multiline mode, the following line was shown as part of the
1573 match. This seems wrong, so I have changed it.
1574
1575 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1576 the check for further matches in the same line (so they could be coloured)
1577 to overrun the end of the current line. If another match was found, it was
1578 incorrectly shown (and then shown again when found in the next line).
1579
1580 7. If pcregrep was compiled under Windows, there was a reference to the
1581 function pcregrep_exit() before it was defined. I am assuming this was
1582 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1583 reported by a user. I've moved the definition above the reference.
1584
1585
1586 Version 8.11 10-Dec-2010
1587 ------------------------
1588
1589 1. (*THEN) was not working properly if there were untried alternatives prior
1590 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1591 backtracked to try for "b" instead of moving to the next alternative branch
1592 at the same level (in this case, to look for "c"). The Perl documentation
1593 is clear that when (*THEN) is backtracked onto, it goes to the "next
1594 alternative in the innermost enclosing group".
1595
1596 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1597 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1598 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1599 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1600 (*THEN).
1601
1602 3. If \s appeared in a character class, it removed the VT character from
1603 the class, even if it had been included by some previous item, for example
1604 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1605 of \s, but is part of the POSIX "space" class.)
1606
1607 4. A partial match never returns an empty string (because you can always
1608 match an empty string at the end of the subject); however the checking for
1609 an empty string was starting at the "start of match" point. This has been
1610 changed to the "earliest inspected character" point, because the returned
1611 data for a partial match starts at this character. This means that, for
1612 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1613 (previously it gave "no match").
1614
1615 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1616 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1617 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1618 has an implication that the given string is incomplete (because a partial
1619 match is preferred over a full match). For this reason, these items now
1620 give a partial match in this situation. [Aside: previously, the one case
1621 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1622 match rather than a full match, which was wrong by the old rules, but is
1623 now correct.]
1624
1625 6. There was a bug in the handling of #-introduced comments, recognized when
1626 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1627 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1628 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1629 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1630 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1631 places in pcre_compile().
1632
1633 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1634 comments when looking ahead for named forward references to subpatterns,
1635 the only newline sequence it recognized was NL. It now handles newlines
1636 according to the set newline convention.
1637
1638 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1639 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1640 cater for a lack of strerror(). These oversights have been fixed.
1641
1642 9. Added --match-limit and --recursion-limit to pcregrep.
1643
1644 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1645
1646 11. When the -o option was used, pcregrep was setting a return code of 1, even
1647 when matches were found, and --line-buffered was not being honoured.
1648
1649 12. Added an optional parentheses number to the -o and --only-matching options
1650 of pcregrep.
1651
1652 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1653 can match an empty string. The code to do it in pcretest and pcredemo
1654 needed fixing:
1655
1656 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1657 only one byte after an empty string match just before CRLF (this case
1658 just got forgotten; "any" and "anycrlf" were OK).
1659
1660 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1661 mode when an empty string match preceded an ASCII character followed by
1662 a non-ASCII character. (The code for advancing by one character rather
1663 than one byte was nonsense.)
1664
1665 (c) The pcredemo.c sample program did not have any code at all to handle
1666 the cases when CRLF is a valid newline sequence.
1667
1668 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1669 as a starting offset was within the subject string. There is now a new
1670 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1671 negative or greater than the length of the string. In order to test this,
1672 pcretest is extended to allow the setting of negative starting offsets.
1673
1674 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1675 starting offset points to the beginning of a UTF-8 character was
1676 unnecessarily clumsy. I tidied it up.
1677
1678 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1679 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1680
1681 17. Nobody had reported that the --include_dir option, which was added in
1682 release 7.7 should have been called --include-dir (hyphen, not underscore)
1683 for compatibility with GNU grep. I have changed it to --include-dir, but
1684 left --include_dir as an undocumented synonym, and the same for
1685 --exclude-dir, though that is not available in GNU grep, at least as of
1686 release 2.5.4.
1687
1688 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1689 characters from a string of bytes) have been redefined so as not to use
1690 loops, in order to improve performance in some environments. At the same
1691 time, I abstracted some of the common code into auxiliary macros to save
1692 repetition (this should not affect the compiled code).
1693
1694 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1695 compile-time error is now given if \c is not followed by an ASCII
1696 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1697 different, and any byte value is allowed.)
1698
1699 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1700 START_OPTIMIZE option, which is now allowed at compile time - but just
1701 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1702 to pcregrep and other applications that have no direct access to PCRE
1703 options. The new /Y option in pcretest sets this option when calling
1704 pcre_compile().
1705
1706 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1707 back references. Groups containing recursive back references were forced to
1708 be atomic by that change, but in the case of named groups, the amount of
1709 memory required was incorrectly computed, leading to "Failed: internal
1710 error: code overflow". This has been fixed.
1711
1712 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1713 pcretest.c, to avoid build problems in some Borland environments.
1714
1715
1716 Version 8.10 25-Jun-2010
1717 ------------------------
1718
1719 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1720 THEN.
1721
1722 2. (*ACCEPT) was not working when inside an atomic group.
1723
1724 3. Inside a character class, \B is treated as a literal by default, but
1725 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1726 causes the error). The code is unchanged, but I tidied the documentation.
1727
1728 4. Inside a character class, PCRE always treated \R and \X as literals,
1729 whereas Perl faults them if its -w option is set. I have changed PCRE so
1730 that it faults them when PCRE_EXTRA is set.
1731
1732 5. Added support for \N, which always matches any character other than
1733 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1734
1735 6. When compiling pcregrep with newer versions of gcc which may have
1736 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1737 declared with attribute warn_unused_result" were given. Just casting the
1738 result to (void) does not stop the warnings; a more elaborate fudge is
1739 needed. I've used a macro to implement this.
1740
1741 7. Minor change to pcretest.c to avoid a compiler warning.
1742
1743 8. Added four artifical Unicode properties to help with an option to make
1744 \s etc use properties (see next item). The new properties are: Xan
1745 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1746
1747 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1748 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1749 this option. Modified pcretest to add /W to test this facility. Added
1750 REG_UCP to make it available via the POSIX interface.
1751
1752 10. Added --line-buffered to pcregrep.
1753
1754 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1755 studied, and the match started with a letter with a code point greater than
1756 127 whose first byte was different to the first byte of the other case of
1757 the letter, the other case of this starting letter was not recognized
1758 (#976).
1759
1760 12. If a pattern that was studied started with a repeated Unicode property
1761 test, for example, \p{Nd}+, there was the theoretical possibility of
1762 setting up an incorrect bitmap of starting bytes, but fortunately it could
1763 not have actually happened in practice until change 8 above was made (it
1764 added property types that matched character-matching opcodes).
1765
1766 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1767 possible starting bytes for non-anchored patterns.
1768
1769 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1770 \R, and also a number of cases that involve Unicode properties, both
1771 explicit and implicit when PCRE_UCP is set.
1772
1773 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1774 input, it could crash or give wrong results if characters with values
1775 greater than 0xc0 were present in the subject string. (Detail: it assumed
1776 UTF-8 input when processing these items.)
1777
1778 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1779 size_t is 64-bit (#991).
1780
1781 17. Added a check for running out of memory when PCRE is compiled with
1782 --disable-stack-for-recursion (#990).
1783
1784 18. If the last data line in a file for pcretest does not have a newline on
1785 the end, a newline was missing in the output.
1786
1787 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1788 less than 128) in its various bitmaps. However, there is a facility for
1789 generating tables according to the current locale when PCRE is compiled. It
1790 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1791 characters, are recognized by isspace() and therefore were getting set in
1792 these tables, and indeed these tables seem to approximate to ISO 8859. This
1793 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1794 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1795 which of course cannot start UTF-8 characters. I have changed the code so
1796 that only real ASCII characters (less than 128) and the correct starting
1797 bytes for UTF-8 encodings are set for characters greater than 127 when in
1798 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1799 altogether.)
1800
1801 20. Added the /T option to pcretest so as to be able to run tests with non-
1802 standard character tables, thus making it possible to include the tests
1803 used for 19 above in the standard set of tests.
1804
1805 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1806 reference to a subpattern the other side of a comment that contains an
1807 opening parenthesis caused either an internal compiling error, or a
1808 reference to the wrong subpattern.
1809
1810
1811 Version 8.02 19-Mar-2010
1812 ------------------------
1813
1814 1. The Unicode data tables have been updated to Unicode 5.2.0.
1815
1816 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1817 configured.
1818
1819 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1820 original author of that file, following a query about its status.
1821
1822 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1823 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1824
1825 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1826 quantifier applied to a forward-referencing subroutine call, could compile
1827 incorrect code or give the error "internal error: previously-checked
1828 referenced subpattern not found".
1829
1830 6. Both MS Visual Studio and Symbian OS have problems with initializing
1831 variables to point to external functions. For these systems, therefore,
1832 pcre_malloc etc. are now initialized to local functions that call the
1833 relevant global functions.
1834
1835 7. There were two entries missing in the vectors called coptable and poptable
1836 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1837 I've fixed the data, and added a kludgy way of testing at compile time that
1838 the lengths are correct (equal to the number of opcodes).
1839
1840 8. Following on from 7, I added a similar kludge to check the length of the
1841 eint vector in pcreposix.c.
1842
1843 9. Error texts for pcre_compile() are held as one long string to avoid too
1844 much relocation at load time. To find a text, the string is searched,
1845 counting zeros. There was no check for running off the end of the string,
1846 which could happen if a new error number was added without updating the
1847 string.
1848
1849 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1850
1851 11. \K was not working if it appeared in an atomic group or in a group that
1852 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1853 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1854 the assertion is positive, but not if it is negative.
1855
1856 12. Change 11 fortuitously reduced the size of the stack frame used in the
1857 "match()" function of pcre_exec.c by one pointer. Forthcoming
1858 implementation of support for (*MARK) will need an extra pointer on the
1859 stack; I have reserved it now, so that the stack frame size does not
1860 decrease.
1861
1862 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1863 item in branch that calls a recursion is a subroutine call - as in the
1864 second branch in the above example - was incorrectly given the compile-
1865 time error "recursive call could loop indefinitely" because pcre_compile()
1866 was not correctly checking the subroutine for matching a non-empty string.
1867
1868 14. The checks for overrunning compiling workspace could trigger after an
1869 overrun had occurred. This is a "should never occur" error, but it can be
1870 triggered by pathological patterns such as hundreds of nested parentheses.
1871 The checks now trigger 100 bytes before the end of the workspace.
1872
1873 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1874
1875
1876 Version 8.01 19-Jan-2010
1877 ------------------------
1878
1879 1. If a pattern contained a conditional subpattern with only one branch (in
1880 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1881 computed the wrong minimum data length (which is of course zero for such
1882 subpatterns). This could cause incorrect "no match" results.
1883
1884 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1885 the pattern is reset in the first branch, pcre_compile() failed with
1886 "internal error: code overflow at offset...". This happened only when
1887 the reset was to the original external option setting. (An optimization
1888 abstracts leading options settings into an external setting, which was the
1889 cause of this.)
1890
1891 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1892 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1893 assertion pattern did not match (meaning that the assertion was true), it
1894 was incorrectly treated as false if the SKIP had been reached during the
1895 matching. This also applied to assertions used as conditions.
1896
1897 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1898 assertion subpattern, including such a pattern used as a condition,
1899 unpredictable results occurred, instead of the error return
1900 PCRE_ERROR_DFA_UITEM.
1901
1902 5. The C++ GlobalReplace function was not working like Perl for the special
1903 situation when an empty string is matched. It now does the fancy magic
1904 stuff that is necessary.
1905
1906 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1907 removed. (These were left over from very, very early versions of PCRE.)
1908
1909 7. Some cosmetic changes to the code to make life easier when compiling it
1910 as part of something else:
1911
1912 (a) Change DEBUG to PCRE_DEBUG.
1913
1914 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1915 called "current" as "current_branch", to prevent a collision with the
1916 Linux macro when compiled as a kernel module.
1917
1918 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1919 prevent a collision with the Linux macro when compiled as a kernel
1920 module.
1921
1922 8. In pcre_compile() there are some checks for integer overflows that used to
1923 cast potentially large values to (double). This has been changed to that
1924 when building, a check for int64_t is made, and if it is found, it is used
1925 instead, thus avoiding the use of floating point arithmetic. (There is no
1926 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1927 double.
1928
1929 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1930 2005 (difference between two addresses compared to an unsigned value).
1931
1932 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1933 custom one, because of the following reported problem in Windows:
1934
1935 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1936 under Win32.
1937 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1938 therefore missing the function definition.
1939 - The compiler thus generates a "C" signature for the test function.
1940 - The linker fails to find the "C" function.
1941 - PCRE fails to configure if asked to do so against libbz2.
1942
1943 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1944 messages were output:
1945
1946 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1947 rerunning libtoolize, to keep the correct libtool macros in-tree.
1948 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1949
1950 I have done both of these things.
1951
1952 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1953 most of the time, it *can* run out if it is given a pattern that contains a
1954 runaway infinite recursion. I updated the discussion in the pcrestack man
1955 page.
1956
1957 13. Now that we have gone to the x.xx style of version numbers, the minor
1958 version may start with zero. Using 08 or 09 is a bad idea because users
1959 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1960 interpreted as invalid octal numbers. I've updated the previous comment in
1961 configure.ac, and also added a check that gives an error if 08 or 09 are
1962 used.
1963
1964 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1965 causing partial matching to fail when the end of the subject matched \W
1966 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1967
1968 15. There were some discrepancies between the declarations in pcre_internal.h
1969 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1970 their definitions. The declarations used "const uschar *" and the
1971 definitions used USPTR. Even though USPTR is normally defined as "const
1972 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1973 reported that: "This difference in casting confuses some C++ compilers, for
1974 example, SunCC recognizes above declarations as different functions and
1975 generates broken code for hbpcre." I have changed the declarations to use
1976 USPTR.
1977
1978 16. GNU libtool is named differently on some systems. The autogen.sh script now
1979 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1980 (FreeBSD).
1981
1982 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1983 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1984 comment: "Figure out how to create a longlong from a string: strtoll and
1985 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1986 instance, but it only takes 2 args instead of 3!"
1987
1988 18. A subtle bug concerned with back references has been fixed by a change of
1989 specification, with a corresponding code fix. A pattern such as
1990 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1991 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1992 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1993 same bug. Such groups have to be quantified to be useful, or contained
1994 inside another quantified group. (If there's no repetition, the reference
1995 can never match.) The problem arises because, having left the group and
1996 moved on to the rest of the pattern, a later failure that backtracks into
1997 the group uses the captured value from the final iteration of the group
1998 rather than the correct earlier one. I have fixed this in PCRE by forcing
1999 any group that contains a reference to itself to be an atomic group; that
2000 is, there cannot be any backtracking into it once it has completed. This is
2001 similar to recursive and subroutine calls.
2002
2003
2004 Version 8.00 19-Oct-09
2005 ----------------------
2006
2007 1. The table for translating pcre_compile() error codes into POSIX error codes
2008 was out-of-date, and there was no check on the pcre_compile() error code
2009 being within the table. This could lead to an OK return being given in
2010 error.
2011
2012 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2013 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2014 in a Windows environment.
2015
2016 3. The pcregrep --count option prints the count for each file even when it is
2017 zero, as does GNU grep. However, pcregrep was also printing all files when
2018 --files-with-matches was added. Now, when both options are given, it prints
2019 counts only for those files that have at least one match. (GNU grep just
2020 prints the file name in this circumstance, but including the count seems
2021 more useful - otherwise, why use --count?) Also ensured that the
2022 combination -clh just lists non-zero counts, with no names.
2023
2024 4. The long form of the pcregrep -F option was incorrectly implemented as
2025 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2026 but it seems right to fix it, and I didn't think it was worth preserving
2027 the old behaviour.
2028
2029 5. The command line items --regex=pattern and --regexp=pattern were not
2030 recognized by pcregrep, which required --regex pattern or --regexp pattern
2031 (with a space rather than an '='). The man page documented the '=' forms,
2032 which are compatible with GNU grep; these now work.
2033
2034 6. No libpcreposix.pc file was created for pkg-config; there was just
2035 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2036
2037 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2038 when UCP support is not needed, by modifying the Python script that
2039 generates it from Unicode data files. This should not matter if the module
2040 is correctly used as a library, but I received one complaint about 50K of
2041 unwanted data. My guess is that the person linked everything into his
2042 program rather than using a library. Anyway, it does no harm.
2043
2044 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2045 was a minimum greater than 1 for a wide character in a possessive
2046 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2047 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2048 character. Chaos in the form of incorrect output or a compiling loop could
2049 result.
2050
2051 9. The restrictions on what a pattern can contain when partial matching is
2052 requested for pcre_exec() have been removed. All patterns can now be
2053 partially matched by this function. In addition, if there are at least two
2054 slots in the offset vector, the offset of the earliest inspected character
2055 for the match and the offset of the end of the subject are set in them when
2056 PCRE_ERROR_PARTIAL is returned.
2057
2058 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2059 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2060 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2061 and may be more useful for multi-segment matching.
2062
2063 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2064 used to be given if ever the end of the subject was reached; now it is
2065 given only if matching could not proceed because another character was
2066 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2067 string "Z", which now yields "no match" instead of "partial match". In the
2068 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2069 final character ended with (*FAIL).
2070
2071 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2072 if the pattern had a "must contain" character that was already found in the
2073 earlier partial match, unless partial matching was again requested. For
2074 example, with the pattern /dog.(body)?/, the "must contain" character is
2075 "g". If the first part-match was for the string "dog", restarting with
2076 "sbody" failed. This bug has been fixed.
2077
2078 13. The string returned by pcre_dfa_exec() after a partial match has been
2079 changed so that it starts at the first inspected character rather than the
2080 first character of the match. This makes a difference only if the pattern
2081 starts with a lookbehind assertion or \b or \B (\K is not supported by
2082 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2083 matching functions compatible, and I think it's the right thing to do.
2084
2085 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2086 so that the demonstration program is easily available in environments where
2087 PCRE has not been installed from source.
2088
2089 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2090 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2091 library.
2092
2093 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2094 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2095 is not the first non-POSIX option to be added. Clearly some people find
2096 these options useful.
2097
2098 17. If a caller to the POSIX matching function regexec() passes a non-zero
2099 value for nmatch with a NULL value for pmatch, the value of
2100 nmatch is forced to zero.
2101
2102 18. RunGrepTest did not have a test for the availability of the -u option of
2103 the diff command, as RunTest does. It now checks in the same way as
2104 RunTest, and also checks for the -b option.
2105
2106 19. If an odd number of negated classes containing just a single character
2107 interposed, within parentheses, between a forward reference to a named
2108 subpattern and the definition of the subpattern, compilation crashed with
2109 an internal error, complaining that it could not find the referenced
2110 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2111 [The bug was that it was starting one character too far in when skipping
2112 over the character class, thus treating the ] as data rather than
2113 terminating the class. This meant it could skip too much.]
2114
2115 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2116 /g option in pcretest when the pattern contains \K, which makes it possible
2117 to have an empty string match not at the start, even when the pattern is
2118 anchored. Updated pcretest and pcredemo to use this option.
2119
2120 21. If the maximum number of capturing subpatterns in a recursion was greater
2121 than the maximum at the outer level, the higher number was returned, but
2122 with unset values at the outer level. The correct (outer level) value is
2123 now given.
2124
2125 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2126 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2127 make it do so. The string so far is captured, making this feature
2128 compatible with Perl.
2129
2130 23. The tests have been re-organized, adding tests 11 and 12, to make it
2131 possible to check the Perl 5.10 features against Perl 5.10.
2132
2133 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2134 pattern matches a fixed length string. PCRE did not allow this; now it
2135 does. Neither allows recursion.
2136
2137 25. I finally figured out how to implement a request to provide the minimum
2138 length of subject string that was needed in order to match a given pattern.
2139 (It was back references and recursion that I had previously got hung up
2140 on.) This code has now been added to pcre_study(); it finds a lower bound
2141 to the length of subject needed. It is not necessarily the greatest lower
2142 bound, but using it to avoid searching strings that are too short does give
2143 some useful speed-ups. The value is available to calling programs via
2144 pcre_fullinfo().
2145
2146 26. While implementing 25, I discovered to my embarrassment that pcretest had
2147 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2148 study optimizations had never been tested with that matching function.
2149 Oops. What is worse, even when it was passed study data, there was a bug in
2150 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2151 were also very few tests of studied patterns with pcre_dfa_exec().
2152
2153 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2154 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2155 on the other side of the coin, they are no longer allowed to have different
2156 names, because these cannot be distinguished in PCRE, and this has caused
2157 confusion. (This is a difference from Perl.)
2158
2159 28. When duplicate subpattern names are present (necessarily with different
2160 numbers, as required by 27 above), and a test is made by name in a
2161 conditional pattern, either for a subpattern having been matched, or for
2162 recursion in such a pattern, all the associated numbered subpatterns are
2163 tested, and the overall condition is true if the condition is true for any
2164 one of them. This is the way Perl works, and is also more like the way
2165 testing by number works.
2166
2167
2168 Version 7.9 11-Apr-09
2169 ---------------------
2170
2171 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2172 (pcretest), all targets were linked against these libraries. This included
2173 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2174 libraries. This caused unwanted dependencies to be created. This problem
2175 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2176 pcretest is linked with readline.
2177
2178 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2179 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2180 moved outside it again, because FALSE and TRUE are already defined in AIX,
2181 but BOOL is not.
2182
2183 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2184 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2185
2186 4. The pcregrep documentation said spaces were inserted as well as colons (or
2187 hyphens) following file names and line numbers when outputting matching
2188 lines. This is not true; no spaces are inserted. I have also clarified the
2189 wording for the --colour (or --color) option.
2190
2191 5. In pcregrep, when --colour was used with -o, the list of matching strings
2192 was not coloured; this is different to GNU grep, so I have changed it to be
2193 the same.
2194
2195 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2196 each matching line was coloured. Now it goes on to look for further matches
2197 of any of the test patterns, which is the same behaviour as GNU grep.
2198
2199 7. A pattern that could match an empty string could cause pcregrep to loop; it
2200 doesn't make sense to accept an empty string match in pcregrep, so I have
2201 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2202 seems to be how GNU grep behaves. [But see later change 40 for release
2203 8.33.]
2204
2205 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2206 start or after a newline", because the conditional assertion was not being
2207 correctly handled. The rule now is that both the assertion and what follows
2208 in the first alternative must satisfy the test.
2209
2210 9. If auto-callout was enabled in a pattern with a conditional group whose
2211 condition was an assertion, PCRE could crash during matching, both with
2212 pcre_exec() and pcre_dfa_exec().
2213
2214 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2215 used for matching.
2216
2217 11. Unicode property support in character classes was not working for
2218 characters (bytes) greater than 127 when not in UTF-8 mode.
2219
2220 12. Added the -M command line option to pcretest.
2221
2222 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2223
2224 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2225
2226 16. Added comments and documentation about mis-use of no_arg in the C++
2227 wrapper.
2228
2229 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2230 from Martin Jerabek that uses macro names for all relevant character and
2231 string constants.
2232
2233 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2234 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2235 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2236 these, but not everybody uses configure.
2237
2238 19. A conditional group that had only one branch was not being correctly
2239 recognized as an item that could match an empty string. This meant that an
2240 enclosing group might also not be so recognized, causing infinite looping
2241 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2242 with the subject "ab", where knowledge that the repeated group can match
2243 nothing is needed in order to break the loop.
2244
2245 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2246 exec(), but without supplying a callout function, matching went wrong.
2247
2248 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2249 leak if the size of the offset vector was greater than 30. When the vector
2250 is smaller, the saved offsets during recursion go onto a local stack
2251 vector, but for larger vectors malloc() is used. It was failing to free
2252 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2253 error, in fact).
2254
2255 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2256 heapframe that is used only when UTF-8 support is enabled. This caused no
2257 problem, but was untidy.
2258
2259 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2260 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2261 included within another project.
2262
2263 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2264 slightly modified by me:
2265
2266 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2267 not building pcregrep.
2268
2269 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2270 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2271
2272 25. Forward references, both numeric and by name, in patterns that made use of
2273 duplicate group numbers, could behave incorrectly or give incorrect errors,
2274 because when scanning forward to find the reference group, PCRE was not
2275 taking into account the duplicate group numbers. A pattern such as
2276 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2277
2278 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2279 the feature of a custom pointer more persuasive (as requested by a user).
2280
2281 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2282 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2283 pre-defined.
2284
2285 28. Added support for (*UTF8) at the start of a pattern.
2286
2287 29. Arrange for flags added by the "release type" setting in CMake to be shown
2288 in the configuration summary.
2289
2290
2291 Version 7.8 05-Sep-08
2292 ---------------------
2293
2294 1. Replaced UCP searching code with optimized version as implemented for Ad
2295 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2296 stage table and inline lookup instead of a function, giving speed ups of 2
2297 to 5 times on some simple patterns that I tested. Permission was given to
2298 distribute the MultiStage2.py script that generates the tables (it's not in
2299 the tarball, but is in the Subversion repository).
2300
2301 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2302 scripts.
2303
2304 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2305 a group with a zero qualifier. The result of the study could be incorrect,
2306 or the function might crash, depending on the pattern.
2307
2308 4. Caseless matching was not working for non-ASCII characters in back
2309 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2310 It now works when Unicode Property Support is available.
2311
2312 5. In pcretest, an escape such as \x{de} in the data was always generating
2313 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2314 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2315 truncation.
2316
2317 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2318
2319 7. Added two (int) casts to pcregrep when printing the difference of two
2320 pointers, in case they are 64-bit values.
2321
2322 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2323 test 2 if it fails.
2324
2325 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2326 and a #define of that name to empty if it is not externally set. This is to
2327 allow users of MSVC to set it if necessary.
2328
2329 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2330 the convenience functions in the pcre_get.c source file.
2331
2332 11. An option change at the start of a pattern that had top-level alternatives
2333 could cause overwriting and/or a crash. This command provoked a crash in
2334 some environments:
2335
2336 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2337
2338 This potential security problem was recorded as CVE-2008-2371.
2339
2340 12. For a pattern where the match had to start at the beginning or immediately
2341 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2342 pcre_dfa_exec() could read past the end of the passed subject if there was
2343 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2344 pcretest so that it places the subject at the end of its malloc-ed buffer.
2345
2346 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2347 exec() might read past the end of the data buffer in UTF-8 mode.
2348
2349 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2350 the data contained the byte 0x85 as part of a UTF-8 character within its
2351 first line. This applied both to normal and DFA matching.
2352
2353 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2354 /^[^d]*?$/8 failed to match "abc".
2355
2356 16. Added a missing copyright notice to pcrecpp_internal.h.
2357
2358 17. Make it more clear in the documentation that values returned from
2359 pcre_exec() in ovector are byte offsets, not character counts.
2360
2361 18. Tidied a few places to stop certain compilers from issuing warnings.
2362
2363 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2364 supplied by Stefan Weber. I made a further small update for 7.8 because
2365 there is a change of source arrangements: the pcre_searchfuncs.c module is
2366 replaced by pcre_ucd.c.
2367
2368
2369 Version 7.7 07-May-08
2370 ---------------------
2371
2372 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2373 a string to a long long, pretend we don't even have a long long." This is
2374 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2375
2376 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2377 pre-7.6 versions, which defined a global no_arg variable instead of putting
2378 it in the RE class. (See also #8 below.)
2379
2380 3. Remove a line of dead code, identified by coverity and reported by Nuno
2381 Lopes.
2382
2383 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2384
2385 (1) The include/exclude patterns were being applied to the whole pathnames
2386 of files, instead of just to the final components.
2387
2388 (2) If there was more than one level of directory, the subdirectories were
2389 skipped unless they satisfied the include/exclude conditions. This is
2390 inconsistent with GNU grep (and could even be seen as contrary to the
2391 pcregrep specification - which I improved to make it absolutely clear).
2392 The action now is always to scan all levels of directory, and just
2393 apply the include/exclude patterns to regular files.
2394
2395 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2396 --exclude_dir in the tests to avoid scanning .svn directories.
2397
2398 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2399 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2400 doesn't support NULs in patterns.
2401
2402 7. Added some missing "const"s to declarations of static tables in
2403 pcre_compile.c and pcre_dfa_exec.c.
2404
2405 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2406 caused by fix #2 above. (Subsequently also a second patch to fix the
2407 first patch. And a third patch - this was a messy problem.)
2408
2409 9. Applied Craig's patch to remove the use of push_back().
2410
2411 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2412 matching function regexec().
2413
2414 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2415 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2416 references. PCRE supports relative numbers with this syntax (I don't think
2417 Oniguruma does).
2418
2419 12. Previously, a group with a zero repeat such as (...){0} was completely
2420 omitted from the compiled regex. However, this means that if the group
2421 was called as a subroutine from elsewhere in the pattern, things went wrong
2422 (an internal error was given). Such groups are now left in the compiled
2423 pattern, with a new opcode that causes them to be skipped at execution
2424 time.
2425
2426 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2427 to the way PCRE behaves:
2428
2429 (a) A lone ] character is dis-allowed (Perl treats it as data).
2430
2431 (b) A back reference to an unmatched subpattern matches an empty string
2432 (Perl fails the current match path).
2433
2434 (c) A data ] in a character class must be notated as \] because if the
2435 first data character in a class is ], it defines an empty class. (In
2436 Perl it is not possible to have an empty class.) The empty class []
2437 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2438 The negative empty class [^] matches any one character, independently
2439 of the DOTALL setting.
2440
2441 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2442 non-existent subpattern following a character class starting with ']' and
2443 containing () gave an internal compiling error instead of "reference to
2444 non-existent subpattern". Fortunately, when the pattern did exist, the
2445 compiled code was correct. (When scanning forwards to check for the
2446 existence of the subpattern, it was treating the data ']' as terminating
2447 the class, so got the count wrong. When actually compiling, the reference
2448 was subsequently set up correctly.)
2449
2450 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2451 it was being rejected as not supported by pcre_dfa_exec(), even though
2452 other assertions are supported. I have made pcre_dfa_exec() support
2453 (*FAIL).
2454
2455 16. The implementation of 13c above involved the invention of a new opcode,
2456 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2457 cannot be changed at match time, I realized I could make a small
2458 improvement to matching performance by compiling OP_ALLANY instead of
2459 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2460 on the OP_ANY path.
2461
2462 17. Compiling pcretest on Windows with readline support failed without the
2463 following two fixes: (1) Make the unistd.h include conditional on
2464 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2465
2466 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2467 ncurses library to be included for pcretest when ReadLine support is
2468 requested, but also to allow for it to be overridden. This patch came from
2469 Daniel Bergström.
2470
2471 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2472 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2473 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2474 spotting this.
2475
2476
2477 Version 7.6 28-Jan-08
2478 ---------------------
2479
2480 1. A character class containing a very large number of characters with
2481 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2482 overflow.
2483
2484 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2485 HAVE_LONG_LONG is not defined.
2486
2487 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2488 bring them up to date and include new features. This patch includes:
2489
2490 - Fixed PH's badly added libz and libbz2 support.
2491 - Fixed a problem with static linking.
2492 - Added pcredemo. [But later removed - see 7 below.]
2493 - Fixed dftables problem and added an option.
2494 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2495 HAVE_LONG_LONG.
2496 - Added readline support for pcretest.
2497 - Added an listing of the option settings after cmake has run.
2498
2499 4. A user submitted a patch to Makefile that makes it easy to create
2500 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2501 Makefile.am that cause it to include this special target, without
2502 affecting anything else. Note that the same mingw target plus all
2503 the other distribution libraries and programs are now supported
2504 when configuring with CMake (see 6 below) instead of with
2505 Configure/Make.
2506
2507 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2508 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2509 exported in the Windows port". It has not yet been confirmed that the patch
2510 solves the problem, but it does no harm.
2511
2512 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2513 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2514 with CMake, and also correct the comment about stack recursion.
2515
2516 7. Remove the automatic building of pcredemo from the ./configure system and
2517 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2518 of a program that users should build themselves after PCRE is installed, so
2519 building it automatically is not really right. What is more, it gave
2520 trouble in some build environments.
2521
2522 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2523
2524
2525 Version 7.5 10-Jan-08
2526 ---------------------
2527
2528 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2529 values in parens when parsing an RE using the C++ wrapper."
2530
2531 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2532 Characters greater than 255 were excluded from the class instead of being
2533 included.
2534
2535 3. The same bug as (2) above applied to negated POSIX classes such as
2536 [:^space:].
2537
2538 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2539 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2540 I have changed it.
2541
2542 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2543 first named subpattern) and a construct such as (?&a) would reference the
2544 first named subpattern whose name started with "a" (in other words, the
2545 length check was missing). Both these problems are fixed. "Subpattern name
2546 expected" is now given for (?&) (a zero-length name), and this patch also
2547 makes it give the same error for \k'' (previously it complained that that
2548 was a reference to a non-existent subpattern).
2549
2550 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2551 this is right because (?- can be followed by option settings as well as by
2552 digits. I have, however, made the messages clearer.
2553
2554 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2555 than the number used in the conditional) now cause a compile-time error.
2556 This is actually not compatible with Perl, which accepts such patterns, but
2557 treats the conditional as always being FALSE (as PCRE used to), but it
2558 seems to me that giving a diagnostic is better.
2559
2560 8. Change "alphameric" to the more common word "alphanumeric" in comments
2561 and messages.
2562
2563 9. Fix two occurrences of "backslash" in comments that should have been
2564 "backspace".
2565
2566 10. Remove two redundant lines of code that can never be obeyed (their function
2567 was moved elsewhere).
2568
2569 11. The program that makes PCRE's Unicode character property table had a bug
2570 which caused it to generate incorrect table entries for sequences of
2571 characters that have the same character type, but are in different scripts.
2572 It amalgamated them into a single range, with the script of the first of
2573 them. In other words, some characters were in the wrong script. There were
2574 thirteen such cases, affecting characters in the following ranges:
2575
2576 U+002b0 - U+002c1
2577 U+0060c - U+0060d
2578 U+0061e - U+00612
2579 U+0064b - U+0065e
2580 U+0074d - U+0076d
2581 U+01800 - U+01805
2582 U+01d00 - U+01d77
2583 U+01d9b - U+01dbf
2584 U+0200b - U+0200f
2585 U+030fc - U+030fe
2586 U+03260 - U+0327f
2587 U+0fb46 - U+0fbb1
2588 U+10450 - U+1049d
2589
2590 12. The -o option (show only the matching part of a line) for pcregrep was not
2591 compatible with GNU grep in that, if there was more than one match in a
2592 line, it showed only the first of them. It now behaves in the same way as
2593 GNU grep.
2594
2595 13. If the -o and -v options were combined for pcregrep, it printed a blank
2596 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2597 does the same. The return code can be used to tell if there were any
2598 non-matching lines.
2599
2600 14. Added --file-offsets and --line-offsets to pcregrep.
2601
2602 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2603 infinitely looping recursion. The bug was that positive lookaheads were not
2604 being skipped when checking for a possible empty match (negative lookaheads
2605 and both kinds of lookbehind were skipped).
2606
2607 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2608 inclusion of <windows.h> to before rather than after the definition of
2609 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2610
2611 17. Specifying a possessive quantifier with a specific limit for a Unicode
2612 character property caused pcre_compile() to compile bad code, which led at
2613 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2614 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2615 caused the error; without that there was no problem.
2616
2617 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2618
2619 19. Added --enable-pcretest-libreadline.
2620
2621 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2622 RE::GlobalReplace(). As a result, the number of replacements returned was
2623 double what it should be. I removed one of the increments, but Craig sent a
2624 later patch that removed the other one (the right fix) and added unit tests
2625 that check the return values (which was not done before).
2626
2627 21. Several CMake things:
2628
2629 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2630 the names libpcre and libpcreposix, not just pcre and pcreposix.
2631
2632 (2) The above change means that pcretest and pcregrep are now correctly
2633 linked with the newly-built libraries, not previously installed ones.
2634
2635 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2636
2637 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2638 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2639 UTF-8 newline character). The key issue is that the pattern starts .*;
2640 this means that the match must be either at the beginning, or after a
2641 newline. The bug was in the code for advancing after a failed match and
2642 checking that the new position followed a newline. It was not taking
2643 account of UTF-8 characters correctly.
2644
2645 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2646 character classes. PCRE was not treating the sequence [:...:] as a
2647 character class unless the ... were all letters. Perl, however, seems to
2648 allow any characters between [: and :], though of course it rejects as
2649 unknown any "names" that contain non-letters, because all the known class
2650 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2651 for example, whereas PCRE did not - it did not recognize a POSIX character
2652 class. This seemed a bit dangerous, so the code has been changed to be
2653 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2654 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2655 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2656 Perl does, and where it didn't before.
2657
2658 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2659 Windows environments %n is disabled by default.
2660
2661
2662 Version 7.4 21-Sep-07
2663 ---------------------
2664
2665 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2666 means that a class such as [\s] counted as "explicit reference to CR or
2667 LF". That isn't really right - the whole point of the change was to try to
2668 help when there was an actual mention of one of the two characters. So now
2669 the change happens only if \r or \n (or a literal CR or LF) character is
2670 encountered.
2671
2672 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2673 of both had grown to the point where there were only 3 bits left.
2674 Fortunately, there was spare space in the data structure, and so I have
2675 moved the internal flags into a new 16-bit field to free up more option
2676 bits.
2677
2678 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2679 but did not set the internal JCHANGED flag - either of these is enough to
2680 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2681 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2682 start sets both bits.
2683
2684 4. Added options (at build time, compile time, exec time) to change \R from
2685 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2686
2687 5. doc/pcresyntax.html was missing from the distribution.
2688
2689 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2690 compatibility, even though it is no longer used.
2691
2692 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2693 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2694 windows.h file is present (where different names are used). [This was
2695 reversed later after testing - see 16 below.]
2696
2697 8. Changed all #include <config.h> to #include "config.h". There were also
2698 some further <pcre.h> cases that I changed to "pcre.h".
2699
2700 9. When pcregrep was used with the --colour option, it missed the line ending
2701 sequence off the lines that it output.
2702
2703 10. It was pointed out to me that arrays of string pointers cause lots of
2704 relocations when a shared library is dynamically loaded. A technique of
2705 using a single long string with a table of offsets can drastically reduce
2706 these. I have refactored PCRE in four places to do this. The result is
2707 dramatic:
2708
2709 Originally: 290
2710 After changing UCP table: 187
2711 After changing error message table: 43
2712 After changing table of "verbs" 36
2713 After changing table of Posix names 22
2714
2715 Thanks to the folks working on Gregex for glib for this insight.
2716
2717 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2718 unicode-properties was also set.
2719
2720 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2721
2722 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2723 checked only for CRLF.
2724
2725 14. Added casts to pcretest.c to avoid compiler warnings.
2726
2727 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2728
2729 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2730 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2731 entirely. This removes changes made in 7 above.
2732
2733 17. The CMake files have been updated, and there is now more information about
2734 building with CMake in the NON-UNIX-USE document.
2735
2736
2737 Version 7.3 28-Aug-07
2738 ---------------------
2739
2740 1. In the rejigging of the build system that eventually resulted in 7.1, the
2741 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2742 brackets there is not right, since it causes compilers to look for an
2743 installed pcre.h, not the version that is in the source that is being
2744 compiled (which of course may be different). I have changed it back to:
2745
2746 #include "pcre.h"
2747
2748 I have a vague recollection that the change was concerned with compiling in
2749 different directories, but in the new build system, that is taken care of
2750 by the VPATH setting the Makefile.
2751
2752 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2753 when the subject happened to end in the byte 0x85 (e.g. if the last
2754 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2755 characters but of course it shouldn't be taken as a newline when it is part
2756 of another character. The bug was that, for an unlimited repeat of . in
2757 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2758 characters when looking for a newline.
2759
2760 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2761
2762 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2763 in debug output.
2764
2765 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2766 long printing in the pcrecpp unittest when running under MinGW.
2767
2768 6. ESC_K was left out of the EBCDIC table.
2769
2770 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2771 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2772 limit also applies to "virtual nesting" when a pattern is recursive, and in
2773 this case 1000 isn't so big. I have been able to remove this limit at the
2774 expense of backing off one optimization in certain circumstances. Normally,
2775 when pcre_exec() would call its internal match() function recursively and
2776 immediately return the result unconditionally, it uses a "tail recursion"
2777 feature to save stack. However, when a subpattern that can match an empty
2778 string has an unlimited repetition quantifier, it no longer makes this
2779 optimization. That gives it a stack frame in which to save the data for
2780 checking that an empty string has been matched. Previously this was taken
2781 from the 1000-entry workspace that had been reserved. So now there is no
2782 explicit limit, but more stack is used.
2783
2784 8. Applied Daniel's patches to solve problems with the import/export magic
2785 syntax that is required for Windows, and which was going wrong for the
2786 pcreposix and pcrecpp parts of the library. These were overlooked when this
2787 problem was solved for the main library.
2788
2789 9. There were some crude static tests to avoid integer overflow when computing
2790 the size of patterns that contain repeated groups with explicit upper
2791 limits. As the maximum quantifier is 65535, the maximum group length was
2792 set at 30,000 so that the product of these two numbers did not overflow a
2793 32-bit integer. However, it turns out that people want to use groups that
2794 are longer than 30,000 bytes (though not repeat them that many times).
2795 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2796 made it possible to implement the integer overflow checks in a much more
2797 dynamic way, which I have now done. The artificial limitation on group
2798 length has been removed - we now have only the limit on the total length of
2799 the compiled pattern, which depends on the LINK_SIZE setting.
2800
2801 10. Fixed a bug in the documentation for get/copy named substring when
2802 duplicate names are permitted. If none of the named substrings are set, the
2803 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2804 empty string.
2805
2806 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2807 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2808 because the ] is interpreted as the first data character and the
2809 terminating ] is not found. PCRE has been made compatible with Perl in this
2810 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2811 cause memory overwriting.
2812
2813 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2814 string has been matched (to stop an infinite loop). It was not recognizing
2815 a conditional subpattern that could match an empty string if that
2816 subpattern was within another subpattern. For example, it looped when
2817 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2818 condition was not nested. This bug has been fixed.
2819
2820 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2821 past the start of the subject in the presence of bytes with the top bit
2822 set, for example "\x8aBCD".
2823
2824 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2825 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2826
2827 14. Optimized (?!) to (*FAIL).
2828
2829 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2830 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2831 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2832 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2833 does: it's just the validity check that is more restrictive.
2834
2835 16. Inserted checks for integer overflows during escape sequence (backslash)
2836 processing, and also fixed erroneous offset values for syntax errors during
2837 backslash processing.
2838
2839 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2840 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2841
2842 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2843 caused an overrun.
2844
2845 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2846 something other than just ASCII characters) inside a group that had an
2847 unlimited repeat caused a loop at compile time (while checking to see
2848 whether the group could match an empty string).
2849
2850 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2851 [\P{Any}] did so. (Error in the code for printing property names.)
2852
2853 21. An orphan \E inside a character class could cause a crash.
2854
2855 22. A repeated capturing bracket such as (A)? could cause a wild memory
2856 reference during compilation.
2857
2858 23. There are several functions in pcre_compile() that scan along a compiled
2859 expression for various reasons (e.g. to see if it's fixed length for look
2860 behind). There were bugs in these functions when a repeated \p or \P was
2861 present in the pattern. These operators have additional parameters compared
2862 with \d, etc, and these were not being taken into account when moving along
2863 the compiled data. Specifically:
2864
2865 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2866 length.
2867
2868 (b) An item such as \pL+ within a repeated group could cause crashes or
2869 loops.
2870
2871 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2872 "reference to non-existent subpattern" error.
2873
2874 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2875
2876 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2877 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2878
2879 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2880
2881 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2882 character were causing crashes (broken optimization).
2883
2884 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2885 \p or \P) caused a compile-time loop.
2886
2887 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2888 break. For example, the unstudied pattern [\r\n]A does not match the string
2889 "\r\nA" because change 7.0/46 below moves the current point on by two
2890 characters after failing to match at the start. However, the pattern \nA
2891 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2892 the same is true. There doesn't seem any very clean way out of this, but
2893 what I have chosen to do makes the common cases work: PCRE now takes note
2894 of whether there can be an explicit match for \r or \n anywhere in the
2895 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2896 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2897 pattern has explicit CR or LF references.
2898
2899 29. Added (*CR) etc for changing newline setting at start of pattern.
2900
2901
2902 Version 7.2 19-Jun-07
2903 ---------------------
2904
2905 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2906 which is apparently normally available under Windows.
2907
2908 2. Re-jig the pcregrep tests with different newline settings in an attempt
2909 to make them independent of the local environment's newline setting.
2910
2911 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2912
2913 4. Some of the "internals" tests were previously cut out when the link size
2914 was not 2, because the output contained actual offsets. The recent new
2915 "Z" feature of pcretest means that these can be cut out, making the tests
2916 usable with all link sizes.
2917
2918 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2919 stack recursion. This gives a massive performance boost under BSD, but just
2920 a small improvement under Linux. However, it saves one field in the frame
2921 in all cases.
2922
2923 6. Added more features from the forthcoming Perl 5.10:
2924
2925 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2926 recursion call. It refers to the nth most recently opened parentheses.
2927
2928 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2929 to be opened parentheses.
2930
2931 (c) Conditions that refer to capturing parentheses can be specified
2932 relatively, for example, (?(-2)... or (?(+3)...
2933
2934 (d) \K resets the start of the current match so that everything before
2935 is not part of it.
2936
2937 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2938
2939 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2940 reference syntax.
2941
2942 (g) (?| introduces a group in which the numbering of parentheses in each
2943 alternative starts with the same number.
2944
2945 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2946
2947 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2948 PCRE_INFO_JCHANGED.
2949
2950 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2951 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2952 for detecting groups that can match an empty string.
2953
2954 9. A pattern with a very large number of alternatives (more than several
2955 hundred) was running out of internal workspace during the pre-compile
2956 phase, where pcre_compile() figures out how much memory will be needed. A
2957 bit of new cunning has reduced the workspace needed for groups with
2958 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2959 workspace instead of running out of the 4096 that are available.
2960
2961 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2962
2963 11. Applied patch from Google to remove an optimization that didn't quite work.
2964 The report of the bug said:
2965
2966 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2967 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2968 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2969
2970 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2971 it matched the wrong number of bytes.
2972
2973
2974 Version 7.1 24-Apr-07
2975 ---------------------
2976
2977 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2978 that is more "standard", making use of automake and other Autotools. There
2979 is some re-arrangement of the files and adjustment of comments consequent
2980 on this.
2981
2982 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2983 for recursive directory scanning broke on some systems because the files
2984 are not scanned in any specific order and on different systems the order
2985 was different. A call to "sort" has been inserted into RunGrepTest for the
2986 approprate test as a short-term fix. In the longer term there may be an
2987 alternative.
2988
2989 3. I had an email from Eric Raymond about problems translating some of PCRE's
2990 man pages to HTML (despite the fact that I distribute HTML pages, some
2991 people do their own conversions for various reasons). The problems
2992 concerned the use of low-level troff macros .br and .in. I have therefore
2993 removed all such uses from the man pages (some were redundant, some could
2994 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2995 HTML has been updated to handle .nf/.fi and to complain if it encounters
2996 .br or .in.
2997
2998 4. Updated comments in configure.ac that get placed in config.h.in and also
2999 arranged for config.h to be included in the distribution, with the name
3000 config.h.generic, for the benefit of those who have to compile without
3001 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3002
3003 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3004 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3005 makevp.bat for the current PCRE, using the additional files
3006 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3007
3008 6. A Windows user reported a minor discrepancy with test 2, which turned out
3009 to be caused by a trailing space on an input line that had got lost in his
3010 copy. The trailing space was an accident, so I've just removed it.
3011
3012 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3013 that is needed.
3014
3015 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3016 as "const" (a) because they are and (b) because it helps the PHP
3017 maintainers who have recently made a script to detect big data structures
3018 in the php code that should be moved to the .rodata section. I remembered
3019 to update Builducptable as well, so it won't revert if ucptable.h is ever
3020 re-created.
3021
3022 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3023 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3024 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3025 support is not required. This saves 1.5-2K of code, which is important in
3026 some applications.
3027
3028 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3029 so as not to refer to the tables, even though these functions will never be
3030 called when UTF-8 support is disabled. Otherwise there are problems with a
3031 shared library.
3032
3033 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3034
3035 (a) It was defining its arguments as char * instead of void *.
3036
3037 (b) It was assuming that all moves were upwards in memory; this was true
3038 a long time ago when I wrote it, but is no longer the case.
3039
3040 The emulated memove() is provided for those environments that have neither
3041 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3042 is clearly not the case, as these two bugs were recently reported.
3043
3044 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3045 and Detrail to create the HTML documentation, the .txt form of the man
3046 pages, and it removes trailing spaces from listed files. It also creates
3047 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3048 case, it wraps all the #defines with #ifndefs. This script should be run
3049 before "make dist".
3050
3051 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3052 with Unicode property support.
3053
3054 (a) For a maximizing quantifier, if the two different cases of the
3055 character were of different lengths in their UTF-8 codings (there are
3056 some cases like this - I found 11), and the matching function had to
3057 back up over a mixture of the two cases, it incorrectly assumed they
3058 were both the same length.
3059
3060 (b) When PCRE was configured to use the heap rather than the stack for
3061 recursion during matching, it was not correctly preserving the data for
3062 the other case of a UTF-8 character when checking ahead for a match
3063 while processing a minimizing repeat. If the check also involved
3064 matching a wide character, but failed, corruption could cause an
3065 erroneous result when trying to check for a repeat of the original
3066 character.
3067
3068 13. Some tidying changes to the testing mechanism:
3069
3070 (a) The RunTest script now detects the internal link size and whether there
3071 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3072 values substituted by "configure". (The RunGrepTest script already did
3073 this for UTF-8.) The configure.ac script no longer substitutes the
3074 relevant variables.
3075
3076 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3077 with length and offset values. This means that the output is different
3078 for different internal link sizes. Test 2 is skipped for link sizes
3079 other than 2 because of this, bypassing the problem. Unfortunately,
3080 there was also a test in test 3 (the locale tests) that used /B and
3081 failed for link sizes other than 2. Rather than cut the whole test out,
3082 I have added a new /Z option to pcretest that replaces the length and
3083 offset values with spaces. This is now used to make test 3 independent
3084 of link size. (Test 2 will be tidied up later.)
3085
3086 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3087 segmentation fault instead of returning the appropriate error message.
3088
3089 15. In multiline mode when the newline sequence was set to "any", the pattern
3090 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3091 This doesn't seem right; it now treats the CRLF combination as the line
3092 ending, and so does not match in that case. It's only a pattern such as ^$
3093 that would hit this one: something like ^ABC$ would have failed after \r
3094 and then tried again after \r\n.
3095
3096 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3097 in an attempt to make files that differ only in their line terminators
3098 compare equal. This works on Linux.
3099
3100 17. Under certain error circumstances pcregrep might try to free random memory
3101 as it exited. This is now fixed, thanks to valgrind.
3102
3103 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3104 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3105 was because its rules for how to advance for /g after matching an empty
3106 string at the end of a line did not allow for this case. They now check for
3107 it specially.
3108
3109 20. pcretest is supposed to handle patterns and data of any length, by
3110 extending its buffers when necessary. It was getting this wrong when the
3111 buffer for a data line had to be extended.
3112
3113 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3114 CRLF as a newline sequence.
3115
3116 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3117 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3118 I have nevertheless tidied it up.
3119
3120 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3121
3122 24. Added a man page for pcre-config.
3123
3124
3125 Version 7.0 19-Dec-06
3126 ---------------------
3127
3128 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3129 moving to gcc 4.1.1.
3130
3131 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3132 sys/time.h, which is documented as needed for this function. It doesn't
3133 seem to matter on Linux, but it showed up on some releases of OS X.
3134
3135 3. It seems that there are systems where bytes whose values are greater than
3136 127 match isprint() in the "C" locale. The "C" locale should be the
3137 default when a C program starts up. In most systems, only ASCII printing
3138 characters match isprint(). This difference caused the output from pcretest
3139 to vary, making some of the tests fail. I have changed pcretest so that:
3140
3141 (a) When it is outputting text in the compiled version of a pattern, bytes
3142 other than 32-126 are always shown as hex escapes.
3143
3144 (b) When it is outputting text that is a matched part of a subject string,
3145 it does the same, unless a different locale has been set for the match
3146 (using the /L modifier). In this case, it uses isprint() to decide.
3147
3148 4. Fixed a major bug that caused incorrect computation of the amount of memory
3149 required for a compiled pattern when options that changed within the
3150 pattern affected the logic of the preliminary scan that determines the
3151 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3152 that the computed length was too small. The symptoms of this bug were
3153 either the PCRE error "internal error: code overflow" from pcre_compile(),
3154 or a glibc crash with a message such as "pcretest: free(): invalid next
3155 size (fast)". Examples of patterns that provoked this bug (shown in
3156 pcretest format) are:
3157
3158 /(?-x: )/x
3159 /(?x)(?-x: \s*#\s*)/
3160 /((?i)[\x{c0}])/8
3161 /(?i:[\x{c0}])/8
3162
3163 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3164 is now done differently.
3165
3166 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3167 wrapper classes; (b) implement a new function in the C++ scanner that is
3168 more efficient than the old way of doing things because it avoids levels of
3169 recursion in the regex matching; (c) add a paragraph to the documentation
3170 for the FullMatch() function.
3171
3172 6. The escape sequence \n was being treated as whatever was defined as
3173 "newline". Not only was this contrary to the documentation, which states
3174 that \n is character 10 (hex 0A), but it also went horribly wrong when
3175 "newline" was defined as CRLF. This has been fixed.
3176
3177 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3178 was being set to -1 for the "end of line" case (supposedly a value that no
3179 character can have). Though this value is never used (the check for end of
3180 line is "zero bytes in current character"), it caused compiler complaints.
3181 I've changed it to 0xffffffff.
3182
3183 8. In pcre_version.c, the version string was being built by a sequence of
3184 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3185 string (as it is for production releases) called a macro with an empty
3186 argument. The C standard says the result of this is undefined. The gcc
3187 compiler treats it as an empty string (which was what was wanted) but it is
3188 reported that Visual C gives an error. The source has been hacked around to
3189 avoid this problem.
3190
3191 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3192 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3193 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3194 of them did).
3195
3196 10. Originally, pcretest opened its input and output without "b"; then I was
3197 told that "b" was needed in some environments, so it was added for release
3198 5.0 to both the input and output. (It makes no difference on Unix-like
3199 systems.) Later I was told that it is wrong for the input on Windows. I've
3200 now abstracted the modes into two macros, to make it easier to fiddle with
3201 them, and removed "b" from the input mode under Windows.
3202
3203 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3204
3205 12. Added -help and --help to pcretest as an official way of being reminded
3206 of the options.
3207
3208 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3209 and pcrecpp.cc because they annoy compilers at high warning levels.
3210
3211 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3212
3213 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3214 scripts are not C programs :-) and which was not noticed because it works
3215 on Linux.
3216
3217 16. pcretest is supposed to handle any length of pattern and data line (as one
3218 line or as a continued sequence of lines) by extending its input buffer if
3219 necessary. This feature was broken for very long pattern lines, leading to
3220 a string of junk being passed to pcre_compile() if the pattern was longer
3221 than about 50K.
3222
3223 17. I have done a major re-factoring of the way pcre_compile() computes the
3224 amount of memory needed for a compiled pattern. Previously, there was code
3225 that made a preliminary scan of the pattern in order to do this. That was
3226 OK when PCRE was new, but as the facilities have expanded, it has become
3227 harder and harder to keep it in step with the real compile phase, and there
3228 have been a number of bugs (see for example, 4 above). I have now found a
3229 cunning way of running the real compile function in a "fake" mode that
3230 enables it to compute how much memory it would need, while actually only
3231 ever using a few hundred bytes of working memory and without too many
3232 tests of the mode. This should make future maintenance and development
3233 easier. A side effect of this work is that the limit of 200 on the nesting
3234 depth of parentheses has been removed (though this was never a serious
3235 limitation, I suspect). However, there is a downside: pcre_compile() now
3236 runs more slowly than before (30% or more, depending on the pattern). I
3237 hope this isn't a big issue. There is no effect on runtime performance.
3238
3239 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3240 newline (only possible for the last line of a file) and it was a
3241 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3242
3243 19. Added additional timing features to pcretest. (1) The -tm option now times
3244 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3245 separate command line item, by a number that specifies the number of
3246 repeats to use when timing. The default is 50000; this gives better
3247 precision, but takes uncomfortably long for very large patterns.
3248
3249 20. Extended pcre_study() to be more clever in cases where a branch of a
3250 subpattern has no definite first character. For example, (a*|b*)[cd] would
3251 previously give no result from pcre_study(). Now it recognizes that the
3252 first character must be a, b, c, or d.
3253
3254 21. There was an incorrect error "recursive call could loop indefinitely" if
3255 a subpattern (or the entire pattern) that was being tested for matching an
3256 empty string contained only one non-empty item after a nested subpattern.
3257 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3258 incorrectly, because the \d was being skipped in the check.
3259
3260 22. The pcretest program now has a new pattern option /B and a command line
3261 option -b, which is equivalent to adding /B to every pattern. This causes
3262 it to show the compiled bytecode, without the additional information that
3263 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3264 is the same as /B/I).
3265
3266 23. A new optimization is now able automatically to treat some sequences such
3267 as a*b as a*+b. More specifically, if something simple (such as a character
3268 or a simple class like \d) has an unlimited quantifier, and is followed by
3269 something that cannot possibly match the quantified thing, the quantifier
3270 is automatically "possessified".
3271
3272 24. A recursive reference to a subpattern whose number was greater than 39
3273 went wrong under certain circumstances in UTF-8 mode. This bug could also
3274 have affected the operation of pcre_study().
3275
3276 25. Realized that a little bit of performance could be had by replacing
3277 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3278
3279 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3280
3281 27. Possessive quantifiers such as a++ were previously implemented by turning
3282 them into atomic groups such as ($>a+). Now they have their own opcodes,
3283 which improves performance. This includes the automatically created ones
3284 from 23 above.
3285
3286 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3287 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3288 the first matched character to be a colon. This applied both to named and
3289 numbered groups.
3290
3291 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3292
3293 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3294 building PCRE on Windows easier, so I have included it in the distribution.
3295
3296 31. There is now a check in pcretest against a ridiculously large number being
3297 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3298 loop, the loop is abandoned.
3299
3300 32. Forward references to subpatterns in conditions such as (?(2)...) where
3301 subpattern 2 is defined later cause pcre_compile() to search forwards in
3302 the pattern for the relevant set of parentheses. This search went wrong
3303 when there were unescaped parentheses in a character class, parentheses
3304 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3305
3306 33. "Subroutine" calls and backreferences were previously restricted to
3307 referencing subpatterns earlier in the regex. This restriction has now
3308 been removed.
3309
3310 34. Added a number of extra features that are going to be in Perl 5.10. On the
3311 whole, these are just syntactic alternatives for features that PCRE had
3312 previously implemented using the Python syntax or my own invention. The
3313 other formats are all retained for compatibility.
3314
3315 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3316 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3317 also .NET compatible.
3318
3319 (b) A recursion or subroutine call to a named group can now be defined as
3320 (?&name) as well as (?P>name).
3321
3322 (c) A backreference to a named group can now be defined as \k<name> or
3323 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3324 5.10, are also .NET compatible.
3325
3326 (d) A conditional reference to a named group can now use the syntax
3327 (?(<name>) or (?('name') as well as (?(name).
3328
3329 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3330 groups (named and numbered) that are never evaluated inline, but can be
3331 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3332 is always false. There may be only one alternative in such a group.
3333
3334 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3335 as the simple (?(R). The condition is true only if the most recent
3336 recursion is that of the given number or name. It does not search out
3337 through the entire recursion stack.
3338
3339 (g) The escape \gN or \g{N} has been added, where N is a positive or
3340 negative number, specifying an absolute or relative reference.
3341
3342 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3343 some "unreachable code" warnings.
3344
3345 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3346 things, this adds five new scripts.
3347
3348 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3349 There were also incompatibilities regarding the handling of \Q..\E inside
3350 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3351 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3352
3353 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3354 matches an empty string, and forcibly breaks the loop. There were bugs in
3355 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3356 against aaaa the result was just "a" rather than "aaaa", for example. Two
3357 separate and independent bugs (that affected different cases) have been
3358 fixed.
3359
3360 39. Refactored the code to abolish the use of different opcodes for small
3361 capturing bracket numbers. This is a tidy that I avoided doing when I
3362 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3363 The new approach is not only tidier, it makes it possible to reduce the
3364 memory needed to fix the previous bug (38).
3365
3366 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3367 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3368 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3369 mode.
3370
3371 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3372 report.
3373
3374 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3375 copy construction and assignment in the C++ wrapper.
3376
3377 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3378 couple of bugs that could have given wrong results in the "--newline=crlf"
3379 case.
3380
3381 44. Added a number of casts and did some reorganization of signed/unsigned int
3382 variables following suggestions from Dair Grant. Also renamed the variable
3383 "this" as "item" because it is a C++ keyword.
3384
3385 45. Arranged for dftables to add
3386
3387 #include "pcre_internal.h"
3388
3389 to pcre_chartables.c because without it, gcc 4.x may remove the array
3390 definition from the final binary if PCRE is built into a static library and
3391 dead code stripping is activated.
3392
3393 46. For an unanchored pattern, if a match attempt fails at the start of a
3394 newline sequence, and the newline setting is CRLF or ANY, and the next two
3395 characters are CRLF, advance by two characters instead of one.
3396
3397
3398 Version 6.7 04-Jul-06
3399 ---------------------
3400
3401 1. In order to handle tests when input lines are enormously long, pcretest has
3402 been re-factored so that it automatically extends its buffers when
3403 necessary. The code is crude, but this _is_ just a test program. The
3404 default size has been increased from 32K to 50K.
3405
3406 2. The code in pcre_study() was using the value of the re argument before
3407 testing it for NULL. (Of course, in any sensible call of the function, it
3408 won't be NULL.)
3409
3410 3. The memmove() emulation function in pcre_internal.h, which is used on
3411 systems that lack both memmove() and bcopy() - that is, hardly ever -
3412 was missing a "static" storage class specifier.
3413
3414 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3415 containing an extended class (one that cannot be represented by a bitmap
3416 because it contains high-valued characters or Unicode property items, e.g.
3417 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3418 pattern, but PCRE should not loop if you do not (it no longer does).
3419 [Detail: two cases were found: (a) a repeated subpattern containing an
3420 extended class; (b) a recursive reference to a subpattern that followed a
3421 previous extended class. It wasn't skipping over the extended class
3422 correctly when UTF-8 mode was not set.]
3423
3424 5. A negated single-character class was not being recognized as fixed-length
3425 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3426 compile error "lookbehind assertion is not fixed length".
3427
3428 6. The RunPerlTest auxiliary script was showing an unexpected difference
3429 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3430 write a Perl script that can interpret lines of an input file either as
3431 byte characters or as UTF-8, which is what "perltest" was being required to
3432 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3433 can't do is switch easily at run time between having the "use utf8;" pragma
3434 or not. In the end, I fudged it by using the RunPerlTest script to insert
3435 "use utf8;" explicitly for the UTF-8 tests.
3436
3437 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3438 the end of the subject string, contrary to the documentation and to what
3439 Perl does. This was true of both matching functions. Now it matches only at
3440 the start of the subject and immediately after *internal* newlines.
3441
3442 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3443 a pointer to an int instead of a pointer to an unsigned long int. This
3444 caused problems on 64-bit systems.
3445
3446 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3447 instance of the 'standard' template library not being so standard".
3448
3449 10. There was no check on the number of named subpatterns nor the maximum
3450 length of a subpattern name. The product of these values is used to compute
3451 the size of the memory block for a compiled pattern. By supplying a very
3452 long subpattern name and a large number of named subpatterns, the size
3453 computation could be caused to overflow. This is now prevented by limiting
3454 the length of names to 32 characters, and the number of named subpatterns
3455 to 10,000.
3456
3457 11. Subpatterns that are repeated with specific counts have to be replicated in
3458 the compiled pattern. The size of memory for this was computed from the
3459 length of the subpattern and the repeat count. The latter is limited to
3460 65535, but there was no limit on the former, meaning that integer overflow
3461 could in principle occur. The compiled length of a repeated subpattern is
3462 now limited to 30,000 bytes in order to prevent this.
3463
3464 12. Added the optional facility to have named substrings with the same name.
3465
3466 13. Added the ability to use a named substring as a condition, using the
3467 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3468 are numbers (not recommended). Forward references are permitted.
3469
3470 14. Added forward references in named backreferences (if you see what I mean).
3471
3472 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3473 pattern could run off the end of the subject. For example, the pattern
3474 "(?s)(.{1,5})"8 did this with the subject "ab".
3475
3476 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3477 PCRE_CASELESS was set when matching characters that were quantified with ?
3478 or *.
3479
3480 17. A character class other than a single negated character that had a minimum
3481 but no maximum quantifier - for example [ab]{6,} - was not handled
3482 correctly by pce_dfa_exec(). It would match only one character.
3483
3484 18. A valid (though odd) pattern that looked like a POSIX character
3485 class but used an invalid character after [ (for example [[,abc,]]) caused
3486 pcre_compile() to give the error "Failed: internal error: code overflow" or
3487 in some cases to crash with a glibc free() error. This could even happen if
3488 the pattern terminated after [[ but there just happened to be a sequence of
3489 letters, a binary zero, and a closing ] in the memory that followed.
3490
3491 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3492 over the years. Originally (before any Unicode support), just the bottom 8
3493 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3494 output from "man perlunicode" includes this:
3495
3496 The regular expression compiler produces polymorphic opcodes. That
3497 is, the pattern adapts to the data and automatically switches to
3498 the Unicode character scheme when presented with Unicode data--or
3499 instead uses a traditional byte scheme when presented with byte
3500 data.
3501
3502 Sadly, a wide octal escape does not cause a switch, and in a string with
3503 no other multibyte characters, these octal escapes are treated as before.
3504 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3505 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3506 Unicode string.
3507
3508 I have not perpetrated such confusion in PCRE. Up till now, it took just
3509 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3510 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3511 translate to the appropriate multibyte character.
3512
3513 29. Applied some refactoring to reduce the number of warnings from Microsoft
3514 and Borland compilers. This has included removing the fudge introduced
3515 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3516 a warning about an unused variable.
3517
3518 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3519 characters since release 4.0, because Perl (from release 5.004) does not.
3520 [Or at least, is documented not to: some releases seem to be in conflict
3521 with the documentation.] However, when a pattern was studied with
3522 pcre_study() and all its branches started with \s, PCRE still included VT
3523 as a possible starting character. Of course, this did no harm; it just
3524 caused an unnecessary match attempt.
3525
3526 22. Removed a now-redundant internal flag bit that recorded the fact that case
3527 dependency changed within the pattern. This was once needed for "required
3528 byte" processing, but is no longer used. This recovers a now-scarce options
3529 bit. Also moved the least significant internal flag bit to the most-
3530 significant bit of the word, which was not previously used (hangover from
3531 the days when it was an int rather than a uint) to free up another bit for
3532 the future.
3533
3534 23. Added support for CRLF line endings as well as CR and LF. As well as the
3535 default being selectable at build time, it can now be changed at runtime
3536 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3537 specify that it is scanning data with non-default line endings.
3538
3539 24. Changed the definition of CXXLINK to make it agree with the definition of
3540 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3541
3542 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3543 recursions. This makes a big different to stack usage for some patterns.
3544
3545 26. If a subpattern containing a named recursion or subroutine reference such
3546 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3547 the space required for the compiled pattern went wrong and gave too small a
3548 value. Depending on the environment, this could lead to "Failed: internal
3549 error: code overflow at offset 49" or "glibc detected double free or
3550 corruption" errors.
3551
3552 27. Applied patches from Google (a) to support the new newline modes and (b) to
3553 advance over multibyte UTF-8 characters in GlobalReplace.
3554
3555 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3556 difference for some implementation of PCRE in some Windows version.
3557
3558 29. Added some extra testing facilities to pcretest:
3559
3560 \q<number> in a data line sets the "match limit" value
3561 \Q<number> in a data line sets the "match recursion limt" value
3562 -S <number> sets the stack size, where <number> is in megabytes
3563
3564 The -S option isn't available for Windows.
3565
3566
3567 Version 6.6 06-Feb-06
3568 ---------------------
3569
3570 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3571 in pcreposix.h. I have copied the definition from pcre.h.
3572
3573 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3574 because pcre.h is no longer a built file.
3575
3576 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3577 not normally included in the compiled code.
3578
3579
3580 Version 6.5 01-Feb-06
3581 ---------------------
3582
3583 1. When using the partial match feature with pcre_dfa_exec(), it was not
3584 anchoring the second and subsequent partial matches at the new starting
3585 point. This could lead to incorrect results. For example, with the pattern
3586 /1234/, partially matching against "123" and then "a4" gave a match.
3587
3588 2. Changes to pcregrep:
3589
3590 (a) All non-match returns from pcre_exec() were being treated as failures
3591 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3592 error message is output. Some extra information is given for the
3593 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3594 probably the only errors that are likely to be caused by users (by
3595 specifying a regex that has nested indefinite repeats, for instance).
3596 If there are more than 20 of these errors, pcregrep is abandoned.
3597
3598 (b) A binary zero was treated as data while matching, but terminated the
3599 output line if it was written out. This has been fixed: binary zeroes
3600 are now no different to any other data bytes.
3601
3602 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3603 used to set a locale for matching. The --locale=xxxx long option has
3604 been added (no short equivalent) to specify a locale explicitly on the
3605 pcregrep command, overriding the environment variables.
3606
3607 (d) When -B was used with -n, some line numbers in the output were one less
3608 than they should have been.
3609
3610 (e) Added the -o (--only-matching) option.
3611
3612 (f) If -A or -C was used with -c (count only), some lines of context were
3613 accidentally printed for the final match.
3614
3615 (g) Added the -H (--with-filename) option.
3616
3617 (h) The combination of options -rh failed to suppress file names for files
3618 that were found from directory arguments.
3619
3620 (i) Added the -D (--devices) and -d (--directories) options.
3621
3622 (j) Added the -F (--fixed-strings) option.
3623
3624 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3625
3626 (l) Added the --colo(u)r option.
3627
3628 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3629 is not present by default.
3630
3631 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3632 items such as (?R) or (?1), when the recursion could match a number of
3633 alternatives. If it matched one of the alternatives, but subsequently,
3634 outside the recursion, there was a failure, the code tried to back up into
3635 the recursion. However, because of the way PCRE is implemented, this is not
3636 possible, and the result was an incorrect result from the match.
3637
3638 In order to prevent this happening, the specification of recursion has
3639 been changed so that all such subpatterns are automatically treated as
3640 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3641
3642 4. I had overlooked the fact that, in some locales, there are characters for
3643 which isalpha() is true but neither isupper() nor islower() are true. In
3644 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3645 and ordfeminine) are like this. This affected the treatment of \w and \W
3646 when they appeared in character classes, but not when they appeared outside
3647 a character class. The bit map for "word" characters is now created
3648 separately from the results of isalnum() instead of just taking it from the
3649 upper, lower, and digit maps. (Plus the underscore character, of course.)
3650
3651 5. The above bug also affected the handling of POSIX character classes such as
3652 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3653 permanent tables. Instead, the bit maps for such a class were previously
3654 created as the appropriate unions of the upper, lower, and digit bitmaps.
3655 Now they are created by subtraction from the [[:word:]] class, which has
3656 its own bitmap.
3657
3658 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3659 It is created by subtracting the vertical space characters (\x09, \x0a,
3660 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3661 subtraction was done in the overall bitmap for a character class, meaning
3662 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3663 be recognized. This bug has been fixed.
3664
3665 7. Patches from the folks at Google:
3666
3667 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3668 real life, but is still worth protecting against".
3669
3670 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3671 regular expressions".
3672
3673 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3674 have it.
3675
3676 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3677 "configure" and the latter not, in order to fix a problem somebody had
3678 with compiling the Arg class on HP-UX.
3679
3680 (e) Improve the error-handling of the C++ wrapper a little bit.
3681
3682 (f) New tests for checking recursion limiting.
3683
3684 8. The pcre_memmove() function, which is used only if the environment does not
3685 have a standard memmove() function (and is therefore rarely compiled),
3686 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3687 returning a result (though PCRE never actually uses the result).
3688
3689 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3690 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3691 returned instead of calling malloc() with an overflowing number that would
3692 most likely cause subsequent chaos.
3693
3694 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3695
3696 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3697 with this option is matched, the nmatch and pmatch options of regexec() are
3698 ignored.
3699
3700 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3701 provided in case anyone wants to the the POSIX interface with UTF-8
3702 strings.
3703
3704 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3705 C++ linking (needed for some HP-UX environments).
3706
3707 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3708 (unused parameter) and in the pcre_printint() function (omitted "default"
3709 switch label when the default is to do nothing).
3710
3711 15. Added some code to make it possible, when PCRE is compiled as a C++
3712 library, to replace subject pointers for pcre_exec() with a smart pointer
3713 class, thus making it possible to process discontinuous strings.
3714
3715 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3716 much the same function. They were added by different people who were trying
3717 to make PCRE easy to compile on non-Unix systems. It has been suggested
3718 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3719 for compiling on Windows systems. I have therefore replaced it with
3720 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3721 defaults to "extern" for C or "extern C" for C++, which works fine on
3722 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3723 SCOPE with something explicit in config.h. In addition:
3724
3725 (a) pcreposix.h still had just "extern" instead of either of these macros;
3726 I have replaced it with PCRE_DATA_SCOPE.
3727
3728 (b) Functions such as _pcre_xclass(), which are internal to the library,
3729 but external in the C sense, all had PCRE_EXPORT in their definitions.
3730 This is apparently wrong for the Windows case, so I have removed it.
3731 (It makes no difference on Unix-like systems.)
3732
3733 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3734 of recursive calls to match(). This is different to MATCH_LIMIT because
3735 that limits the total number of calls to match(), not all of which increase
3736 the depth of recursion. Limiting the recursion depth limits the amount of
3737 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3738 when PCRE is compiled, and changed at run time. A patch from Google adds
3739 this functionality to the C++ interface.
3740
3741 18. Changes to the handling of Unicode character properties:
3742
3743 (a) Updated the table to Unicode 4.1.0.
3744
3745 (b) Recognize characters that are not in the table as "Cn" (undefined).
3746
3747 (c) I revised the way the table is implemented to a much improved format
3748 which includes recognition of ranges. It now supports the ranges that
3749 are defined in UnicodeData.txt, and it also amalgamates other
3750 characters into ranges. This has reduced the number of entries in the
3751 table from around 16,000 to around 3,000, thus reducing its size
3752 considerably. I realized I did not need to use a tree structure after
3753 all - a binary chop search is just as efficient. Having reduced the
3754 number of entries, I extended their size from 6 bytes to 8 bytes to
3755 allow for more data.
3756
3757 (d) Added support for Unicode script names via properties such as \p{Han}.
3758
3759 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3760 matching that character.
3761
3762 20. When matching a repeated Unicode property with a minimum greater than zero,
3763 (for example \pL{2,}), PCRE could look past the end of the subject if it
3764 reached it while seeking the minimum number of characters. This could
3765 happen only if some of the characters were more than one byte long, because
3766 there is a check for at least the minimum number of bytes.
3767
3768 21. Refactored the implementation of \p and \P so as to be more general, to
3769 allow for more different types of property in future. This has changed the
3770 compiled form incompatibly. Anybody with saved compiled patterns that use
3771 \p or \P will have to recompile them.
3772
3773 22. Added "Any" and "L&" to the supported property types.
3774
3775 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3776 but give a compile time error if the value is greater than 0xff.
3777
3778 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3779 accidentally not being installed or uninstalled.
3780
3781 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3782 made were to insert the current release number. This seemed silly, because
3783 it made things harder for people building PCRE on systems that don't run
3784 "configure". I have turned pcre.h into a distributed file, no longer built
3785 by "configure", with the version identification directly included. There is
3786 no longer a pcre.h.in file.
3787
3788 However, this change necessitated a change to the pcre-config script as
3789 well. It is built from pcre-config.in, and one of the substitutions was the
3790 release number. I have updated configure.ac so that ./configure now finds
3791 the release number by grepping pcre.h.
3792
3793 26. Added the ability to run the tests under valgrind.
3794
3795
3796 Version 6.4 05-Sep-05
3797 ---------------------
3798
3799 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3800 "--" to be printed when multiple files were scanned, even when none of the
3801 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3802 consider it to be a bug, and have restored the previous behaviour.
3803
3804 2. A couple of code tidies to get rid of compiler warnings.
3805
3806 3. The pcretest program used to cheat by referring to symbols in the library
3807 whose names begin with _pcre_. These are internal symbols that are not
3808 really supposed to be visible externally, and in some environments it is
3809 possible to suppress them. The cheating is now confined to including
3810 certain files from the library's source, which is a bit cleaner.
3811
3812 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3813 file's purpose clearer.
3814
3815 5. Reorganized pcre_ucp_findchar().
3816
3817
3818 Version 6.3 15-Aug-05
3819 ---------------------
3820
3821 1. The file libpcre.pc.in did not have general read permission in the tarball.
3822
3823 2. There were some problems when building without C++ support:
3824
3825 (a) If C++ support was not built, "make install" and "make test" still
3826 tried to test it.
3827
3828 (b) There were problems when the value of CXX was explicitly set. Some
3829 changes have been made to try to fix these, and ...
3830
3831 (c) --disable-cpp can now be used to explicitly disable C++ support.
3832
3833 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3834 backslash in a target when C++ was disabled. This confuses some
3835 versions of "make", apparently. Using an intermediate variable solves
3836 this. (Same for CPP_LOBJ.)
3837
3838 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3839 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3840 necessary on certain architectures.
3841
3842 4. Added a setting of -export-symbols-regex to the link command to remove
3843 those symbols that are exported in the C sense, but actually are local
3844 within the library, and not documented. Their names all begin with
3845 "_pcre_". This is not a perfect job, because (a) we have to except some
3846 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3847 available (and never for static libraries). I have made a note to try to
3848 find a way round (a) in the future.
3849
3850
3851 Version 6.2 01-Aug-05
3852 ---------------------
3853
3854 1. There was no test for integer overflow of quantifier values. A construction
3855 such as {1111111111111111} would give undefined results. What is worse, if
3856 a minimum quantifier for a parenthesized subpattern overflowed and became
3857 negative, the calculation of the memory size went wrong. This could have
3858 led to memory overwriting.
3859
3860 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3861
3862 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3863 operating environments where this matters.
3864
3865 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3866 PCRE options from within the C++ wrapper.
3867
3868 5. Named capturing subpatterns were not being correctly counted when a pattern
3869 was compiled. This caused two problems: (a) If there were more than 100
3870 such subpatterns, the calculation of the memory needed for the whole
3871 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3872 back references of the form \12, where the number was greater than 9, were
3873 not recognized as back references, even though there were sufficient
3874 previous subpatterns.
3875
3876 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3877 versions of gcc, e.g. 2.95.4.
3878
3879
3880 Version 6.1 21-Jun-05
3881 ---------------------
3882
3883 1. There was one reference to the variable "posix" in pcretest.c that was not
3884 surrounded by "#if !defined NOPOSIX".
3885
3886 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3887 the cross-check on the old pcre_info() function, for the benefit of the
3888 cut-down version of PCRE that is currently imported into Exim.
3889
3890 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3891 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3892 patterns that start (?i) but I don't think that matters. The use of (?i) is
3893 just an example; this all applies to the other options as well.
3894
3895 4. Since libtool seems to echo the compile commands it is issuing, the output
3896 from "make" can be reduced a bit by putting "@" in front of each libtool
3897 compile command.
3898
3899 5. Patch from the folks at Google for configure.in to be a bit more thorough
3900 in checking for a suitable C++ installation before trying to compile the
3901 C++ stuff. This should fix a reported problem when a compiler was present,
3902 but no suitable headers.
3903
3904 6. The man pages all had just "PCRE" as their title. I have changed them to
3905 be the relevant file name. I have also arranged that these names are
3906 retained in the file doc/pcre.txt, which is a concatenation in text format
3907 of all the man pages except the little individual ones for each function.
3908
3909 7. The NON-UNIX-USE file had not been updated for the different set of source
3910 files that come with release 6. I also added a few comments about the C++
3911 wrapper.
3912
3913
3914 Version 6.0 07-Jun-05
3915 ---------------------
3916
3917 1. Some minor internal re-organization to help with my DFA experiments.
3918
3919 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3920 didn't matter for the library itself when fully configured, but did matter
3921 when compiling without UCP support, or within Exim, where the ucp files are
3922 not imported.
3923
3924 3. Refactoring of the library code to split up the various functions into
3925 different source modules. The addition of the new DFA matching code (see
3926 below) to a single monolithic source would have made it really too
3927 unwieldy, quite apart from causing all the code to be include in a
3928 statically linked application, when only some functions are used. This is
3929 relevant even without the DFA addition now that patterns can be compiled in
3930 one application and matched in another.
3931
3932 The downside of splitting up is that there have to be some external
3933 functions and data tables that are used internally in different modules of
3934 the library but which are not part of the API. These have all had their
3935 names changed to start with "_pcre_" so that they are unlikely to clash
3936 with other external names.
3937
3938 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3939 a different (DFA) algorithm. Although it is slower than the original
3940 function, it does have some advantages for certain types of matching
3941 problem.
3942
3943 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3944 including restarting after a partial match.
3945
3946 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3947 defined when compiling for Windows was sent to me. I have put it into the
3948 code, though I have no means of testing or verifying it.
3949
3950 7. Added the pcre_refcount() auxiliary function.
3951
3952 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3953 match before or at the first newline in the subject string. In pcretest,
3954 the /f option on a pattern can be used to set this.
3955
3956 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3957 would behave wrongly. This has been present in PCRE since release 4.0.
3958
3959 10. A number of changes to the pcregrep command:
3960
3961 (a) Refactored how -x works; insert ^(...)$ instead of setting
3962 PCRE_ANCHORED and checking the length, in preparation for adding
3963 something similar for -w.
3964
3965 (b) Added the -w (match as a word) option.
3966
3967 (c) Refactored the way lines are read and buffered so as to have more
3968 than one at a time available.
3969
3970 (d) Implemented a pcregrep test script.
3971
3972 (e) Added the -M (multiline match) option. This allows patterns to match
3973 over several lines of the subject. The buffering ensures that at least
3974 8K, or the rest of the document (whichever is the shorter) is available
3975 for matching (and similarly the previous 8K for lookbehind assertions).
3976
3977 (f) Changed the --help output so that it now says
3978
3979 -w, --word-regex(p)
3980
3981 instead of two lines, one with "regex" and the other with "regexp"
3982 because that confused at least one person since the short forms are the
3983 same. (This required a bit of code, as the output is generated
3984 automatically from a table. It wasn't just a text change.)
3985
3986 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3987 option but starts with a hyphen. Could be a pattern or a path name
3988 starting with a hyphen, for instance.
3989
3990 (h) "-" can be given as a file name to represent stdin.
3991
3992 (i) When file names are being printed, "(standard input)" is used for
3993 the standard input, for compatibility with GNU grep. Previously
3994 "<stdin>" was used.
3995
3996 (j) The option --label=xxx can be used to supply a name to be used for
3997 stdin when file names are being printed. There is no short form.
3998
3999 (k) Re-factored the options decoding logic because we are going to add
4000 two more options that take data. Such options can now be given in four
4001 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4002
4003 (l) Added the -A, -B, and -C options for requesting that lines of context
4004 around matches be printed.
4005
4006 (m) Added the -L option to print the names of files that do not contain
4007 any matching lines, that is, the complement of -l.
4008
4009 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4010 continue to scan other files.
4011
4012 (o) The -s option was incorrectly implemented. For compatibility with other
4013 greps, it now suppresses the error message for a non-existent or non-
4014 accessible file (but not the return code). There is a new option called
4015 -q that suppresses the output of matching lines, which was what -s was
4016 previously doing.
4017
4018 (p) Added --include and --exclude options to specify files for inclusion
4019 and exclusion when recursing.
4020
4021 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4022 Hopefully, it now does.
4023
4024 12. Missing cast in pcre_study().
4025
4026 13. Added an "uninstall" target to the makefile.
4027
4028 14. Replaced "extern" in the function prototypes in Makefile.in with
4029 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4030 world, but is set differently for Windows.
4031
4032 15. Added a second compiling function called pcre_compile2(). The only
4033 difference is that it has an extra argument, which is a pointer to an
4034 integer error code. When there is a compile-time failure, this is set
4035 non-zero, in addition to the error test pointer being set to point to an
4036 error message. The new argument may be NULL if no error number is required
4037 (but then you may as well call pcre_compile(), which is now just a
4038 wrapper). This facility is provided because some applications need a
4039 numeric error indication, but it has also enabled me to tidy up the way
4040 compile-time errors are handled in the POSIX wrapper.
4041
4042 16. Added VPATH=.libs to the makefile; this should help when building with one
4043 prefix path and installing with another. (Or so I'm told by someone who
4044 knows more about this stuff than I do.)
4045
4046 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4047 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4048 match everything, including newlines. This is not POSIX-compatible, but
4049 somebody wanted the feature. From pcretest it can be activated by using
4050 both the P and the s flags.
4051
4052 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4053
4054 19. libpcre.pc was being incorrectly installed as executable.
4055
4056 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4057 it now also looks for '\r' so that it will work unmodified on Windows.
4058
4059 21. Added Google's contributed C++ wrapper to the distribution.
4060
4061 22. Added some untidy missing memory free() calls in pcretest, to keep
4062 Electric Fence happy when testing.
4063
4064
4065
4066 Version 5.0 13-Sep-04
4067 ---------------------
4068
4069 1. Internal change: literal characters are no longer packed up into items
4070 containing multiple characters in a single byte-string. Each character
4071 is now matched using a separate opcode. However, there may be more than one
4072 byte in the character in UTF-8 mode.
4073
4074 2. The pcre_callout_block structure has two new fields: pattern_position and
4075 next_item_length. These contain the offset in the pattern to the next match
4076 item, and its length, respectively.
4077
4078 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4079 insertion of callouts before each pattern item. Added the /C option to
4080 pcretest to make use of this.
4081
4082 4. On the advice of a Windows user, the lines
4083
4084 #if defined(_WIN32) || defined(WIN32)
4085 _setmode( _fileno( stdout ), 0x8000 );
4086 #endif /* defined(_WIN32) || defined(WIN32) */
4087
4088 have been added to the source of pcretest. This apparently does useful
4089 magic in relation to line terminators.
4090
4091 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4092 for the benefit of those environments where the "b" makes a difference.
4093
4094 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4095 to know about it. I have put a hack into configure.in that adds in code
4096 to set GCC=yes if CC=icc. This seems to end up at a point in the
4097 generated configure script that is early enough to affect the setting of
4098 compiler options, which is what is needed, but I have no means of testing
4099 whether it really works. (The user who reported this had patched the
4100 generated configure script, which of course I cannot do.)
4101
4102 LATER: After change 22 below (new libtool files), the configure script
4103 seems to know about icc (and also ecc). Therefore, I have commented out
4104 this hack in configure.in.
4105
4106 7. Added support for pkg-config (2 patches were sent in).
4107
4108 8. Negated POSIX character classes that used a combination of internal tables
4109 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4110 [[:^ascii]]. Typically, they would match almost any characters. The other
4111 POSIX classes were not broken in this way.
4112
4113 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4114 to find the match, as PCRE was deluded into thinking that the match had to
4115 start at the start point or following a newline. The same bug applied to
4116 patterns with negative forward assertions or any backward assertions
4117 preceding ".*" at the start, unless the pattern required a fixed first
4118 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4119
4120 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4121 starting at the last subject character, bytes beyond the end of the subject
4122 string were read.
4123
4124 11. Renamed the variable "class" as "classbits" to make life easier for C++
4125 users. (Previously there was a macro definition, but it apparently wasn't
4126 enough.)
4127
4128 12. Added the new field "tables" to the extra data so that tables can be passed
4129 in at exec time, or the internal tables can be re-selected. This allows
4130 a compiled regex to be saved and re-used at a later time by a different
4131 program that might have everything at different addresses.
4132
4133 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4134 -R library as well as a -L library.
4135
4136 14. The debugging options of pcretest (-d on the command line or D on a
4137 pattern) showed incorrect output for anything following an extended class
4138 that contained multibyte characters and which was followed by a quantifier.
4139
4140 15. Added optional support for general category Unicode character properties
4141 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4142 support. It adds about 90K to the size of the library. The meanings of the
4143 inbuilt class escapes such as \d and \s have NOT been changed.
4144
4145 16. Updated pcredemo.c to include calls to free() to release the memory for the
4146 compiled pattern.
4147
4148 17. The generated file chartables.c was being created in the source directory
4149 instead of in the building directory. This caused the build to fail if the
4150 source directory was different from the building directory, and was
4151 read-only.
4152
4153 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4154 file. No doubt somebody will tell me if they don't make sense... Also added
4155 Dan Mooney's comments about building on OpenVMS.
4156
4157 19. Added support for partial matching via the PCRE_PARTIAL option for
4158 pcre_exec() and the \P data escape in pcretest.
4159
4160 20. Extended pcretest with 3 new pattern features:
4161
4162 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4163 write the compiled pattern to the file whose name is "rest-of-line".
4164 This is a straight binary dump of the data, with the saved pointer to
4165 the character tables forced to be NULL. The study data, if any, is
4166 written too. After writing, pcretest reads a new pattern.
4167
4168 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4169 compiled pattern from the given file. There must not be any
4170 occurrences of "<" in the file name (pretty unlikely); if there are,
4171 pcretest will instead treat the initial "<" as a pattern delimiter.
4172 After reading in the pattern, pcretest goes on to read data lines as
4173 usual.
4174
4175 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4176 and 16-bit fields in a compiled pattern, to simulate a pattern that
4177 was compiled on a host of opposite endianness.
4178
4179 21. The pcre-exec() function can now cope with patterns that were compiled on
4180 hosts of opposite endianness, with this restriction:
4181
4182 As for any compiled expression that is saved and used later, the tables
4183 pointer field cannot be preserved; the extra_data field in the arguments
4184 to pcre_exec() should be used to pass in a tables address if a value
4185 other than the default internal tables were used at compile time.
4186
4187 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4188 now diagnosed as an error. Previously, most of the time, a negative number
4189 would have been treated as zero, but if in addition "ovector" was passed as
4190 NULL, a crash could occur.
4191
4192 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4193 new versions from the libtool 1.5 distribution (the last one is a copy of
4194 a file called libtool.m4). This seems to have fixed the need to patch
4195 "configure" to support Darwin 1.3 (which I used to do). However, I still
4196 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4197 workstation).
4198
4199 24. Changed the PCRE licence to be the more standard "BSD" licence.
4200
4201
4202 Version 4.5 01-Dec-03
4203 ---------------------
4204
4205 1. There has been some re-arrangement of the code for the match() function so
4206 that it can be compiled in a version that does not call itself recursively.
4207 Instead, it keeps those local variables that need separate instances for
4208 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4209 needs to "recurse". Keeping track of where control must go is done by means
4210 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4211 hide most of the details from the main code, and operates only if
4212 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4213 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4214 operating.
4215
4216 To make it easier for callers to provide specially tailored get/free
4217 functions for this usage, two new functions, pcre_stack_malloc, and
4218 pcre_stack_free, are used. They are always called in strict stacking order,
4219 and the size of block requested is always the same.
4220
4221 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4222 PCRE has been compiled to use the stack or the heap for recursion. The
4223 -C option of pcretest uses this to show which version is compiled.
4224
4225 A new data escape \S, is added to pcretest; it causes the amounts of store
4226 obtained and freed by both kinds of malloc/free at match time to be added
4227 to the output.
4228
4229 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4230 what's available on my current Linux desktop machine.
4231
4232 3. When matching a UTF-8 string, the test for a valid string at the start has
4233 been extended. If start_offset is not zero, PCRE now checks that it points
4234 to a byte that is the start of a UTF-8 character. If not, it returns
4235 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4236 this is necessary because there may be backward assertions in the pattern.
4237 When matching the same subject several times, it may save resources to use
4238 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4239
4240 4. The code for checking the validity of UTF-8 strings has been tightened so
4241 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4242 containing "overlong sequences".
4243
4244 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4245 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4246 should have been "&", but it just so happened that all the cases this let
4247 through by mistake were picked up later in the function.
4248
4249 6. I had used a variable called "isblank" - this is a C99 function, causing
4250 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4251
4252 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4253 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4254 so the version is shown; (c) stop "make test" echoing "./RunTest".
4255
4256 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4257
4258 9. The prototype for memmove() for systems that don't have it was using
4259 size_t, but the inclusion of the header that defines size_t was later. I've
4260 moved the #includes for the C headers earlier to avoid this.
4261
4262 10. Added some adjustments to the code to make it easier to compiler on certain
4263 special systems:
4264
4265 (a) Some "const" qualifiers were missing.
4266 (b) Added the macro EXPORT before all exported functions; by default this
4267 is defined to be empty.
4268 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4269 that it reads its output file name as an argument instead of writing
4270 to the standard output and assuming this can be redirected.
4271
4272 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4273 class containing characters with values greater than 255, PCRE compilation
4274 went into a loop.
4275
4276 12. A recursive reference to a subpattern that was within another subpattern
4277 that had a minimum quantifier of zero caused PCRE to crash. For example,
4278 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4279 recursion. If the recursively-called subpattern itself had a zero repeat,
4280 that was OK.
4281
4282 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4283 buffer into which it was copied (for escape processing) was still set at
4284 1024, so long lines caused crashes.
4285
4286 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4287 "internal error: code overflow...". This applied to any character class
4288 that was followed by a possessive quantifier.
4289
4290 15. Modified the Makefile to add libpcre.la as a prerequisite for
4291 libpcreposix.la because I was told this is needed for a parallel build to
4292 work.
4293
4294 16. If a pattern that contained .* following optional items at the start was
4295 studied, the wrong optimizing data was generated, leading to matching
4296 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4297 matching string must start with a or b or c. The correct conclusion for
4298 this pattern is that a match can start with any character.
4299
4300
4301 Version 4.4 13-Aug-03
4302 ---------------------
4303
4304 1. In UTF-8 mode, a character class containing characters with values between
4305 127 and 255 was not handled correctly if the compiled pattern was studied.
4306 In fixing this, I have also improved the studying algorithm for such
4307 classes (slightly).
4308
4309 2. Three internal functions had redundant arguments passed to them. Removal
4310 might give a very teeny performance improvement.
4311
4312 3. Documentation bug: the value of the capture_top field in a callout is *one
4313 more than* the number of the hightest numbered captured substring.
4314
4315 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4316 in incorrectly linking with a previously installed version. They now link
4317 explicitly with libpcre.la.
4318
4319 5. configure.in no longer needs to recognize Cygwin specially.
4320
4321 6. A problem in pcre.in for Windows platforms is fixed.
4322
4323 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4324 pcretest, it used to include the size of the study block as part of its
4325 output. Unfortunately, the structure contains a field that has a different
4326 size on different hardware architectures. This meant that the tests that
4327 showed this size failed. As the block is currently always of a fixed size,
4328 this information isn't actually particularly useful in pcretest output, so
4329 I have just removed it.
4330
4331 8. Three pre-processor statements accidentally did not start in column 1.
4332 Sadly, there are *still* compilers around that complain, even though
4333 standard C has not required this for well over a decade. Sigh.
4334
4335 9. In pcretest, the code for checking callouts passed small integers in the
4336 callout_data field, which is a void * field. However, some picky compilers
4337 complained about the casts involved for this on 64-bit systems. Now
4338 pcretest passes the address of the small integer instead, which should get
4339 rid of the warnings.
4340
4341 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4342 both compile and run time, and gives an error if an invalid UTF-8 sequence
4343 is found. There is a option for disabling this check in cases where the
4344 string is known to be correct and/or the maximum performance is wanted.
4345
4346 11. In response to a bug report, I changed one line in Makefile.in from
4347
4348 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4349 to
4350 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4351
4352 to look similar to other lines, but I have no way of telling whether this
4353 is the right thing to do, as I do not use Windows. No doubt I'll get told
4354 if it's wrong...
4355
4356
4357 Version 4.3 21-May-03
4358 ---------------------
4359
4360 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4361 Makefile.
4362
4363 2. Some refactoring to improve the quality of the code:
4364
4365 (i) The utf8_table... variables are now declared "const".
4366
4367 (ii) The code for \cx, which used the "case flipping" table to upper case
4368 lower case letters, now just substracts 32. This is ASCII-specific,
4369 but the whole concept of \cx is ASCII-specific, so it seems
4370 reasonable.
4371
4372 (iii) PCRE was using its character types table to recognize decimal and
4373 hexadecimal digits in the pattern. This is silly, because it handles
4374 only 0-9, a-f, and A-F, but the character types table is locale-
4375 specific, which means strange things might happen. A private
4376 table is now used for this - though it costs 256 bytes, a table is
4377 much faster than multiple explicit tests. Of course, the standard
4378 character types table is still used for matching digits in subject
4379 strings against \d.
4380
4381 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4382 ending in _t are). So I've renamed it as ESC_tee.
4383
4384 3. The first argument for regexec() in the POSIX wrapper should have been
4385 defined as "const".
4386
4387 4. Changed pcretest to use malloc() for its buffers so that they can be
4388 Electric Fenced for debugging.
4389
4390 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4391 to read one or more bytes before the start of the subject string. Often this
4392 had no effect on PCRE's behaviour, but in some circumstances it could
4393 provoke a segmentation fault.
4394
4395 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4396 to try to read one or more bytes before the start of the subject string.
4397
4398 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4399 UTF-8 support could misbehave in various ways if the subject string
4400 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4401 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4402 back over UTF-8 characters.)
4403
4404
4405 Version 4.2 14-Apr-03
4406 ---------------------
4407
4408 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4409
4410 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4411 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4412 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4413 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4414 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4415 and BUILD_EXEEXT
4416 Note: automatic setting of the BUILD variables is not yet working
4417 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4418 compile-time but not at link-time
4419 [LINK]: use for linking executables only
4420 make different versions for Windows and non-Windows
4421 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4422 libraries
4423 [LINK_FOR_BUILD]: new variable
4424 [OBJEXT]: use throughout
4425 [EXEEXT]: use throughout
4426 <winshared>: new target
4427 <wininstall>: new target
4428 <dftables.o>: use native compiler
4429 <dftables>: use native linker
4430 <install>: handle Windows platform correctly
4431 <clean>: ditto
4432 <check>: ditto
4433 copy DLL to top builddir before testing
4434
4435 As part of these changes, -no-undefined was removed again. This was reported
4436 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4437 in any case.
4438
4439 3. Some tidies to get rid of compiler warnings:
4440
4441 . In the match_data structure, match_limit was an unsigned long int, whereas
4442 match_call_count was an int. I've made them both unsigned long ints.
4443
4444 . In pcretest the fact that a const uschar * doesn't automatically cast to
4445 a void * provoked a warning.
4446
4447 . Turning on some more compiler warnings threw up some "shadow" variables
4448 and a few more missing casts.
4449
4450 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4451 option, a class that contained a single character with a value between 128
4452 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4453
4454 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4455 option, a class that contained several characters, but with at least one
4456 whose value was between 128 and 255 caused PCRE to crash.
4457
4458
4459 Version 4.1 12-Mar-03
4460 ---------------------
4461
4462 1. Compiling with gcc -pedantic found a couple of places where casts were
4463 needed, and a string in dftables.c that was longer than standard compilers are
4464 required to support.
4465
4466 2. Compiling with Sun's compiler found a few more places where the code could
4467 be tidied up in order to avoid warnings.
4468
4469 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4470 first of these names is deprecated in the latest Autoconf in favour of the name
4471 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4472 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4473 analogy I have changed it to CFLAGS_FOR_BUILD.
4474
4475 4. Added -no-undefined to the linking command in the Makefile, because this is
4476 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4477 linking step for the pcreposix library.
4478
4479 5. PCRE was failing to diagnose the case of two named groups with the same
4480 name.
4481
4482 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4483 literal character that is needed in the subject for a match, and scans along to
4484 ensure that it is present before embarking on the full matching process. This
4485 saves time in cases of nested unlimited repeats that are never going to match.
4486 Problem: the scan can take a lot of time if the subject is very long (e.g.
4487 megabytes), thus penalizing straightforward matches. It is now done only if the
4488 amount of subject to be scanned is less than 1000 bytes.
4489
4490 7. A lesser problem with the same optimization is that it was recording the
4491 first character of an anchored pattern as "needed", thus provoking a search
4492 right along the subject, even when the first match of the pattern was going to
4493 fail. The "needed" character is now not set for anchored patterns, unless it
4494 follows something in the pattern that is of non-fixed length. Thus, it still
4495 fulfils its original purpose of finding quick non-matches in cases of nested
4496 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4497
4498
4499 Version 4.0 17-Feb-03
4500 ---------------------
4501
4502 1. If a comment in an extended regex that started immediately after a meta-item
4503 extended to the end of string, PCRE compiled incorrect data. This could lead to
4504 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4505
4506 2. Moved to autoconf 2.53 and libtool 1.4.2.
4507
4508 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4509 the special perltest8 script is no longer needed - all the tests can be run
4510 from a single perltest script.
4511
4512 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4513 by \s. It has now been removed in PCRE. This means it isn't recognized as
4514 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4515 class [:space:] *does* include VT, thereby creating a mess.
4516
4517 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4518 space and tab.
4519
4520 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4521 its new features into the main test script, reducing the number of scripts.
4522
4523 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4524 were backward compatible, and made the (?i) apply to the whole pattern, as if
4525 /i were given. Now it behaves more logically, and applies the option setting
4526 only to what follows. PCRE has been changed to follow suit. However, if it
4527 finds options settings right at the start of the pattern, it extracts them into
4528 the global options, as before. Thus, they show up in the info data.
4529
4530 8. Added support for the \Q...\E escape sequence. Characters in between are
4531 treated as literals. This is slightly different from Perl in that $ and @ are
4532 also handled as literals inside the quotes. In Perl, they will cause variable
4533 interpolation. Note the following examples:
4534
4535 Pattern PCRE matches Perl matches
4536
4537 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4538 \Qabc\$xyz\E abc\$xyz abc\$xyz
4539 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4540
4541 For compatibility with Perl, \Q...\E sequences are recognized inside character
4542 classes as well as outside them.
4543
4544 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4545 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4546 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4547 signed/unsigned warnings.
4548
4549 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4550 option for pcretest, so I've replaced it by a simple function that does just
4551 that job.
4552
4553 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4554 "pcregrep -".
4555
4556 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4557 Java package. This provides some syntactic sugar for simple cases of what my
4558 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4559 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4560 item, you can use this simplified notation. Note that only makes sense with
4561 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4562 greediness, whatever the setting of the PCRE_UNGREEDY option.
4563
4564 13. A change of greediness default within a pattern was not taking effect at
4565 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4566 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4567 was abstracted outside.
4568
4569 14. PCRE now supports the \G assertion. It is true when the current matching
4570 position is at the start point of the match. This differs from \A when the
4571 starting offset is non-zero. Used with the /g option of pcretest (or similar
4572 code), it works in the same way as it does for Perl's /g option. If all
4573 alternatives of a regex begin with \G, the expression is anchored to the start
4574 match position, and the "anchored" flag is set in the compiled expression.
4575
4576 15. Some bugs concerning the handling of certain option changes within patterns
4577 have been fixed. These applied to options other than (?ims). For example,
4578 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4579 been the other way round. Some of this was related to change 7 above.
4580
4581 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4582 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4583 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4584 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4585
4586 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4587 mode. Unlike ".", it always matches newline, whatever the setting of
4588 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4589 assertions. Perl allows it, but it doesn't (in general) work because it can't
4590 calculate the length of the lookbehind. At least, that's the case for Perl
4591 5.8.0 - I've been told they are going to document that it doesn't work in
4592 future.
4593
4594 18. Added an error diagnosis for escapes that PCRE does not support: these are
4595 \L, \l, \N, \P, \p, \U, \u, and \X.
4596
4597 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4598 reading past the end of the pattern in cases such as /[abcd/.
4599
4600 20. PCRE was getting more memory than necessary for patterns with classes that
4601 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4602
4603 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4604 compiling PCRE for use with Virtual Pascal.
4605
4606 22. Small fix to the Makefile to make it work properly if the build is done
4607 outside the source tree.
4608
4609 23. Added a new extension: a condition to go with recursion. If a conditional
4610 subpattern starts with (?(R) the "true" branch is used if recursion has
4611 happened, whereas the "false" branch is used only at the top level.
4612
4613 24. When there was a very long string of literal characters (over 255 bytes
4614 without UTF support, over 250 bytes with UTF support), the computation of how
4615 much memory was required could be incorrect, leading to segfaults or other
4616 strange effects.
4617
4618 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4619 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4620 there was a subsequent back reference to those brackets. This meant that, for
4621 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4622 possible to check for precisely this case. All we can do is abandon the
4623 optimization if .* occurs inside capturing brackets when there are any back
4624 references whatsoever. (See below for a better fix that came later.)
4625
4626 26. The handling of the optimization for finding the first character of a
4627 non-anchored pattern, and for finding a character that is required later in the
4628 match were failing in some cases. This didn't break the matching; it just
4629 failed to optimize when it could. The way this is done has been re-implemented.
4630
4631 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4632
4633 28. Added a new feature that provides some of the functionality that Perl
4634 provides with (?{...}). The facility is termed a "callout". The way it is done
4635 in PCRE is for the caller to provide an optional function, by setting
4636 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4637 global variable. By default it is unset, which disables all calling out. To get
4638 the function called, the regex must include (?C) at appropriate points. This
4639 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4640 This provides a means of identifying different callout points. When PCRE
4641 reaches such a point in the regex, if pcre_callout has been set, the external
4642 function is called. It is provided with data in a structure called
4643 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4644 matching continues; if it returns a non-zero value, the match at the current
4645 point fails. However, backtracking will occur if possible. [This was changed
4646 later and other features added - see item 49 below.]
4647
4648 29. pcretest is upgraded to test the callout functionality. It provides a
4649 callout function that displays information. By default, it shows the start of
4650 the match and the current position in the text. There are some new data escapes
4651 to vary what happens:
4652
4653 \C+ in addition, show current contents of captured substrings
4654 \C- do not supply a callout function
4655 \C!n return 1 when callout number n is reached
4656 \C!n!m return 1 when callout number n is reached for the mth time
4657
4658 30. If pcregrep was called with the -l option and just a single file name, it
4659 output "<stdin>" if a match was found, instead of the file name.
4660
4661 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4662 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4663 pcre_exec(). This saves a malloc/free per call. The default value of
4664 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4665 when configuring.
4666
4667 32. The default maximum size of a compiled pattern is 64K. There have been a
4668 few cases of people hitting this limit. The code now uses macros to handle the
4669 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4670 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4671 configuring. Tests 2 and 5 work only with 2-byte links because they output
4672 debugging information about compiled patterns.
4673
4674 33. Internal code re-arrangements:
4675
4676 (a) Moved the debugging function for printing out a compiled regex into
4677 its own source file (printint.c) and used #include to pull it into
4678 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4679 separate copies.
4680
4681 (b) Defined the list of op-code names for debugging as a macro in
4682 internal.h so that it is next to the definition of the opcodes.
4683
4684 (c) Defined a table of op-code lengths for simpler skipping along compiled
4685 code. This is again a macro in internal.h so that it is next to the
4686 definition of the opcodes.
4687
4688 34. Added support for recursive calls to individual subpatterns, along the
4689 lines of Robin Houston's patch (but implemented somewhat differently).
4690
4691 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4692 allow it to read and process whole directories in Win32. This code was
4693 contributed by Lionel Fourquaux; it has not been tested by me.
4694
4695 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4696 used to name a group. Names consist of alphanumerics and underscores, and must
4697 be unique. Back references use the syntax (?P=name) and recursive calls use
4698 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4699 numbers. The function pcre_fullinfo() can be used after compilation to extract
4700 a name/number map. There are three relevant calls:
4701
4702 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4703 PCRE_INFO_NAMECOUNT yields the number of entries
4704 PCRE_INFO_NAMETABLE yields a pointer to the map.
4705
4706 The map is a vector of fixed-size entries. The size of each entry depends on
4707 the length of the longest name used. The first two bytes of each entry are the
4708 group number, most significant byte first. There follows the corresponding
4709 name, zero terminated. The names are in alphabetical order.
4710
4711 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4712 case instead of 255. Making it the same both with and without UTF-8 support
4713 means that the same test output works with both.
4714
4715 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4716 calling malloc() with a zero argument.
4717
4718 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4719 optimization. I've improved things by keeping a bitmap of backreferences with
4720 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4721 fact referenced, the optimization can be applied. It is unlikely that a
4722 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4723 the match to follow \n) will appear inside brackets with a number greater than
4724 31, but if it does, any back reference > 31 suppresses the optimization.
4725
4726 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4727 of disabling numbered capturing parentheses. Any opening parenthesis that is
4728 not followed by ? behaves as if it were followed by ?: but named parentheses
4729 can still be used for capturing (and they will acquire numbers in the usual
4730 way).
4731
4732 41. Redesigned the return codes from the match() function into yes/no/error so
4733 that errors can be passed back from deep inside the nested calls. A malloc
4734 failure while inside a recursive subpattern call now causes the
4735 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4736
4737 42. It is now possible to set a limit on the number of times the match()
4738 function is called in a call to pcre_exec(). This facility makes it possible to
4739 limit the amount of recursion and backtracking, though not in a directly
4740 obvious way, because the match() function is used in a number of different
4741 circumstances. The count starts from zero for each position in the subject
4742 string (for non-anchored patterns). The default limit is, for compatibility, a
4743 large number, namely 10 000 000. You can change this in two ways:
4744
4745 (a) When configuring PCRE before making, you can use --with-match-limit=n
4746 to set a default value for the compiled library.
4747
4748 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4749 a different value is set. See 45 below.
4750
4751 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4752
4753 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4754 of things that can be changed at compile time. The first argument specifies
4755 what is wanted and the second points to where the information is to be placed.
4756 The current list of available information is:
4757
4758 PCRE_CONFIG_UTF8
4759
4760 The output is an integer that is set to one if UTF-8 support is available;
4761 otherwise it is set to zero.
4762
4763 PCRE_CONFIG_NEWLINE
4764
4765 The output is an integer that it set to the value of the code that is used for
4766 newline. It is either LF (10) or CR (13).
4767
4768 PCRE_CONFIG_LINK_SIZE
4769
4770 The output is an integer that contains the number of bytes used for internal
4771 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4772
4773 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4774
4775 The output is an integer that contains the threshold above which the POSIX
4776 interface uses malloc() for output vectors. See item 31 above.
4777
4778 PCRE_CONFIG_MATCH_LIMIT
4779
4780 The output is an unsigned integer that contains the default limit of the number
4781 of match() calls in a pcre_exec() execution. See 42 above.
4782
4783 44. pcretest has been upgraded by the addition of the -C option. This causes it
4784 to extract all the available output from the new pcre_config() function, and to
4785 output it. The program then exits immediately.
4786
4787 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4788 order to support additional features. One way would have been to define
4789 pcre_exec2() (for example) with extra arguments, but this would not have been
4790 extensible, and would also have required all calls to the original function to
4791 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4792 is used for passing in "extra" data from pcre_study().
4793
4794 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4795 contains the following fields:
4796
4797 flags a bitmap indicating which of the following fields are set
4798 study_data opaque data from pcre_study()
4799 match_limit a way of specifying a limit on match() calls for a specific
4800 call to pcre_exec()
4801 callout_data data for callouts (see 49 below)
4802
4803 The flag bits are also defined in pcre.h, and are
4804
4805 PCRE_EXTRA_STUDY_DATA
4806 PCRE_EXTRA_MATCH_LIMIT
4807 PCRE_EXTRA_CALLOUT_DATA
4808
4809 The pcre_study() function now returns one of these new pcre_extra blocks, with
4810 the actual study data pointed to by the study_data field, and the
4811 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4812 before. That is, this change is entirely upwards-compatible and requires no
4813 change to existing code.
4814
4815 If you want to pass in additional data to pcre_exec(), you can either place it
4816 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4817 block.
4818
4819 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4820 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4821 times with different match limits, until it finds the minimum value needed for
4822 pcre_exec() to complete. The value is then output. This can be instructive; for
4823 most simple matches the number is quite small, but for pathological cases it
4824 gets very large very quickly.
4825
4826 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4827 returns the size of the data block pointed to by the study_data field in a
4828 pcre_extra block, that is, the value that was passed as the argument to
4829 pcre_malloc() when PCRE was getting memory in which to place the information
4830 created by pcre_study(). The fourth argument should point to a size_t variable.
4831 pcretest has been extended so that this information is shown after a successful
4832 pcre_study() call when information about the compiled regex is being displayed.
4833
4834 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4835 because what follows is always an absolute path. (Later: it turns out that this
4836 is more than cosmetic for MinGW, because it doesn't like empty path
4837 components.)
4838
4839 49. Some changes have been made to the callout feature (see 28 above):
4840
4841 (i) A callout function now has three choices for what it returns:
4842
4843 0 => success, carry on matching
4844 > 0 => failure at this point, but backtrack if possible
4845 < 0 => serious error, return this value from pcre_exec()
4846
4847 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4848 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4849 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4850 use by callout functions. It will never be used by PCRE itself.
4851
4852 (ii) The pcre_extra structure (see 45 above) has a void * field called
4853 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4854 pcre_callout_block structure has a field of the same name. The contents of
4855 the field passed in the pcre_extra structure are passed to the callout
4856 function in the corresponding field in the callout block. This makes it
4857 easier to use the same callout-containing regex from multiple threads. For
4858 testing, the pcretest program has a new data escape
4859
4860 \C*n pass the number n (may be negative) as callout_data
4861
4862 If the callout function in pcretest receives a non-zero value as
4863 callout_data, it returns that value.
4864
4865 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4866 there were some redundant $(CFLAGS) in commands that are now specified as
4867 $(LINK), which already includes $(CFLAGS).
4868
4869 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4870 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4871 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4872 one-byte characters throughout. Note that case-insensitive matching applies
4873 only to characters whose values are less than 256. PCRE doesn't support the
4874 notion of cases for higher-valued characters.
4875
4876 (i) A character class whose characters are all within 0-255 is handled as
4877 a bit map, and the map is inverted for negative classes. Previously, a
4878 character > 255 always failed to match such a class; however it should
4879 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4880
4881 (ii) A negated character class with a single character < 255 is coded as
4882 "not this character" (OP_NOT). This wasn't working properly when the test
4883 character was multibyte, either singly or repeated.
4884
4885 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4886 mode, for example: \x{100}{2,3}.
4887
4888 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4889 singly or repeated) now correctly test multibyte characters. However,
4890 PCRE doesn't recognize any characters with values greater than 255 as
4891 digits, spaces, or word characters. Such characters always match \D, \S,
4892 and \W, and never match \d, \s, or \w.
4893
4894 (v) Classes may now contain characters and character ranges with values
4895 greater than 255. For example: [ab\x{100}-\x{400}].
4896
4897 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4898 PCRE in UTF-8 mode.
4899
4900 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4901 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4902 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4903 value.)
4904
4905 53. The single man page has become too large. I have therefore split it up into
4906 a number of separate man pages. These also give rise to individual HTML pages;
4907 these are now put in a separate directory, and there is an index.html page that
4908 lists them all. Some hyperlinking between the pages has been installed.
4909
4910 54. Added convenience functions for handling named capturing parentheses.
4911
4912 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4913 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4914 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4915 are faulted.
4916
4917 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4918 calling configure. These values are used when compiling the dftables.c program
4919 which is run to generate the source of the default character tables. They
4920 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4921 you will need to set these values.
4922
4923 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4924
4925
4926 Version 3.9 02-Jan-02
4927 ---------------------
4928
4929 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4930
4931 2. If --disable-static was given, the building process failed when trying to
4932 build pcretest and pcregrep. (For some reason it was using libtool to compile
4933 them, which is not right, as they aren't part of the library.)
4934
4935
4936 Version 3.8 18-Dec-01
4937 ---------------------
4938
4939 1. The experimental UTF-8 code was completely screwed up. It was packing the
4940 bytes in the wrong order. How dumb can you get?
4941
4942
4943 Version 3.7 29-Oct-01
4944 ---------------------
4945
4946 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4947 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4948 this didn't happen under Solaris 8, where I normally test things.
4949
4950 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4951 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4952 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4953 this fix an hour or so after the initial 3.7 release.)
4954
4955
4956 Version 3.6 23-Oct-01
4957 ---------------------
4958
4959 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4960 offsets passed as NULL with zero offset count.
4961
4962 2. The config.guess and config.sub files had not been updated when I moved to
4963 the latest autoconf.
4964
4965
4966 Version 3.5 15-Aug-01
4967 ---------------------
4968
4969 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4970 had been forgotten.
4971
4972 2. By using declared but undefined structures, we can avoid using "void"
4973 definitions in pcre.h while keeping the internal definitions of the structures
4974 private.
4975
4976 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4977 user point of view, this means that both static and shared libraries are built
4978 by default, but this can be individually controlled. More of the work of
4979 handling this static/shared cases is now inside libtool instead of PCRE's make
4980 file.
4981
4982 4. The pcretest utility is now installed along with pcregrep because it is
4983 useful for users (to test regexs) and by doing this, it automatically gets
4984 relinked by libtool. The documentation has been turned into a man page, so
4985 there are now .1, .txt, and .html versions in /doc.
4986
4987 5. Upgrades to pcregrep:
4988 (i) Added long-form option names like gnu grep.
4989 (ii) Added --help to list all options with an explanatory phrase.
4990 (iii) Added -r, --recursive to recurse into sub-directories.
4991 (iv) Added -f, --file to read patterns from a file.
4992
4993 6. pcre_exec() was referring to its "code" argument before testing that
4994 argument for NULL (and giving an error if it was NULL).
4995
4996 7. Upgraded Makefile.in to allow for compiling in a different directory from
4997 the source directory.
4998
4999 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5000 options bits, the pointer it was passed was to an int instead of to an unsigned
5001 long int. This mattered only on 64-bit systems.
5002
5003 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5004 generated) instead of pcre.in, which it its source. Also made the same change
5005 in several of the .c files.
5006
5007 10. A new release of gcc defines printf() as a macro, which broke pcretest
5008 because it had an ifdef in the middle of a string argument for printf(). Fixed
5009 by using separate calls to printf().
5010
5011 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5012 script, to force use of CR or LF instead of \n in the source. On non-Unix
5013 systems, the value can be set in config.h.
5014
5015 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5016 absolute limit. Changed the text of the error message to make this clear, and
5017 likewise updated the man page.
5018
5019 13. The limit of 99 on the number of capturing subpatterns has been removed.
5020 The new limit is 65535, which I hope will not be a "real" limit.
5021
5022
5023 Version 3.4 22-Aug-00
5024 ---------------------
5025
5026 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5027
5028 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5029
5030
5031 Version 3.3 01-Aug-00
5032 ---------------------
5033
5034 1. If an octal character was given, but the value was greater than \377, it
5035 was not getting masked to the least significant bits, as documented. This could
5036 lead to crashes in some systems.
5037
5038 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5039 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5040
5041 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5042 These just pass their arguments on to (pcre_free)(), but they are provided
5043 because some uses of PCRE bind it to non-C systems that can call its functions,
5044 but cannot call free() or pcre_free() directly.
5045
5046 4. Add "make test" as a synonym for "make check". Corrected some comments in
5047 the Makefile.
5048
5049 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5050 Makefile.
5051
5052 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5053 command called pgrep for grepping around the active processes.
5054
5055 7. Added the beginnings of support for UTF-8 character strings.
5056
5057 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5058 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5059 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5060 out for the ar command.)
5061
5062
5063 Version 3.2 12-May-00
5064 ---------------------
5065
5066 This is purely a bug fixing release.
5067
5068 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5069 of ZA. This was just one example of several cases that could provoke this bug,
5070 which was introduced by change 9 of version 2.00. The code for breaking
5071 infinite loops after an iteration that matches an empty string was't working
5072 correctly.
5073
5074 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5075 when matched against abbab (for example). After matching an empty string, it
5076 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5077 caused it to match further down the string than it should.
5078
5079 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5080 was there because it doesn't seem to be needed, and it causes trouble on some
5081 systems, as it is not a Standard C header. It has been removed.
5082
5083 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5084 were reported on the Macintosh. The changes were from
5085
5086 while ((c = *(++ptr)) != 0 && c != '\n');
5087 to
5088 while ((c = *(++ptr)) != 0 && c != '\n') ;
5089
5090 Totally extraordinary, but if that's what it takes...
5091
5092 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5093 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5094 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5095 assumes the way PCRE uses memmove() (always moving upwards).
5096
5097 6. PCRE is being used in one environment where strchr() is not available. There
5098 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5099 faster code anyway.
5100
5101
5102 Version 3.1 09-Feb-00
5103 ---------------------
5104
5105 The only change in this release is the fixing of some bugs in Makefile.in for
5106 the "install" target:
5107
5108 (1) It was failing to install pcreposix.h.
5109
5110 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5111
5112
5113 Version 3.0 01-Feb-00
5114 ---------------------
5115
5116 1. Add support for the /+ modifier to perltest (to output $` like it does in
5117 pcretest).
5118
5119 2. Add support for the /g modifier to perltest.
5120
5121 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5122 matches null strings.
5123
5124 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5125 pattern. Perl treats empty patterns specially - it reuses the most recent
5126 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5127 effect.
5128
5129 5. The POSIX interface was broken in that it was just handing over the POSIX
5130 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5131 required a bigger vector, with some working space on the end. This means that
5132 the POSIX wrapper now has to get and free some memory, and copy the results.
5133
5134 6. Added some simple autoconf support, placing the test data and the
5135 documentation in separate directories, re-organizing some of the
5136 information files, and making it build pcre-config (a GNU standard). Also added
5137 libtool support for building PCRE as a shared library, which is now the
5138 default.
5139
5140 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5141 09 are not valid octal constants. Single digits will be used for minor values
5142 less than 10.
5143
5144 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5145 existing programs that set these in the POSIX interface can use PCRE without
5146 modification.
5147
5148 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5149 return all that pcre_info() returns, plus additional data. The pcre_info()
5150 function is retained for compatibility, but is considered to be obsolete.
5151
5152 10. Added experimental recursion feature (?R) to handle one common case that
5153 Perl 5.6 will be able to do with (?p{...}).
5154
5155 11. Added support for POSIX character classes like [:alpha:], which Perl is
5156 adopting.
5157
5158
5159 Version 2.08 31-Aug-99
5160 ----------------------
5161
5162 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5163 trying to match at the startoffset position, but instead was moving forward to
5164 the next newline as if a previous match had failed.
5165
5166 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5167 and could get into a loop if a null string was matched other than at the start
5168 of the subject.
5169
5170 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5171 be distinguished at compile time, and for completeness also added PCRE_DATE.
5172
5173 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5174 in GnuWin32 environments.
5175
5176
5177 Version 2.07 29-Jul-99
5178 ----------------------
5179
5180 1. The documentation is now supplied in plain text form and HTML as well as in
5181 the form of man page sources.
5182
5183 2. C++ compilers don't like assigning (void *) values to other pointer types.
5184 In particular this affects malloc(). Although there is no problem in Standard
5185 C, I've put in casts to keep C++ compilers happy.
5186
5187 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5188 should be (const char *).
5189
5190 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5191 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5192 However, I haven't made this a standard facility. The documentation doesn't
5193 mention it, and the Makefile doesn't support it.
5194
5195 5. The Makefile now contains an "install" target, with editable destinations at
5196 the top of the file. The pcretest program is not installed.
5197
5198 6. pgrep -V now gives the PCRE version number and date.
5199
5200 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5201 causing the entire string to be ignored, instead of just the last character.
5202
5203 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5204 non-matching string, it can take a very, very long time, even for strings of
5205 quite modest length, because of the nested recursion. PCRE now does better in
5206 some of these cases. It does this by remembering the last required literal
5207 character in the pattern, and pre-searching the subject to ensure it is present
5208 before running the real match. In other words, it applies a heuristic to detect
5209 some types of certain failure quickly, and in the above example, if presented
5210 with a string that has no trailing " it gives "no match" very quickly.
5211
5212 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5213 other alternatives are tried instead.
5214
5215
5216 Version 2.06 09-Jun-99
5217 ----------------------
5218
5219 1. Change pcretest's output for amount of store used to show just the code
5220 space, because the remainder (the data block) varies in size between 32-bit and
5221 64-bit systems.
5222
5223 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5224 start matching at. This allows lookbehinds to work when searching for multiple
5225 occurrences in a string.
5226
5227 3. Added additional options to pcretest for testing multiple occurrences:
5228
5229 /+ outputs the rest of the string that follows a match
5230 /g loops for multiple occurrences, using the new startoffset argument
5231 /G loops for multiple occurrences by passing an incremented pointer
5232
5233 4. PCRE wasn't doing the "first character" optimization for patterns starting
5234 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5235 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5236 the letter 'x'. On long subject strings, this gives a significant speed-up.
5237
5238
5239 Version 2.05 21-Apr-99
5240 ----------------------
5241
5242 1. Changed the type of magic_number from int to long int so that it works
5243 properly on 16-bit systems.
5244
5245 2. Fixed a bug which caused patterns starting with .* not to work correctly
5246 when the subject string contained newline characters. PCRE was assuming
5247 anchoring for such patterns in all cases, which is not correct because .* will
5248 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5249 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5250 must be retried after every newline in the subject.
5251
5252
5253 Version 2.04 18-Feb-99
5254 ----------------------
5255
5256 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5257 computation of the store needed to hold the pattern was incorrect (too large).
5258 If such patterns were nested a few deep, this could multiply and become a real
5259 problem.
5260
5261 2. Added /M option to pcretest to show the memory requirement of a specific
5262 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5263
5264 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5265 compiled in such a way that the backtracking after subsequent failure was
5266 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5267 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5268
5269
5270 Version 2.03 02-Feb-99
5271 ----------------------
5272
5273 1. Fixed typo and small mistake in man page.
5274
5275 2. Added 4th condition (GPL supersedes if conflict) and created separate
5276 LICENCE file containing the conditions.
5277
5278 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5279 Perl, that is the internal \ allows the delimiter to be included in the
5280 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5281 the final delimiter, add \ to the end of the pattern (to test the error).
5282
5283 4. Added the convenience functions for extracting substrings after a successful
5284 match. Updated pcretest to make it able to test these functions.
5285
5286
5287 Version 2.02 14-Jan-99
5288 ----------------------
5289
5290 1. Initialized the working variables associated with each extraction so that
5291 their saving and restoring doesn't refer to uninitialized store.
5292
5293 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5294 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5295 fix the problem.
5296
5297 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5298 calls, and wasn't printing the correct value for compiling calls. Increased the
5299 default value of LOOPREPEAT, and the number of significant figures in the
5300 times.
5301
5302 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5303
5304 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5305 a building problem on Windows NT with a FAT file system.
5306
5307
5308 Version 2.01 21-Oct-98
5309 ----------------------
5310
5311 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5312 to character tables built by pcre_maketables() in the current locale. If NULL
5313 is passed, the default tables are used.
5314
5315
5316 Version 2.00 24-Sep-98
5317 ----------------------
5318
5319 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5320 it any more.
5321
5322 2. Allow quantification of (?>) groups, and make it work correctly.
5323
5324 3. The first character computation wasn't working for (?>) groups.
5325
5326 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5327 end of the subject) and add 5.005's \z, which really does match only at the
5328 very end of the subject.
5329
5330 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5331
5332 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5333 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5334 localized options. All options to pcre_study() were also removed.
5335
5336 7. Add other new features from 5.005:
5337
5338 $(?<= positive lookbehind
5339 $(?<! negative lookbehind
5340 (?imsx-imsx) added the unsetting capability
5341 such a setting is global if at outer level; local otherwise
5342 (?imsx-imsx:) non-capturing groups with option setting
5343 (?(cond)re|re) conditional pattern matching
5344
5345 A backreference to itself in a repeated group matches the previous
5346 captured string.
5347
5348 8. General tidying up of studying (both automatic and via "study")
5349 consequential on the addition of new assertions.
5350
5351 9. As in 5.005, unlimited repeated groups that could match an empty substring
5352 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5353 runtime if any iteration does actually match an empty substring.
5354
5355 10. Include the RunTest script in the distribution.
5356
5357 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5358 discrepancies, some of which were old and were also with respect to 5.004. They
5359 have now been fixed.
5360
5361
5362 Version 1.09 28-Apr-98
5363 ----------------------
5364
5365 1. A negated single character class followed by a quantifier with a minimum
5366 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5367 program crashes, or just wrong answers. This did not apply to negated classes
5368 containing more than one character, or to minima other than one.
5369
5370
5371 Version 1.08 27-Mar-98
5372 ----------------------
5373
5374 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5375
5376 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5377 latter must appear before anything that relies on it in the pattern.
5378
5379
5380 Version 1.07 16-Feb-98
5381 ----------------------
5382
5383 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5384 repeat of a potentially empty string).
5385
5386
5387 Version 1.06 23-Jan-98
5388 ----------------------
5389
5390 1. Added Markus Oberhumer's little patches for C++.
5391
5392 2. Literal strings longer than 255 characters were broken.
5393
5394
5395 Version 1.05 23-Dec-97
5396 ----------------------
5397
5398 1. Negated character classes containing more than one character were failing if
5399 PCRE_CASELESS was set at run time.
5400
5401
5402 Version 1.04 19-Dec-97
5403 ----------------------
5404
5405 1. Corrected the man page, where some "const" qualifiers had been omitted.
5406
5407 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5408 input syntax.
5409
5410 3. Fixed memory leak which occurred when a regex with back references was
5411 matched with an offsets vector that wasn't big enough. The temporary memory
5412 that is used in this case wasn't being freed if the match failed.
5413
5414 4. Tidied pcretest to ensure it frees memory that it gets.
5415
5416 5. Temporary memory was being obtained in the case where the passed offsets
5417 vector was exactly big enough.
5418
5419 6. Corrected definition of offsetof() from change 5 below.
5420
5421 7. I had screwed up change 6 below and broken the rules for the use of
5422 setjmp(). Now fixed.
5423
5424
5425 Version 1.03 18-Dec-97
5426 ----------------------
5427
5428 1. A erroneous regex with a missing opening parenthesis was correctly
5429 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5430 on some systems.
5431
5432 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5433 it was reported that one broken compiler failed on the former because "code" is
5434 also an independent variable.
5435
5436 3. The erroneous regex a[]b caused an array overrun reference.
5437
5438 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5439 fail on data ending with that character. (It was going on too far, and checking
5440 the next character, typically a binary zero.) This was specific to the
5441 optimized code for single-character negative classes.
5442
5443 5. Added a contributed patch from the TIN world which does the following:
5444
5445 + Add an undef for memmove, in case the the system defines a macro for it.
5446
5447 + Add a definition of offsetof(), in case there isn't one. (I don't know
5448 the reason behind this - offsetof() is part of the ANSI standard - but
5449 it does no harm).
5450
5451 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5452 most of the places where whitespace preceded '#'. I have given up and
5453 allowed the remaining 2 cases to be at the margin.
5454
5455 + Rename some variables in pcre to eliminate shadowing. This seems very
5456 pedantic, but does no harm, of course.
5457
5458 6. Moved the call to setjmp() into its own function, to get rid of warnings
5459 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5460
5461 7. Constructs such as \d{8,} were compiling into the equivalent of
5462 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5463 outcome, but in this particular case used more store than had been allocated,
5464 which caused the bug to be discovered because it threw up an internal error.
5465
5466 8. The debugging code in both pcre and pcretest for outputting the compiled
5467 form of a regex was going wrong in the case of back references followed by
5468 curly-bracketed repeats.
5469
5470
5471 Version 1.02 12-Dec-97
5472 ----------------------
5473
5474 1. Typos in pcre.3 and comments in the source fixed.
5475
5476 2. Applied a contributed patch to get rid of places where it used to remove
5477 'const' from variables, and fixed some signed/unsigned and uninitialized
5478 variable warnings.
5479
5480 3. Added the "runtest" target to Makefile.
5481
5482 4. Set default compiler flag to -O2 rather than just -O.
5483
5484
5485 Version 1.01 19-Nov-97
5486 ----------------------
5487
5488 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5489 like /([ab]*)*/, that is, for classes with more than one character in them.
5490
5491 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5492 as /((?>a*))*/ (a PCRE_EXTRA facility).
5493
5494
5495 Version 1.00 18-Nov-97
5496 ----------------------
5497
5498 1. Added compile-time macros to support systems such as SunOS4 which don't have
5499 memmove() or strerror() but have other things that can be used instead.
5500
5501 2. Arranged that "make clean" removes the executables.
5502
5503
5504 Version 0.99 27-Oct-97
5505 ----------------------
5506
5507 1. Fixed bug in code for optimizing classes with only one character. It was
5508 initializing a 32-byte map regardless, which could cause it to run off the end
5509 of the memory it had got.
5510
5511 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5512
5513
5514 Version 0.98 22-Oct-97
5515 ----------------------
5516
5517 1. Fixed bug in code for handling temporary memory usage when there are more
5518 back references than supplied space in the ovector. This could cause segfaults.
5519
5520
5521 Version 0.97 21-Oct-97
5522 ----------------------
5523
5524 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5525
5526 2. Optimized negated single characters not to use a bit map.
5527
5528 3. Brought error texts together as macro definitions; clarified some of them;
5529 fixed one that was wrong - it said "range out of order" when it meant "invalid
5530 escape sequence".
5531
5532 4. Changed some char * arguments to const char *.
5533
5534 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5535
5536 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5537 pcretest.
5538
5539
5540 Version 0.96 16-Oct-97
5541 ----------------------
5542
5543 1. Added a simple "pgrep" utility to the distribution.
5544
5545 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5546 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5547 where "ddd" means "one or more decimal digits".
5548
5549 3. Fixed serious bug. If a pattern had a back reference, but the call to
5550 pcre_exec() didn't supply a large enough ovector to record the related
5551 identifying subpattern, the match always failed. PCRE now remembers the number
5552 of the largest back reference, and gets some temporary memory in which to save
5553 the offsets during matching if necessary, in order to ensure that
5554 backreferences always work.
5555
5556 4. Increased the compatibility with Perl in a number of ways:
5557
5558 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5559 to request this handling. The option can be set at compile or exec time.
5560
5561 (b) $ matches before a terminating newline by default; an option
5562 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5563 mode). The option can be set at compile or exec time.
5564
5565 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5566 the same as Perl's. If the decimal number it represents is less than 10
5567 or there aren't that many previous left capturing parentheses, an octal
5568 escape is read. Inside a character class, it's always an octal escape,
5569 even if it is a single digit.
5570
5571 (d) An escaped but undefined alphabetic character is taken as a literal,
5572 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5573 escapes.
5574
5575 (e) {0} is now permitted. (The previous item is removed from the compiled
5576 pattern).
5577
5578 5. Changed all the names of code files so that the basic parts are no longer
5579 than 10 characters, and abolished the teeny "globals.c" file.
5580
5581 6. Changed the handling of character classes; they are now done with a 32-byte
5582 bit map always.
5583
5584 7. Added the -d and /D options to pcretest to make it possible to look at the
5585 internals of compilation without having to recompile pcre.
5586
5587
5588 Version 0.95 23-Sep-97
5589 ----------------------
5590
5591 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5592 \x20 at the start of a run of normal characters. These were being treated as
5593 real characters, instead of the source characters being re-checked.
5594
5595
5596 Version 0.94 18-Sep-97
5597 ----------------------
5598
5599 1. The functions are now thread-safe, with the caveat that the global variables
5600 containing pointers to malloc() and free() or alternative functions are the
5601 same for all threads.
5602
5603 2. Get pcre_study() to generate a bitmap of initial characters for non-
5604 anchored patterns when this is possible, and use it if passed to pcre_exec().
5605
5606
5607 Version 0.93 15-Sep-97
5608 ----------------------
5609
5610 1. /(b)|(:+)/ was computing an incorrect first character.
5611
5612 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5613 but not actually doing anything yet.
5614
5615 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5616 as Perl does (e.g. [-az] or [az-]).
5617
5618 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5619 all possible positions.
5620
5621 5. Split up into different modules to avoid including unneeded functions in a
5622 compiled binary. However, compile and exec are still in one module. The "study"
5623 function is split off.
5624
5625 6. The character tables are now in a separate module whose source is generated
5626 by an auxiliary program - but can then be edited by hand if required. There are
5627 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5628 toupper() in the code.
5629
5630 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5631 make them global. Abolish the function for setting them, as the caller can now
5632 set them directly.
5633
5634
5635 Version 0.92 11-Sep-97
5636 ----------------------
5637
5638 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5639 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5640
5641 2. Caseless matching was not working in character classes if the characters in
5642 the pattern were in upper case.
5643
5644 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5645
5646 4. Make PCRE_ANCHORED public and accept as a compile option.
5647
5648 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5649 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5650 pass them.
5651
5652 6. Give an error if bad option bits passed at compile or run time.
5653
5654 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5655 pcretest to cause it to pass that flag.
5656
5657 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5658 options, and the first character, if set.
5659
5660 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5661
5662
5663 Version 0.91 10-Sep-97
5664 ----------------------
5665
5666 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5667 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5668
5669 2. PCRE was looping on encountering an indefinitely repeated back reference to
5670 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5671 Perl does - treats the match as successful.
5672
5673 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5