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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1534 - (show annotations)
Tue Mar 24 10:33:21 2015 UTC (4 years, 8 months ago) by ph10
File size: 274313 byte(s)
Error occurred while calculating annotation data.
Fix bugs caused by (?!) as a condition (which is converted to OP_FAIL).
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.37 xx-xxx-2015
5 ------------------------
6
7 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
8 for those parentheses to be closed with whatever has been captured so far.
9 However, it was failing to mark any other groups between the hightest
10 capture so far and the currrent group as "unset". Thus, the ovector for
11 those groups contained whatever was previously there. An example is the
12 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
13
14 2. If an assertion condition was quantified with a minimum of zero (an odd
15 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
16
17 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
18 unrecognized modifier, a crash could occur.
19
20 4. An attempt to do global matching in pcretest with a zero-length ovector
21 caused a crash.
22
23 5. Fixed a memory leak during matching that could occur for a subpattern
24 subroutine call (recursive or otherwise) if the number of captured groups
25 that had to be saved was greater than ten.
26
27 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
28 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
29 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
30
31 7. A UTF pattern containing a "not" match of a non-ASCII character and a
32 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
33
34 8. When a pattern is compiled, it remembers the highest back reference so that
35 when matching, if the ovector is too small, extra memory can be obtained to
36 use instead. A conditional subpattern whose condition is a check on a
37 capture having happened, such as, for example in the pattern
38 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
39 setting the highest backreference number. This mattered only if pcre_exec()
40 was called with an ovector that was too small to hold the capture, and there
41 was no other kind of back reference (a situation which is probably quite
42 rare). The effect of the bug was that the condition was always treated as
43 FALSE when the capture could not be consulted, leading to a incorrect
44 behaviour by pcre2_match(). This bug has been fixed.
45
46 9. A reference to a duplicated named group (either a back reference or a test
47 for being set in a conditional) that occurred in a part of the pattern where
48 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
49 to be incorrectly calculated, leading to overwriting.
50
51 10. A mutually recursive set of back references such as (\2)(\1) caused a
52 segfault at study time (while trying to find the minimum matching length).
53 The infinite loop is now broken (with the minimum length unset, that is,
54 zero).
55
56 11. If an assertion that was used as a condition was quantified with a minimum
57 of zero, matching went wrong. In particular, if the whole group had
58 unlimited repetition and could match an empty string, a segfault was
59 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
60 assertions to be quantified, but not if they are being used as conditions,
61 so the above pattern is faulted by Perl. PCRE has now been changed so that
62 it also rejects such patterns.
63
64 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
65 failed to allow the zero-repeat case if pcre2_exec() was called with an
66 ovector too small to capture the group.
67
68 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
69 Red Hat Product Security:
70
71 (a) A crash if /K and /F were both set with the option to save the compiled
72 pattern.
73
74 (b) Another crash if the option to print captured substrings in a callout
75 was combined with setting a null ovector, for example \O\C+ as a subject
76 string.
77
78 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
79 forward reference repeated a large (but limited) number of times within a
80 repeated outer group that has a zero minimum quantifier, caused incorrect
81 code to be compiled, leading to the error "internal error:
82 previously-checked referenced subpattern not found" when an incorrect
83 memory address was read. This bug was reported as "heap overflow",
84 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
85 CVE-2015-2325.
86
87 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
88 call within a group that also contained a recursive back reference caused
89 incorrect code to be compiled. This bug was reported as "heap overflow",
90 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
91 number CVE-2015-2326.
92
93 24. Computing the size of the JIT read-only data in advance has been a source
94 of various issues, and new ones are still appear unfortunately. To fix
95 existing and future issues, size computation is eliminated from the code,
96 and replaced by on-demand memory allocation.
97
98 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
99 adjacent to the end of the range, and the range contained characters with
100 more than one other case, caused incorrect behaviour when compiled in UTF
101 mode. In that example, the range a-j was left out of the class.
102
103 26. Fix JIT compilation of conditional blocks, which assertion
104 is converted to (*FAIL). E.g: /(?(?!))/.
105
106 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
107 discovered by the LLVM fuzzer.
108
109 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
110 when this assertion was used as a condition, for example (?(?!)a|b). In
111 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
112 error about an unsupported item.
113
114
115 Version 8.36 26-September-2014
116 ------------------------------
117
118 1. Got rid of some compiler warnings in the C++ modules that were shown up by
119 -Wmissing-field-initializers and -Wunused-parameter.
120
121 2. The tests for quantifiers being too big (greater than 65535) were being
122 applied after reading the number, and stupidly assuming that integer
123 overflow would give a negative number. The tests are now applied as the
124 numbers are read.
125
126 3. Tidy code in pcre_exec.c where two branches that used to be different are
127 now the same.
128
129 4. The JIT compiler did not generate match limit checks for certain
130 bracketed expressions with quantifiers. This may lead to exponential
131 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
132 issue should be resolved now.
133
134 5. Fixed an issue, which occures when nested alternatives are optimized
135 with table jumps.
136
137 6. Inserted two casts and changed some ints to size_t in the light of some
138 reported 64-bit compiler warnings (Bugzilla 1477).
139
140 7. Fixed a bug concerned with zero-minimum possessive groups that could match
141 an empty string, which sometimes were behaving incorrectly in the
142 interpreter (though correctly in the JIT matcher). This pcretest input is
143 an example:
144
145 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
146 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
147
148 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
149 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
150 for an empty string was breaking the inner loop and carrying on at a lower
151 level, when possessive repeated groups should always return to a higher
152 level as they have no backtrack points in them. The empty string test now
153 occurs at the outer level.
154
155 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
156 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
157
158 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
159 doesn't).
160
161 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
162 to hold a pattern to be incorrectly computed (too small) when there were
163 named back references to duplicated names. This could cause "internal
164 error: code overflow" or "double free or corruption" or other memory
165 handling errors.
166
167 11. When named subpatterns had the same prefixes, back references could be
168 confused. For example, in this pattern:
169
170 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
171
172 the reference to 'Name' was incorrectly treated as a reference to a
173 duplicate name.
174
175 12. A pattern such as /^s?c/mi8 where the optional character has more than
176 one "other case" was incorrectly compiled such that it would only try to
177 match starting at "c".
178
179 13. When a pattern starting with \s was studied, VT was not included in the
180 list of possible starting characters; this should have been part of the
181 8.34/18 patch.
182
183 14. If a character class started [\Qx]... where x is any character, the class
184 was incorrectly terminated at the ].
185
186 15. If a pattern that started with a caseless match for a character with more
187 than one "other case" was studied, PCRE did not set up the starting code
188 unit bit map for the list of possible characters. Now it does. This is an
189 optimization improvement, not a bug fix.
190
191 16. The Unicode data tables have been updated to Unicode 7.0.0.
192
193 17. Fixed a number of memory leaks in pcregrep.
194
195 18. Avoid a compiler warning (from some compilers) for a function call with
196 a cast that removes "const" from an lvalue by using an intermediate
197 variable (to which the compiler does not object).
198
199 19. Incorrect code was compiled if a group that contained an internal recursive
200 back reference was optional (had quantifier with a minimum of zero). This
201 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
202 caused segmentation faults because of stack overflows at compile time.
203
204 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
205 group that is quantified with an indefinite repeat, caused a compile-time
206 loop which used up all the system stack and provoked a segmentation fault.
207 This was not the same bug as 19 above.
208
209 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
210 Patch by Mike Frysinger.
211
212
213 Version 8.35 04-April-2014
214 --------------------------
215
216 1. A new flag is set, when property checks are present in an XCLASS.
217 When this flag is not set, PCRE can perform certain optimizations
218 such as studying these XCLASS-es.
219
220 2. The auto-possessification of character sets were improved: a normal
221 and an extended character set can be compared now. Furthermore
222 the JIT compiler optimizes more character set checks.
223
224 3. Got rid of some compiler warnings for potentially uninitialized variables
225 that show up only when compiled with -O2.
226
227 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
228 of a match later then the end of the match. The pcretest program was not
229 handling the case sensibly - it was outputting from the start to the next
230 binary zero. It now reports this situation in a message, and outputs the
231 text from the end to the start.
232
233 5. Fast forward search is improved in JIT. Instead of the first three
234 characters, any three characters with fixed position can be searched.
235 Search order: first, last, middle.
236
237 6. Improve character range checks in JIT. Characters are read by an inprecise
238 function now, which returns with an unknown value if the character code is
239 above a certain threshold (e.g: 256). The only limitation is that the value
240 must be bigger than the threshold as well. This function is useful when
241 the characters above the threshold are handled in the same way.
242
243 7. The macros whose names start with RAWUCHAR are placeholders for a future
244 mode in which only the bottom 21 bits of 32-bit data items are used. To
245 make this more memorable for those maintaining the code, the names have
246 been changed to start with UCHAR21, and an extensive comment has been added
247 to their definition.
248
249 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
250 to the export list in Makefile.am (they were accidentally omitted from the
251 8.34 tarball).
252
253 9. The informational output from pcretest used the phrase "starting byte set"
254 which is inappropriate for the 16-bit and 32-bit libraries. As the output
255 for "first char" and "need char" really means "non-UTF-char", I've changed
256 "byte" to "char", and slightly reworded the output. The documentation about
257 these values has also been (I hope) clarified.
258
259 10. Another JIT related optimization: use table jumps for selecting the correct
260 backtracking path, when more than four alternatives are present inside a
261 bracket.
262
263 11. Empty match is not possible, when the minimum length is greater than zero,
264 and there is no \K in the pattern. JIT should avoid empty match checks in
265 such cases.
266
267 12. In a caseless character class with UCP support, when a character with more
268 than one alternative case was not the first character of a range, not all
269 the alternative cases were added to the class. For example, s and \x{17f}
270 are both alternative cases for S: the class [RST] was handled correctly,
271 but [R-T] was not.
272
273 13. The configure.ac file always checked for pthread support when JIT was
274 enabled. This is not used in Windows, so I have put this test inside a
275 check for the presence of windows.h (which was already tested for).
276
277 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
278 The algorithm provides a way to skip certain starting offsets, and usually
279 faster than linear prefix searches.
280
281 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
282 as for 'fr_FR' and 'french'. For some reason, however, it then used the
283 Windows-specific input and output files, which have 'french' screwed in.
284 So this could never have worked. One of the problems with locales is that
285 they aren't always the same. I have now updated RunTest so that it checks
286 the output of the locale test (test 3) against three different output
287 files, and it allows the test to pass if any one of them matches. With luck
288 this should make the test pass on some versions of Solaris where it was
289 failing. Because of the uncertainty, the script did not used to stop if
290 test 3 failed; it now does. If further versions of a French locale ever
291 come to light, they can now easily be added.
292
293 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
294 there was a message during ./configure, but it did not stop. This now
295 provokes an error. The invalid example in README has been corrected.
296 If a value less than the minimum is given, the minimum value has always
297 been used, but now a warning is given.
298
299 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
300 was a bug in the test system, which is now fixed. Also, the list of various
301 configurations that are tested for each release did not have one with both
302 16/32 bits and --enable-bar-anycrlf. It now does.
303
304 18. pcretest was missing "-C bsr" for displaying the \R default setting.
305
306 19. Little endian PowerPC systems are supported now by the JIT compiler.
307
308 20. The fast forward newline mechanism could enter to an infinite loop on
309 certain invalid UTF-8 input. Although we don't support these cases
310 this issue can be fixed by a performance optimization.
311
312 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
313 not take account if existing stack usage. There is now a new global
314 variable called pcre_stack_guard that can be set to point to an external
315 function to check stack availability. It is called at the start of
316 processing every parenthesized group.
317
318 22. A typo in the code meant that in ungreedy mode the max/min qualifier
319 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
320 not match "ab".
321
322 23. When UTF was disabled, the JIT program reported some incorrect compile
323 errors. These messages are silenced now.
324
325 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
326 compiler.
327
328 25. Change all the temporary files used in RunGrepTest to be different to those
329 used by RunTest so that the tests can be run simultaneously, for example by
330 "make -j check".
331
332
333 Version 8.34 15-December-2013
334 -----------------------------
335
336 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
337 executable memory. Patch inspired by Carsten Klein.
338
339 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
340 this macro is never tested and has no effect, because the work to support
341 coverage involves only compiling and linking options and special targets in
342 the Makefile. The comment in config.h implied that defining the macro would
343 enable coverage support, which is totally false. There was also support for
344 setting this macro in the CMake files (my fault, I just copied it from
345 configure). SUPPORT_GCOV has now been removed.
346
347 3. Make a small performance improvement in strlen16() and strlen32() in
348 pcretest.
349
350 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
351 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
352 these cases). There was also one in pcretest.c.
353
354 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
355
356 6. In UTF mode, the code for checking whether a group could match an empty
357 string (which is used for indefinitely repeated groups to allow for
358 breaking an infinite loop) was broken when the group contained a repeated
359 negated single-character class with a character that occupied more than one
360 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
361 UTF-8 mode). The effect was undefined: the group might or might not be
362 deemed as matching an empty string, or the program might have crashed.
363
364 7. The code for checking whether a group could match an empty string was not
365 recognizing that \h, \H, \v, \V, and \R must match a character.
366
367 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
368 an empty string. If it can, pcretest shows this in its information output.
369
370 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
371 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
372 matched by pcre_exec() without using JIT:
373
374 (a) If the rest of the pattern did not match after a maximal run of
375 grapheme clusters, the code for backing up to try with fewer of them
376 did not always back up over a full grapheme when characters that do not
377 have the modifier quality were involved, e.g. Hangul syllables.
378
379 (b) If the match point in a subject started with modifier character, and
380 there was no match, the code could incorrectly back up beyond the match
381 point, and potentially beyond the first character in the subject,
382 leading to a segfault or an incorrect match result.
383
384 10. A conditional group with an assertion condition could lead to PCRE
385 recording an incorrect first data item for a match if no other first data
386 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
387 first data item, and therefore matched "ca" after "c" instead of at the
388 start.
389
390 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
391 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
392
393 12. The source of pcregrep now includes z/OS-specific code so that it can be
394 compiled for z/OS as part of the special z/OS distribution.
395
396 13. Added the -T and -TM options to pcretest.
397
398 14. The code in pcre_compile.c for creating the table of named capturing groups
399 has been refactored. Instead of creating the table dynamically during the
400 actual compiling pass, the information is remembered during the pre-compile
401 pass (on the stack unless there are more than 20 named groups, in which
402 case malloc() is used) and the whole table is created before the actual
403 compile happens. This has simplified the code (it is now nearly 150 lines
404 shorter) and prepared the way for better handling of references to groups
405 with duplicate names.
406
407 15. A back reference to a named subpattern when there is more than one of the
408 same name now checks them in the order in which they appear in the pattern.
409 The first one that is set is used for the reference. Previously only the
410 first one was inspected. This change makes PCRE more compatible with Perl.
411
412 16. Unicode character properties were updated from Unicode 6.3.0.
413
414 17. The compile-time code for auto-possessification has been refactored, based
415 on a patch by Zoltan Herczeg. It now happens after instead of during
416 compilation. The code is cleaner, and more cases are handled. The option
417 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
418 options in pcretest are provided to set it. It can also be set by
419 (*NO_AUTO_POSSESS) at the start of a pattern.
420
421 18. The character VT has been added to the default ("C" locale) set of
422 characters that match \s and are generally treated as white space,
423 following this same change in Perl 5.18. There is now no difference between
424 "Perl space" and "POSIX space". Whether VT is treated as white space in
425 other locales depends on the locale.
426
427 19. The code for checking named groups as conditions, either for being set or
428 for being recursed, has been refactored (this is related to 14 and 15
429 above). Processing unduplicated named groups should now be as fast at
430 numerical groups, and processing duplicated groups should be faster than
431 before.
432
433 20. Two patches to the CMake build system, by Alexander Barkov:
434
435 (1) Replace the "source" command by "." in CMakeLists.txt because
436 "source" is a bash-ism.
437
438 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
439 without these the CMake build does not work on Solaris.
440
441 21. Perl has changed its handling of \8 and \9. If there is no previously
442 encountered capturing group of those numbers, they are treated as the
443 literal characters 8 and 9 instead of a binary zero followed by the
444 literals. PCRE now does the same.
445
446 22. Following Perl, added \o{} to specify codepoints in octal, making it
447 possible to specify values greater than 0777 and also making them
448 unambiguous.
449
450 23. Perl now gives an error for missing closing braces after \x{... instead of
451 treating the string as literal. PCRE now does the same.
452
453 24. RunTest used to grumble if an inappropriate test was selected explicitly,
454 but just skip it when running all tests. This make it awkward to run ranges
455 of tests when one of them was inappropriate. Now it just skips any
456 inappropriate tests, as it always did when running all tests.
457
458 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
459 character types such as \d or \w, too many callouts were inserted, and the
460 data that they returned was rubbish.
461
462 26. In UCP mode, \s was not matching two of the characters that Perl matches,
463 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
464 were matched by \h. The code has now been refactored so that the lists of
465 the horizontal and vertical whitespace characters used for \h and \v (which
466 are defined only in one place) are now also used for \s.
467
468 27. Add JIT support for the 64 bit TileGX architecture.
469 Patch by Jiong Wang (Tilera Corporation).
470
471 28. Possessive quantifiers for classes (both explicit and automatically
472 generated) now use special opcodes instead of wrapping in ONCE brackets.
473
474 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
475 (because it's meaningless), this was not happening when PCRE_CASELESS was
476 set. Not wrong, but inefficient.
477
478 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
479 properties for \w, \d, etc) is present in a test regex. Otherwise if the
480 test contains no characters greater than 255, Perl doesn't realise it
481 should be using Unicode semantics.
482
483 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
484 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
485 does in Unicode mode.
486
487 32. Added the "forbid" facility to pcretest so that putting tests into the
488 wrong test files can sometimes be quickly detected.
489
490 33. There is now a limit (default 250) on the depth of nesting of parentheses.
491 This limit is imposed to control the amount of system stack used at compile
492 time. It can be changed at build time by --with-parens-nest-limit=xxx or
493 the equivalent in CMake.
494
495 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
496 errors. Perl warns for these when in warning mode, but PCRE has no facility
497 for giving warnings.
498
499 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
500 However, this was not working for (?!) because it is optimized to (*FAIL),
501 for which PCRE does not allow quantifiers. The optimization is now disabled
502 when a quantifier follows (?!). I can't see any use for this, but it makes
503 things uniform.
504
505 36. Perl no longer allows group names to start with digits, so I have made this
506 change also in PCRE. It simplifies the code a bit.
507
508 37. In extended mode, Perl ignores spaces before a + that indicates a
509 possessive quantifier. PCRE allowed a space before the quantifier, but not
510 before the possessive +. It now does.
511
512 38. The use of \K (reset reported match start) within a repeated possessive
513 group such as (a\Kb)*+ was not working.
514
515 40. Document that the same character tables must be used at compile time and
516 run time, and that the facility to pass tables to pcre_exec() and
517 pcre_dfa_exec() is for use only with saved/restored patterns.
518
519 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
520 features for Builds with MSVC:
521
522 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
523 builds). The .rc files can be used to set FileDescription and many other
524 attributes.
525
526 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
527 This allows higher-level build scripts which want .pdb files to avoid
528 hard-coding the exact files needed."
529
530 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
531 mean "start of word" and "end of word", respectively, as a transition aid.
532
533 43. A minimizing repeat of a class containing codepoints greater than 255 in
534 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
535 compiled to use the heap for recursion.
536
537 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
538 is configured.
539
540
541 Version 8.33 28-May-2013
542 ------------------------
543
544 1. Added 'U' to some constants that are compared to unsigned integers, to
545 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
546 variables that are added to signed variables, to ensure the result is
547 signed and can be negated.
548
549 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
550 CMake config files.
551
552 3. Revise the creation of config.h.generic so that all boolean macros are
553 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
554 overriding via -D on the command line possible.
555
556 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
557 to unsigned int is reported to make a quite noticeable speed difference in
558 a specific Windows environment. Testing on Linux did also appear to show
559 some benefit (and it is clearly not harmful). Also fixed the definition of
560 Xop which should be unsigned.
561
562 5. Related to (4), changing the definition of the intermediate variable cc
563 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
564 improvements.
565
566 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
567 unnecessary spaces.
568
569 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
570 and later.
571
572 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
573
574 9. Optimizing fast_forward_start_bits in JIT.
575
576 10. Adding support for callouts in JIT, and fixing some issues revealed
577 during this work. Namely:
578
579 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
580
581 (b) Minimum length was not checked before the matching is started.
582
583 11. The value of capture_last that is passed to callouts was incorrect in some
584 cases when there was a capture on one path that was subsequently abandoned
585 after a backtrack. Also, the capture_last value is now reset after a
586 recursion, since all captures are also reset in this case.
587
588 12. The interpreter no longer returns the "too many substrings" error in the
589 case when an overflowing capture is in a branch that is subsequently
590 abandoned after a backtrack.
591
592 13. In the pathological case when an offset vector of size 2 is used, pcretest
593 now prints out the matched string after a yield of 0 or 1.
594
595 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
596 Only supported by the JIT compiler at the moment.
597
598 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
599
600 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
601 offset of the starting point of the matching process, provided the offsets
602 vector is large enough.
603
604 17. The \A escape now records a lookbehind value of 1, though its execution
605 does not actually inspect the previous character. This is to ensure that,
606 in partial multi-segment matching, at least one character from the old
607 segment is retained when a new segment is processed. Otherwise, if there
608 are no lookbehinds in the pattern, \A might match incorrectly at the start
609 of a new segment.
610
611 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
612
613 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
614 gives some modest performance improvement in 8-bit mode.
615
616 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
617 be expressed in certain programming languages using Universal Character
618 Names.
619
620 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
621 which points out that "non characters" are not "characters that may not
622 appear in Unicode strings" but rather "characters that are reserved for
623 internal use and have only local meaning".
624
625 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
626 there was a conditional group that depended on an assertion, if the
627 assertion was false, the callout that immediately followed the alternation
628 in the condition was skipped when pcre_exec() was used for matching.
629
630 23. Allow an explicit callout to be inserted before an assertion that is the
631 condition for a conditional group, for compatibility with automatic
632 callouts, which always insert a callout at this point.
633
634 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
635 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
636
637 25. (*PRUNE) is now supported by the JIT compiler.
638
639 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
640
641 27. Fix the case where there are two or more SKIPs with arguments that may be
642 ignored.
643
644 28. (*SKIP) is now supported by the JIT compiler.
645
646 29. (*THEN) is now supported by the JIT compiler.
647
648 30. Update RunTest with additional test selector options.
649
650 31. The way PCRE handles backtracking verbs has been changed in two ways.
651
652 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
653 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
654 backtracking. In some cases this makes it more Perl-compatible, but Perl's
655 rather obscure rules do not always do the same thing.
656
657 (2) Previously, backtracking verbs were confined within assertions. This is
658 no longer the case for positive assertions, except for (*ACCEPT). Again,
659 this sometimes improves Perl compatibility, and sometimes does not.
660
661 32. A number of tests that were in test 2 because Perl did things differently
662 have been moved to test 1, because either Perl or PCRE has changed, and
663 these tests are now compatible.
664
665 32. Backtracking control verbs are now handled in the same way in JIT and
666 interpreter.
667
668 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
669 contained a forward subroutine reference caused a compile error.
670
671 34. Auto-detect and optimize limited repetitions in JIT.
672
673 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
674 blocking (*UTF) etc.
675
676 36. In the interpreter, maximizing pattern repetitions for characters and
677 character types now use tail recursion, which reduces stack usage.
678
679 37. The value of the max lookbehind was not correctly preserved if a compiled
680 and saved regex was reloaded on a host of different endianness.
681
682 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
683 of the compiled pattern block, expand the flags field from 16 to 32 bits
684 because it was almost full.
685
686 39. Try madvise first before posix_madvise.
687
688 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
689 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
690 original change locked out all matches of empty strings. This has been
691 changed so that one match of an empty string per line is recognized.
692 Subsequent searches on the same line (for colouring or for --only-matching,
693 for example) do not recognize empty strings.
694
695 41. Applied a user patch to fix a number of spelling mistakes in comments.
696
697 42. Data lines longer than 65536 caused pcretest to crash.
698
699 43. Clarified the data type for length and startoffset arguments for pcre_exec
700 and pcre_dfa_exec in the function-specific man pages, where they were
701 explicitly stated to be in bytes, never having been updated. I also added
702 some clarification to the pcreapi man page.
703
704 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
705 a segmentation fault.
706
707
708 Version 8.32 30-November-2012
709 -----------------------------
710
711 1. Improved JIT compiler optimizations for first character search and single
712 character iterators.
713
714 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
715 Patch by Daniel Richard G.
716
717 3. Single character iterator optimizations in the JIT compiler.
718
719 4. Improved JIT compiler optimizations for character ranges.
720
721 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
722 Reported by Giuseppe D'Angelo.
723
724 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
725 of a line, was being set incorrectly in cases where .* appeared inside
726 atomic brackets at the start of a pattern, or where there was a subsequent
727 *PRUNE or *SKIP.
728
729 7. Improved instruction cache flush for POWER/PowerPC.
730 Patch by Daniel Richard G.
731
732 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
733 grep:
734
735 (a) There is now no limit to the number of patterns to be matched.
736
737 (b) An error is given if a pattern is too long.
738
739 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
740 are now supported.
741
742 (d) --exclude-from and --include-from (multiple use) have been added.
743
744 (e) Exclusions and inclusions now apply to all files and directories, not
745 just to those obtained from scanning a directory recursively.
746
747 (f) Multiple uses of -f and --file-list are now supported.
748
749 (g) In a Windows environment, the default for -d has been changed from
750 "read" (the GNU grep default) to "skip", because otherwise the presence
751 of a directory in the file list provokes an error.
752
753 (h) The documentation has been revised and clarified in places.
754
755 9. Improve the matching speed of capturing brackets.
756
757 10. Changed the meaning of \X so that it now matches a Unicode extended
758 grapheme cluster.
759
760 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
761 out POSIX threads when JIT support is configured.
762
763 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
764
765 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
766 structure could go wrong in environments where size_t is not the same size
767 as int.
768
769 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
770 set.
771
772 15. The EBCDIC support had decayed; later updates to the code had included
773 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
774 general tidy up of EBCDIC-related issues, and the documentation was also
775 not quite right. There is now a test that can be run on ASCII systems to
776 check some of the EBCDIC-related things (but is it not a full test).
777
778 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
779 in a small tidy to the code.
780
781 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
782
783 18. If the --only-matching (-o) option in pcregrep is specified multiple
784 times, each one causes appropriate output. For example, -o1 -o2 outputs the
785 substrings matched by the 1st and 2nd capturing parentheses. A separating
786 string can be specified by --om-separator (default empty).
787
788 19. Improving the first n character searches.
789
790 20. Turn case lists for horizontal and vertical white space into macros so that
791 they are defined only once.
792
793 21. This set of changes together give more compatible Unicode case-folding
794 behaviour for characters that have more than one other case when UCP
795 support is available.
796
797 (a) The Unicode property table now has offsets into a new table of sets of
798 three or more characters that are case-equivalent. The MultiStage2.py
799 script that generates these tables (the pcre_ucd.c file) now scans
800 CaseFolding.txt instead of UnicodeData.txt for character case
801 information.
802
803 (b) The code for adding characters or ranges of characters to a character
804 class has been abstracted into a generalized function that also handles
805 case-independence. In UTF-mode with UCP support, this uses the new data
806 to handle characters with more than one other case.
807
808 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
809 whose other case is greater than 256 are now correctly matched
810 caselessly. Previously, the high codepoint matched the low one, but not
811 vice versa.
812
813 (d) The processing of \h, \H, \v, and \ in character classes now makes use
814 of the new class addition function, using character lists defined as
815 macros alongside the case definitions of 20 above.
816
817 (e) Caseless back references now work with characters that have more than
818 one other case.
819
820 (f) General caseless matching of characters with more than one other case
821 is supported.
822
823 22. Unicode character properties were updated from Unicode 6.2.0
824
825 23. Improved CMake support under Windows. Patch by Daniel Richard G.
826
827 24. Add support for 32-bit character strings, and UTF-32
828
829 25. Major JIT compiler update (code refactoring and bugfixing).
830 Experimental Sparc 32 support is added.
831
832 26. Applied a modified version of Daniel Richard G's patch to create
833 pcre.h.generic and config.h.generic by "make" instead of in the
834 PrepareRelease script.
835
836 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
837 pcre_compile.c when checking for a zero character.
838
839 28. Introducing a native interface for JIT. Through this interface, the compiled
840 machine code can be directly executed. The purpose of this interface is to
841 provide fast pattern matching, so several sanity checks are not performed.
842 However, feature tests are still performed. The new interface provides
843 1.4x speedup compared to the old one.
844
845 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
846 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
847 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
848
849 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
850 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
851 the "old" RFC 2279). Instead, it ended up passing a negative length to
852 pcre_exec().
853
854 31. Add support for GCC's visibility feature to hide internal functions.
855
856 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
857 "unknown -C option" after outputting 0 or 1.
858
859 33. There is now support for generating a code coverage report for the test
860 suite in environments where gcc is the compiler and lcov is installed. This
861 is mainly for the benefit of the developers.
862
863 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
864 unaddressable using valgrind annotations, allowing valgrind to detect
865 invalid memory accesses. This is mainly for the benefit of the developers.
866
867 25. (*UTF) can now be used to start a pattern in any of the three libraries.
868
869 26. Give configure error if --enable-cpp but no C++ compiler found.
870
871
872 Version 8.31 06-July-2012
873 -------------------------
874
875 1. Fixing a wrong JIT test case and some compiler warnings.
876
877 2. Removed a bashism from the RunTest script.
878
879 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
880 to unsigned type, result still unsigned" that was given by an MS compiler
881 on encountering the code "-sizeof(xxx)".
882
883 4. Partial matching support is added to the JIT compiler.
884
885 5. Fixed several bugs concerned with partial matching of items that consist
886 of more than one character:
887
888 (a) /^(..)\1/ did not partially match "aba" because checking references was
889 done on an "all or nothing" basis. This also applied to repeated
890 references.
891
892 (b) \R did not give a hard partial match if \r was found at the end of the
893 subject.
894
895 (c) \X did not give a hard partial match after matching one or more
896 characters at the end of the subject.
897
898 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
899 a partial match for the string "\r".
900
901 (e) When newline was set to CRLF, the metacharacter "." did not recognize
902 a partial match for a CR character at the end of the subject string.
903
904 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
905 running pcretest, the text "(JIT)" added to the output whenever JIT is
906 actually used to run the match.
907
908 7. Individual JIT compile options can be set in pcretest by following -s+[+]
909 or /S+[+] with a digit between 1 and 7.
910
911 8. OP_NOT now supports any UTF character not just single-byte ones.
912
913 9. (*MARK) control verb is now supported by the JIT compiler.
914
915 10. The command "./RunTest list" lists the available tests without actually
916 running any of them. (Because I keep forgetting what they all are.)
917
918 11. Add PCRE_INFO_MAXLOOKBEHIND.
919
920 12. Applied a (slightly modified) user-supplied patch that improves performance
921 when the heap is used for recursion (compiled with --disable-stack-for-
922 recursion). Instead of malloc and free for each heap frame each time a
923 logical recursion happens, frames are retained on a chain and re-used where
924 possible. This sometimes gives as much as 30% improvement.
925
926 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
927 call.
928
929 14. As documented, (*COMMIT) is now confined to within a positive assertion.
930
931 15. It is now possible to link pcretest with libedit as an alternative to
932 libreadline.
933
934 16. (*COMMIT) control verb is now supported by the JIT compiler.
935
936 17. The Unicode data tables have been updated to Unicode 6.1.0.
937
938 18. Added --file-list option to pcregrep.
939
940 19. Added binary file support to pcregrep, including the -a, --binary-files,
941 -I, and --text options.
942
943 20. The madvise function is renamed for posix_madvise for QNX compatibility
944 reasons. Fixed by Giuseppe D'Angelo.
945
946 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
947
948 22. Changed the option for creating symbolic links for 16-bit man pages from
949 -s to -sf so that re-installing does not cause issues.
950
951 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
952
953 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
954 matches in certain environments (the workspace was not being correctly
955 retained). Also added to pcre_dfa_exec() a simple plausibility check on
956 some of the workspace data at the beginning of a restart.
957
958 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
959 was not doing so when it should - probably a typo introduced by SVN 528
960 (change 8.10/14).
961
962 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
963 \w+ when the character tables indicated that \x{c4} was a word character.
964 There were several related cases, all because the tests for doing a table
965 lookup were testing for characters less than 127 instead of 255.
966
967 27. If a pattern contains capturing parentheses that are not used in a match,
968 their slots in the ovector are set to -1. For those that are higher than
969 any matched groups, this happens at the end of processing. In the case when
970 there were back references that the ovector was too small to contain
971 (causing temporary malloc'd memory to be used during matching), and the
972 highest capturing number was not used, memory off the end of the ovector
973 was incorrectly being set to -1. (It was using the size of the temporary
974 memory instead of the true size.)
975
976 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
977 ovector size, it uses memory at the end of the block that it has got.
978
979 29. Check for an overlong MARK name and give an error at compile time. The
980 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
981
982 30. JIT compiler update.
983
984 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
985 Rill for the patch.
986
987 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
988
989 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
990
991 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
992 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
993 was enabled.
994
995 35. Improve JIT code generation for greedy plus quantifier.
996
997 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
998 1 to "aa" instead of to an empty string. The bug affected repeated groups
999 that could potentially match an empty string.
1000
1001 37. Optimizing single character iterators in JIT.
1002
1003 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1004 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1005 codepoints that are too big for the mode are faulted, and in a UTF mode,
1006 disallowed codepoints are also faulted.
1007
1008 39. If PCRE was compiled with UTF support, in three places in the DFA
1009 matcher there was code that should only have been obeyed in UTF mode, but
1010 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1011 processing when bytes with values greater than 127 were present. In 16-bit
1012 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1013 both cases the values are those that cannot be the first data item in a UTF
1014 character. The three items that might have provoked this were recursions,
1015 possessively repeated groups, and atomic groups.
1016
1017 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1018 and pcregrep, because some OS require shared objects to be explicitly
1019 passed to ld, causing the link step to fail if they are not.
1020
1021 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1022 mode, patterns that started with \h* or \R* might be incorrectly matched.
1023
1024
1025 Version 8.30 04-February-2012
1026 -----------------------------
1027
1028 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1029 name is defined in ctype.h.
1030
1031 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1032 only in quite long subpatterns.
1033
1034 3. Removed the function pcre_info(), which has been obsolete and deprecated
1035 since it was replaced by pcre_fullinfo() in February 2000.
1036
1037 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1038 match a (*MARK), and the match failed at the start of the subject, a
1039 reference to memory before the start of the subject could occur. This bug
1040 was introduced by fix 17 of release 8.21.
1041
1042 5. A reference to an unset group with zero minimum repetition was giving
1043 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1044 /(another)?(\1?)test/ matched against "hello world test". This bug was
1045 introduced in release 8.13.
1046
1047 6. Add support for 16-bit character strings (a large amount of work involving
1048 many changes and refactorings).
1049
1050 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1051 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1052 from a file.
1053
1054 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1055 rounding is not applied in this particular case).
1056
1057 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1058 if they appear, or are escaped, in patterns.
1059
1060 10. Get rid of a number of -Wunused-but-set-variable warnings.
1061
1062 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1063 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1064 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1065 also returns the mark "x". This bug applied to capturing parentheses,
1066 non-capturing parentheses, and atomic parentheses. It also applied to some
1067 assertions.
1068
1069 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1070 information out of configure.ac instead of relying on pcre.h.generic, which
1071 is not stored in the repository.
1072
1073 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1074 -lreadline.
1075
1076 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1077
1078 15. Applied Graycode's patch to put the top-level frame on the stack rather
1079 than the heap when not using the stack for recursion. This gives a
1080 performance improvement in many cases when recursion is not deep.
1081
1082 16. Experimental code added to "pcretest -C" to output the stack frame size.
1083
1084
1085 Version 8.21 12-Dec-2011
1086 ------------------------
1087
1088 1. Updating the JIT compiler.
1089
1090 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1091 are added as well.
1092
1093 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1094 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1095 calling _pcre_jit_exec. Some extra comments are added.
1096
1097 4. (*MARK) settings inside atomic groups that do not contain any capturing
1098 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1099 was introduced by change 18 for 8.20.
1100
1101 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1102 ECMA-262 standard.
1103
1104 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1105 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1106 This bug was probably introduced by change 9 of 8.13.
1107
1108 7. While fixing 6 above, I noticed that a number of other items were being
1109 incorrectly rejected as "not fixed length". This arose partly because newer
1110 opcodes had not been added to the fixed-length checking code. I have (a)
1111 corrected the bug and added tests for these items, and (b) arranged for an
1112 error to occur if an unknown opcode is encountered while checking for fixed
1113 length instead of just assuming "not fixed length". The items that were
1114 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1115 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1116 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1117
1118 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1119 being incorrectly compiled and would have given unpredicatble results.
1120
1121 9. A possessively repeated subpattern with minimum repeat count greater than
1122 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1123 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1124 the first (A) could occur when it should not.
1125
1126 10. Add a cast and remove a redundant test from the code.
1127
1128 11. JIT should use pcre_malloc/pcre_free for allocation.
1129
1130 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1131 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1132 is anything other than /usr, -L is still shown).
1133
1134 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1135
1136 14. Perl does not support \N without a following name in a [] class; PCRE now
1137 also gives an error.
1138
1139 15. If a forward reference was repeated with an upper limit of around 2000,
1140 it caused the error "internal error: overran compiling workspace". The
1141 maximum number of forward references (including repeats) was limited by the
1142 internal workspace, and dependent on the LINK_SIZE. The code has been
1143 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1144 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1145 of around 200,000 forward references. While doing this, I also speeded up
1146 the filling in of repeated forward references.
1147
1148 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1149 incorrectly expecting the subject to contain another "a" after the start.
1150
1151 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1152 in the match, the SKIP should be ignored. This was not happening; instead
1153 the SKIP was being treated as NOMATCH. For patterns such as
1154 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1155 tested.
1156
1157 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1158 now much more compatible with Perl, in particular in cases where the result
1159 is a non-match for a non-anchored pattern. For example, if
1160 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1161 "m", where previously it did not return a name. A side effect of this
1162 change is that for partial matches, the last encountered mark name is
1163 returned, as for non matches. A number of tests that were previously not
1164 Perl-compatible have been moved into the Perl-compatible test files. The
1165 refactoring has had the pleasing side effect of removing one argument from
1166 the match() function, thus reducing its stack requirements.
1167
1168 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1169 subsequent uses of /S (without +) incorrectly behaved like /S+.
1170
1171 21. Retrieve executable code size support for the JIT compiler and fixing
1172 some warnings.
1173
1174 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1175 not work when the shorter character appeared right at the end of the
1176 subject string.
1177
1178 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1179 systems.
1180
1181 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1182 output it when the /M option is used in pcretest.
1183
1184 25. The CheckMan script was not being included in the distribution. Also, added
1185 an explicit "perl" to run Perl scripts from the PrepareRelease script
1186 because this is reportedly needed in Windows.
1187
1188 26. If study data was being save in a file and studying had not found a set of
1189 "starts with" bytes for the pattern, the data written to the file (though
1190 never used) was taken from uninitialized memory and so caused valgrind to
1191 complain.
1192
1193 27. Updated RunTest.bat as provided by Sheri Pierce.
1194
1195 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1196
1197 29. Computation of memory usage for the table of capturing group names was
1198 giving an unnecessarily large value.
1199
1200
1201 Version 8.20 21-Oct-2011
1202 ------------------------
1203
1204 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1205 a POSIX class. After further experiments with Perl, which convinced me that
1206 Perl has bugs and confusions, a closing square bracket is no longer allowed
1207 in a POSIX name. This bug also affected patterns with classes that started
1208 with full stops.
1209
1210 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1211 captured substring, but while checking the failing first alternative,
1212 substring 1 is temporarily captured. If the output vector supplied to
1213 pcre_exec() was not big enough for this capture, the yield of the function
1214 was still zero ("insufficient space for captured substrings"). This cannot
1215 be totally fixed without adding another stack variable, which seems a lot
1216 of expense for a edge case. However, I have improved the situation in cases
1217 such as /(a)(b)x|abc/ matched against "abc", where the return code
1218 indicates that fewer than the maximum number of slots in the ovector have
1219 been set.
1220
1221 3. Related to (2) above: when there are more back references in a pattern than
1222 slots in the output vector, pcre_exec() uses temporary memory during
1223 matching, and copies in the captures as far as possible afterwards. It was
1224 using the entire output vector, but this conflicts with the specification
1225 that only 2/3 is used for passing back captured substrings. Now it uses
1226 only the first 2/3, for compatibility. This is, of course, another edge
1227 case.
1228
1229 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1230 main code base, and can be used by building with --enable-jit. When this is
1231 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1232 runtime --no-jit option is given.
1233
1234 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1235 ovector, the return from the function was zero, implying that there were
1236 other matches that did not fit. The correct "exactly full" value is now
1237 returned.
1238
1239 6. If a subpattern that was called recursively or as a subroutine contained
1240 (*PRUNE) or any other control that caused it to give a non-standard return,
1241 invalid errors such as "Error -26 (nested recursion at the same subject
1242 position)" or even infinite loops could occur.
1243
1244 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1245 computing the minimum length on reaching *ACCEPT, and so ended up with the
1246 wrong value of 1 rather than 0. Further investigation indicates that
1247 computing a minimum subject length in the presence of *ACCEPT is difficult
1248 (think back references, subroutine calls), and so I have changed the code
1249 so that no minimum is registered for a pattern that contains *ACCEPT.
1250
1251 8. If (*THEN) was present in the first (true) branch of a conditional group,
1252 it was not handled as intended. [But see 16 below.]
1253
1254 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1255 Sheri Pierce.
1256
1257 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1258 the first byte in a match must be "a".
1259
1260 11. Change 17 for 8.13 increased the recursion depth for patterns like
1261 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1262 pattern contains any instances of (*THEN). If it does not, the old
1263 optimizations are restored. It would be nice to do this on a per-group
1264 basis, but at the moment that is not feasible.
1265
1266 12. In some environments, the output of pcretest -C is CRLF terminated. This
1267 broke RunTest's code that checks for the link size. A single white space
1268 character after the value is now allowed for.
1269
1270 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1271 For "fr", it uses the Windows-specific input and output files.
1272
1273 14. If (*THEN) appeared in a group that was called recursively or as a
1274 subroutine, it did not work as intended. [But see next item.]
1275
1276 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1277 pattern fragments (but not containing any | characters). If A and B are
1278 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1279 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1280 D. In other words, Perl considers parentheses that do not contain any |
1281 characters to be part of a surrounding alternative, whereas PCRE was
1282 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1283 differently. PCRE now behaves in the same way as Perl, except in the case
1284 of subroutine/recursion calls such as (?1) which have in any case always
1285 been different (but PCRE had them first :-).
1286
1287 16. Related to 15 above: Perl does not treat the | in a conditional group as
1288 creating alternatives. Such a group is treated in the same way as an
1289 ordinary group without any | characters when processing (*THEN). PCRE has
1290 been changed to match Perl's behaviour.
1291
1292 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1293 RunGrepTest script failed.
1294
1295 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1296 inevitable for groups that contain captures, but it can lead to a lot of
1297 stack use in large patterns. The old behaviour has been restored for atomic
1298 groups that do not contain any capturing parentheses.
1299
1300 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1301 suppress the check for a minimum subject length at run time. (If it was
1302 given to pcre_exec() or pcre_dfa_exec() it did work.)
1303
1304 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1305 fail to work when decoding hex characters in data strings in EBCDIC
1306 environments.
1307
1308 21. It appears that in at least one Mac OS environment, the isxdigit() function
1309 is implemented as a macro that evaluates to its argument more than once,
1310 contravening the C 90 Standard (I haven't checked a later standard). There
1311 was an instance in pcretest which caused it to go wrong when processing
1312 \x{...} escapes in subject strings. The has been rewritten to avoid using
1313 things like p++ in the argument of isxdigit().
1314
1315
1316 Version 8.13 16-Aug-2011
1317 ------------------------
1318
1319 1. The Unicode data tables have been updated to Unicode 6.0.0.
1320
1321 2. Two minor typos in pcre_internal.h have been fixed.
1322
1323 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1324 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1325 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1326
1327 4. There were a number of related bugs in the code for matching backrefences
1328 caselessly in UTF-8 mode when codes for the characters concerned were
1329 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1330 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1331 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1332 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1333 2-byte code at the end of the subject (it thought there wasn't enough data
1334 left).
1335
1336 5. Comprehensive information about what went wrong is now returned by
1337 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1338 as the output vector has at least 2 elements. The offset of the start of
1339 the failing character and a reason code are placed in the vector.
1340
1341 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1342 now returned is for the first byte of the failing character, instead of the
1343 last byte inspected. This is an incompatible change, but I hope it is small
1344 enough not to be a problem. It makes the returned offset consistent with
1345 pcre_exec() and pcre_dfa_exec().
1346
1347 7. pcretest now gives a text phrase as well as the error number when
1348 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1349 failure, the offset and reason code are output.
1350
1351 8. When \R was used with a maximizing quantifier it failed to skip backwards
1352 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1353 back over a single character (\n). This seems wrong (because it treated the
1354 two characters as a single entity when going forwards), conflicts with the
1355 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1356 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1357 has been changed.
1358
1359 9. Some internal refactoring has changed the processing so that the handling
1360 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1361 time (the PCRE_DOTALL option was changed this way some time ago: version
1362 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1363 which was always a bit of a fudge. It also means that there is one less
1364 argument for the match() function, which reduces its stack requirements
1365 slightly. This change also fixes an incompatibility with Perl: the pattern
1366 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1367
1368 10. More internal refactoring has drastically reduced the number of recursive
1369 calls to match() for possessively repeated groups such as (abc)++ when
1370 using pcre_exec().
1371
1372 11. While implementing 10, a number of bugs in the handling of groups were
1373 discovered and fixed:
1374
1375 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1376 (a|)*(?1) gave a compile-time internal error.
1377 ((a|)+)+ did not notice that the outer group could match an empty string.
1378 (^a|^)+ was not marked as anchored.
1379 (.*a|.*)+ was not marked as matching at start or after a newline.
1380
1381 12. Yet more internal refactoring has removed another argument from the match()
1382 function. Special calls to this function are now indicated by setting a
1383 value in a variable in the "match data" data block.
1384
1385 13. Be more explicit in pcre_study() instead of relying on "default" for
1386 opcodes that mean there is no starting character; this means that when new
1387 ones are added and accidentally left out of pcre_study(), testing should
1388 pick them up.
1389
1390 14. The -s option of pcretest has been documented for ages as being an old
1391 synonym of -m (show memory usage). I have changed it to mean "force study
1392 for every regex", that is, assume /S for every regex. This is similar to -i
1393 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1394 using it. It makes it easier to run collections of tests with and without
1395 study enabled, and thereby test pcre_study() more easily. All the standard
1396 tests are now run with and without -s (but some patterns can be marked as
1397 "never study" - see 20 below).
1398
1399 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1400 restoration of the capturing data to the outer values was not happening
1401 correctly.
1402
1403 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1404 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1405 pattern had matched an empty string, and so incorrectly returned a no
1406 match.
1407
1408 17. There was optimizing code for the last branch of non-capturing parentheses,
1409 and also for the obeyed branch of a conditional subexpression, which used
1410 tail recursion to cut down on stack usage. Unfortunately, now that there is
1411 the possibility of (*THEN) occurring in these branches, tail recursion is
1412 no longer possible because the return has to be checked for (*THEN). These
1413 two optimizations have therefore been removed. [But see 8.20/11 above.]
1414
1415 18. If a pattern containing \R was studied, it was assumed that \R always
1416 matched two bytes, thus causing the minimum subject length to be
1417 incorrectly computed because \R can also match just one byte.
1418
1419 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1420 was incorrectly computed.
1421
1422 20. If /S is present twice on a test pattern in pcretest input, it now
1423 *disables* studying, thereby overriding the use of -s on the command line
1424 (see 14 above). This is necessary for one or two tests to keep the output
1425 identical in both cases.
1426
1427 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1428 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1429
1430 22. When an atomic group that contained a capturing parenthesis was
1431 successfully matched, but the branch in which it appeared failed, the
1432 capturing was not being forgotten if a higher numbered group was later
1433 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1434 group 1 to "a", when in fact it should be unset. This applied to multi-
1435 branched capturing and non-capturing groups, repeated or not, and also to
1436 positive assertions (capturing in negative assertions does not happen
1437 in PCRE) and also to nested atomic groups.
1438
1439 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1440 subject after a captured substring, to make it easier to tell which of a
1441 number of identical substrings has been captured.
1442
1443 24. The way atomic groups are processed by pcre_exec() has been changed so that
1444 if they are repeated, backtracking one repetition now resets captured
1445 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1446 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1447 "aaa". Previously, it would have been "a". As part of this code
1448 refactoring, the way recursive calls are handled has also been changed.
1449
1450 25. If an assertion condition captured any substrings, they were not passed
1451 back unless some other capturing happened later. For example, if
1452 (?(?=(a))a) was matched against "a", no capturing was returned.
1453
1454 26. When studying a pattern that contained subroutine calls or assertions,
1455 the code for finding the minimum length of a possible match was handling
1456 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1457 group 1 called group 2 while simultaneously a separate group 2 called group
1458 1). A stack overflow occurred in this case. I have fixed this by limiting
1459 the recursion depth to 10.
1460
1461 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1462 Fortmann. This supports explicit test numbers on the command line, and has
1463 argument validation and error reporting.
1464
1465 28. An instance of \X with an unlimited repeat could fail if at any point the
1466 first character it looked at was a mark character.
1467
1468 29. Some minor code refactoring concerning Unicode properties and scripts
1469 should reduce the stack requirement of match() slightly.
1470
1471 30. Added the '=' option to pcretest to check the setting of unused capturing
1472 slots at the end of the pattern, which are documented as being -1, but are
1473 not included in the return count.
1474
1475 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1476 compiled something random. Now it gives a compile-time error (as does
1477 Perl).
1478
1479 32. A *MARK encountered during the processing of a positive assertion is now
1480 recorded and passed back (compatible with Perl).
1481
1482 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1483 had alternative anchored branches, the search for a second match in a line
1484 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1485 matched the line "0102" twice. The same bug affected patterns that started
1486 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1487 twice.
1488
1489 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1490 does, and because of capturing effects, quantifying parenthesized
1491 assertions may at times be useful. Quantifiers are now allowed for
1492 parenthesized assertions.
1493
1494 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1495
1496 36. \g was being checked for fancy things in a character class, when it should
1497 just be a literal "g".
1498
1499 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1500 appearance of a nested POSIX class supersedes an apparent external class.
1501 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1502 unescaped square brackets may also appear as part of class names. For
1503 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1504 more like Perl. (But see 8.20/1 above.)
1505
1506 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1507 was because it thought it was \N{name}, which is not supported).
1508
1509 39. Add minix to OS list not supporting the -S option in pcretest.
1510
1511 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1512 cannot analyze patterns in sufficient detail to catch mutual recursions
1513 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1514 subgroup is called recursively as a subpattern for a second time at the
1515 same position in the subject string. In previous releases this might have
1516 been caught by the recursion limit, or it might have run out of stack.
1517
1518 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1519 happen only once. PCRE was, however incorrectly giving a compile time error
1520 "recursive call could loop indefinitely" because it cannot analyze the
1521 pattern in sufficient detail. The compile time test no longer happens when
1522 PCRE is compiling a conditional subpattern, but actual runaway loops are
1523 now caught at runtime (see 40 above).
1524
1525 42. It seems that Perl allows any characters other than a closing parenthesis
1526 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1527 has been changed to be the same.
1528
1529 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1530 as not to get warnings when autogen.sh is called. Also changed
1531 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1532
1533 44. To help people who use pcregrep to scan files containing exceedingly long
1534 lines, the following changes have been made:
1535
1536 (a) The default value of the buffer size parameter has been increased from
1537 8K to 20K. (The actual buffer used is three times this size.)
1538
1539 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1540 PCRE is built.
1541
1542 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1543 to be set at run time.
1544
1545 (d) Numerical values in pcregrep options can be followed by K or M, for
1546 example --buffer-size=50K.
1547
1548 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1549 given and the return code is set to 2.
1550
1551 45. Add a pointer to the latest mark to the callout data block.
1552
1553 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1554 partial match of an empty string instead of no match. This was specific to
1555 the use of ".".
1556
1557 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1558 complete match instead of a partial match. This bug was dependent on both
1559 the PCRE_UTF8 and PCRE_DOTALL options being set.
1560
1561 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1562 starting byte set, because \b was not being ignored.
1563
1564
1565 Version 8.12 15-Jan-2011
1566 ------------------------
1567
1568 1. Fixed some typos in the markup of the man pages, and wrote a script that
1569 checks for such things as part of the documentation building process.
1570
1571 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1572 --match-limit and --recursion-limit options (added for 8.11). In
1573 particular, this made one of the standard tests fail. (The integer value
1574 went into the wrong half of a long int.)
1575
1576 3. If the --colour option was given to pcregrep with -v (invert match), it
1577 did strange things, either producing crazy output, or crashing. It should,
1578 of course, ignore a request for colour when reporting lines that do not
1579 match.
1580
1581 4. Another pcregrep bug caused similar problems if --colour was specified with
1582 -M (multiline) and the pattern match finished with a line ending.
1583
1584 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1585 matched in multiline mode, the following line was shown as part of the
1586 match. This seems wrong, so I have changed it.
1587
1588 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1589 the check for further matches in the same line (so they could be coloured)
1590 to overrun the end of the current line. If another match was found, it was
1591 incorrectly shown (and then shown again when found in the next line).
1592
1593 7. If pcregrep was compiled under Windows, there was a reference to the
1594 function pcregrep_exit() before it was defined. I am assuming this was
1595 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1596 reported by a user. I've moved the definition above the reference.
1597
1598
1599 Version 8.11 10-Dec-2010
1600 ------------------------
1601
1602 1. (*THEN) was not working properly if there were untried alternatives prior
1603 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1604 backtracked to try for "b" instead of moving to the next alternative branch
1605 at the same level (in this case, to look for "c"). The Perl documentation
1606 is clear that when (*THEN) is backtracked onto, it goes to the "next
1607 alternative in the innermost enclosing group".
1608
1609 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1610 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1611 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1612 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1613 (*THEN).
1614
1615 3. If \s appeared in a character class, it removed the VT character from
1616 the class, even if it had been included by some previous item, for example
1617 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1618 of \s, but is part of the POSIX "space" class.)
1619
1620 4. A partial match never returns an empty string (because you can always
1621 match an empty string at the end of the subject); however the checking for
1622 an empty string was starting at the "start of match" point. This has been
1623 changed to the "earliest inspected character" point, because the returned
1624 data for a partial match starts at this character. This means that, for
1625 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1626 (previously it gave "no match").
1627
1628 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1629 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1630 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1631 has an implication that the given string is incomplete (because a partial
1632 match is preferred over a full match). For this reason, these items now
1633 give a partial match in this situation. [Aside: previously, the one case
1634 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1635 match rather than a full match, which was wrong by the old rules, but is
1636 now correct.]
1637
1638 6. There was a bug in the handling of #-introduced comments, recognized when
1639 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1640 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1641 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1642 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1643 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1644 places in pcre_compile().
1645
1646 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1647 comments when looking ahead for named forward references to subpatterns,
1648 the only newline sequence it recognized was NL. It now handles newlines
1649 according to the set newline convention.
1650
1651 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1652 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1653 cater for a lack of strerror(). These oversights have been fixed.
1654
1655 9. Added --match-limit and --recursion-limit to pcregrep.
1656
1657 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1658
1659 11. When the -o option was used, pcregrep was setting a return code of 1, even
1660 when matches were found, and --line-buffered was not being honoured.
1661
1662 12. Added an optional parentheses number to the -o and --only-matching options
1663 of pcregrep.
1664
1665 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1666 can match an empty string. The code to do it in pcretest and pcredemo
1667 needed fixing:
1668
1669 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1670 only one byte after an empty string match just before CRLF (this case
1671 just got forgotten; "any" and "anycrlf" were OK).
1672
1673 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1674 mode when an empty string match preceded an ASCII character followed by
1675 a non-ASCII character. (The code for advancing by one character rather
1676 than one byte was nonsense.)
1677
1678 (c) The pcredemo.c sample program did not have any code at all to handle
1679 the cases when CRLF is a valid newline sequence.
1680
1681 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1682 as a starting offset was within the subject string. There is now a new
1683 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1684 negative or greater than the length of the string. In order to test this,
1685 pcretest is extended to allow the setting of negative starting offsets.
1686
1687 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1688 starting offset points to the beginning of a UTF-8 character was
1689 unnecessarily clumsy. I tidied it up.
1690
1691 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1692 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1693
1694 17. Nobody had reported that the --include_dir option, which was added in
1695 release 7.7 should have been called --include-dir (hyphen, not underscore)
1696 for compatibility with GNU grep. I have changed it to --include-dir, but
1697 left --include_dir as an undocumented synonym, and the same for
1698 --exclude-dir, though that is not available in GNU grep, at least as of
1699 release 2.5.4.
1700
1701 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1702 characters from a string of bytes) have been redefined so as not to use
1703 loops, in order to improve performance in some environments. At the same
1704 time, I abstracted some of the common code into auxiliary macros to save
1705 repetition (this should not affect the compiled code).
1706
1707 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1708 compile-time error is now given if \c is not followed by an ASCII
1709 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1710 different, and any byte value is allowed.)
1711
1712 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1713 START_OPTIMIZE option, which is now allowed at compile time - but just
1714 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1715 to pcregrep and other applications that have no direct access to PCRE
1716 options. The new /Y option in pcretest sets this option when calling
1717 pcre_compile().
1718
1719 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1720 back references. Groups containing recursive back references were forced to
1721 be atomic by that change, but in the case of named groups, the amount of
1722 memory required was incorrectly computed, leading to "Failed: internal
1723 error: code overflow". This has been fixed.
1724
1725 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1726 pcretest.c, to avoid build problems in some Borland environments.
1727
1728
1729 Version 8.10 25-Jun-2010
1730 ------------------------
1731
1732 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1733 THEN.
1734
1735 2. (*ACCEPT) was not working when inside an atomic group.
1736
1737 3. Inside a character class, \B is treated as a literal by default, but
1738 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1739 causes the error). The code is unchanged, but I tidied the documentation.
1740
1741 4. Inside a character class, PCRE always treated \R and \X as literals,
1742 whereas Perl faults them if its -w option is set. I have changed PCRE so
1743 that it faults them when PCRE_EXTRA is set.
1744
1745 5. Added support for \N, which always matches any character other than
1746 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1747
1748 6. When compiling pcregrep with newer versions of gcc which may have
1749 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1750 declared with attribute warn_unused_result" were given. Just casting the
1751 result to (void) does not stop the warnings; a more elaborate fudge is
1752 needed. I've used a macro to implement this.
1753
1754 7. Minor change to pcretest.c to avoid a compiler warning.
1755
1756 8. Added four artifical Unicode properties to help with an option to make
1757 \s etc use properties (see next item). The new properties are: Xan
1758 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1759
1760 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1761 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1762 this option. Modified pcretest to add /W to test this facility. Added
1763 REG_UCP to make it available via the POSIX interface.
1764
1765 10. Added --line-buffered to pcregrep.
1766
1767 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1768 studied, and the match started with a letter with a code point greater than
1769 127 whose first byte was different to the first byte of the other case of
1770 the letter, the other case of this starting letter was not recognized
1771 (#976).
1772
1773 12. If a pattern that was studied started with a repeated Unicode property
1774 test, for example, \p{Nd}+, there was the theoretical possibility of
1775 setting up an incorrect bitmap of starting bytes, but fortunately it could
1776 not have actually happened in practice until change 8 above was made (it
1777 added property types that matched character-matching opcodes).
1778
1779 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1780 possible starting bytes for non-anchored patterns.
1781
1782 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1783 \R, and also a number of cases that involve Unicode properties, both
1784 explicit and implicit when PCRE_UCP is set.
1785
1786 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1787 input, it could crash or give wrong results if characters with values
1788 greater than 0xc0 were present in the subject string. (Detail: it assumed
1789 UTF-8 input when processing these items.)
1790
1791 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1792 size_t is 64-bit (#991).
1793
1794 17. Added a check for running out of memory when PCRE is compiled with
1795 --disable-stack-for-recursion (#990).
1796
1797 18. If the last data line in a file for pcretest does not have a newline on
1798 the end, a newline was missing in the output.
1799
1800 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1801 less than 128) in its various bitmaps. However, there is a facility for
1802 generating tables according to the current locale when PCRE is compiled. It
1803 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1804 characters, are recognized by isspace() and therefore were getting set in
1805 these tables, and indeed these tables seem to approximate to ISO 8859. This
1806 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1807 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1808 which of course cannot start UTF-8 characters. I have changed the code so
1809 that only real ASCII characters (less than 128) and the correct starting
1810 bytes for UTF-8 encodings are set for characters greater than 127 when in
1811 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1812 altogether.)
1813
1814 20. Added the /T option to pcretest so as to be able to run tests with non-
1815 standard character tables, thus making it possible to include the tests
1816 used for 19 above in the standard set of tests.
1817
1818 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1819 reference to a subpattern the other side of a comment that contains an
1820 opening parenthesis caused either an internal compiling error, or a
1821 reference to the wrong subpattern.
1822
1823
1824 Version 8.02 19-Mar-2010
1825 ------------------------
1826
1827 1. The Unicode data tables have been updated to Unicode 5.2.0.
1828
1829 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1830 configured.
1831
1832 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1833 original author of that file, following a query about its status.
1834
1835 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1836 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1837
1838 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1839 quantifier applied to a forward-referencing subroutine call, could compile
1840 incorrect code or give the error "internal error: previously-checked
1841 referenced subpattern not found".
1842
1843 6. Both MS Visual Studio and Symbian OS have problems with initializing
1844 variables to point to external functions. For these systems, therefore,
1845 pcre_malloc etc. are now initialized to local functions that call the
1846 relevant global functions.
1847
1848 7. There were two entries missing in the vectors called coptable and poptable
1849 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1850 I've fixed the data, and added a kludgy way of testing at compile time that
1851 the lengths are correct (equal to the number of opcodes).
1852
1853 8. Following on from 7, I added a similar kludge to check the length of the
1854 eint vector in pcreposix.c.
1855
1856 9. Error texts for pcre_compile() are held as one long string to avoid too
1857 much relocation at load time. To find a text, the string is searched,
1858 counting zeros. There was no check for running off the end of the string,
1859 which could happen if a new error number was added without updating the
1860 string.
1861
1862 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1863
1864 11. \K was not working if it appeared in an atomic group or in a group that
1865 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1866 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1867 the assertion is positive, but not if it is negative.
1868
1869 12. Change 11 fortuitously reduced the size of the stack frame used in the
1870 "match()" function of pcre_exec.c by one pointer. Forthcoming
1871 implementation of support for (*MARK) will need an extra pointer on the
1872 stack; I have reserved it now, so that the stack frame size does not
1873 decrease.
1874
1875 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1876 item in branch that calls a recursion is a subroutine call - as in the
1877 second branch in the above example - was incorrectly given the compile-
1878 time error "recursive call could loop indefinitely" because pcre_compile()
1879 was not correctly checking the subroutine for matching a non-empty string.
1880
1881 14. The checks for overrunning compiling workspace could trigger after an
1882 overrun had occurred. This is a "should never occur" error, but it can be
1883 triggered by pathological patterns such as hundreds of nested parentheses.
1884 The checks now trigger 100 bytes before the end of the workspace.
1885
1886 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1887
1888
1889 Version 8.01 19-Jan-2010
1890 ------------------------
1891
1892 1. If a pattern contained a conditional subpattern with only one branch (in
1893 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1894 computed the wrong minimum data length (which is of course zero for such
1895 subpatterns). This could cause incorrect "no match" results.
1896
1897 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1898 the pattern is reset in the first branch, pcre_compile() failed with
1899 "internal error: code overflow at offset...". This happened only when
1900 the reset was to the original external option setting. (An optimization
1901 abstracts leading options settings into an external setting, which was the
1902 cause of this.)
1903
1904 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1905 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1906 assertion pattern did not match (meaning that the assertion was true), it
1907 was incorrectly treated as false if the SKIP had been reached during the
1908 matching. This also applied to assertions used as conditions.
1909
1910 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1911 assertion subpattern, including such a pattern used as a condition,
1912 unpredictable results occurred, instead of the error return
1913 PCRE_ERROR_DFA_UITEM.
1914
1915 5. The C++ GlobalReplace function was not working like Perl for the special
1916 situation when an empty string is matched. It now does the fancy magic
1917 stuff that is necessary.
1918
1919 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1920 removed. (These were left over from very, very early versions of PCRE.)
1921
1922 7. Some cosmetic changes to the code to make life easier when compiling it
1923 as part of something else:
1924
1925 (a) Change DEBUG to PCRE_DEBUG.
1926
1927 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1928 called "current" as "current_branch", to prevent a collision with the
1929 Linux macro when compiled as a kernel module.
1930
1931 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1932 prevent a collision with the Linux macro when compiled as a kernel
1933 module.
1934
1935 8. In pcre_compile() there are some checks for integer overflows that used to
1936 cast potentially large values to (double). This has been changed to that
1937 when building, a check for int64_t is made, and if it is found, it is used
1938 instead, thus avoiding the use of floating point arithmetic. (There is no
1939 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1940 double.
1941
1942 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1943 2005 (difference between two addresses compared to an unsigned value).
1944
1945 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1946 custom one, because of the following reported problem in Windows:
1947
1948 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1949 under Win32.
1950 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1951 therefore missing the function definition.
1952 - The compiler thus generates a "C" signature for the test function.
1953 - The linker fails to find the "C" function.
1954 - PCRE fails to configure if asked to do so against libbz2.
1955
1956 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1957 messages were output:
1958
1959 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1960 rerunning libtoolize, to keep the correct libtool macros in-tree.
1961 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1962
1963 I have done both of these things.
1964
1965 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1966 most of the time, it *can* run out if it is given a pattern that contains a
1967 runaway infinite recursion. I updated the discussion in the pcrestack man
1968 page.
1969
1970 13. Now that we have gone to the x.xx style of version numbers, the minor
1971 version may start with zero. Using 08 or 09 is a bad idea because users
1972 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1973 interpreted as invalid octal numbers. I've updated the previous comment in
1974 configure.ac, and also added a check that gives an error if 08 or 09 are
1975 used.
1976
1977 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1978 causing partial matching to fail when the end of the subject matched \W
1979 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1980
1981 15. There were some discrepancies between the declarations in pcre_internal.h
1982 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1983 their definitions. The declarations used "const uschar *" and the
1984 definitions used USPTR. Even though USPTR is normally defined as "const
1985 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1986 reported that: "This difference in casting confuses some C++ compilers, for
1987 example, SunCC recognizes above declarations as different functions and
1988 generates broken code for hbpcre." I have changed the declarations to use
1989 USPTR.
1990
1991 16. GNU libtool is named differently on some systems. The autogen.sh script now
1992 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1993 (FreeBSD).
1994
1995 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1996 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1997 comment: "Figure out how to create a longlong from a string: strtoll and
1998 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1999 instance, but it only takes 2 args instead of 3!"
2000
2001 18. A subtle bug concerned with back references has been fixed by a change of
2002 specification, with a corresponding code fix. A pattern such as
2003 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2004 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2005 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2006 same bug. Such groups have to be quantified to be useful, or contained
2007 inside another quantified group. (If there's no repetition, the reference
2008 can never match.) The problem arises because, having left the group and
2009 moved on to the rest of the pattern, a later failure that backtracks into
2010 the group uses the captured value from the final iteration of the group
2011 rather than the correct earlier one. I have fixed this in PCRE by forcing
2012 any group that contains a reference to itself to be an atomic group; that
2013 is, there cannot be any backtracking into it once it has completed. This is
2014 similar to recursive and subroutine calls.
2015
2016
2017 Version 8.00 19-Oct-09
2018 ----------------------
2019
2020 1. The table for translating pcre_compile() error codes into POSIX error codes
2021 was out-of-date, and there was no check on the pcre_compile() error code
2022 being within the table. This could lead to an OK return being given in
2023 error.
2024
2025 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2026 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2027 in a Windows environment.
2028
2029 3. The pcregrep --count option prints the count for each file even when it is
2030 zero, as does GNU grep. However, pcregrep was also printing all files when
2031 --files-with-matches was added. Now, when both options are given, it prints
2032 counts only for those files that have at least one match. (GNU grep just
2033 prints the file name in this circumstance, but including the count seems
2034 more useful - otherwise, why use --count?) Also ensured that the
2035 combination -clh just lists non-zero counts, with no names.
2036
2037 4. The long form of the pcregrep -F option was incorrectly implemented as
2038 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2039 but it seems right to fix it, and I didn't think it was worth preserving
2040 the old behaviour.
2041
2042 5. The command line items --regex=pattern and --regexp=pattern were not
2043 recognized by pcregrep, which required --regex pattern or --regexp pattern
2044 (with a space rather than an '='). The man page documented the '=' forms,
2045 which are compatible with GNU grep; these now work.
2046
2047 6. No libpcreposix.pc file was created for pkg-config; there was just
2048 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2049
2050 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2051 when UCP support is not needed, by modifying the Python script that
2052 generates it from Unicode data files. This should not matter if the module
2053 is correctly used as a library, but I received one complaint about 50K of
2054 unwanted data. My guess is that the person linked everything into his
2055 program rather than using a library. Anyway, it does no harm.
2056
2057 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2058 was a minimum greater than 1 for a wide character in a possessive
2059 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2060 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2061 character. Chaos in the form of incorrect output or a compiling loop could
2062 result.
2063
2064 9. The restrictions on what a pattern can contain when partial matching is
2065 requested for pcre_exec() have been removed. All patterns can now be
2066 partially matched by this function. In addition, if there are at least two
2067 slots in the offset vector, the offset of the earliest inspected character
2068 for the match and the offset of the end of the subject are set in them when
2069 PCRE_ERROR_PARTIAL is returned.
2070
2071 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2072 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2073 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2074 and may be more useful for multi-segment matching.
2075
2076 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2077 used to be given if ever the end of the subject was reached; now it is
2078 given only if matching could not proceed because another character was
2079 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2080 string "Z", which now yields "no match" instead of "partial match". In the
2081 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2082 final character ended with (*FAIL).
2083
2084 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2085 if the pattern had a "must contain" character that was already found in the
2086 earlier partial match, unless partial matching was again requested. For
2087 example, with the pattern /dog.(body)?/, the "must contain" character is
2088 "g". If the first part-match was for the string "dog", restarting with
2089 "sbody" failed. This bug has been fixed.
2090
2091 13. The string returned by pcre_dfa_exec() after a partial match has been
2092 changed so that it starts at the first inspected character rather than the
2093 first character of the match. This makes a difference only if the pattern
2094 starts with a lookbehind assertion or \b or \B (\K is not supported by
2095 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2096 matching functions compatible, and I think it's the right thing to do.
2097
2098 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2099 so that the demonstration program is easily available in environments where
2100 PCRE has not been installed from source.
2101
2102 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2103 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2104 library.
2105
2106 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2107 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2108 is not the first non-POSIX option to be added. Clearly some people find
2109 these options useful.
2110
2111 17. If a caller to the POSIX matching function regexec() passes a non-zero
2112 value for nmatch with a NULL value for pmatch, the value of
2113 nmatch is forced to zero.
2114
2115 18. RunGrepTest did not have a test for the availability of the -u option of
2116 the diff command, as RunTest does. It now checks in the same way as
2117 RunTest, and also checks for the -b option.
2118
2119 19. If an odd number of negated classes containing just a single character
2120 interposed, within parentheses, between a forward reference to a named
2121 subpattern and the definition of the subpattern, compilation crashed with
2122 an internal error, complaining that it could not find the referenced
2123 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2124 [The bug was that it was starting one character too far in when skipping
2125 over the character class, thus treating the ] as data rather than
2126 terminating the class. This meant it could skip too much.]
2127
2128 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2129 /g option in pcretest when the pattern contains \K, which makes it possible
2130 to have an empty string match not at the start, even when the pattern is
2131 anchored. Updated pcretest and pcredemo to use this option.
2132
2133 21. If the maximum number of capturing subpatterns in a recursion was greater
2134 than the maximum at the outer level, the higher number was returned, but
2135 with unset values at the outer level. The correct (outer level) value is
2136 now given.
2137
2138 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2139 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2140 make it do so. The string so far is captured, making this feature
2141 compatible with Perl.
2142
2143 23. The tests have been re-organized, adding tests 11 and 12, to make it
2144 possible to check the Perl 5.10 features against Perl 5.10.
2145
2146 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2147 pattern matches a fixed length string. PCRE did not allow this; now it
2148 does. Neither allows recursion.
2149
2150 25. I finally figured out how to implement a request to provide the minimum
2151 length of subject string that was needed in order to match a given pattern.
2152 (It was back references and recursion that I had previously got hung up
2153 on.) This code has now been added to pcre_study(); it finds a lower bound
2154 to the length of subject needed. It is not necessarily the greatest lower
2155 bound, but using it to avoid searching strings that are too short does give
2156 some useful speed-ups. The value is available to calling programs via
2157 pcre_fullinfo().
2158
2159 26. While implementing 25, I discovered to my embarrassment that pcretest had
2160 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2161 study optimizations had never been tested with that matching function.
2162 Oops. What is worse, even when it was passed study data, there was a bug in
2163 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2164 were also very few tests of studied patterns with pcre_dfa_exec().
2165
2166 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2167 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2168 on the other side of the coin, they are no longer allowed to have different
2169 names, because these cannot be distinguished in PCRE, and this has caused
2170 confusion. (This is a difference from Perl.)
2171
2172 28. When duplicate subpattern names are present (necessarily with different
2173 numbers, as required by 27 above), and a test is made by name in a
2174 conditional pattern, either for a subpattern having been matched, or for
2175 recursion in such a pattern, all the associated numbered subpatterns are
2176 tested, and the overall condition is true if the condition is true for any
2177 one of them. This is the way Perl works, and is also more like the way
2178 testing by number works.
2179
2180
2181 Version 7.9 11-Apr-09
2182 ---------------------
2183
2184 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2185 (pcretest), all targets were linked against these libraries. This included
2186 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2187 libraries. This caused unwanted dependencies to be created. This problem
2188 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2189 pcretest is linked with readline.
2190
2191 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2192 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2193 moved outside it again, because FALSE and TRUE are already defined in AIX,
2194 but BOOL is not.
2195
2196 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2197 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2198
2199 4. The pcregrep documentation said spaces were inserted as well as colons (or
2200 hyphens) following file names and line numbers when outputting matching
2201 lines. This is not true; no spaces are inserted. I have also clarified the
2202 wording for the --colour (or --color) option.
2203
2204 5. In pcregrep, when --colour was used with -o, the list of matching strings
2205 was not coloured; this is different to GNU grep, so I have changed it to be
2206 the same.
2207
2208 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2209 each matching line was coloured. Now it goes on to look for further matches
2210 of any of the test patterns, which is the same behaviour as GNU grep.
2211
2212 7. A pattern that could match an empty string could cause pcregrep to loop; it
2213 doesn't make sense to accept an empty string match in pcregrep, so I have
2214 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2215 seems to be how GNU grep behaves. [But see later change 40 for release
2216 8.33.]
2217
2218 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2219 start or after a newline", because the conditional assertion was not being
2220 correctly handled. The rule now is that both the assertion and what follows
2221 in the first alternative must satisfy the test.
2222
2223 9. If auto-callout was enabled in a pattern with a conditional group whose
2224 condition was an assertion, PCRE could crash during matching, both with
2225 pcre_exec() and pcre_dfa_exec().
2226
2227 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2228 used for matching.
2229
2230 11. Unicode property support in character classes was not working for
2231 characters (bytes) greater than 127 when not in UTF-8 mode.
2232
2233 12. Added the -M command line option to pcretest.
2234
2235 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2236
2237 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2238
2239 16. Added comments and documentation about mis-use of no_arg in the C++
2240 wrapper.
2241
2242 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2243 from Martin Jerabek that uses macro names for all relevant character and
2244 string constants.
2245
2246 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2247 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2248 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2249 these, but not everybody uses configure.
2250
2251 19. A conditional group that had only one branch was not being correctly
2252 recognized as an item that could match an empty string. This meant that an
2253 enclosing group might also not be so recognized, causing infinite looping
2254 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2255 with the subject "ab", where knowledge that the repeated group can match
2256 nothing is needed in order to break the loop.
2257
2258 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2259 exec(), but without supplying a callout function, matching went wrong.
2260
2261 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2262 leak if the size of the offset vector was greater than 30. When the vector
2263 is smaller, the saved offsets during recursion go onto a local stack
2264 vector, but for larger vectors malloc() is used. It was failing to free
2265 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2266 error, in fact).
2267
2268 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2269 heapframe that is used only when UTF-8 support is enabled. This caused no
2270 problem, but was untidy.
2271
2272 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2273 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2274 included within another project.
2275
2276 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2277 slightly modified by me:
2278
2279 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2280 not building pcregrep.
2281
2282 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2283 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2284
2285 25. Forward references, both numeric and by name, in patterns that made use of
2286 duplicate group numbers, could behave incorrectly or give incorrect errors,
2287 because when scanning forward to find the reference group, PCRE was not
2288 taking into account the duplicate group numbers. A pattern such as
2289 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2290
2291 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2292 the feature of a custom pointer more persuasive (as requested by a user).
2293
2294 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2295 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2296 pre-defined.
2297
2298 28. Added support for (*UTF8) at the start of a pattern.
2299
2300 29. Arrange for flags added by the "release type" setting in CMake to be shown
2301 in the configuration summary.
2302
2303
2304 Version 7.8 05-Sep-08
2305 ---------------------
2306
2307 1. Replaced UCP searching code with optimized version as implemented for Ad
2308 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2309 stage table and inline lookup instead of a function, giving speed ups of 2
2310 to 5 times on some simple patterns that I tested. Permission was given to
2311 distribute the MultiStage2.py script that generates the tables (it's not in
2312 the tarball, but is in the Subversion repository).
2313
2314 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2315 scripts.
2316
2317 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2318 a group with a zero qualifier. The result of the study could be incorrect,
2319 or the function might crash, depending on the pattern.
2320
2321 4. Caseless matching was not working for non-ASCII characters in back
2322 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2323 It now works when Unicode Property Support is available.
2324
2325 5. In pcretest, an escape such as \x{de} in the data was always generating
2326 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2327 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2328 truncation.
2329
2330 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2331
2332 7. Added two (int) casts to pcregrep when printing the difference of two
2333 pointers, in case they are 64-bit values.
2334
2335 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2336 test 2 if it fails.
2337
2338 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2339 and a #define of that name to empty if it is not externally set. This is to
2340 allow users of MSVC to set it if necessary.
2341
2342 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2343 the convenience functions in the pcre_get.c source file.
2344
2345 11. An option change at the start of a pattern that had top-level alternatives
2346 could cause overwriting and/or a crash. This command provoked a crash in
2347 some environments:
2348
2349 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2350
2351 This potential security problem was recorded as CVE-2008-2371.
2352
2353 12. For a pattern where the match had to start at the beginning or immediately
2354 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2355 pcre_dfa_exec() could read past the end of the passed subject if there was
2356 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2357 pcretest so that it places the subject at the end of its malloc-ed buffer.
2358
2359 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2360 exec() might read past the end of the data buffer in UTF-8 mode.
2361
2362 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2363 the data contained the byte 0x85 as part of a UTF-8 character within its
2364 first line. This applied both to normal and DFA matching.
2365
2366 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2367 /^[^d]*?$/8 failed to match "abc".
2368
2369 16. Added a missing copyright notice to pcrecpp_internal.h.
2370
2371 17. Make it more clear in the documentation that values returned from
2372 pcre_exec() in ovector are byte offsets, not character counts.
2373
2374 18. Tidied a few places to stop certain compilers from issuing warnings.
2375
2376 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2377 supplied by Stefan Weber. I made a further small update for 7.8 because
2378 there is a change of source arrangements: the pcre_searchfuncs.c module is
2379 replaced by pcre_ucd.c.
2380
2381
2382 Version 7.7 07-May-08
2383 ---------------------
2384
2385 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2386 a string to a long long, pretend we don't even have a long long." This is
2387 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2388
2389 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2390 pre-7.6 versions, which defined a global no_arg variable instead of putting
2391 it in the RE class. (See also #8 below.)
2392
2393 3. Remove a line of dead code, identified by coverity and reported by Nuno
2394 Lopes.
2395
2396 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2397
2398 (1) The include/exclude patterns were being applied to the whole pathnames
2399 of files, instead of just to the final components.
2400
2401 (2) If there was more than one level of directory, the subdirectories were
2402 skipped unless they satisfied the include/exclude conditions. This is
2403 inconsistent with GNU grep (and could even be seen as contrary to the
2404 pcregrep specification - which I improved to make it absolutely clear).
2405 The action now is always to scan all levels of directory, and just
2406 apply the include/exclude patterns to regular files.
2407
2408 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2409 --exclude_dir in the tests to avoid scanning .svn directories.
2410
2411 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2412 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2413 doesn't support NULs in patterns.
2414
2415 7. Added some missing "const"s to declarations of static tables in
2416 pcre_compile.c and pcre_dfa_exec.c.
2417
2418 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2419 caused by fix #2 above. (Subsequently also a second patch to fix the
2420 first patch. And a third patch - this was a messy problem.)
2421
2422 9. Applied Craig's patch to remove the use of push_back().
2423
2424 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2425 matching function regexec().
2426
2427 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2428 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2429 references. PCRE supports relative numbers with this syntax (I don't think
2430 Oniguruma does).
2431
2432 12. Previously, a group with a zero repeat such as (...){0} was completely
2433 omitted from the compiled regex. However, this means that if the group
2434 was called as a subroutine from elsewhere in the pattern, things went wrong
2435 (an internal error was given). Such groups are now left in the compiled
2436 pattern, with a new opcode that causes them to be skipped at execution
2437 time.
2438
2439 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2440 to the way PCRE behaves:
2441
2442 (a) A lone ] character is dis-allowed (Perl treats it as data).
2443
2444 (b) A back reference to an unmatched subpattern matches an empty string
2445 (Perl fails the current match path).
2446
2447 (c) A data ] in a character class must be notated as \] because if the
2448 first data character in a class is ], it defines an empty class. (In
2449 Perl it is not possible to have an empty class.) The empty class []
2450 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2451 The negative empty class [^] matches any one character, independently
2452 of the DOTALL setting.
2453
2454 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2455 non-existent subpattern following a character class starting with ']' and
2456 containing () gave an internal compiling error instead of "reference to
2457 non-existent subpattern". Fortunately, when the pattern did exist, the
2458 compiled code was correct. (When scanning forwards to check for the
2459 existence of the subpattern, it was treating the data ']' as terminating
2460 the class, so got the count wrong. When actually compiling, the reference
2461 was subsequently set up correctly.)
2462
2463 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2464 it was being rejected as not supported by pcre_dfa_exec(), even though
2465 other assertions are supported. I have made pcre_dfa_exec() support
2466 (*FAIL).
2467
2468 16. The implementation of 13c above involved the invention of a new opcode,
2469 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2470 cannot be changed at match time, I realized I could make a small
2471 improvement to matching performance by compiling OP_ALLANY instead of
2472 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2473 on the OP_ANY path.
2474
2475 17. Compiling pcretest on Windows with readline support failed without the
2476 following two fixes: (1) Make the unistd.h include conditional on
2477 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2478
2479 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2480 ncurses library to be included for pcretest when ReadLine support is
2481 requested, but also to allow for it to be overridden. This patch came from
2482 Daniel Bergström.
2483
2484 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2485 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2486 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2487 spotting this.
2488
2489
2490 Version 7.6 28-Jan-08
2491 ---------------------
2492
2493 1. A character class containing a very large number of characters with
2494 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2495 overflow.
2496
2497 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2498 HAVE_LONG_LONG is not defined.
2499
2500 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2501 bring them up to date and include new features. This patch includes:
2502
2503 - Fixed PH's badly added libz and libbz2 support.
2504 - Fixed a problem with static linking.
2505 - Added pcredemo. [But later removed - see 7 below.]
2506 - Fixed dftables problem and added an option.
2507 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2508 HAVE_LONG_LONG.
2509 - Added readline support for pcretest.
2510 - Added an listing of the option settings after cmake has run.
2511
2512 4. A user submitted a patch to Makefile that makes it easy to create
2513 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2514 Makefile.am that cause it to include this special target, without
2515 affecting anything else. Note that the same mingw target plus all
2516 the other distribution libraries and programs are now supported
2517 when configuring with CMake (see 6 below) instead of with
2518 Configure/Make.
2519
2520 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2521 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2522 exported in the Windows port". It has not yet been confirmed that the patch
2523 solves the problem, but it does no harm.
2524
2525 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2526 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2527 with CMake, and also correct the comment about stack recursion.
2528
2529 7. Remove the automatic building of pcredemo from the ./configure system and
2530 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2531 of a program that users should build themselves after PCRE is installed, so
2532 building it automatically is not really right. What is more, it gave
2533 trouble in some build environments.
2534
2535 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2536
2537
2538 Version 7.5 10-Jan-08
2539 ---------------------
2540
2541 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2542 values in parens when parsing an RE using the C++ wrapper."
2543
2544 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2545 Characters greater than 255 were excluded from the class instead of being
2546 included.
2547
2548 3. The same bug as (2) above applied to negated POSIX classes such as
2549 [:^space:].
2550
2551 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2552 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2553 I have changed it.
2554
2555 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2556 first named subpattern) and a construct such as (?&a) would reference the
2557 first named subpattern whose name started with "a" (in other words, the
2558 length check was missing). Both these problems are fixed. "Subpattern name
2559 expected" is now given for (?&) (a zero-length name), and this patch also
2560 makes it give the same error for \k'' (previously it complained that that
2561 was a reference to a non-existent subpattern).
2562
2563 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2564 this is right because (?- can be followed by option settings as well as by
2565 digits. I have, however, made the messages clearer.
2566
2567 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2568 than the number used in the conditional) now cause a compile-time error.
2569 This is actually not compatible with Perl, which accepts such patterns, but
2570 treats the conditional as always being FALSE (as PCRE used to), but it
2571 seems to me that giving a diagnostic is better.
2572
2573 8. Change "alphameric" to the more common word "alphanumeric" in comments
2574 and messages.
2575
2576 9. Fix two occurrences of "backslash" in comments that should have been
2577 "backspace".
2578
2579 10. Remove two redundant lines of code that can never be obeyed (their function
2580 was moved elsewhere).
2581
2582 11. The program that makes PCRE's Unicode character property table had a bug
2583 which caused it to generate incorrect table entries for sequences of
2584 characters that have the same character type, but are in different scripts.
2585 It amalgamated them into a single range, with the script of the first of
2586 them. In other words, some characters were in the wrong script. There were
2587 thirteen such cases, affecting characters in the following ranges:
2588
2589 U+002b0 - U+002c1
2590 U+0060c - U+0060d
2591 U+0061e - U+00612
2592 U+0064b - U+0065e
2593 U+0074d - U+0076d
2594 U+01800 - U+01805
2595 U+01d00 - U+01d77
2596 U+01d9b - U+01dbf
2597 U+0200b - U+0200f
2598 U+030fc - U+030fe
2599 U+03260 - U+0327f
2600 U+0fb46 - U+0fbb1
2601 U+10450 - U+1049d
2602
2603 12. The -o option (show only the matching part of a line) for pcregrep was not
2604 compatible with GNU grep in that, if there was more than one match in a
2605 line, it showed only the first of them. It now behaves in the same way as
2606 GNU grep.
2607
2608 13. If the -o and -v options were combined for pcregrep, it printed a blank
2609 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2610 does the same. The return code can be used to tell if there were any
2611 non-matching lines.
2612
2613 14. Added --file-offsets and --line-offsets to pcregrep.
2614
2615 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2616 infinitely looping recursion. The bug was that positive lookaheads were not
2617 being skipped when checking for a possible empty match (negative lookaheads
2618 and both kinds of lookbehind were skipped).
2619
2620 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2621 inclusion of <windows.h> to before rather than after the definition of
2622 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2623
2624 17. Specifying a possessive quantifier with a specific limit for a Unicode
2625 character property caused pcre_compile() to compile bad code, which led at
2626 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2627 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2628 caused the error; without that there was no problem.
2629
2630 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2631
2632 19. Added --enable-pcretest-libreadline.
2633
2634 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2635 RE::GlobalReplace(). As a result, the number of replacements returned was
2636 double what it should be. I removed one of the increments, but Craig sent a
2637 later patch that removed the other one (the right fix) and added unit tests
2638 that check the return values (which was not done before).
2639
2640 21. Several CMake things:
2641
2642 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2643 the names libpcre and libpcreposix, not just pcre and pcreposix.
2644
2645 (2) The above change means that pcretest and pcregrep are now correctly
2646 linked with the newly-built libraries, not previously installed ones.
2647
2648 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2649
2650 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2651 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2652 UTF-8 newline character). The key issue is that the pattern starts .*;
2653 this means that the match must be either at the beginning, or after a
2654 newline. The bug was in the code for advancing after a failed match and
2655 checking that the new position followed a newline. It was not taking
2656 account of UTF-8 characters correctly.
2657
2658 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2659 character classes. PCRE was not treating the sequence [:...:] as a
2660 character class unless the ... were all letters. Perl, however, seems to
2661 allow any characters between [: and :], though of course it rejects as
2662 unknown any "names" that contain non-letters, because all the known class
2663 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2664 for example, whereas PCRE did not - it did not recognize a POSIX character
2665 class. This seemed a bit dangerous, so the code has been changed to be
2666 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2667 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2668 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2669 Perl does, and where it didn't before.
2670
2671 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2672 Windows environments %n is disabled by default.
2673
2674
2675 Version 7.4 21-Sep-07
2676 ---------------------
2677
2678 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2679 means that a class such as [\s] counted as "explicit reference to CR or
2680 LF". That isn't really right - the whole point of the change was to try to
2681 help when there was an actual mention of one of the two characters. So now
2682 the change happens only if \r or \n (or a literal CR or LF) character is
2683 encountered.
2684
2685 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2686 of both had grown to the point where there were only 3 bits left.
2687 Fortunately, there was spare space in the data structure, and so I have
2688 moved the internal flags into a new 16-bit field to free up more option
2689 bits.
2690
2691 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2692 but did not set the internal JCHANGED flag - either of these is enough to
2693 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2694 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2695 start sets both bits.
2696
2697 4. Added options (at build time, compile time, exec time) to change \R from
2698 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2699
2700 5. doc/pcresyntax.html was missing from the distribution.
2701
2702 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2703 compatibility, even though it is no longer used.
2704
2705 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2706 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2707 windows.h file is present (where different names are used). [This was
2708 reversed later after testing - see 16 below.]
2709
2710 8. Changed all #include <config.h> to #include "config.h". There were also
2711 some further <pcre.h> cases that I changed to "pcre.h".
2712
2713 9. When pcregrep was used with the --colour option, it missed the line ending
2714 sequence off the lines that it output.
2715
2716 10. It was pointed out to me that arrays of string pointers cause lots of
2717 relocations when a shared library is dynamically loaded. A technique of
2718 using a single long string with a table of offsets can drastically reduce
2719 these. I have refactored PCRE in four places to do this. The result is
2720 dramatic:
2721
2722 Originally: 290
2723 After changing UCP table: 187
2724 After changing error message table: 43
2725 After changing table of "verbs" 36
2726 After changing table of Posix names 22
2727
2728 Thanks to the folks working on Gregex for glib for this insight.
2729
2730 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2731 unicode-properties was also set.
2732
2733 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2734
2735 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2736 checked only for CRLF.
2737
2738 14. Added casts to pcretest.c to avoid compiler warnings.
2739
2740 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2741
2742 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2743 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2744 entirely. This removes changes made in 7 above.
2745
2746 17. The CMake files have been updated, and there is now more information about
2747 building with CMake in the NON-UNIX-USE document.
2748
2749
2750 Version 7.3 28-Aug-07
2751 ---------------------
2752
2753 1. In the rejigging of the build system that eventually resulted in 7.1, the
2754 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2755 brackets there is not right, since it causes compilers to look for an
2756 installed pcre.h, not the version that is in the source that is being
2757 compiled (which of course may be different). I have changed it back to:
2758
2759 #include "pcre.h"
2760
2761 I have a vague recollection that the change was concerned with compiling in
2762 different directories, but in the new build system, that is taken care of
2763 by the VPATH setting the Makefile.
2764
2765 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2766 when the subject happened to end in the byte 0x85 (e.g. if the last
2767 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2768 characters but of course it shouldn't be taken as a newline when it is part
2769 of another character. The bug was that, for an unlimited repeat of . in
2770 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2771 characters when looking for a newline.
2772
2773 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2774
2775 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2776 in debug output.
2777
2778 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2779 long printing in the pcrecpp unittest when running under MinGW.
2780
2781 6. ESC_K was left out of the EBCDIC table.
2782
2783 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2784 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2785 limit also applies to "virtual nesting" when a pattern is recursive, and in
2786 this case 1000 isn't so big. I have been able to remove this limit at the
2787 expense of backing off one optimization in certain circumstances. Normally,
2788 when pcre_exec() would call its internal match() function recursively and
2789 immediately return the result unconditionally, it uses a "tail recursion"
2790 feature to save stack. However, when a subpattern that can match an empty
2791 string has an unlimited repetition quantifier, it no longer makes this
2792 optimization. That gives it a stack frame in which to save the data for
2793 checking that an empty string has been matched. Previously this was taken
2794 from the 1000-entry workspace that had been reserved. So now there is no
2795 explicit limit, but more stack is used.
2796
2797 8. Applied Daniel's patches to solve problems with the import/export magic
2798 syntax that is required for Windows, and which was going wrong for the
2799 pcreposix and pcrecpp parts of the library. These were overlooked when this
2800 problem was solved for the main library.
2801
2802 9. There were some crude static tests to avoid integer overflow when computing
2803 the size of patterns that contain repeated groups with explicit upper
2804 limits. As the maximum quantifier is 65535, the maximum group length was
2805 set at 30,000 so that the product of these two numbers did not overflow a
2806 32-bit integer. However, it turns out that people want to use groups that
2807 are longer than 30,000 bytes (though not repeat them that many times).
2808 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2809 made it possible to implement the integer overflow checks in a much more
2810 dynamic way, which I have now done. The artificial limitation on group
2811 length has been removed - we now have only the limit on the total length of
2812 the compiled pattern, which depends on the LINK_SIZE setting.
2813
2814 10. Fixed a bug in the documentation for get/copy named substring when
2815 duplicate names are permitted. If none of the named substrings are set, the
2816 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2817 empty string.
2818
2819 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2820 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2821 because the ] is interpreted as the first data character and the
2822 terminating ] is not found. PCRE has been made compatible with Perl in this
2823 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2824 cause memory overwriting.
2825
2826 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2827 string has been matched (to stop an infinite loop). It was not recognizing
2828 a conditional subpattern that could match an empty string if that
2829 subpattern was within another subpattern. For example, it looped when
2830 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2831 condition was not nested. This bug has been fixed.
2832
2833 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2834 past the start of the subject in the presence of bytes with the top bit
2835 set, for example "\x8aBCD".
2836
2837 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2838 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2839
2840 14. Optimized (?!) to (*FAIL).
2841
2842 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2843 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2844 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2845 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2846 does: it's just the validity check that is more restrictive.
2847
2848 16. Inserted checks for integer overflows during escape sequence (backslash)
2849 processing, and also fixed erroneous offset values for syntax errors during
2850 backslash processing.
2851
2852 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2853 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2854
2855 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2856 caused an overrun.
2857
2858 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2859 something other than just ASCII characters) inside a group that had an
2860 unlimited repeat caused a loop at compile time (while checking to see
2861 whether the group could match an empty string).
2862
2863 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2864 [\P{Any}] did so. (Error in the code for printing property names.)
2865
2866 21. An orphan \E inside a character class could cause a crash.
2867
2868 22. A repeated capturing bracket such as (A)? could cause a wild memory
2869 reference during compilation.
2870
2871 23. There are several functions in pcre_compile() that scan along a compiled
2872 expression for various reasons (e.g. to see if it's fixed length for look
2873 behind). There were bugs in these functions when a repeated \p or \P was
2874 present in the pattern. These operators have additional parameters compared
2875 with \d, etc, and these were not being taken into account when moving along
2876 the compiled data. Specifically:
2877
2878 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2879 length.
2880
2881 (b) An item such as \pL+ within a repeated group could cause crashes or
2882 loops.
2883
2884 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2885 "reference to non-existent subpattern" error.
2886
2887 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2888
2889 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2890 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2891
2892 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2893
2894 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2895 character were causing crashes (broken optimization).
2896
2897 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2898 \p or \P) caused a compile-time loop.
2899
2900 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2901 break. For example, the unstudied pattern [\r\n]A does not match the string
2902 "\r\nA" because change 7.0/46 below moves the current point on by two
2903 characters after failing to match at the start. However, the pattern \nA
2904 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2905 the same is true. There doesn't seem any very clean way out of this, but
2906 what I have chosen to do makes the common cases work: PCRE now takes note
2907 of whether there can be an explicit match for \r or \n anywhere in the
2908 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2909 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2910 pattern has explicit CR or LF references.
2911
2912 29. Added (*CR) etc for changing newline setting at start of pattern.
2913
2914
2915 Version 7.2 19-Jun-07
2916 ---------------------
2917
2918 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2919 which is apparently normally available under Windows.
2920
2921 2. Re-jig the pcregrep tests with different newline settings in an attempt
2922 to make them independent of the local environment's newline setting.
2923
2924 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2925
2926 4. Some of the "internals" tests were previously cut out when the link size
2927 was not 2, because the output contained actual offsets. The recent new
2928 "Z" feature of pcretest means that these can be cut out, making the tests
2929 usable with all link sizes.
2930
2931 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2932 stack recursion. This gives a massive performance boost under BSD, but just
2933 a small improvement under Linux. However, it saves one field in the frame
2934 in all cases.
2935
2936 6. Added more features from the forthcoming Perl 5.10:
2937
2938 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2939 recursion call. It refers to the nth most recently opened parentheses.
2940
2941 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2942 to be opened parentheses.
2943
2944 (c) Conditions that refer to capturing parentheses can be specified
2945 relatively, for example, (?(-2)... or (?(+3)...
2946
2947 (d) \K resets the start of the current match so that everything before
2948 is not part of it.
2949
2950 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2951
2952 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2953 reference syntax.
2954
2955 (g) (?| introduces a group in which the numbering of parentheses in each
2956 alternative starts with the same number.
2957
2958 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2959
2960 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2961 PCRE_INFO_JCHANGED.
2962
2963 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2964 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2965 for detecting groups that can match an empty string.
2966
2967 9. A pattern with a very large number of alternatives (more than several
2968 hundred) was running out of internal workspace during the pre-compile
2969 phase, where pcre_compile() figures out how much memory will be needed. A
2970 bit of new cunning has reduced the workspace needed for groups with
2971 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2972 workspace instead of running out of the 4096 that are available.
2973
2974 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2975
2976 11. Applied patch from Google to remove an optimization that didn't quite work.
2977 The report of the bug said:
2978
2979 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2980 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2981 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2982
2983 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2984 it matched the wrong number of bytes.
2985
2986
2987 Version 7.1 24-Apr-07
2988 ---------------------
2989
2990 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2991 that is more "standard", making use of automake and other Autotools. There
2992 is some re-arrangement of the files and adjustment of comments consequent
2993 on this.
2994
2995 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2996 for recursive directory scanning broke on some systems because the files
2997 are not scanned in any specific order and on different systems the order
2998 was different. A call to "sort" has been inserted into RunGrepTest for the
2999 approprate test as a short-term fix. In the longer term there may be an
3000 alternative.
3001
3002 3. I had an email from Eric Raymond about problems translating some of PCRE's
3003 man pages to HTML (despite the fact that I distribute HTML pages, some
3004 people do their own conversions for various reasons). The problems
3005 concerned the use of low-level troff macros .br and .in. I have therefore
3006 removed all such uses from the man pages (some were redundant, some could
3007 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3008 HTML has been updated to handle .nf/.fi and to complain if it encounters
3009 .br or .in.
3010
3011 4. Updated comments in configure.ac that get placed in config.h.in and also
3012 arranged for config.h to be included in the distribution, with the name
3013 config.h.generic, for the benefit of those who have to compile without
3014 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3015
3016 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3017 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3018 makevp.bat for the current PCRE, using the additional files
3019 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3020
3021 6. A Windows user reported a minor discrepancy with test 2, which turned out
3022 to be caused by a trailing space on an input line that had got lost in his
3023 copy. The trailing space was an accident, so I've just removed it.
3024
3025 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3026 that is needed.
3027
3028 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3029 as "const" (a) because they are and (b) because it helps the PHP
3030 maintainers who have recently made a script to detect big data structures
3031 in the php code that should be moved to the .rodata section. I remembered
3032 to update Builducptable as well, so it won't revert if ucptable.h is ever
3033 re-created.
3034
3035 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3036 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3037 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3038 support is not required. This saves 1.5-2K of code, which is important in
3039 some applications.
3040
3041 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3042 so as not to refer to the tables, even though these functions will never be
3043 called when UTF-8 support is disabled. Otherwise there are problems with a
3044 shared library.
3045
3046 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3047
3048 (a) It was defining its arguments as char * instead of void *.
3049
3050 (b) It was assuming that all moves were upwards in memory; this was true
3051 a long time ago when I wrote it, but is no longer the case.
3052
3053 The emulated memove() is provided for those environments that have neither
3054 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3055 is clearly not the case, as these two bugs were recently reported.
3056
3057 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3058 and Detrail to create the HTML documentation, the .txt form of the man
3059 pages, and it removes trailing spaces from listed files. It also creates
3060 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3061 case, it wraps all the #defines with #ifndefs. This script should be run
3062 before "make dist".
3063
3064 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3065 with Unicode property support.
3066
3067 (a) For a maximizing quantifier, if the two different cases of the
3068 character were of different lengths in their UTF-8 codings (there are
3069 some cases like this - I found 11), and the matching function had to
3070 back up over a mixture of the two cases, it incorrectly assumed they
3071 were both the same length.
3072
3073 (b) When PCRE was configured to use the heap rather than the stack for
3074 recursion during matching, it was not correctly preserving the data for
3075 the other case of a UTF-8 character when checking ahead for a match
3076 while processing a minimizing repeat. If the check also involved
3077 matching a wide character, but failed, corruption could cause an
3078 erroneous result when trying to check for a repeat of the original
3079 character.
3080
3081 13. Some tidying changes to the testing mechanism:
3082
3083 (a) The RunTest script now detects the internal link size and whether there
3084 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3085 values substituted by "configure". (The RunGrepTest script already did
3086 this for UTF-8.) The configure.ac script no longer substitutes the
3087 relevant variables.
3088
3089 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3090 with length and offset values. This means that the output is different
3091 for different internal link sizes. Test 2 is skipped for link sizes
3092 other than 2 because of this, bypassing the problem. Unfortunately,
3093 there was also a test in test 3 (the locale tests) that used /B and
3094 failed for link sizes other than 2. Rather than cut the whole test out,
3095 I have added a new /Z option to pcretest that replaces the length and
3096 offset values with spaces. This is now used to make test 3 independent
3097 of link size. (Test 2 will be tidied up later.)
3098
3099 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3100 segmentation fault instead of returning the appropriate error message.
3101
3102 15. In multiline mode when the newline sequence was set to "any", the pattern
3103 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3104 This doesn't seem right; it now treats the CRLF combination as the line
3105 ending, and so does not match in that case. It's only a pattern such as ^$
3106 that would hit this one: something like ^ABC$ would have failed after \r
3107 and then tried again after \r\n.
3108
3109 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3110 in an attempt to make files that differ only in their line terminators
3111 compare equal. This works on Linux.
3112
3113 17. Under certain error circumstances pcregrep might try to free random memory
3114 as it exited. This is now fixed, thanks to valgrind.
3115
3116 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3117 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3118 was because its rules for how to advance for /g after matching an empty
3119 string at the end of a line did not allow for this case. They now check for
3120 it specially.
3121
3122 20. pcretest is supposed to handle patterns and data of any length, by
3123 extending its buffers when necessary. It was getting this wrong when the
3124 buffer for a data line had to be extended.
3125
3126 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3127 CRLF as a newline sequence.
3128
3129 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3130 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3131 I have nevertheless tidied it up.
3132
3133 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3134
3135 24. Added a man page for pcre-config.
3136
3137
3138 Version 7.0 19-Dec-06
3139 ---------------------
3140
3141 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3142 moving to gcc 4.1.1.
3143
3144 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3145 sys/time.h, which is documented as needed for this function. It doesn't
3146 seem to matter on Linux, but it showed up on some releases of OS X.
3147
3148 3. It seems that there are systems where bytes whose values are greater than
3149 127 match isprint() in the "C" locale. The "C" locale should be the
3150 default when a C program starts up. In most systems, only ASCII printing
3151 characters match isprint(). This difference caused the output from pcretest
3152 to vary, making some of the tests fail. I have changed pcretest so that:
3153
3154 (a) When it is outputting text in the compiled version of a pattern, bytes
3155 other than 32-126 are always shown as hex escapes.
3156
3157 (b) When it is outputting text that is a matched part of a subject string,
3158 it does the same, unless a different locale has been set for the match
3159 (using the /L modifier). In this case, it uses isprint() to decide.
3160
3161 4. Fixed a major bug that caused incorrect computation of the amount of memory
3162 required for a compiled pattern when options that changed within the
3163 pattern affected the logic of the preliminary scan that determines the
3164 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3165 that the computed length was too small. The symptoms of this bug were
3166 either the PCRE error "internal error: code overflow" from pcre_compile(),
3167 or a glibc crash with a message such as "pcretest: free(): invalid next
3168 size (fast)". Examples of patterns that provoked this bug (shown in
3169 pcretest format) are:
3170
3171 /(?-x: )/x
3172 /(?x)(?-x: \s*#\s*)/
3173 /((?i)[\x{c0}])/8
3174 /(?i:[\x{c0}])/8
3175
3176 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3177 is now done differently.
3178
3179 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3180 wrapper classes; (b) implement a new function in the C++ scanner that is
3181 more efficient than the old way of doing things because it avoids levels of
3182 recursion in the regex matching; (c) add a paragraph to the documentation
3183 for the FullMatch() function.
3184
3185 6. The escape sequence \n was being treated as whatever was defined as
3186 "newline". Not only was this contrary to the documentation, which states
3187 that \n is character 10 (hex 0A), but it also went horribly wrong when
3188 "newline" was defined as CRLF. This has been fixed.
3189
3190 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3191 was being set to -1 for the "end of line" case (supposedly a value that no
3192 character can have). Though this value is never used (the check for end of
3193 line is "zero bytes in current character"), it caused compiler complaints.
3194 I've changed it to 0xffffffff.
3195
3196 8. In pcre_version.c, the version string was being built by a sequence of
3197 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3198 string (as it is for production releases) called a macro with an empty
3199 argument. The C standard says the result of this is undefined. The gcc
3200 compiler treats it as an empty string (which was what was wanted) but it is
3201 reported that Visual C gives an error. The source has been hacked around to
3202 avoid this problem.
3203
3204 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3205 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3206 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3207 of them did).
3208
3209 10. Originally, pcretest opened its input and output without "b"; then I was
3210 told that "b" was needed in some environments, so it was added for release
3211 5.0 to both the input and output. (It makes no difference on Unix-like
3212 systems.) Later I was told that it is wrong for the input on Windows. I've
3213 now abstracted the modes into two macros, to make it easier to fiddle with
3214 them, and removed "b" from the input mode under Windows.
3215
3216 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3217
3218 12. Added -help and --help to pcretest as an official way of being reminded
3219 of the options.
3220
3221 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3222 and pcrecpp.cc because they annoy compilers at high warning levels.
3223
3224 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3225
3226 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3227 scripts are not C programs :-) and which was not noticed because it works
3228 on Linux.
3229
3230 16. pcretest is supposed to handle any length of pattern and data line (as one
3231 line or as a continued sequence of lines) by extending its input buffer if
3232 necessary. This feature was broken for very long pattern lines, leading to
3233 a string of junk being passed to pcre_compile() if the pattern was longer
3234 than about 50K.
3235
3236 17. I have done a major re-factoring of the way pcre_compile() computes the
3237 amount of memory needed for a compiled pattern. Previously, there was code
3238 that made a preliminary scan of the pattern in order to do this. That was
3239 OK when PCRE was new, but as the facilities have expanded, it has become
3240 harder and harder to keep it in step with the real compile phase, and there
3241 have been a number of bugs (see for example, 4 above). I have now found a
3242 cunning way of running the real compile function in a "fake" mode that
3243 enables it to compute how much memory it would need, while actually only
3244 ever using a few hundred bytes of working memory and without too many
3245 tests of the mode. This should make future maintenance and development
3246 easier. A side effect of this work is that the limit of 200 on the nesting
3247 depth of parentheses has been removed (though this was never a serious
3248 limitation, I suspect). However, there is a downside: pcre_compile() now
3249 runs more slowly than before (30% or more, depending on the pattern). I
3250 hope this isn't a big issue. There is no effect on runtime performance.
3251
3252 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3253 newline (only possible for the last line of a file) and it was a
3254 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3255
3256 19. Added additional timing features to pcretest. (1) The -tm option now times
3257 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3258 separate command line item, by a number that specifies the number of
3259 repeats to use when timing. The default is 50000; this gives better
3260 precision, but takes uncomfortably long for very large patterns.
3261
3262 20. Extended pcre_study() to be more clever in cases where a branch of a
3263 subpattern has no definite first character. For example, (a*|b*)[cd] would
3264 previously give no result from pcre_study(). Now it recognizes that the
3265 first character must be a, b, c, or d.
3266
3267 21. There was an incorrect error "recursive call could loop indefinitely" if
3268 a subpattern (or the entire pattern) that was being tested for matching an
3269 empty string contained only one non-empty item after a nested subpattern.
3270 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3271 incorrectly, because the \d was being skipped in the check.
3272
3273 22. The pcretest program now has a new pattern option /B and a command line
3274 option -b, which is equivalent to adding /B to every pattern. This causes
3275 it to show the compiled bytecode, without the additional information that
3276 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3277 is the same as /B/I).
3278
3279 23. A new optimization is now able automatically to treat some sequences such
3280 as a*b as a*+b. More specifically, if something simple (such as a character
3281 or a simple class like \d) has an unlimited quantifier, and is followed by
3282 something that cannot possibly match the quantified thing, the quantifier
3283 is automatically "possessified".
3284
3285 24. A recursive reference to a subpattern whose number was greater than 39
3286 went wrong under certain circumstances in UTF-8 mode. This bug could also
3287 have affected the operation of pcre_study().
3288
3289 25. Realized that a little bit of performance could be had by replacing
3290 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3291
3292 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3293
3294 27. Possessive quantifiers such as a++ were previously implemented by turning
3295 them into atomic groups such as ($>a+). Now they have their own opcodes,
3296 which improves performance. This includes the automatically created ones
3297 from 23 above.
3298
3299 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3300 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3301 the first matched character to be a colon. This applied both to named and
3302 numbered groups.
3303
3304 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3305
3306 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3307 building PCRE on Windows easier, so I have included it in the distribution.
3308
3309 31. There is now a check in pcretest against a ridiculously large number being
3310 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3311 loop, the loop is abandoned.
3312
3313 32. Forward references to subpatterns in conditions such as (?(2)...) where
3314 subpattern 2 is defined later cause pcre_compile() to search forwards in
3315 the pattern for the relevant set of parentheses. This search went wrong
3316 when there were unescaped parentheses in a character class, parentheses
3317 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3318
3319 33. "Subroutine" calls and backreferences were previously restricted to
3320 referencing subpatterns earlier in the regex. This restriction has now
3321 been removed.
3322
3323 34. Added a number of extra features that are going to be in Perl 5.10. On the
3324 whole, these are just syntactic alternatives for features that PCRE had
3325 previously implemented using the Python syntax or my own invention. The
3326 other formats are all retained for compatibility.
3327
3328 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3329 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3330 also .NET compatible.
3331
3332 (b) A recursion or subroutine call to a named group can now be defined as
3333 (?&name) as well as (?P>name).
3334
3335 (c) A backreference to a named group can now be defined as \k<name> or
3336 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3337 5.10, are also .NET compatible.
3338
3339 (d) A conditional reference to a named group can now use the syntax
3340 (?(<name>) or (?('name') as well as (?(name).
3341
3342 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3343 groups (named and numbered) that are never evaluated inline, but can be
3344 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3345 is always false. There may be only one alternative in such a group.
3346
3347 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3348 as the simple (?(R). The condition is true only if the most recent
3349 recursion is that of the given number or name. It does not search out
3350 through the entire recursion stack.
3351
3352 (g) The escape \gN or \g{N} has been added, where N is a positive or
3353 negative number, specifying an absolute or relative reference.
3354
3355 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3356 some "unreachable code" warnings.
3357
3358 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3359 things, this adds five new scripts.
3360
3361 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3362 There were also incompatibilities regarding the handling of \Q..\E inside
3363 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3364 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3365
3366 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3367 matches an empty string, and forcibly breaks the loop. There were bugs in
3368 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3369 against aaaa the result was just "a" rather than "aaaa", for example. Two
3370 separate and independent bugs (that affected different cases) have been
3371 fixed.
3372
3373 39. Refactored the code to abolish the use of different opcodes for small
3374 capturing bracket numbers. This is a tidy that I avoided doing when I
3375 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3376 The new approach is not only tidier, it makes it possible to reduce the
3377 memory needed to fix the previous bug (38).
3378
3379 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3380 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3381 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3382 mode.
3383
3384 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3385 report.
3386
3387 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3388 copy construction and assignment in the C++ wrapper.
3389
3390 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3391 couple of bugs that could have given wrong results in the "--newline=crlf"
3392 case.
3393
3394 44. Added a number of casts and did some reorganization of signed/unsigned int
3395 variables following suggestions from Dair Grant. Also renamed the variable
3396 "this" as "item" because it is a C++ keyword.
3397
3398 45. Arranged for dftables to add
3399
3400 #include "pcre_internal.h"
3401
3402 to pcre_chartables.c because without it, gcc 4.x may remove the array
3403 definition from the final binary if PCRE is built into a static library and
3404 dead code stripping is activated.
3405
3406 46. For an unanchored pattern, if a match attempt fails at the start of a
3407 newline sequence, and the newline setting is CRLF or ANY, and the next two
3408 characters are CRLF, advance by two characters instead of one.
3409
3410
3411 Version 6.7 04-Jul-06
3412 ---------------------
3413
3414 1. In order to handle tests when input lines are enormously long, pcretest has
3415 been re-factored so that it automatically extends its buffers when
3416 necessary. The code is crude, but this _is_ just a test program. The
3417 default size has been increased from 32K to 50K.
3418
3419 2. The code in pcre_study() was using the value of the re argument before
3420 testing it for NULL. (Of course, in any sensible call of the function, it
3421 won't be NULL.)
3422
3423 3. The memmove() emulation function in pcre_internal.h, which is used on
3424 systems that lack both memmove() and bcopy() - that is, hardly ever -
3425 was missing a "static" storage class specifier.
3426
3427 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3428 containing an extended class (one that cannot be represented by a bitmap
3429 because it contains high-valued characters or Unicode property items, e.g.
3430 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3431 pattern, but PCRE should not loop if you do not (it no longer does).
3432 [Detail: two cases were found: (a) a repeated subpattern containing an
3433 extended class; (b) a recursive reference to a subpattern that followed a
3434 previous extended class. It wasn't skipping over the extended class
3435 correctly when UTF-8 mode was not set.]
3436
3437 5. A negated single-character class was not being recognized as fixed-length
3438 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3439 compile error "lookbehind assertion is not fixed length".
3440
3441 6. The RunPerlTest auxiliary script was showing an unexpected difference
3442 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3443 write a Perl script that can interpret lines of an input file either as
3444 byte characters or as UTF-8, which is what "perltest" was being required to
3445 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3446 can't do is switch easily at run time between having the "use utf8;" pragma
3447 or not. In the end, I fudged it by using the RunPerlTest script to insert
3448 "use utf8;" explicitly for the UTF-8 tests.
3449
3450 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3451 the end of the subject string, contrary to the documentation and to what
3452 Perl does. This was true of both matching functions. Now it matches only at
3453 the start of the subject and immediately after *internal* newlines.
3454
3455 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3456 a pointer to an int instead of a pointer to an unsigned long int. This
3457 caused problems on 64-bit systems.
3458
3459 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3460 instance of the 'standard' template library not being so standard".
3461
3462 10. There was no check on the number of named subpatterns nor the maximum
3463 length of a subpattern name. The product of these values is used to compute
3464 the size of the memory block for a compiled pattern. By supplying a very
3465 long subpattern name and a large number of named subpatterns, the size
3466 computation could be caused to overflow. This is now prevented by limiting
3467 the length of names to 32 characters, and the number of named subpatterns
3468 to 10,000.
3469
3470 11. Subpatterns that are repeated with specific counts have to be replicated in
3471 the compiled pattern. The size of memory for this was computed from the
3472 length of the subpattern and the repeat count. The latter is limited to
3473 65535, but there was no limit on the former, meaning that integer overflow
3474 could in principle occur. The compiled length of a repeated subpattern is
3475 now limited to 30,000 bytes in order to prevent this.
3476
3477 12. Added the optional facility to have named substrings with the same name.
3478
3479 13. Added the ability to use a named substring as a condition, using the
3480 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3481 are numbers (not recommended). Forward references are permitted.
3482
3483 14. Added forward references in named backreferences (if you see what I mean).
3484
3485 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3486 pattern could run off the end of the subject. For example, the pattern
3487 "(?s)(.{1,5})"8 did this with the subject "ab".
3488
3489 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3490 PCRE_CASELESS was set when matching characters that were quantified with ?
3491 or *.
3492
3493 17. A character class other than a single negated character that had a minimum
3494 but no maximum quantifier - for example [ab]{6,} - was not handled
3495 correctly by pce_dfa_exec(). It would match only one character.
3496
3497 18. A valid (though odd) pattern that looked like a POSIX character
3498 class but used an invalid character after [ (for example [[,abc,]]) caused
3499 pcre_compile() to give the error "Failed: internal error: code overflow" or
3500 in some cases to crash with a glibc free() error. This could even happen if
3501 the pattern terminated after [[ but there just happened to be a sequence of
3502 letters, a binary zero, and a closing ] in the memory that followed.
3503
3504 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3505 over the years. Originally (before any Unicode support), just the bottom 8
3506 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3507 output from "man perlunicode" includes this:
3508
3509 The regular expression compiler produces polymorphic opcodes. That
3510 is, the pattern adapts to the data and automatically switches to
3511 the Unicode character scheme when presented with Unicode data--or
3512 instead uses a traditional byte scheme when presented with byte
3513 data.
3514
3515 Sadly, a wide octal escape does not cause a switch, and in a string with
3516 no other multibyte characters, these octal escapes are treated as before.
3517 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3518 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3519 Unicode string.
3520
3521 I have not perpetrated such confusion in PCRE. Up till now, it took just
3522 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3523 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3524 translate to the appropriate multibyte character.
3525
3526 29. Applied some refactoring to reduce the number of warnings from Microsoft
3527 and Borland compilers. This has included removing the fudge introduced
3528 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3529 a warning about an unused variable.
3530
3531 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3532 characters since release 4.0, because Perl (from release 5.004) does not.
3533 [Or at least, is documented not to: some releases seem to be in conflict
3534 with the documentation.] However, when a pattern was studied with
3535 pcre_study() and all its branches started with \s, PCRE still included VT
3536 as a possible starting character. Of course, this did no harm; it just
3537 caused an unnecessary match attempt.
3538
3539 22. Removed a now-redundant internal flag bit that recorded the fact that case
3540 dependency changed within the pattern. This was once needed for "required
3541 byte" processing, but is no longer used. This recovers a now-scarce options
3542 bit. Also moved the least significant internal flag bit to the most-
3543 significant bit of the word, which was not previously used (hangover from
3544 the days when it was an int rather than a uint) to free up another bit for
3545 the future.
3546
3547 23. Added support for CRLF line endings as well as CR and LF. As well as the
3548 default being selectable at build time, it can now be changed at runtime
3549 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3550 specify that it is scanning data with non-default line endings.
3551
3552 24. Changed the definition of CXXLINK to make it agree with the definition of
3553 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3554
3555 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3556 recursions. This makes a big different to stack usage for some patterns.
3557
3558 26. If a subpattern containing a named recursion or subroutine reference such
3559 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3560 the space required for the compiled pattern went wrong and gave too small a
3561 value. Depending on the environment, this could lead to "Failed: internal
3562 error: code overflow at offset 49" or "glibc detected double free or
3563 corruption" errors.
3564
3565 27. Applied patches from Google (a) to support the new newline modes and (b) to
3566 advance over multibyte UTF-8 characters in GlobalReplace.
3567
3568 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3569 difference for some implementation of PCRE in some Windows version.
3570
3571 29. Added some extra testing facilities to pcretest:
3572
3573 \q<number> in a data line sets the "match limit" value
3574 \Q<number> in a data line sets the "match recursion limt" value
3575 -S <number> sets the stack size, where <number> is in megabytes
3576
3577 The -S option isn't available for Windows.
3578
3579
3580 Version 6.6 06-Feb-06
3581 ---------------------
3582
3583 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3584 in pcreposix.h. I have copied the definition from pcre.h.
3585
3586 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3587 because pcre.h is no longer a built file.
3588
3589 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3590 not normally included in the compiled code.
3591
3592
3593 Version 6.5 01-Feb-06
3594 ---------------------
3595
3596 1. When using the partial match feature with pcre_dfa_exec(), it was not
3597 anchoring the second and subsequent partial matches at the new starting
3598 point. This could lead to incorrect results. For example, with the pattern
3599 /1234/, partially matching against "123" and then "a4" gave a match.
3600
3601 2. Changes to pcregrep:
3602
3603 (a) All non-match returns from pcre_exec() were being treated as failures
3604 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3605 error message is output. Some extra information is given for the
3606 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3607 probably the only errors that are likely to be caused by users (by
3608 specifying a regex that has nested indefinite repeats, for instance).
3609 If there are more than 20 of these errors, pcregrep is abandoned.
3610
3611 (b) A binary zero was treated as data while matching, but terminated the
3612 output line if it was written out. This has been fixed: binary zeroes
3613 are now no different to any other data bytes.
3614
3615 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3616 used to set a locale for matching. The --locale=xxxx long option has
3617 been added (no short equivalent) to specify a locale explicitly on the
3618 pcregrep command, overriding the environment variables.
3619
3620 (d) When -B was used with -n, some line numbers in the output were one less
3621 than they should have been.
3622
3623 (e) Added the -o (--only-matching) option.
3624
3625 (f) If -A or -C was used with -c (count only), some lines of context were
3626 accidentally printed for the final match.
3627
3628 (g) Added the -H (--with-filename) option.
3629
3630 (h) The combination of options -rh failed to suppress file names for files
3631 that were found from directory arguments.
3632
3633 (i) Added the -D (--devices) and -d (--directories) options.
3634
3635 (j) Added the -F (--fixed-strings) option.
3636
3637 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3638
3639 (l) Added the --colo(u)r option.
3640
3641 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3642 is not present by default.
3643
3644 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3645 items such as (?R) or (?1), when the recursion could match a number of
3646 alternatives. If it matched one of the alternatives, but subsequently,
3647 outside the recursion, there was a failure, the code tried to back up into
3648 the recursion. However, because of the way PCRE is implemented, this is not
3649 possible, and the result was an incorrect result from the match.
3650
3651 In order to prevent this happening, the specification of recursion has
3652 been changed so that all such subpatterns are automatically treated as
3653 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3654
3655 4. I had overlooked the fact that, in some locales, there are characters for
3656 which isalpha() is true but neither isupper() nor islower() are true. In
3657 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3658 and ordfeminine) are like this. This affected the treatment of \w and \W
3659 when they appeared in character classes, but not when they appeared outside
3660 a character class. The bit map for "word" characters is now created
3661 separately from the results of isalnum() instead of just taking it from the
3662 upper, lower, and digit maps. (Plus the underscore character, of course.)
3663
3664 5. The above bug also affected the handling of POSIX character classes such as
3665 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3666 permanent tables. Instead, the bit maps for such a class were previously
3667 created as the appropriate unions of the upper, lower, and digit bitmaps.
3668 Now they are created by subtraction from the [[:word:]] class, which has
3669 its own bitmap.
3670
3671 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3672 It is created by subtracting the vertical space characters (\x09, \x0a,
3673 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3674 subtraction was done in the overall bitmap for a character class, meaning
3675 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3676 be recognized. This bug has been fixed.
3677
3678 7. Patches from the folks at Google:
3679
3680 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3681 real life, but is still worth protecting against".
3682
3683 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3684 regular expressions".
3685
3686 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3687 have it.
3688
3689 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3690 "configure" and the latter not, in order to fix a problem somebody had
3691 with compiling the Arg class on HP-UX.
3692
3693 (e) Improve the error-handling of the C++ wrapper a little bit.
3694
3695 (f) New tests for checking recursion limiting.
3696
3697 8. The pcre_memmove() function, which is used only if the environment does not
3698 have a standard memmove() function (and is therefore rarely compiled),
3699 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3700 returning a result (though PCRE never actually uses the result).
3701
3702 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3703 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3704 returned instead of calling malloc() with an overflowing number that would
3705 most likely cause subsequent chaos.
3706
3707 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3708
3709 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3710 with this option is matched, the nmatch and pmatch options of regexec() are
3711 ignored.
3712
3713 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3714 provided in case anyone wants to the the POSIX interface with UTF-8
3715 strings.
3716
3717 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3718 C++ linking (needed for some HP-UX environments).
3719
3720 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3721 (unused parameter) and in the pcre_printint() function (omitted "default"
3722 switch label when the default is to do nothing).
3723
3724 15. Added some code to make it possible, when PCRE is compiled as a C++
3725 library, to replace subject pointers for pcre_exec() with a smart pointer
3726 class, thus making it possible to process discontinuous strings.
3727
3728 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3729 much the same function. They were added by different people who were trying
3730 to make PCRE easy to compile on non-Unix systems. It has been suggested
3731 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3732 for compiling on Windows systems. I have therefore replaced it with
3733 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3734 defaults to "extern" for C or "extern C" for C++, which works fine on
3735 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3736 SCOPE with something explicit in config.h. In addition:
3737
3738 (a) pcreposix.h still had just "extern" instead of either of these macros;
3739 I have replaced it with PCRE_DATA_SCOPE.
3740
3741 (b) Functions such as _pcre_xclass(), which are internal to the library,
3742 but external in the C sense, all had PCRE_EXPORT in their definitions.
3743 This is apparently wrong for the Windows case, so I have removed it.
3744 (It makes no difference on Unix-like systems.)
3745
3746 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3747 of recursive calls to match(). This is different to MATCH_LIMIT because
3748 that limits the total number of calls to match(), not all of which increase
3749 the depth of recursion. Limiting the recursion depth limits the amount of
3750 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3751 when PCRE is compiled, and changed at run time. A patch from Google adds
3752 this functionality to the C++ interface.
3753
3754 18. Changes to the handling of Unicode character properties:
3755
3756 (a) Updated the table to Unicode 4.1.0.
3757
3758 (b) Recognize characters that are not in the table as "Cn" (undefined).
3759
3760 (c) I revised the way the table is implemented to a much improved format
3761 which includes recognition of ranges. It now supports the ranges that
3762 are defined in UnicodeData.txt, and it also amalgamates other
3763 characters into ranges. This has reduced the number of entries in the
3764 table from around 16,000 to around 3,000, thus reducing its size
3765 considerably. I realized I did not need to use a tree structure after
3766 all - a binary chop search is just as efficient. Having reduced the
3767 number of entries, I extended their size from 6 bytes to 8 bytes to
3768 allow for more data.
3769
3770 (d) Added support for Unicode script names via properties such as \p{Han}.
3771
3772 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3773 matching that character.
3774
3775 20. When matching a repeated Unicode property with a minimum greater than zero,
3776 (for example \pL{2,}), PCRE could look past the end of the subject if it
3777 reached it while seeking the minimum number of characters. This could
3778 happen only if some of the characters were more than one byte long, because
3779 there is a check for at least the minimum number of bytes.
3780
3781 21. Refactored the implementation of \p and \P so as to be more general, to
3782 allow for more different types of property in future. This has changed the
3783 compiled form incompatibly. Anybody with saved compiled patterns that use
3784 \p or \P will have to recompile them.
3785
3786 22. Added "Any" and "L&" to the supported property types.
3787
3788 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3789 but give a compile time error if the value is greater than 0xff.
3790
3791 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3792 accidentally not being installed or uninstalled.
3793
3794 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3795 made were to insert the current release number. This seemed silly, because
3796 it made things harder for people building PCRE on systems that don't run
3797 "configure". I have turned pcre.h into a distributed file, no longer built
3798 by "configure", with the version identification directly included. There is
3799 no longer a pcre.h.in file.
3800
3801 However, this change necessitated a change to the pcre-config script as
3802 well. It is built from pcre-config.in, and one of the substitutions was the
3803 release number. I have updated configure.ac so that ./configure now finds
3804 the release number by grepping pcre.h.
3805
3806 26. Added the ability to run the tests under valgrind.
3807
3808
3809 Version 6.4 05-Sep-05
3810 ---------------------
3811
3812 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3813 "--" to be printed when multiple files were scanned, even when none of the
3814 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3815 consider it to be a bug, and have restored the previous behaviour.
3816
3817 2. A couple of code tidies to get rid of compiler warnings.
3818
3819 3. The pcretest program used to cheat by referring to symbols in the library
3820 whose names begin with _pcre_. These are internal symbols that are not
3821 really supposed to be visible externally, and in some environments it is
3822 possible to suppress them. The cheating is now confined to including
3823 certain files from the library's source, which is a bit cleaner.
3824
3825 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3826 file's purpose clearer.
3827
3828 5. Reorganized pcre_ucp_findchar().
3829
3830
3831 Version 6.3 15-Aug-05
3832 ---------------------
3833
3834 1. The file libpcre.pc.in did not have general read permission in the tarball.
3835
3836 2. There were some problems when building without C++ support:
3837
3838 (a) If C++ support was not built, "make install" and "make test" still
3839 tried to test it.
3840
3841 (b) There were problems when the value of CXX was explicitly set. Some
3842 changes have been made to try to fix these, and ...
3843
3844 (c) --disable-cpp can now be used to explicitly disable C++ support.
3845
3846 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3847 backslash in a target when C++ was disabled. This confuses some
3848 versions of "make", apparently. Using an intermediate variable solves
3849 this. (Same for CPP_LOBJ.)
3850
3851 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3852 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3853 necessary on certain architectures.
3854
3855 4. Added a setting of -export-symbols-regex to the link command to remove
3856 those symbols that are exported in the C sense, but actually are local
3857 within the library, and not documented. Their names all begin with
3858 "_pcre_". This is not a perfect job, because (a) we have to except some
3859 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3860 available (and never for static libraries). I have made a note to try to
3861 find a way round (a) in the future.
3862
3863
3864 Version 6.2 01-Aug-05
3865 ---------------------
3866
3867 1. There was no test for integer overflow of quantifier values. A construction
3868 such as {1111111111111111} would give undefined results. What is worse, if
3869 a minimum quantifier for a parenthesized subpattern overflowed and became
3870 negative, the calculation of the memory size went wrong. This could have
3871 led to memory overwriting.
3872
3873 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3874
3875 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3876 operating environments where this matters.
3877
3878 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3879 PCRE options from within the C++ wrapper.
3880
3881 5. Named capturing subpatterns were not being correctly counted when a pattern
3882 was compiled. This caused two problems: (a) If there were more than 100
3883 such subpatterns, the calculation of the memory needed for the whole
3884 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3885 back references of the form \12, where the number was greater than 9, were
3886 not recognized as back references, even though there were sufficient
3887 previous subpatterns.
3888
3889 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3890 versions of gcc, e.g. 2.95.4.
3891
3892
3893 Version 6.1 21-Jun-05
3894 ---------------------
3895
3896 1. There was one reference to the variable "posix" in pcretest.c that was not
3897 surrounded by "#if !defined NOPOSIX".
3898
3899 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3900 the cross-check on the old pcre_info() function, for the benefit of the
3901 cut-down version of PCRE that is currently imported into Exim.
3902
3903 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3904 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3905 patterns that start (?i) but I don't think that matters. The use of (?i) is
3906 just an example; this all applies to the other options as well.
3907
3908 4. Since libtool seems to echo the compile commands it is issuing, the output
3909 from "make" can be reduced a bit by putting "@" in front of each libtool
3910 compile command.
3911
3912 5. Patch from the folks at Google for configure.in to be a bit more thorough
3913 in checking for a suitable C++ installation before trying to compile the
3914 C++ stuff. This should fix a reported problem when a compiler was present,
3915 but no suitable headers.
3916
3917 6. The man pages all had just "PCRE" as their title. I have changed them to
3918 be the relevant file name. I have also arranged that these names are
3919 retained in the file doc/pcre.txt, which is a concatenation in text format
3920 of all the man pages except the little individual ones for each function.
3921
3922 7. The NON-UNIX-USE file had not been updated for the different set of source
3923 files that come with release 6. I also added a few comments about the C++
3924 wrapper.
3925
3926
3927 Version 6.0 07-Jun-05
3928 ---------------------
3929
3930 1. Some minor internal re-organization to help with my DFA experiments.
3931
3932 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3933 didn't matter for the library itself when fully configured, but did matter
3934 when compiling without UCP support, or within Exim, where the ucp files are
3935 not imported.
3936
3937 3. Refactoring of the library code to split up the various functions into
3938 different source modules. The addition of the new DFA matching code (see
3939 below) to a single monolithic source would have made it really too
3940 unwieldy, quite apart from causing all the code to be include in a
3941 statically linked application, when only some functions are used. This is
3942 relevant even without the DFA addition now that patterns can be compiled in
3943 one application and matched in another.
3944
3945 The downside of splitting up is that there have to be some external
3946 functions and data tables that are used internally in different modules of
3947 the library but which are not part of the API. These have all had their
3948 names changed to start with "_pcre_" so that they are unlikely to clash
3949 with other external names.
3950
3951 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3952 a different (DFA) algorithm. Although it is slower than the original
3953 function, it does have some advantages for certain types of matching
3954 problem.
3955
3956 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3957 including restarting after a partial match.
3958
3959 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3960 defined when compiling for Windows was sent to me. I have put it into the
3961 code, though I have no means of testing or verifying it.
3962
3963 7. Added the pcre_refcount() auxiliary function.
3964
3965 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3966 match before or at the first newline in the subject string. In pcretest,
3967 the /f option on a pattern can be used to set this.
3968
3969 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3970 would behave wrongly. This has been present in PCRE since release 4.0.
3971
3972 10. A number of changes to the pcregrep command:
3973
3974 (a) Refactored how -x works; insert ^(...)$ instead of setting
3975 PCRE_ANCHORED and checking the length, in preparation for adding
3976 something similar for -w.
3977
3978 (b) Added the -w (match as a word) option.
3979
3980 (c) Refactored the way lines are read and buffered so as to have more
3981 than one at a time available.
3982
3983 (d) Implemented a pcregrep test script.
3984
3985 (e) Added the -M (multiline match) option. This allows patterns to match
3986 over several lines of the subject. The buffering ensures that at least
3987 8K, or the rest of the document (whichever is the shorter) is available
3988 for matching (and similarly the previous 8K for lookbehind assertions).
3989
3990 (f) Changed the --help output so that it now says
3991
3992 -w, --word-regex(p)
3993
3994 instead of two lines, one with "regex" and the other with "regexp"
3995 because that confused at least one person since the short forms are the
3996 same. (This required a bit of code, as the output is generated
3997 automatically from a table. It wasn't just a text change.)
3998
3999 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4000 option but starts with a hyphen. Could be a pattern or a path name
4001 starting with a hyphen, for instance.
4002
4003 (h) "-" can be given as a file name to represent stdin.
4004
4005 (i) When file names are being printed, "(standard input)" is used for
4006 the standard input, for compatibility with GNU grep. Previously
4007 "<stdin>" was used.
4008
4009 (j) The option --label=xxx can be used to supply a name to be used for
4010 stdin when file names are being printed. There is no short form.
4011
4012 (k) Re-factored the options decoding logic because we are going to add
4013 two more options that take data. Such options can now be given in four
4014 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4015
4016 (l) Added the -A, -B, and -C options for requesting that lines of context
4017 around matches be printed.
4018
4019 (m) Added the -L option to print the names of files that do not contain
4020 any matching lines, that is, the complement of -l.
4021
4022 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4023 continue to scan other files.
4024
4025 (o) The -s option was incorrectly implemented. For compatibility with other
4026 greps, it now suppresses the error message for a non-existent or non-
4027 accessible file (but not the return code). There is a new option called
4028 -q that suppresses the output of matching lines, which was what -s was
4029 previously doing.
4030
4031 (p) Added --include and --exclude options to specify files for inclusion
4032 and exclusion when recursing.
4033
4034 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4035 Hopefully, it now does.
4036
4037 12. Missing cast in pcre_study().
4038
4039 13. Added an "uninstall" target to the makefile.
4040
4041 14. Replaced "extern" in the function prototypes in Makefile.in with
4042 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4043 world, but is set differently for Windows.
4044
4045 15. Added a second compiling function called pcre_compile2(). The only
4046 difference is that it has an extra argument, which is a pointer to an
4047 integer error code. When there is a compile-time failure, this is set
4048 non-zero, in addition to the error test pointer being set to point to an
4049 error message. The new argument may be NULL if no error number is required
4050 (but then you may as well call pcre_compile(), which is now just a
4051 wrapper). This facility is provided because some applications need a
4052 numeric error indication, but it has also enabled me to tidy up the way
4053 compile-time errors are handled in the POSIX wrapper.
4054
4055 16. Added VPATH=.libs to the makefile; this should help when building with one
4056 prefix path and installing with another. (Or so I'm told by someone who
4057 knows more about this stuff than I do.)
4058
4059 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4060 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4061 match everything, including newlines. This is not POSIX-compatible, but
4062 somebody wanted the feature. From pcretest it can be activated by using
4063 both the P and the s flags.
4064
4065 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4066
4067 19. libpcre.pc was being incorrectly installed as executable.
4068
4069 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4070 it now also looks for '\r' so that it will work unmodified on Windows.
4071
4072 21. Added Google's contributed C++ wrapper to the distribution.
4073
4074 22. Added some untidy missing memory free() calls in pcretest, to keep
4075 Electric Fence happy when testing.
4076
4077
4078
4079 Version 5.0 13-Sep-04
4080 ---------------------
4081
4082 1. Internal change: literal characters are no longer packed up into items
4083 containing multiple characters in a single byte-string. Each character
4084 is now matched using a separate opcode. However, there may be more than one
4085 byte in the character in UTF-8 mode.
4086
4087 2. The pcre_callout_block structure has two new fields: pattern_position and
4088 next_item_length. These contain the offset in the pattern to the next match
4089 item, and its length, respectively.
4090
4091 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4092 insertion of callouts before each pattern item. Added the /C option to
4093 pcretest to make use of this.
4094
4095 4. On the advice of a Windows user, the lines
4096
4097 #if defined(_WIN32) || defined(WIN32)
4098 _setmode( _fileno( stdout ), 0x8000 );
4099 #endif /* defined(_WIN32) || defined(WIN32) */
4100
4101 have been added to the source of pcretest. This apparently does useful
4102 magic in relation to line terminators.
4103
4104 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4105 for the benefit of those environments where the "b" makes a difference.
4106
4107 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4108 to know about it. I have put a hack into configure.in that adds in code
4109 to set GCC=yes if CC=icc. This seems to end up at a point in the
4110 generated configure script that is early enough to affect the setting of
4111 compiler options, which is what is needed, but I have no means of testing
4112 whether it really works. (The user who reported this had patched the
4113 generated configure script, which of course I cannot do.)
4114
4115 LATER: After change 22 below (new libtool files), the configure script
4116 seems to know about icc (and also ecc). Therefore, I have commented out
4117 this hack in configure.in.
4118
4119 7. Added support for pkg-config (2 patches were sent in).
4120
4121 8. Negated POSIX character classes that used a combination of internal tables
4122 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4123 [[:^ascii]]. Typically, they would match almost any characters. The other
4124 POSIX classes were not broken in this way.
4125
4126 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4127 to find the match, as PCRE was deluded into thinking that the match had to
4128 start at the start point or following a newline. The same bug applied to
4129 patterns with negative forward assertions or any backward assertions
4130 preceding ".*" at the start, unless the pattern required a fixed first
4131 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4132
4133 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4134 starting at the last subject character, bytes beyond the end of the subject
4135 string were read.
4136
4137 11. Renamed the variable "class" as "classbits" to make life easier for C++
4138 users. (Previously there was a macro definition, but it apparently wasn't
4139 enough.)
4140
4141 12. Added the new field "tables" to the extra data so that tables can be passed
4142 in at exec time, or the internal tables can be re-selected. This allows
4143 a compiled regex to be saved and re-used at a later time by a different
4144 program that might have everything at different addresses.
4145
4146 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4147 -R library as well as a -L library.
4148
4149 14. The debugging options of pcretest (-d on the command line or D on a
4150 pattern) showed incorrect output for anything following an extended class
4151 that contained multibyte characters and which was followed by a quantifier.
4152
4153 15. Added optional support for general category Unicode character properties
4154 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4155 support. It adds about 90K to the size of the library. The meanings of the
4156 inbuilt class escapes such as \d and \s have NOT been changed.
4157
4158 16. Updated pcredemo.c to include calls to free() to release the memory for the
4159 compiled pattern.
4160
4161 17. The generated file chartables.c was being created in the source directory
4162 instead of in the building directory. This caused the build to fail if the
4163 source directory was different from the building directory, and was
4164 read-only.
4165
4166 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4167 file. No doubt somebody will tell me if they don't make sense... Also added
4168 Dan Mooney's comments about building on OpenVMS.
4169
4170 19. Added support for partial matching via the PCRE_PARTIAL option for
4171 pcre_exec() and the \P data escape in pcretest.
4172
4173 20. Extended pcretest with 3 new pattern features:
4174
4175 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4176 write the compiled pattern to the file whose name is "rest-of-line".
4177 This is a straight binary dump of the data, with the saved pointer to
4178 the character tables forced to be NULL. The study data, if any, is
4179 written too. After writing, pcretest reads a new pattern.
4180
4181 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4182 compiled pattern from the given file. There must not be any
4183 occurrences of "<" in the file name (pretty unlikely); if there are,
4184 pcretest will instead treat the initial "<" as a pattern delimiter.
4185 After reading in the pattern, pcretest goes on to read data lines as
4186 usual.
4187
4188 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4189 and 16-bit fields in a compiled pattern, to simulate a pattern that
4190 was compiled on a host of opposite endianness.
4191
4192 21. The pcre-exec() function can now cope with patterns that were compiled on
4193 hosts of opposite endianness, with this restriction:
4194
4195 As for any compiled expression that is saved and used later, the tables
4196 pointer field cannot be preserved; the extra_data field in the arguments
4197 to pcre_exec() should be used to pass in a tables address if a value
4198 other than the default internal tables were used at compile time.
4199
4200 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4201 now diagnosed as an error. Previously, most of the time, a negative number
4202 would have been treated as zero, but if in addition "ovector" was passed as
4203 NULL, a crash could occur.
4204
4205 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4206 new versions from the libtool 1.5 distribution (the last one is a copy of
4207 a file called libtool.m4). This seems to have fixed the need to patch
4208 "configure" to support Darwin 1.3 (which I used to do). However, I still
4209 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4210 workstation).
4211
4212 24. Changed the PCRE licence to be the more standard "BSD" licence.
4213
4214
4215 Version 4.5 01-Dec-03
4216 ---------------------
4217
4218 1. There has been some re-arrangement of the code for the match() function so
4219 that it can be compiled in a version that does not call itself recursively.
4220 Instead, it keeps those local variables that need separate instances for
4221 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4222 needs to "recurse". Keeping track of where control must go is done by means
4223 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4224 hide most of the details from the main code, and operates only if
4225 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4226 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4227 operating.
4228
4229 To make it easier for callers to provide specially tailored get/free
4230 functions for this usage, two new functions, pcre_stack_malloc, and
4231 pcre_stack_free, are used. They are always called in strict stacking order,
4232 and the size of block requested is always the same.
4233
4234 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4235 PCRE has been compiled to use the stack or the heap for recursion. The
4236 -C option of pcretest uses this to show which version is compiled.
4237
4238 A new data escape \S, is added to pcretest; it causes the amounts of store
4239 obtained and freed by both kinds of malloc/free at match time to be added
4240 to the output.
4241
4242 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4243 what's available on my current Linux desktop machine.
4244
4245 3. When matching a UTF-8 string, the test for a valid string at the start has
4246 been extended. If start_offset is not zero, PCRE now checks that it points
4247 to a byte that is the start of a UTF-8 character. If not, it returns
4248 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4249 this is necessary because there may be backward assertions in the pattern.
4250 When matching the same subject several times, it may save resources to use
4251 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4252
4253 4. The code for checking the validity of UTF-8 strings has been tightened so
4254 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4255 containing "overlong sequences".
4256
4257 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4258 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4259 should have been "&", but it just so happened that all the cases this let
4260 through by mistake were picked up later in the function.
4261
4262 6. I had used a variable called "isblank" - this is a C99 function, causing
4263 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4264
4265 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4266 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4267 so the version is shown; (c) stop "make test" echoing "./RunTest".
4268
4269 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4270
4271 9. The prototype for memmove() for systems that don't have it was using
4272 size_t, but the inclusion of the header that defines size_t was later. I've
4273 moved the #includes for the C headers earlier to avoid this.
4274
4275 10. Added some adjustments to the code to make it easier to compiler on certain
4276 special systems:
4277
4278 (a) Some "const" qualifiers were missing.
4279 (b) Added the macro EXPORT before all exported functions; by default this
4280 is defined to be empty.
4281 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4282 that it reads its output file name as an argument instead of writing
4283 to the standard output and assuming this can be redirected.
4284
4285 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4286 class containing characters with values greater than 255, PCRE compilation
4287 went into a loop.
4288
4289 12. A recursive reference to a subpattern that was within another subpattern
4290 that had a minimum quantifier of zero caused PCRE to crash. For example,
4291 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4292 recursion. If the recursively-called subpattern itself had a zero repeat,
4293 that was OK.
4294
4295 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4296 buffer into which it was copied (for escape processing) was still set at
4297 1024, so long lines caused crashes.
4298
4299 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4300 "internal error: code overflow...". This applied to any character class
4301 that was followed by a possessive quantifier.
4302
4303 15. Modified the Makefile to add libpcre.la as a prerequisite for
4304 libpcreposix.la because I was told this is needed for a parallel build to
4305 work.
4306
4307 16. If a pattern that contained .* following optional items at the start was
4308 studied, the wrong optimizing data was generated, leading to matching
4309 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4310 matching string must start with a or b or c. The correct conclusion for
4311 this pattern is that a match can start with any character.
4312
4313
4314 Version 4.4 13-Aug-03
4315 ---------------------
4316
4317 1. In UTF-8 mode, a character class containing characters with values between
4318 127 and 255 was not handled correctly if the compiled pattern was studied.
4319 In fixing this, I have also improved the studying algorithm for such
4320 classes (slightly).
4321
4322 2. Three internal functions had redundant arguments passed to them. Removal
4323 might give a very teeny performance improvement.
4324
4325 3. Documentation bug: the value of the capture_top field in a callout is *one
4326 more than* the number of the hightest numbered captured substring.
4327
4328 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4329 in incorrectly linking with a previously installed version. They now link
4330 explicitly with libpcre.la.
4331
4332 5. configure.in no longer needs to recognize Cygwin specially.
4333
4334 6. A problem in pcre.in for Windows platforms is fixed.
4335
4336 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4337 pcretest, it used to include the size of the study block as part of its
4338 output. Unfortunately, the structure contains a field that has a different
4339 size on different hardware architectures. This meant that the tests that
4340 showed this size failed. As the block is currently always of a fixed size,
4341 this information isn't actually particularly useful in pcretest output, so
4342 I have just removed it.
4343
4344 8. Three pre-processor statements accidentally did not start in column 1.
4345 Sadly, there are *still* compilers around that complain, even though
4346 standard C has not required this for well over a decade. Sigh.
4347
4348 9. In pcretest, the code for checking callouts passed small integers in the
4349 callout_data field, which is a void * field. However, some picky compilers
4350 complained about the casts involved for this on 64-bit systems. Now
4351 pcretest passes the address of the small integer instead, which should get
4352 rid of the warnings.
4353
4354 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4355 both compile and run time, and gives an error if an invalid UTF-8 sequence
4356 is found. There is a option for disabling this check in cases where the
4357 string is known to be correct and/or the maximum performance is wanted.
4358
4359 11. In response to a bug report, I changed one line in Makefile.in from
4360
4361 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4362 to
4363 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4364
4365 to look similar to other lines, but I have no way of telling whether this
4366 is the right thing to do, as I do not use Windows. No doubt I'll get told
4367 if it's wrong...
4368
4369
4370 Version 4.3 21-May-03
4371 ---------------------
4372
4373 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4374 Makefile.
4375
4376 2. Some refactoring to improve the quality of the code:
4377
4378 (i) The utf8_table... variables are now declared "const".
4379
4380 (ii) The code for \cx, which used the "case flipping" table to upper case
4381 lower case letters, now just substracts 32. This is ASCII-specific,
4382 but the whole concept of \cx is ASCII-specific, so it seems
4383 reasonable.
4384
4385 (iii) PCRE was using its character types table to recognize decimal and
4386 hexadecimal digits in the pattern. This is silly, because it handles
4387 only 0-9, a-f, and A-F, but the character types table is locale-
4388 specific, which means strange things might happen. A private
4389 table is now used for this - though it costs 256 bytes, a table is
4390 much faster than multiple explicit tests. Of course, the standard
4391 character types table is still used for matching digits in subject
4392 strings against \d.
4393
4394 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4395 ending in _t are). So I've renamed it as ESC_tee.
4396
4397 3. The first argument for regexec() in the POSIX wrapper should have been
4398 defined as "const".
4399
4400 4. Changed pcretest to use malloc() for its buffers so that they can be
4401 Electric Fenced for debugging.
4402
4403 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4404 to read one or more bytes before the start of the subject string. Often this
4405 had no effect on PCRE's behaviour, but in some circumstances it could
4406 provoke a segmentation fault.
4407
4408 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4409 to try to read one or more bytes before the start of the subject string.
4410
4411 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4412 UTF-8 support could misbehave in various ways if the subject string
4413 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4414 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4415 back over UTF-8 characters.)
4416
4417
4418 Version 4.2 14-Apr-03
4419 ---------------------
4420
4421 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4422
4423 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4424 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4425 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4426 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4427 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4428 and BUILD_EXEEXT
4429 Note: automatic setting of the BUILD variables is not yet working
4430 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4431 compile-time but not at link-time
4432 [LINK]: use for linking executables only
4433 make different versions for Windows and non-Windows
4434 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4435 libraries
4436 [LINK_FOR_BUILD]: new variable
4437 [OBJEXT]: use throughout
4438 [EXEEXT]: use throughout
4439 <winshared>: new target
4440 <wininstall>: new target
4441 <dftables.o>: use native compiler
4442 <dftables>: use native linker
4443 <install>: handle Windows platform correctly
4444 <clean>: ditto
4445 <check>: ditto
4446 copy DLL to top builddir before testing
4447
4448 As part of these changes, -no-undefined was removed again. This was reported
4449 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4450 in any case.
4451
4452 3. Some tidies to get rid of compiler warnings:
4453
4454 . In the match_data structure, match_limit was an unsigned long int, whereas
4455 match_call_count was an int. I've made them both unsigned long ints.
4456
4457 . In pcretest the fact that a const uschar * doesn't automatically cast to
4458 a void * provoked a warning.
4459
4460 . Turning on some more compiler warnings threw up some "shadow" variables
4461 and a few more missing casts.
4462
4463 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4464 option, a class that contained a single character with a value between 128
4465 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4466
4467 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4468 option, a class that contained several characters, but with at least one
4469 whose value was between 128 and 255 caused PCRE to crash.
4470
4471
4472 Version 4.1 12-Mar-03
4473 ---------------------
4474
4475 1. Compiling with gcc -pedantic found a couple of places where casts were
4476 needed, and a string in dftables.c that was longer than standard compilers are
4477 required to support.
4478
4479 2. Compiling with Sun's compiler found a few more places where the code could
4480 be tidied up in order to avoid warnings.
4481
4482 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4483 first of these names is deprecated in the latest Autoconf in favour of the name
4484 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4485 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4486 analogy I have changed it to CFLAGS_FOR_BUILD.
4487
4488 4. Added -no-undefined to the linking command in the Makefile, because this is
4489 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4490 linking step for the pcreposix library.
4491
4492 5. PCRE was failing to diagnose the case of two named groups with the same
4493 name.
4494
4495 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4496 literal character that is needed in the subject for a match, and scans along to
4497 ensure that it is present before embarking on the full matching process. This
4498 saves time in cases of nested unlimited repeats that are never going to match.
4499 Problem: the scan can take a lot of time if the subject is very long (e.g.
4500 megabytes), thus penalizing straightforward matches. It is now done only if the
4501 amount of subject to be scanned is less than 1000 bytes.
4502
4503 7. A lesser problem with the same optimization is that it was recording the
4504 first character of an anchored pattern as "needed", thus provoking a search
4505 right along the subject, even when the first match of the pattern was going to
4506 fail. The "needed" character is now not set for anchored patterns, unless it
4507 follows something in the pattern that is of non-fixed length. Thus, it still
4508 fulfils its original purpose of finding quick non-matches in cases of nested
4509 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4510
4511
4512 Version 4.0 17-Feb-03
4513 ---------------------
4514
4515 1. If a comment in an extended regex that started immediately after a meta-item
4516 extended to the end of string, PCRE compiled incorrect data. This could lead to
4517 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4518
4519 2. Moved to autoconf 2.53 and libtool 1.4.2.
4520
4521 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4522 the special perltest8 script is no longer needed - all the tests can be run
4523 from a single perltest script.
4524
4525 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4526 by \s. It has now been removed in PCRE. This means it isn't recognized as
4527 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4528 class [:space:] *does* include VT, thereby creating a mess.
4529
4530 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4531 space and tab.
4532
4533 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4534 its new features into the main test script, reducing the number of scripts.
4535
4536 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4537 were backward compatible, and made the (?i) apply to the whole pattern, as if
4538 /i were given. Now it behaves more logically, and applies the option setting
4539 only to what follows. PCRE has been changed to follow suit. However, if it
4540 finds options settings right at the start of the pattern, it extracts them into
4541 the global options, as before. Thus, they show up in the info data.
4542
4543 8. Added support for the \Q...\E escape sequence. Characters in between are
4544 treated as literals. This is slightly different from Perl in that $ and @ are
4545 also handled as literals inside the quotes. In Perl, they will cause variable
4546 interpolation. Note the following examples:
4547
4548 Pattern PCRE matches Perl matches
4549
4550 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4551 \Qabc\$xyz\E abc\$xyz abc\$xyz
4552 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4553
4554 For compatibility with Perl, \Q...\E sequences are recognized inside character
4555 classes as well as outside them.
4556
4557 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4558 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4559 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4560 signed/unsigned warnings.
4561
4562 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4563 option for pcretest, so I've replaced it by a simple function that does just
4564 that job.
4565
4566 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4567 "pcregrep -".
4568
4569 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4570 Java package. This provides some syntactic sugar for simple cases of what my
4571 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4572 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4573 item, you can use this simplified notation. Note that only makes sense with
4574 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4575 greediness, whatever the setting of the PCRE_UNGREEDY option.
4576
4577 13. A change of greediness default within a pattern was not taking effect at
4578 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4579 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4580 was abstracted outside.
4581
4582 14. PCRE now supports the \G assertion. It is true when the current matching
4583 position is at the start point of the match. This differs from \A when the
4584 starting offset is non-zero. Used with the /g option of pcretest (or similar
4585 code), it works in the same way as it does for Perl's /g option. If all
4586 alternatives of a regex begin with \G, the expression is anchored to the start
4587 match position, and the "anchored" flag is set in the compiled expression.
4588
4589 15. Some bugs concerning the handling of certain option changes within patterns
4590 have been fixed. These applied to options other than (?ims). For example,
4591 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4592 been the other way round. Some of this was related to change 7 above.
4593
4594 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4595 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4596 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4597 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4598
4599 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4600 mode. Unlike ".", it always matches newline, whatever the setting of
4601 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4602 assertions. Perl allows it, but it doesn't (in general) work because it can't
4603 calculate the length of the lookbehind. At least, that's the case for Perl
4604 5.8.0 - I've been told they are going to document that it doesn't work in
4605 future.
4606
4607 18. Added an error diagnosis for escapes that PCRE does not support: these are
4608 \L, \l, \N, \P, \p, \U, \u, and \X.
4609
4610 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4611 reading past the end of the pattern in cases such as /[abcd/.
4612
4613 20. PCRE was getting more memory than necessary for patterns with classes that
4614 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4615
4616 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4617 compiling PCRE for use with Virtual Pascal.
4618
4619 22. Small fix to the Makefile to make it work properly if the build is done
4620 outside the source tree.
4621
4622 23. Added a new extension: a condition to go with recursion. If a conditional
4623 subpattern starts with (?(R) the "true" branch is used if recursion has
4624 happened, whereas the "false" branch is used only at the top level.
4625
4626 24. When there was a very long string of literal characters (over 255 bytes
4627 without UTF support, over 250 bytes with UTF support), the computation of how
4628 much memory was required could be incorrect, leading to segfaults or other
4629 strange effects.
4630
4631 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4632 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4633 there was a subsequent back reference to those brackets. This meant that, for
4634 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4635 possible to check for precisely this case. All we can do is abandon the
4636 optimization if .* occurs inside capturing brackets when there are any back
4637 references whatsoever. (See below for a better fix that came later.)
4638
4639 26. The handling of the optimization for finding the first character of a
4640 non-anchored pattern, and for finding a character that is required later in the
4641 match were failing in some cases. This didn't break the matching; it just
4642 failed to optimize when it could. The way this is done has been re-implemented.
4643
4644 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4645
4646 28. Added a new feature that provides some of the functionality that Perl
4647 provides with (?{...}). The facility is termed a "callout". The way it is done
4648 in PCRE is for the caller to provide an optional function, by setting
4649 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4650 global variable. By default it is unset, which disables all calling out. To get
4651 the function called, the regex must include (?C) at appropriate points. This
4652 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4653 This provides a means of identifying different callout points. When PCRE
4654 reaches such a point in the regex, if pcre_callout has been set, the external
4655 function is called. It is provided with data in a structure called
4656 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4657 matching continues; if it returns a non-zero value, the match at the current
4658 point fails. However, backtracking will occur if possible. [This was changed
4659 later and other features added - see item 49 below.]
4660
4661 29. pcretest is upgraded to test the callout functionality. It provides a
4662 callout function that displays information. By default, it shows the start of
4663 the match and the current position in the text. There are some new data escapes
4664 to vary what happens:
4665
4666 \C+ in addition, show current contents of captured substrings
4667 \C- do not supply a callout function
4668 \C!n return 1 when callout number n is reached
4669 \C!n!m return 1 when callout number n is reached for the mth time
4670
4671 30. If pcregrep was called with the -l option and just a single file name, it
4672 output "<stdin>" if a match was found, instead of the file name.
4673
4674 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4675 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4676 pcre_exec(). This saves a malloc/free per call. The default value of
4677 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4678 when configuring.
4679
4680 32. The default maximum size of a compiled pattern is 64K. There have been a
4681 few cases of people hitting this limit. The code now uses macros to handle the
4682 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4683 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4684 configuring. Tests 2 and 5 work only with 2-byte links because they output
4685 debugging information about compiled patterns.
4686
4687 33. Internal code re-arrangements:
4688
4689 (a) Moved the debugging function for printing out a compiled regex into
4690 its own source file (printint.c) and used #include to pull it into
4691 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4692 separate copies.
4693
4694 (b) Defined the list of op-code names for debugging as a macro in
4695 internal.h so that it is next to the definition of the opcodes.
4696
4697 (c) Defined a table of op-code lengths for simpler skipping along compiled
4698 code. This is again a macro in internal.h so that it is next to the
4699 definition of the opcodes.
4700
4701 34. Added support for recursive calls to individual subpatterns, along the
4702 lines of Robin Houston's patch (but implemented somewhat differently).
4703
4704 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4705 allow it to read and process whole directories in Win32. This code was
4706 contributed by Lionel Fourquaux; it has not been tested by me.
4707
4708 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4709 used to name a group. Names consist of alphanumerics and underscores, and must
4710 be unique. Back references use the syntax (?P=name) and recursive calls use
4711 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4712 numbers. The function pcre_fullinfo() can be used after compilation to extract
4713 a name/number map. There are three relevant calls:
4714
4715 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4716 PCRE_INFO_NAMECOUNT yields the number of entries
4717 PCRE_INFO_NAMETABLE yields a pointer to the map.
4718
4719 The map is a vector of fixed-size entries. The size of each entry depends on
4720 the length of the longest name used. The first two bytes of each entry are the
4721 group number, most significant byte first. There follows the corresponding
4722 name, zero terminated. The names are in alphabetical order.
4723
4724 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4725 case instead of 255. Making it the same both with and without UTF-8 support
4726 means that the same test output works with both.
4727
4728 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4729 calling malloc() with a zero argument.
4730
4731 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4732 optimization. I've improved things by keeping a bitmap of backreferences with
4733 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4734 fact referenced, the optimization can be applied. It is unlikely that a
4735 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4736 the match to follow \n) will appear inside brackets with a number greater than
4737 31, but if it does, any back reference > 31 suppresses the optimization.
4738
4739 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4740 of disabling numbered capturing parentheses. Any opening parenthesis that is
4741 not followed by ? behaves as if it were followed by ?: but named parentheses
4742 can still be used for capturing (and they will acquire numbers in the usual
4743 way).
4744
4745 41. Redesigned the return codes from the match() function into yes/no/error so
4746 that errors can be passed back from deep inside the nested calls. A malloc
4747 failure while inside a recursive subpattern call now causes the
4748 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4749
4750 42. It is now possible to set a limit on the number of times the match()
4751 function is called in a call to pcre_exec(). This facility makes it possible to
4752 limit the amount of recursion and backtracking, though not in a directly
4753 obvious way, because the match() function is used in a number of different
4754 circumstances. The count starts from zero for each position in the subject
4755 string (for non-anchored patterns). The default limit is, for compatibility, a
4756 large number, namely 10 000 000. You can change this in two ways:
4757
4758 (a) When configuring PCRE before making, you can use --with-match-limit=n
4759 to set a default value for the compiled library.
4760
4761 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4762 a different value is set. See 45 below.
4763
4764 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4765
4766 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4767 of things that can be changed at compile time. The first argument specifies
4768 what is wanted and the second points to where the information is to be placed.
4769 The current list of available information is:
4770
4771 PCRE_CONFIG_UTF8
4772
4773 The output is an integer that is set to one if UTF-8 support is available;
4774 otherwise it is set to zero.
4775
4776 PCRE_CONFIG_NEWLINE
4777
4778 The output is an integer that it set to the value of the code that is used for
4779 newline. It is either LF (10) or CR (13).
4780
4781 PCRE_CONFIG_LINK_SIZE
4782
4783 The output is an integer that contains the number of bytes used for internal
4784 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4785
4786 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4787
4788 The output is an integer that contains the threshold above which the POSIX
4789 interface uses malloc() for output vectors. See item 31 above.
4790
4791 PCRE_CONFIG_MATCH_LIMIT
4792
4793 The output is an unsigned integer that contains the default limit of the number
4794 of match() calls in a pcre_exec() execution. See 42 above.
4795
4796 44. pcretest has been upgraded by the addition of the -C option. This causes it
4797 to extract all the available output from the new pcre_config() function, and to
4798 output it. The program then exits immediately.
4799
4800 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4801 order to support additional features. One way would have been to define
4802 pcre_exec2() (for example) with extra arguments, but this would not have been
4803 extensible, and would also have required all calls to the original function to
4804 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4805 is used for passing in "extra" data from pcre_study().
4806
4807 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4808 contains the following fields:
4809
4810 flags a bitmap indicating which of the following fields are set
4811 study_data opaque data from pcre_study()
4812 match_limit a way of specifying a limit on match() calls for a specific
4813 call to pcre_exec()
4814 callout_data data for callouts (see 49 below)
4815
4816 The flag bits are also defined in pcre.h, and are
4817
4818 PCRE_EXTRA_STUDY_DATA
4819 PCRE_EXTRA_MATCH_LIMIT
4820 PCRE_EXTRA_CALLOUT_DATA
4821
4822 The pcre_study() function now returns one of these new pcre_extra blocks, with
4823 the actual study data pointed to by the study_data field, and the
4824 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4825 before. That is, this change is entirely upwards-compatible and requires no
4826 change to existing code.
4827
4828 If you want to pass in additional data to pcre_exec(), you can either place it
4829 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4830 block.
4831
4832 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4833 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4834 times with different match limits, until it finds the minimum value needed for
4835 pcre_exec() to complete. The value is then output. This can be instructive; for
4836 most simple matches the number is quite small, but for pathological cases it
4837 gets very large very quickly.
4838
4839 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4840 returns the size of the data block pointed to by the study_data field in a
4841 pcre_extra block, that is, the value that was passed as the argument to
4842 pcre_malloc() when PCRE was getting memory in which to place the information
4843 created by pcre_study(). The fourth argument should point to a size_t variable.
4844 pcretest has been extended so that this information is shown after a successful
4845 pcre_study() call when information about the compiled regex is being displayed.
4846
4847 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4848 because what follows is always an absolute path. (Later: it turns out that this
4849 is more than cosmetic for MinGW, because it doesn't like empty path
4850 components.)
4851
4852 49. Some changes have been made to the callout feature (see 28 above):
4853
4854 (i) A callout function now has three choices for what it returns:
4855
4856 0 => success, carry on matching
4857 > 0 => failure at this point, but backtrack if possible
4858 < 0 => serious error, return this value from pcre_exec()
4859
4860 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4861 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4862 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4863 use by callout functions. It will never be used by PCRE itself.
4864
4865 (ii) The pcre_extra structure (see 45 above) has a void * field called
4866 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4867 pcre_callout_block structure has a field of the same name. The contents of
4868 the field passed in the pcre_extra structure are passed to the callout
4869 function in the corresponding field in the callout block. This makes it
4870 easier to use the same callout-containing regex from multiple threads. For
4871 testing, the pcretest program has a new data escape
4872
4873 \C*n pass the number n (may be negative) as callout_data
4874
4875 If the callout function in pcretest receives a non-zero value as
4876 callout_data, it returns that value.
4877
4878 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4879 there were some redundant $(CFLAGS) in commands that are now specified as
4880 $(LINK), which already includes $(CFLAGS).
4881
4882 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4883 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4884 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4885 one-byte characters throughout. Note that case-insensitive matching applies
4886 only to characters whose values are less than 256. PCRE doesn't support the
4887 notion of cases for higher-valued characters.
4888
4889 (i) A character class whose characters are all within 0-255 is handled as
4890 a bit map, and the map is inverted for negative classes. Previously, a
4891 character > 255 always failed to match such a class; however it should
4892 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4893
4894 (ii) A negated character class with a single character < 255 is coded as
4895 "not this character" (OP_NOT). This wasn't working properly when the test
4896 character was multibyte, either singly or repeated.
4897
4898 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4899 mode, for example: \x{100}{2,3}.
4900
4901 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4902 singly or repeated) now correctly test multibyte characters. However,
4903 PCRE doesn't recognize any characters with values greater than 255 as
4904 digits, spaces, or word characters. Such characters always match \D, \S,
4905 and \W, and never match \d, \s, or \w.
4906
4907 (v) Classes may now contain characters and character ranges with values
4908 greater than 255. For example: [ab\x{100}-\x{400}].
4909
4910 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4911 PCRE in UTF-8 mode.
4912
4913 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4914 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4915 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4916 value.)
4917
4918 53. The single man page has become too large. I have therefore split it up into
4919 a number of separate man pages. These also give rise to individual HTML pages;
4920 these are now put in a separate directory, and there is an index.html page that
4921 lists them all. Some hyperlinking between the pages has been installed.
4922
4923 54. Added convenience functions for handling named capturing parentheses.
4924
4925 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4926 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4927 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4928 are faulted.
4929
4930 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4931 calling configure. These values are used when compiling the dftables.c program
4932 which is run to generate the source of the default character tables. They
4933 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4934 you will need to set these values.
4935
4936 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4937
4938
4939 Version 3.9 02-Jan-02
4940 ---------------------
4941
4942 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4943
4944 2. If --disable-static was given, the building process failed when trying to
4945 build pcretest and pcregrep. (For some reason it was using libtool to compile
4946 them, which is not right, as they aren't part of the library.)
4947
4948
4949 Version 3.8 18-Dec-01
4950 ---------------------
4951
4952 1. The experimental UTF-8 code was completely screwed up. It was packing the
4953 bytes in the wrong order. How dumb can you get?
4954
4955
4956 Version 3.7 29-Oct-01
4957 ---------------------
4958
4959 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4960 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4961 this didn't happen under Solaris 8, where I normally test things.
4962
4963 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4964 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4965 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4966 this fix an hour or so after the initial 3.7 release.)
4967
4968
4969 Version 3.6 23-Oct-01
4970 ---------------------
4971
4972 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4973 offsets passed as NULL with zero offset count.
4974
4975 2. The config.guess and config.sub files had not been updated when I moved to
4976 the latest autoconf.
4977
4978
4979 Version 3.5 15-Aug-01
4980 ---------------------
4981
4982 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4983 had been forgotten.
4984
4985 2. By using declared but undefined structures, we can avoid using "void"
4986 definitions in pcre.h while keeping the internal definitions of the structures
4987 private.
4988
4989 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4990 user point of view, this means that both static and shared libraries are built
4991 by default, but this can be individually controlled. More of the work of
4992 handling this static/shared cases is now inside libtool instead of PCRE's make
4993 file.
4994
4995 4. The pcretest utility is now installed along with pcregrep because it is
4996 useful for users (to test regexs) and by doing this, it automatically gets
4997 relinked by libtool. The documentation has been turned into a man page, so
4998 there are now .1, .txt, and .html versions in /doc.
4999
5000 5. Upgrades to pcregrep:
5001 (i) Added long-form option names like gnu grep.
5002 (ii) Added --help to list all options with an explanatory phrase.
5003 (iii) Added -r, --recursive to recurse into sub-directories.
5004 (iv) Added -f, --file to read patterns from a file.
5005
5006 6. pcre_exec() was referring to its "code" argument before testing that
5007 argument for NULL (and giving an error if it was NULL).
5008
5009 7. Upgraded Makefile.in to allow for compiling in a different directory from
5010 the source directory.
5011
5012 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5013 options bits, the pointer it was passed was to an int instead of to an unsigned
5014 long int. This mattered only on 64-bit systems.
5015
5016 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5017 generated) instead of pcre.in, which it its source. Also made the same change
5018 in several of the .c files.
5019
5020 10. A new release of gcc defines printf() as a macro, which broke pcretest
5021 because it had an ifdef in the middle of a string argument for printf(). Fixed
5022 by using separate calls to printf().
5023
5024 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5025 script, to force use of CR or LF instead of \n in the source. On non-Unix
5026 systems, the value can be set in config.h.
5027
5028 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5029 absolute limit. Changed the text of the error message to make this clear, and
5030 likewise updated the man page.
5031
5032 13. The limit of 99 on the number of capturing subpatterns has been removed.
5033 The new limit is 65535, which I hope will not be a "real" limit.
5034
5035
5036 Version 3.4 22-Aug-00
5037 ---------------------
5038
5039 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5040
5041 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5042
5043
5044 Version 3.3 01-Aug-00
5045 ---------------------
5046
5047 1. If an octal character was given, but the value was greater than \377, it
5048 was not getting masked to the least significant bits, as documented. This could
5049 lead to crashes in some systems.
5050
5051 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5052 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5053
5054 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5055 These just pass their arguments on to (pcre_free)(), but they are provided
5056 because some uses of PCRE bind it to non-C systems that can call its functions,
5057 but cannot call free() or pcre_free() directly.
5058
5059 4. Add "make test" as a synonym for "make check". Corrected some comments in
5060 the Makefile.
5061
5062 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5063 Makefile.
5064
5065 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5066 command called pgrep for grepping around the active processes.
5067
5068 7. Added the beginnings of support for UTF-8 character strings.
5069
5070 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5071 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5072 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5073 out for the ar command.)
5074
5075
5076 Version 3.2 12-May-00
5077 ---------------------
5078
5079 This is purely a bug fixing release.
5080
5081 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5082 of ZA. This was just one example of several cases that could provoke this bug,
5083 which was introduced by change 9 of version 2.00. The code for breaking
5084 infinite loops after an iteration that matches an empty string was't working
5085 correctly.
5086
5087 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5088 when matched against abbab (for example). After matching an empty string, it
5089 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5090 caused it to match further down the string than it should.
5091
5092 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5093 was there because it doesn't seem to be needed, and it causes trouble on some
5094 systems, as it is not a Standard C header. It has been removed.
5095
5096 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5097 were reported on the Macintosh. The changes were from
5098
5099 while ((c = *(++ptr)) != 0 && c != '\n');
5100 to
5101 while ((c = *(++ptr)) != 0 && c != '\n') ;
5102
5103 Totally extraordinary, but if that's what it takes...
5104
5105 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5106 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5107 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5108 assumes the way PCRE uses memmove() (always moving upwards).
5109
5110 6. PCRE is being used in one environment where strchr() is not available. There
5111 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5112 faster code anyway.
5113
5114
5115 Version 3.1 09-Feb-00
5116 ---------------------
5117
5118 The only change in this release is the fixing of some bugs in Makefile.in for
5119 the "install" target:
5120
5121 (1) It was failing to install pcreposix.h.
5122
5123 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5124
5125
5126 Version 3.0 01-Feb-00
5127 ---------------------
5128
5129 1. Add support for the /+ modifier to perltest (to output $` like it does in
5130 pcretest).
5131
5132 2. Add support for the /g modifier to perltest.
5133
5134 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5135 matches null strings.
5136
5137 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5138 pattern. Perl treats empty patterns specially - it reuses the most recent
5139 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5140 effect.
5141
5142 5. The POSIX interface was broken in that it was just handing over the POSIX
5143 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5144 required a bigger vector, with some working space on the end. This means that
5145 the POSIX wrapper now has to get and free some memory, and copy the results.
5146
5147 6. Added some simple autoconf support, placing the test data and the
5148 documentation in separate directories, re-organizing some of the
5149 information files, and making it build pcre-config (a GNU standard). Also added
5150 libtool support for building PCRE as a shared library, which is now the
5151 default.
5152
5153 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5154 09 are not valid octal constants. Single digits will be used for minor values
5155 less than 10.
5156
5157 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5158 existing programs that set these in the POSIX interface can use PCRE without
5159 modification.
5160
5161 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5162 return all that pcre_info() returns, plus additional data. The pcre_info()
5163 function is retained for compatibility, but is considered to be obsolete.
5164
5165 10. Added experimental recursion feature (?R) to handle one common case that
5166 Perl 5.6 will be able to do with (?p{...}).
5167
5168 11. Added support for POSIX character classes like [:alpha:], which Perl is
5169 adopting.
5170
5171
5172 Version 2.08 31-Aug-99
5173 ----------------------
5174
5175 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5176 trying to match at the startoffset position, but instead was moving forward to
5177 the next newline as if a previous match had failed.
5178
5179 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5180 and could get into a loop if a null string was matched other than at the start
5181 of the subject.
5182
5183 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5184 be distinguished at compile time, and for completeness also added PCRE_DATE.
5185
5186 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5187 in GnuWin32 environments.
5188
5189
5190 Version 2.07 29-Jul-99
5191 ----------------------
5192
5193 1. The documentation is now supplied in plain text form and HTML as well as in
5194 the form of man page sources.
5195
5196 2. C++ compilers don't like assigning (void *) values to other pointer types.
5197 In particular this affects malloc(). Although there is no problem in Standard
5198 C, I've put in casts to keep C++ compilers happy.
5199
5200 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5201 should be (const char *).
5202
5203 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5204 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5205 However, I haven't made this a standard facility. The documentation doesn't
5206 mention it, and the Makefile doesn't support it.
5207
5208 5. The Makefile now contains an "install" target, with editable destinations at
5209 the top of the file. The pcretest program is not installed.
5210
5211 6. pgrep -V now gives the PCRE version number and date.
5212
5213 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5214 causing the entire string to be ignored, instead of just the last character.
5215
5216 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5217 non-matching string, it can take a very, very long time, even for strings of
5218 quite modest length, because of the nested recursion. PCRE now does better in
5219 some of these cases. It does this by remembering the last required literal
5220 character in the pattern, and pre-searching the subject to ensure it is present
5221 before running the real match. In other words, it applies a heuristic to detect
5222 some types of certain failure quickly, and in the above example, if presented
5223 with a string that has no trailing " it gives "no match" very quickly.
5224
5225 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5226 other alternatives are tried instead.
5227
5228
5229 Version 2.06 09-Jun-99
5230 ----------------------
5231
5232 1. Change pcretest's output for amount of store used to show just the code
5233 space, because the remainder (the data block) varies in size between 32-bit and
5234 64-bit systems.
5235
5236 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5237 start matching at. This allows lookbehinds to work when searching for multiple
5238 occurrences in a string.
5239
5240 3. Added additional options to pcretest for testing multiple occurrences:
5241
5242 /+ outputs the rest of the string that follows a match
5243 /g loops for multiple occurrences, using the new startoffset argument
5244 /G loops for multiple occurrences by passing an incremented pointer
5245
5246 4. PCRE wasn't doing the "first character" optimization for patterns starting
5247 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5248 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5249 the letter 'x'. On long subject strings, this gives a significant speed-up.
5250
5251
5252 Version 2.05 21-Apr-99
5253 ----------------------
5254
5255 1. Changed the type of magic_number from int to long int so that it works
5256 properly on 16-bit systems.
5257
5258 2. Fixed a bug which caused patterns starting with .* not to work correctly
5259 when the subject string contained newline characters. PCRE was assuming
5260 anchoring for such patterns in all cases, which is not correct because .* will
5261 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5262 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5263 must be retried after every newline in the subject.
5264
5265
5266 Version 2.04 18-Feb-99
5267 ----------------------
5268
5269 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5270 computation of the store needed to hold the pattern was incorrect (too large).
5271 If such patterns were nested a few deep, this could multiply and become a real
5272 problem.
5273
5274 2. Added /M option to pcretest to show the memory requirement of a specific
5275 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5276
5277 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5278 compiled in such a way that the backtracking after subsequent failure was
5279 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5280 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5281
5282
5283 Version 2.03 02-Feb-99
5284 ----------------------
5285
5286 1. Fixed typo and small mistake in man page.
5287
5288 2. Added 4th condition (GPL supersedes if conflict) and created separate
5289 LICENCE file containing the conditions.
5290
5291 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5292 Perl, that is the internal \ allows the delimiter to be included in the
5293 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5294 the final delimiter, add \ to the end of the pattern (to test the error).
5295
5296 4. Added the convenience functions for extracting substrings after a successful
5297 match. Updated pcretest to make it able to test these functions.
5298
5299
5300 Version 2.02 14-Jan-99
5301 ----------------------
5302
5303 1. Initialized the working variables associated with each extraction so that
5304 their saving and restoring doesn't refer to uninitialized store.
5305
5306 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5307 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5308 fix the problem.
5309
5310 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5311 calls, and wasn't printing the correct value for compiling calls. Increased the
5312 default value of LOOPREPEAT, and the number of significant figures in the
5313 times.
5314
5315 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5316
5317 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5318 a building problem on Windows NT with a FAT file system.
5319
5320
5321 Version 2.01 21-Oct-98
5322 ----------------------
5323
5324 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5325 to character tables built by pcre_maketables() in the current locale. If NULL
5326 is passed, the default tables are used.
5327
5328
5329 Version 2.00 24-Sep-98
5330 ----------------------
5331
5332 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5333 it any more.
5334
5335 2. Allow quantification of (?>) groups, and make it work correctly.
5336
5337 3. The first character computation wasn't working for (?>) groups.
5338
5339 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5340 end of the subject) and add 5.005's \z, which really does match only at the
5341 very end of the subject.
5342
5343 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5344
5345 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5346 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5347 localized options. All options to pcre_study() were also removed.
5348
5349 7. Add other new features from 5.005:
5350
5351 $(?<= positive lookbehind
5352 $(?<! negative lookbehind
5353 (?imsx-imsx) added the unsetting capability
5354 such a setting is global if at outer level; local otherwise
5355 (?imsx-imsx:) non-capturing groups with option setting
5356 (?(cond)re|re) conditional pattern matching
5357
5358 A backreference to itself in a repeated group matches the previous
5359 captured string.
5360
5361 8. General tidying up of studying (both automatic and via "study")
5362 consequential on the addition of new assertions.
5363
5364 9. As in 5.005, unlimited repeated groups that could match an empty substring
5365 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5366 runtime if any iteration does actually match an empty substring.
5367
5368 10. Include the RunTest script in the distribution.
5369
5370 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5371 discrepancies, some of which were old and were also with respect to 5.004. They
5372 have now been fixed.
5373
5374
5375 Version 1.09 28-Apr-98
5376 ----------------------
5377
5378 1. A negated single character class followed by a quantifier with a minimum
5379 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5380 program crashes, or just wrong answers. This did not apply to negated classes
5381 containing more than one character, or to minima other than one.
5382
5383
5384 Version 1.08 27-Mar-98
5385 ----------------------
5386
5387 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5388
5389 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5390 latter must appear before anything that relies on it in the pattern.
5391
5392
5393 Version 1.07 16-Feb-98
5394 ----------------------
5395
5396 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5397 repeat of a potentially empty string).
5398
5399
5400 Version 1.06 23-Jan-98
5401 ----------------------
5402
5403 1. Added Markus Oberhumer's little patches for C++.
5404
5405 2. Literal strings longer than 255 characters were broken.
5406
5407
5408 Version 1.05 23-Dec-97
5409 ----------------------
5410
5411 1. Negated character classes containing more than one character were failing if
5412 PCRE_CASELESS was set at run time.
5413
5414
5415 Version 1.04 19-Dec-97
5416 ----------------------
5417
5418 1. Corrected the man page, where some "const" qualifiers had been omitted.
5419
5420 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5421 input syntax.
5422
5423 3. Fixed memory leak which occurred when a regex with back references was
5424 matched with an offsets vector that wasn't big enough. The temporary memory
5425 that is used in this case wasn't being freed if the match failed.
5426
5427 4. Tidied pcretest to ensure it frees memory that it gets.
5428
5429 5. Temporary memory was being obtained in the case where the passed offsets
5430 vector was exactly big enough.
5431
5432 6. Corrected definition of offsetof() from change 5 below.
5433
5434 7. I had screwed up change 6 below and broken the rules for the use of
5435 setjmp(). Now fixed.
5436
5437
5438 Version 1.03 18-Dec-97
5439 ----------------------
5440
5441 1. A erroneous regex with a missing opening parenthesis was correctly
5442 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5443 on some systems.
5444
5445 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5446 it was reported that one broken compiler failed on the former because "code" is
5447 also an independent variable.
5448
5449 3. The erroneous regex a[]b caused an array overrun reference.
5450
5451 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5452 fail on data ending with that character. (It was going on too far, and checking
5453 the next character, typically a binary zero.) This was specific to the
5454 optimized code for single-character negative classes.
5455
5456 5. Added a contributed patch from the TIN world which does the following:
5457
5458 + Add an undef for memmove, in case the the system defines a macro for it.
5459
5460 + Add a definition of offsetof(), in case there isn't one. (I don't know
5461 the reason behind this - offsetof() is part of the ANSI standard - but
5462 it does no harm).
5463
5464 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5465 most of the places where whitespace preceded '#'. I have given up and
5466 allowed the remaining 2 cases to be at the margin.
5467
5468 + Rename some variables in pcre to eliminate shadowing. This seems very
5469 pedantic, but does no harm, of course.
5470
5471 6. Moved the call to setjmp() into its own function, to get rid of warnings
5472 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5473
5474 7. Constructs such as \d{8,} were compiling into the equivalent of
5475 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5476 outcome, but in this particular case used more store than had been allocated,
5477 which caused the bug to be discovered because it threw up an internal error.
5478
5479 8. The debugging code in both pcre and pcretest for outputting the compiled
5480 form of a regex was going wrong in the case of back references followed by
5481 curly-bracketed repeats.
5482
5483
5484 Version 1.02 12-Dec-97
5485 ----------------------
5486
5487 1. Typos in pcre.3 and comments in the source fixed.
5488
5489 2. Applied a contributed patch to get rid of places where it used to remove
5490 'const' from variables, and fixed some signed/unsigned and uninitialized
5491 variable warnings.
5492
5493 3. Added the "runtest" target to Makefile.
5494
5495 4. Set default compiler flag to -O2 rather than just -O.
5496
5497
5498 Version 1.01 19-Nov-97
5499 ----------------------
5500
5501 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5502 like /([ab]*)*/, that is, for classes with more than one character in them.
5503
5504 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5505 as /((?>a*))*/ (a PCRE_EXTRA facility).
5506
5507
5508 Version 1.00 18-Nov-97
5509 ----------------------
5510
5511 1. Added compile-time macros to support systems such as SunOS4 which don't have
5512 memmove() or strerror() but have other things that can be used instead.
5513
5514 2. Arranged that "make clean" removes the executables.
5515
5516
5517 Version 0.99 27-Oct-97
5518 ----------------------
5519
5520 1. Fixed bug in code for optimizing classes with only one character. It was
5521 initializing a 32-byte map regardless, which could cause it to run off the end
5522 of the memory it had got.
5523
5524 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5525
5526
5527 Version 0.98 22-Oct-97
5528 ----------------------
5529
5530 1. Fixed bug in code for handling temporary memory usage when there are more
5531 back references than supplied space in the ovector. This could cause segfaults.
5532
5533
5534 Version 0.97 21-Oct-97
5535 ----------------------
5536
5537 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5538
5539 2. Optimized negated single characters not to use a bit map.
5540
5541 3. Brought error texts together as macro definitions; clarified some of them;
5542 fixed one that was wrong - it said "range out of order" when it meant "invalid
5543 escape sequence".
5544
5545 4. Changed some char * arguments to const char *.
5546
5547 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5548
5549 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5550 pcretest.
5551
5552
5553 Version 0.96 16-Oct-97
5554 ----------------------
5555
5556 1. Added a simple "pgrep" utility to the distribution.
5557
5558 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5559 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5560 where "ddd" means "one or more decimal digits".
5561
5562 3. Fixed serious bug. If a pattern had a back reference, but the call to
5563 pcre_exec() didn't supply a large enough ovector to record the related
5564 identifying subpattern, the match always failed. PCRE now remembers the number
5565 of the largest back reference, and gets some temporary memory in which to save
5566 the offsets during matching if necessary, in order to ensure that
5567 backreferences always work.
5568
5569 4. Increased the compatibility with Perl in a number of ways:
5570
5571 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5572 to request this handling. The option can be set at compile or exec time.
5573
5574 (b) $ matches before a terminating newline by default; an option
5575 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5576 mode). The option can be set at compile or exec time.
5577
5578 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5579 the same as Perl's. If the decimal number it represents is less than 10
5580 or there aren't that many previous left capturing parentheses, an octal
5581 escape is read. Inside a character class, it's always an octal escape,
5582 even if it is a single digit.
5583
5584 (d) An escaped but undefined alphabetic character is taken as a literal,
5585 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5586 escapes.
5587
5588 (e) {0} is now permitted. (The previous item is removed from the compiled
5589 pattern).
5590
5591 5. Changed all the names of code files so that the basic parts are no longer
5592 than 10 characters, and abolished the teeny "globals.c" file.
5593
5594 6. Changed the handling of character classes; they are now done with a 32-byte
5595 bit map always.
5596
5597 7. Added the -d and /D options to pcretest to make it possible to look at the
5598 internals of compilation without having to recompile pcre.
5599
5600
5601 Version 0.95 23-Sep-97
5602 ----------------------
5603
5604 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5605 \x20 at the start of a run of normal characters. These were being treated as
5606 real characters, instead of the source characters being re-checked.
5607
5608
5609 Version 0.94 18-Sep-97
5610 ----------------------
5611
5612 1. The functions are now thread-safe, with the caveat that the global variables
5613 containing pointers to malloc() and free() or alternative functions are the
5614 same for all threads.
5615
5616 2. Get pcre_study() to generate a bitmap of initial characters for non-
5617 anchored patterns when this is possible, and use it if passed to pcre_exec().
5618
5619
5620 Version 0.93 15-Sep-97
5621 ----------------------
5622
5623 1. /(b)|(:+)/ was computing an incorrect first character.
5624
5625 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5626 but not actually doing anything yet.
5627
5628 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5629 as Perl does (e.g. [-az] or [az-]).
5630
5631 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5632 all possible positions.
5633
5634 5. Split up into different modules to avoid including unneeded functions in a
5635 compiled binary. However, compile and exec are still in one module. The "study"
5636 function is split off.
5637
5638 6. The character tables are now in a separate module whose source is generated
5639 by an auxiliary program - but can then be edited by hand if required. There are
5640 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5641 toupper() in the code.
5642
5643 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5644 make them global. Abolish the function for setting them, as the caller can now
5645 set them directly.
5646
5647
5648 Version 0.92 11-Sep-97
5649 ----------------------
5650
5651 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5652 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5653
5654 2. Caseless matching was not working in character classes if the characters in
5655 the pattern were in upper case.
5656
5657 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5658
5659 4. Make PCRE_ANCHORED public and accept as a compile option.
5660
5661 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5662 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5663 pass them.
5664
5665 6. Give an error if bad option bits passed at compile or run time.
5666
5667 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5668 pcretest to cause it to pass that flag.
5669
5670 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5671 options, and the first character, if set.
5672
5673 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5674
5675
5676 Version 0.91 10-Sep-97
5677 ----------------------
5678
5679 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5680 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5681
5682 2. PCRE was looping on encountering an indefinitely repeated back reference to
5683 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5684 Perl does - treats the match as successful.
5685
5686 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5