ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1452 - (show annotations)
Tue Jan 28 16:07:52 2014 UTC (5 years, 8 months ago) by zherczeg
File size: 263279 byte(s)
Error occurred while calculating annotation data.
Fix an infinite fast-forward newline on invalid UTF input.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.35-RC1 xx-xxxx-201x
5 -----------------------------
7 1. A new flag is set, when property checks are present in an XCLASS.
8 When this flag is not set, PCRE can perform certain optimizations
9 such as studying these XCLASS-es.
11 2. The auto-possessification of character sets were improved: a normal
12 and an extended character set can be compared now. Furthermore
13 the JIT compiler optimizes more character set checks.
15 3. Got rid of some compiler warnings for potentially uninitialized variables
16 that show up only when compiled with -O2.
18 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
19 of a match later then the end of the match. The pcretest program was not
20 handling the case sensibly - it was outputting from the start to the next
21 binary zero. It now reports this situation in a message, and outputs the
22 text from the end to the start.
24 5. Fast forward search is improved in JIT. Instead of the first three
25 characters, any three characters with fixed position can be searched.
26 Search order: first, last, middle.
28 6. Improve character range checks in JIT. Characters are read by an inprecise
29 function now, which returns with an unknown value if the character code is
30 above a certain treshold (e.g: 256). The only limitation is that the value
31 must be bigger than the treshold as well. This function is useful, when
32 the characters above the treshold are handled in the same way.
34 7. The macros whose names start with RAWUCHAR are placeholders for a future
35 mode in which only the bottom 21 bits of 32-bit data items are used. To
36 make this more memorable for those maintaining the code, the names have
37 been changed to start with UCHAR21, and an extensive comment has been added
38 to their definition.
40 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
41 to the export list in Makefile.am (they were accidentally omitted from the
42 8.34 tarball).
44 9. The informational output from pcretest used the phrase "starting byte set"
45 which is inappropriate for the 16-bit and 32-bit libraries. As the output
46 for "first char" and "need char" really means "non-UTF-char", I've changed
47 "byte" to "char", and slightly reworded the output. The documentation about
48 these values has also been (I hope) clarified.
50 10. Another JIT related optimization: use table jumps for selecting the correct
51 backtracking path, when more than four alternatives are present inside a
52 bracket.
54 11. Empty match is not possible, when the minimum length is greater than zero,
55 and there is no \K in the pattern. JIT should avoid empty match checks in
56 such cases.
58 12. In a caseless character class with UCP support, when a character with more
59 than one alternative case was not the first character of a range, not all
60 the alternative cases were added to the class. For example, s and \x{17f}
61 are both alternative cases for S: the class [RST] was handled correctly,
62 but [R-T] was not.
64 13. The configure.ac file always checked for pthread support when JIT was
65 enabled. This is not used in Windows, so I have put this test inside a
66 check for the presence of windows.h (which was already tested for).
68 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
69 The algorithm provides a way to skip certain starting offsets, and usually
70 faster than linear prefix searches.
72 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
73 as for 'fr_FR' and 'french'. For some reason, however, it then used the
74 Windows-specific input and output files, which have 'french' screwed in.
75 So this could never have worked. One of the problems with locales is that
76 they aren't always the same. I have now updated RunTest so that it checks
77 the output of the locale test (test 3) against three different output
78 files, and it allows the test to pass if any one of them matches. With luck
79 this should make the test pass on some versions of Solaris where it was
80 failing. Because of the uncertainty, the script did not used to stop if
81 test 3 failed; it now does. If further versions of a French locale ever
82 come to light, they can now easily be added.
84 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
85 there was a message during ./configure, but it did not stop. This now
86 provokes an error. The invalid example in README has been corrected.
87 If a value less than the minimum is given, the minimum value has always
88 been used, but now a warning is given.
90 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
91 was a bug in the test system, which is now fixed. Also, the list of various
92 configurations that are tested for each release did not have one with both
93 16/32 bits and --enable-bar-anycrlf. It now does.
95 18. pcretest was missing "-C bsr" for displaying the \R default setting.
97 19. Little endian PowerPC systems are supported now by the JIT compiler.
99 20. The fast forward newline mechanism could enter to an infinite loop on
100 certain invalid UTF-8 input. Although we don't support these cases
101 this issue can be fixed by a performance optimization.
104 Version 8.34 15-December-2013
105 -----------------------------
107 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
108 executable memory. Patch inspired by Carsten Klein.
110 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
111 this macro is never tested and has no effect, because the work to support
112 coverage involves only compiling and linking options and special targets in
113 the Makefile. The comment in config.h implied that defining the macro would
114 enable coverage support, which is totally false. There was also support for
115 setting this macro in the CMake files (my fault, I just copied it from
116 configure). SUPPORT_GCOV has now been removed.
118 3. Make a small performance improvement in strlen16() and strlen32() in
119 pcretest.
121 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
122 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
123 these cases). There was also one in pcretest.c.
125 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
127 6. In UTF mode, the code for checking whether a group could match an empty
128 string (which is used for indefinitely repeated groups to allow for
129 breaking an infinite loop) was broken when the group contained a repeated
130 negated single-character class with a character that occupied more than one
131 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
132 UTF-8 mode). The effect was undefined: the group might or might not be
133 deemed as matching an empty string, or the program might have crashed.
135 7. The code for checking whether a group could match an empty string was not
136 recognizing that \h, \H, \v, \V, and \R must match a character.
138 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
139 an empty string. If it can, pcretest shows this in its information output.
141 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
142 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
143 matched by pcre_exec() without using JIT:
145 (a) If the rest of the pattern did not match after a maximal run of
146 grapheme clusters, the code for backing up to try with fewer of them
147 did not always back up over a full grapheme when characters that do not
148 have the modifier quality were involved, e.g. Hangul syllables.
150 (b) If the match point in a subject started with modifier character, and
151 there was no match, the code could incorrectly back up beyond the match
152 point, and potentially beyond the first character in the subject,
153 leading to a segfault or an incorrect match result.
155 10. A conditional group with an assertion condition could lead to PCRE
156 recording an incorrect first data item for a match if no other first data
157 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
158 first data item, and therefore matched "ca" after "c" instead of at the
159 start.
161 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
162 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
164 12. The source of pcregrep now includes z/OS-specific code so that it can be
165 compiled for z/OS as part of the special z/OS distribution.
167 13. Added the -T and -TM options to pcretest.
169 14. The code in pcre_compile.c for creating the table of named capturing groups
170 has been refactored. Instead of creating the table dynamically during the
171 actual compiling pass, the information is remembered during the pre-compile
172 pass (on the stack unless there are more than 20 named groups, in which
173 case malloc() is used) and the whole table is created before the actual
174 compile happens. This has simplified the code (it is now nearly 150 lines
175 shorter) and prepared the way for better handling of references to groups
176 with duplicate names.
178 15. A back reference to a named subpattern when there is more than one of the
179 same name now checks them in the order in which they appear in the pattern.
180 The first one that is set is used for the reference. Previously only the
181 first one was inspected. This change makes PCRE more compatible with Perl.
183 16. Unicode character properties were updated from Unicode 6.3.0.
185 17. The compile-time code for auto-possessification has been refactored, based
186 on a patch by Zoltan Herczeg. It now happens after instead of during
187 compilation. The code is cleaner, and more cases are handled. The option
188 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
189 options in pcretest are provided to set it. It can also be set by
190 (*NO_AUTO_POSSESS) at the start of a pattern.
192 18. The character VT has been added to the default ("C" locale) set of
193 characters that match \s and are generally treated as white space,
194 following this same change in Perl 5.18. There is now no difference between
195 "Perl space" and "POSIX space". Whether VT is treated as white space in
196 other locales depends on the locale.
198 19. The code for checking named groups as conditions, either for being set or
199 for being recursed, has been refactored (this is related to 14 and 15
200 above). Processing unduplicated named groups should now be as fast at
201 numerical groups, and processing duplicated groups should be faster than
202 before.
204 20. Two patches to the CMake build system, by Alexander Barkov:
206 (1) Replace the "source" command by "." in CMakeLists.txt because
207 "source" is a bash-ism.
209 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
210 without these the CMake build does not work on Solaris.
212 21. Perl has changed its handling of \8 and \9. If there is no previously
213 encountered capturing group of those numbers, they are treated as the
214 literal characters 8 and 9 instead of a binary zero followed by the
215 literals. PCRE now does the same.
217 22. Following Perl, added \o{} to specify codepoints in octal, making it
218 possible to specify values greater than 0777 and also making them
219 unambiguous.
221 23. Perl now gives an error for missing closing braces after \x{... instead of
222 treating the string as literal. PCRE now does the same.
224 24. RunTest used to grumble if an inappropriate test was selected explicitly,
225 but just skip it when running all tests. This make it awkward to run ranges
226 of tests when one of them was inappropriate. Now it just skips any
227 inappropriate tests, as it always did when running all tests.
229 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
230 character types such as \d or \w, too many callouts were inserted, and the
231 data that they returned was rubbish.
233 26. In UCP mode, \s was not matching two of the characters that Perl matches,
234 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
235 were matched by \h. The code has now been refactored so that the lists of
236 the horizontal and vertical whitespace characters used for \h and \v (which
237 are defined only in one place) are now also used for \s.
239 27. Add JIT support for the 64 bit TileGX architecture.
240 Patch by Jiong Wang (Tilera Corporation).
242 28. Possessive quantifiers for classes (both explicit and automatically
243 generated) now use special opcodes instead of wrapping in ONCE brackets.
245 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
246 (because it's meaningless), this was not happening when PCRE_CASELESS was
247 set. Not wrong, but inefficient.
249 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
250 properties for \w, \d, etc) is present in a test regex. Otherwise if the
251 test contains no characters greater than 255, Perl doesn't realise it
252 should be using Unicode semantics.
254 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
255 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
256 does in Unicode mode.
258 32. Added the "forbid" facility to pcretest so that putting tests into the
259 wrong test files can sometimes be quickly detected.
261 33. There is now a limit (default 250) on the depth of nesting of parentheses.
262 This limit is imposed to control the amount of system stack used at compile
263 time. It can be changed at build time by --with-parens-nest-limit=xxx or
264 the equivalent in CMake.
266 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
267 errors. Perl warns for these when in warning mode, but PCRE has no facility
268 for giving warnings.
270 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
271 However, this was not working for (?!) because it is optimized to (*FAIL),
272 for which PCRE does not allow quantifiers. The optimization is now disabled
273 when a quantifier follows (?!). I can't see any use for this, but it makes
274 things uniform.
276 36. Perl no longer allows group names to start with digits, so I have made this
277 change also in PCRE. It simplifies the code a bit.
279 37. In extended mode, Perl ignores spaces before a + that indicates a
280 possessive quantifier. PCRE allowed a space before the quantifier, but not
281 before the possessive +. It now does.
283 38. The use of \K (reset reported match start) within a repeated possessive
284 group such as (a\Kb)*+ was not working.
286 40. Document that the same character tables must be used at compile time and
287 run time, and that the facility to pass tables to pcre_exec() and
288 pcre_dfa_exec() is for use only with saved/restored patterns.
290 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
291 features for Builds with MSVC:
293 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
294 builds). The .rc files can be used to set FileDescription and many other
295 attributes.
297 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
298 This allows higher-level build scripts which want .pdb files to avoid
299 hard-coding the exact files needed."
301 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
302 mean "start of word" and "end of word", respectively, as a transition aid.
304 43. A minimizing repeat of a class containing codepoints greater than 255 in
305 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
306 compiled to use the heap for recursion.
308 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
309 is configured.
312 Version 8.33 28-May-2013
313 ------------------------
315 1. Added 'U' to some constants that are compared to unsigned integers, to
316 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
317 variables that are added to signed variables, to ensure the result is
318 signed and can be negated.
320 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
321 CMake config files.
323 3. Revise the creation of config.h.generic so that all boolean macros are
324 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
325 overriding via -D on the command line possible.
327 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
328 to unsigned int is reported to make a quite noticeable speed difference in
329 a specific Windows environment. Testing on Linux did also appear to show
330 some benefit (and it is clearly not harmful). Also fixed the definition of
331 Xop which should be unsigned.
333 5. Related to (4), changing the definition of the intermediate variable cc
334 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
335 improvements.
337 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
338 unnecessary spaces.
340 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
341 and later.
343 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
345 9. Optimizing fast_forward_start_bits in JIT.
347 10. Adding support for callouts in JIT, and fixing some issues revealed
348 during this work. Namely:
350 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
352 (b) Minimum length was not checked before the matching is started.
354 11. The value of capture_last that is passed to callouts was incorrect in some
355 cases when there was a capture on one path that was subsequently abandoned
356 after a backtrack. Also, the capture_last value is now reset after a
357 recursion, since all captures are also reset in this case.
359 12. The interpreter no longer returns the "too many substrings" error in the
360 case when an overflowing capture is in a branch that is subsequently
361 abandoned after a backtrack.
363 13. In the pathological case when an offset vector of size 2 is used, pcretest
364 now prints out the matched string after a yield of 0 or 1.
366 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
367 Only supported by the JIT compiler at the moment.
369 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
371 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
372 offset of the starting point of the matching process, provided the offsets
373 vector is large enough.
375 17. The \A escape now records a lookbehind value of 1, though its execution
376 does not actually inspect the previous character. This is to ensure that,
377 in partial multi-segment matching, at least one character from the old
378 segment is retained when a new segment is processed. Otherwise, if there
379 are no lookbehinds in the pattern, \A might match incorrectly at the start
380 of a new segment.
382 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
384 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
385 gives some modest performance improvement in 8-bit mode.
387 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
388 be expressed in certain programming languages using Universal Character
389 Names.
391 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
392 which points out that "non characters" are not "characters that may not
393 appear in Unicode strings" but rather "characters that are reserved for
394 internal use and have only local meaning".
396 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
397 there was a conditional group that depended on an assertion, if the
398 assertion was false, the callout that immediately followed the alternation
399 in the condition was skipped when pcre_exec() was used for matching.
401 23. Allow an explicit callout to be inserted before an assertion that is the
402 condition for a conditional group, for compatibility with automatic
403 callouts, which always insert a callout at this point.
405 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
406 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
408 25. (*PRUNE) is now supported by the JIT compiler.
410 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
412 27. Fix the case where there are two or more SKIPs with arguments that may be
413 ignored.
415 28. (*SKIP) is now supported by the JIT compiler.
417 29. (*THEN) is now supported by the JIT compiler.
419 30. Update RunTest with additional test selector options.
421 31. The way PCRE handles backtracking verbs has been changed in two ways.
423 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
424 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
425 backtracking. In some cases this makes it more Perl-compatible, but Perl's
426 rather obscure rules do not always do the same thing.
428 (2) Previously, backtracking verbs were confined within assertions. This is
429 no longer the case for positive assertions, except for (*ACCEPT). Again,
430 this sometimes improves Perl compatibility, and sometimes does not.
432 32. A number of tests that were in test 2 because Perl did things differently
433 have been moved to test 1, because either Perl or PCRE has changed, and
434 these tests are now compatible.
436 32. Backtracking control verbs are now handled in the same way in JIT and
437 interpreter.
439 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
440 contained a forward subroutine reference caused a compile error.
442 34. Auto-detect and optimize limited repetitions in JIT.
444 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
445 blocking (*UTF) etc.
447 36. In the interpreter, maximizing pattern repetitions for characters and
448 character types now use tail recursion, which reduces stack usage.
450 37. The value of the max lookbehind was not correctly preserved if a compiled
451 and saved regex was reloaded on a host of different endianness.
453 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
454 of the compiled pattern block, expand the flags field from 16 to 32 bits
455 because it was almost full.
457 39. Try madvise first before posix_madvise.
459 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
460 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
461 original change locked out all matches of empty strings. This has been
462 changed so that one match of an empty string per line is recognized.
463 Subsequent searches on the same line (for colouring or for --only-matching,
464 for example) do not recognize empty strings.
466 41. Applied a user patch to fix a number of spelling mistakes in comments.
468 42. Data lines longer than 65536 caused pcretest to crash.
470 43. Clarified the data type for length and startoffset arguments for pcre_exec
471 and pcre_dfa_exec in the function-specific man pages, where they were
472 explicitly stated to be in bytes, never having been updated. I also added
473 some clarification to the pcreapi man page.
475 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
476 a segmentation fault.
479 Version 8.32 30-November-2012
480 -----------------------------
482 1. Improved JIT compiler optimizations for first character search and single
483 character iterators.
485 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
486 Patch by Daniel Richard G.
488 3. Single character iterator optimizations in the JIT compiler.
490 4. Improved JIT compiler optimizations for character ranges.
492 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
493 Reported by Giuseppe D'Angelo.
495 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
496 of a line, was being set incorrectly in cases where .* appeared inside
497 atomic brackets at the start of a pattern, or where there was a subsequent
498 *PRUNE or *SKIP.
500 7. Improved instruction cache flush for POWER/PowerPC.
501 Patch by Daniel Richard G.
503 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
504 grep:
506 (a) There is now no limit to the number of patterns to be matched.
508 (b) An error is given if a pattern is too long.
510 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
511 are now supported.
513 (d) --exclude-from and --include-from (multiple use) have been added.
515 (e) Exclusions and inclusions now apply to all files and directories, not
516 just to those obtained from scanning a directory recursively.
518 (f) Multiple uses of -f and --file-list are now supported.
520 (g) In a Windows environment, the default for -d has been changed from
521 "read" (the GNU grep default) to "skip", because otherwise the presence
522 of a directory in the file list provokes an error.
524 (h) The documentation has been revised and clarified in places.
526 9. Improve the matching speed of capturing brackets.
528 10. Changed the meaning of \X so that it now matches a Unicode extended
529 grapheme cluster.
531 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
532 out POSIX threads when JIT support is configured.
534 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
536 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
537 structure could go wrong in environments where size_t is not the same size
538 as int.
540 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
541 set.
543 15. The EBCDIC support had decayed; later updates to the code had included
544 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
545 general tidy up of EBCDIC-related issues, and the documentation was also
546 not quite right. There is now a test that can be run on ASCII systems to
547 check some of the EBCDIC-related things (but is it not a full test).
549 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
550 in a small tidy to the code.
552 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
554 18. If the --only-matching (-o) option in pcregrep is specified multiple
555 times, each one causes appropriate output. For example, -o1 -o2 outputs the
556 substrings matched by the 1st and 2nd capturing parentheses. A separating
557 string can be specified by --om-separator (default empty).
559 19. Improving the first n character searches.
561 20. Turn case lists for horizontal and vertical white space into macros so that
562 they are defined only once.
564 21. This set of changes together give more compatible Unicode case-folding
565 behaviour for characters that have more than one other case when UCP
566 support is available.
568 (a) The Unicode property table now has offsets into a new table of sets of
569 three or more characters that are case-equivalent. The MultiStage2.py
570 script that generates these tables (the pcre_ucd.c file) now scans
571 CaseFolding.txt instead of UnicodeData.txt for character case
572 information.
574 (b) The code for adding characters or ranges of characters to a character
575 class has been abstracted into a generalized function that also handles
576 case-independence. In UTF-mode with UCP support, this uses the new data
577 to handle characters with more than one other case.
579 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
580 whose other case is greater than 256 are now correctly matched
581 caselessly. Previously, the high codepoint matched the low one, but not
582 vice versa.
584 (d) The processing of \h, \H, \v, and \ in character classes now makes use
585 of the new class addition function, using character lists defined as
586 macros alongside the case definitions of 20 above.
588 (e) Caseless back references now work with characters that have more than
589 one other case.
591 (f) General caseless matching of characters with more than one other case
592 is supported.
594 22. Unicode character properties were updated from Unicode 6.2.0
596 23. Improved CMake support under Windows. Patch by Daniel Richard G.
598 24. Add support for 32-bit character strings, and UTF-32
600 25. Major JIT compiler update (code refactoring and bugfixing).
601 Experimental Sparc 32 support is added.
603 26. Applied a modified version of Daniel Richard G's patch to create
604 pcre.h.generic and config.h.generic by "make" instead of in the
605 PrepareRelease script.
607 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
608 pcre_compile.c when checking for a zero character.
610 28. Introducing a native interface for JIT. Through this interface, the compiled
611 machine code can be directly executed. The purpose of this interface is to
612 provide fast pattern matching, so several sanity checks are not performed.
613 However, feature tests are still performed. The new interface provides
614 1.4x speedup compared to the old one.
616 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
617 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
618 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
620 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
621 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
622 the "old" RFC 2279). Instead, it ended up passing a negative length to
623 pcre_exec().
625 31. Add support for GCC's visibility feature to hide internal functions.
627 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
628 "unknown -C option" after outputting 0 or 1.
630 33. There is now support for generating a code coverage report for the test
631 suite in environments where gcc is the compiler and lcov is installed. This
632 is mainly for the benefit of the developers.
634 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
635 unaddressable using valgrind annotations, allowing valgrind to detect
636 invalid memory accesses. This is mainly for the benefit of the developers.
638 25. (*UTF) can now be used to start a pattern in any of the three libraries.
640 26. Give configure error if --enable-cpp but no C++ compiler found.
643 Version 8.31 06-July-2012
644 -------------------------
646 1. Fixing a wrong JIT test case and some compiler warnings.
648 2. Removed a bashism from the RunTest script.
650 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
651 to unsigned type, result still unsigned" that was given by an MS compiler
652 on encountering the code "-sizeof(xxx)".
654 4. Partial matching support is added to the JIT compiler.
656 5. Fixed several bugs concerned with partial matching of items that consist
657 of more than one character:
659 (a) /^(..)\1/ did not partially match "aba" because checking references was
660 done on an "all or nothing" basis. This also applied to repeated
661 references.
663 (b) \R did not give a hard partial match if \r was found at the end of the
664 subject.
666 (c) \X did not give a hard partial match after matching one or more
667 characters at the end of the subject.
669 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
670 a partial match for the string "\r".
672 (e) When newline was set to CRLF, the metacharacter "." did not recognize
673 a partial match for a CR character at the end of the subject string.
675 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
676 running pcretest, the text "(JIT)" added to the output whenever JIT is
677 actually used to run the match.
679 7. Individual JIT compile options can be set in pcretest by following -s+[+]
680 or /S+[+] with a digit between 1 and 7.
682 8. OP_NOT now supports any UTF character not just single-byte ones.
684 9. (*MARK) control verb is now supported by the JIT compiler.
686 10. The command "./RunTest list" lists the available tests without actually
687 running any of them. (Because I keep forgetting what they all are.)
691 12. Applied a (slightly modified) user-supplied patch that improves performance
692 when the heap is used for recursion (compiled with --disable-stack-for-
693 recursion). Instead of malloc and free for each heap frame each time a
694 logical recursion happens, frames are retained on a chain and re-used where
695 possible. This sometimes gives as much as 30% improvement.
697 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
698 call.
700 14. As documented, (*COMMIT) is now confined to within a positive assertion.
702 15. It is now possible to link pcretest with libedit as an alternative to
703 libreadline.
705 16. (*COMMIT) control verb is now supported by the JIT compiler.
707 17. The Unicode data tables have been updated to Unicode 6.1.0.
709 18. Added --file-list option to pcregrep.
711 19. Added binary file support to pcregrep, including the -a, --binary-files,
712 -I, and --text options.
714 20. The madvise function is renamed for posix_madvise for QNX compatibility
715 reasons. Fixed by Giuseppe D'Angelo.
717 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
719 22. Changed the option for creating symbolic links for 16-bit man pages from
720 -s to -sf so that re-installing does not cause issues.
722 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
724 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
725 matches in certain environments (the workspace was not being correctly
726 retained). Also added to pcre_dfa_exec() a simple plausibility check on
727 some of the workspace data at the beginning of a restart.
729 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
730 was not doing so when it should - probably a typo introduced by SVN 528
731 (change 8.10/14).
733 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
734 \w+ when the character tables indicated that \x{c4} was a word character.
735 There were several related cases, all because the tests for doing a table
736 lookup were testing for characters less than 127 instead of 255.
738 27. If a pattern contains capturing parentheses that are not used in a match,
739 their slots in the ovector are set to -1. For those that are higher than
740 any matched groups, this happens at the end of processing. In the case when
741 there were back references that the ovector was too small to contain
742 (causing temporary malloc'd memory to be used during matching), and the
743 highest capturing number was not used, memory off the end of the ovector
744 was incorrectly being set to -1. (It was using the size of the temporary
745 memory instead of the true size.)
747 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
748 ovector size, it uses memory at the end of the block that it has got.
750 29. Check for an overlong MARK name and give an error at compile time. The
751 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
753 30. JIT compiler update.
755 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
756 Rill for the patch.
758 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
760 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
762 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
763 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
764 was enabled.
766 35. Improve JIT code generation for greedy plus quantifier.
768 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
769 1 to "aa" instead of to an empty string. The bug affected repeated groups
770 that could potentially match an empty string.
772 37. Optimizing single character iterators in JIT.
774 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
775 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
776 codepoints that are too big for the mode are faulted, and in a UTF mode,
777 disallowed codepoints are also faulted.
779 39. If PCRE was compiled with UTF support, in three places in the DFA
780 matcher there was code that should only have been obeyed in UTF mode, but
781 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
782 processing when bytes with values greater than 127 were present. In 16-bit
783 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
784 both cases the values are those that cannot be the first data item in a UTF
785 character. The three items that might have provoked this were recursions,
786 possessively repeated groups, and atomic groups.
788 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
789 and pcregrep, because some OS require shared objects to be explicitly
790 passed to ld, causing the link step to fail if they are not.
792 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
793 mode, patterns that started with \h* or \R* might be incorrectly matched.
796 Version 8.30 04-February-2012
797 -----------------------------
799 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
800 name is defined in ctype.h.
802 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
803 only in quite long subpatterns.
805 3. Removed the function pcre_info(), which has been obsolete and deprecated
806 since it was replaced by pcre_fullinfo() in February 2000.
808 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
809 match a (*MARK), and the match failed at the start of the subject, a
810 reference to memory before the start of the subject could occur. This bug
811 was introduced by fix 17 of release 8.21.
813 5. A reference to an unset group with zero minimum repetition was giving
814 totally wrong answers (in non-JavaScript-compatibility mode). For example,
815 /(another)?(\1?)test/ matched against "hello world test". This bug was
816 introduced in release 8.13.
818 6. Add support for 16-bit character strings (a large amount of work involving
819 many changes and refactorings).
821 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
822 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
823 from a file.
825 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
826 rounding is not applied in this particular case).
828 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
829 if they appear, or are escaped, in patterns.
831 10. Get rid of a number of -Wunused-but-set-variable warnings.
833 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
834 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
835 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
836 also returns the mark "x". This bug applied to capturing parentheses,
837 non-capturing parentheses, and atomic parentheses. It also applied to some
838 assertions.
840 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
841 information out of configure.ac instead of relying on pcre.h.generic, which
842 is not stored in the repository.
844 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
845 -lreadline.
847 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
849 15. Applied Graycode's patch to put the top-level frame on the stack rather
850 than the heap when not using the stack for recursion. This gives a
851 performance improvement in many cases when recursion is not deep.
853 16. Experimental code added to "pcretest -C" to output the stack frame size.
856 Version 8.21 12-Dec-2011
857 ------------------------
859 1. Updating the JIT compiler.
861 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
862 are added as well.
864 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
865 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
866 calling _pcre_jit_exec. Some extra comments are added.
868 4. (*MARK) settings inside atomic groups that do not contain any capturing
869 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
870 was introduced by change 18 for 8.20.
872 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
873 ECMA-262 standard.
875 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
876 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
877 This bug was probably introduced by change 9 of 8.13.
879 7. While fixing 6 above, I noticed that a number of other items were being
880 incorrectly rejected as "not fixed length". This arose partly because newer
881 opcodes had not been added to the fixed-length checking code. I have (a)
882 corrected the bug and added tests for these items, and (b) arranged for an
883 error to occur if an unknown opcode is encountered while checking for fixed
884 length instead of just assuming "not fixed length". The items that were
885 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
886 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
887 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
889 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
890 being incorrectly compiled and would have given unpredicatble results.
892 9. A possessively repeated subpattern with minimum repeat count greater than
893 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
894 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
895 the first (A) could occur when it should not.
897 10. Add a cast and remove a redundant test from the code.
899 11. JIT should use pcre_malloc/pcre_free for allocation.
901 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
902 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
903 is anything other than /usr, -L is still shown).
905 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
907 14. Perl does not support \N without a following name in a [] class; PCRE now
908 also gives an error.
910 15. If a forward reference was repeated with an upper limit of around 2000,
911 it caused the error "internal error: overran compiling workspace". The
912 maximum number of forward references (including repeats) was limited by the
913 internal workspace, and dependent on the LINK_SIZE. The code has been
914 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
915 the default depends on LINK_SIZE. There is a new upper limit (for safety)
916 of around 200,000 forward references. While doing this, I also speeded up
917 the filling in of repeated forward references.
919 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
920 incorrectly expecting the subject to contain another "a" after the start.
922 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
923 in the match, the SKIP should be ignored. This was not happening; instead
924 the SKIP was being treated as NOMATCH. For patterns such as
925 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
926 tested.
928 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
929 now much more compatible with Perl, in particular in cases where the result
930 is a non-match for a non-anchored pattern. For example, if
931 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
932 "m", where previously it did not return a name. A side effect of this
933 change is that for partial matches, the last encountered mark name is
934 returned, as for non matches. A number of tests that were previously not
935 Perl-compatible have been moved into the Perl-compatible test files. The
936 refactoring has had the pleasing side effect of removing one argument from
937 the match() function, thus reducing its stack requirements.
939 19. If the /S+ option was used in pcretest to study a pattern using JIT,
940 subsequent uses of /S (without +) incorrectly behaved like /S+.
942 21. Retrieve executable code size support for the JIT compiler and fixing
943 some warnings.
945 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
946 not work when the shorter character appeared right at the end of the
947 subject string.
949 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
950 systems.
952 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
953 output it when the /M option is used in pcretest.
955 25. The CheckMan script was not being included in the distribution. Also, added
956 an explicit "perl" to run Perl scripts from the PrepareRelease script
957 because this is reportedly needed in Windows.
959 26. If study data was being save in a file and studying had not found a set of
960 "starts with" bytes for the pattern, the data written to the file (though
961 never used) was taken from uninitialized memory and so caused valgrind to
962 complain.
964 27. Updated RunTest.bat as provided by Sheri Pierce.
966 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
968 29. Computation of memory usage for the table of capturing group names was
969 giving an unnecessarily large value.
972 Version 8.20 21-Oct-2011
973 ------------------------
975 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
976 a POSIX class. After further experiments with Perl, which convinced me that
977 Perl has bugs and confusions, a closing square bracket is no longer allowed
978 in a POSIX name. This bug also affected patterns with classes that started
979 with full stops.
981 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
982 captured substring, but while checking the failing first alternative,
983 substring 1 is temporarily captured. If the output vector supplied to
984 pcre_exec() was not big enough for this capture, the yield of the function
985 was still zero ("insufficient space for captured substrings"). This cannot
986 be totally fixed without adding another stack variable, which seems a lot
987 of expense for a edge case. However, I have improved the situation in cases
988 such as /(a)(b)x|abc/ matched against "abc", where the return code
989 indicates that fewer than the maximum number of slots in the ovector have
990 been set.
992 3. Related to (2) above: when there are more back references in a pattern than
993 slots in the output vector, pcre_exec() uses temporary memory during
994 matching, and copies in the captures as far as possible afterwards. It was
995 using the entire output vector, but this conflicts with the specification
996 that only 2/3 is used for passing back captured substrings. Now it uses
997 only the first 2/3, for compatibility. This is, of course, another edge
998 case.
1000 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1001 main code base, and can be used by building with --enable-jit. When this is
1002 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1003 runtime --no-jit option is given.
1005 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1006 ovector, the return from the function was zero, implying that there were
1007 other matches that did not fit. The correct "exactly full" value is now
1008 returned.
1010 6. If a subpattern that was called recursively or as a subroutine contained
1011 (*PRUNE) or any other control that caused it to give a non-standard return,
1012 invalid errors such as "Error -26 (nested recursion at the same subject
1013 position)" or even infinite loops could occur.
1015 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1016 computing the minimum length on reaching *ACCEPT, and so ended up with the
1017 wrong value of 1 rather than 0. Further investigation indicates that
1018 computing a minimum subject length in the presence of *ACCEPT is difficult
1019 (think back references, subroutine calls), and so I have changed the code
1020 so that no minimum is registered for a pattern that contains *ACCEPT.
1022 8. If (*THEN) was present in the first (true) branch of a conditional group,
1023 it was not handled as intended. [But see 16 below.]
1025 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1026 Sheri Pierce.
1028 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1029 the first byte in a match must be "a".
1031 11. Change 17 for 8.13 increased the recursion depth for patterns like
1032 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1033 pattern contains any instances of (*THEN). If it does not, the old
1034 optimizations are restored. It would be nice to do this on a per-group
1035 basis, but at the moment that is not feasible.
1037 12. In some environments, the output of pcretest -C is CRLF terminated. This
1038 broke RunTest's code that checks for the link size. A single white space
1039 character after the value is now allowed for.
1041 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1042 For "fr", it uses the Windows-specific input and output files.
1044 14. If (*THEN) appeared in a group that was called recursively or as a
1045 subroutine, it did not work as intended. [But see next item.]
1047 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1048 pattern fragments (but not containing any | characters). If A and B are
1049 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1050 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1051 D. In other words, Perl considers parentheses that do not contain any |
1052 characters to be part of a surrounding alternative, whereas PCRE was
1053 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1054 differently. PCRE now behaves in the same way as Perl, except in the case
1055 of subroutine/recursion calls such as (?1) which have in any case always
1056 been different (but PCRE had them first :-).
1058 16. Related to 15 above: Perl does not treat the | in a conditional group as
1059 creating alternatives. Such a group is treated in the same way as an
1060 ordinary group without any | characters when processing (*THEN). PCRE has
1061 been changed to match Perl's behaviour.
1063 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1064 RunGrepTest script failed.
1066 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1067 inevitable for groups that contain captures, but it can lead to a lot of
1068 stack use in large patterns. The old behaviour has been restored for atomic
1069 groups that do not contain any capturing parentheses.
1071 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1072 suppress the check for a minimum subject length at run time. (If it was
1073 given to pcre_exec() or pcre_dfa_exec() it did work.)
1075 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1076 fail to work when decoding hex characters in data strings in EBCDIC
1077 environments.
1079 21. It appears that in at least one Mac OS environment, the isxdigit() function
1080 is implemented as a macro that evaluates to its argument more than once,
1081 contravening the C 90 Standard (I haven't checked a later standard). There
1082 was an instance in pcretest which caused it to go wrong when processing
1083 \x{...} escapes in subject strings. The has been rewritten to avoid using
1084 things like p++ in the argument of isxdigit().
1087 Version 8.13 16-Aug-2011
1088 ------------------------
1090 1. The Unicode data tables have been updated to Unicode 6.0.0.
1092 2. Two minor typos in pcre_internal.h have been fixed.
1094 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1095 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1096 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1098 4. There were a number of related bugs in the code for matching backrefences
1099 caselessly in UTF-8 mode when codes for the characters concerned were
1100 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1101 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1102 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1103 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1104 2-byte code at the end of the subject (it thought there wasn't enough data
1105 left).
1107 5. Comprehensive information about what went wrong is now returned by
1108 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1109 as the output vector has at least 2 elements. The offset of the start of
1110 the failing character and a reason code are placed in the vector.
1112 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1113 now returned is for the first byte of the failing character, instead of the
1114 last byte inspected. This is an incompatible change, but I hope it is small
1115 enough not to be a problem. It makes the returned offset consistent with
1116 pcre_exec() and pcre_dfa_exec().
1118 7. pcretest now gives a text phrase as well as the error number when
1119 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1120 failure, the offset and reason code are output.
1122 8. When \R was used with a maximizing quantifier it failed to skip backwards
1123 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1124 back over a single character (\n). This seems wrong (because it treated the
1125 two characters as a single entity when going forwards), conflicts with the
1126 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1127 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1128 has been changed.
1130 9. Some internal refactoring has changed the processing so that the handling
1131 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1132 time (the PCRE_DOTALL option was changed this way some time ago: version
1133 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1134 which was always a bit of a fudge. It also means that there is one less
1135 argument for the match() function, which reduces its stack requirements
1136 slightly. This change also fixes an incompatibility with Perl: the pattern
1137 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1139 10. More internal refactoring has drastically reduced the number of recursive
1140 calls to match() for possessively repeated groups such as (abc)++ when
1141 using pcre_exec().
1143 11. While implementing 10, a number of bugs in the handling of groups were
1144 discovered and fixed:
1146 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1147 (a|)*(?1) gave a compile-time internal error.
1148 ((a|)+)+ did not notice that the outer group could match an empty string.
1149 (^a|^)+ was not marked as anchored.
1150 (.*a|.*)+ was not marked as matching at start or after a newline.
1152 12. Yet more internal refactoring has removed another argument from the match()
1153 function. Special calls to this function are now indicated by setting a
1154 value in a variable in the "match data" data block.
1156 13. Be more explicit in pcre_study() instead of relying on "default" for
1157 opcodes that mean there is no starting character; this means that when new
1158 ones are added and accidentally left out of pcre_study(), testing should
1159 pick them up.
1161 14. The -s option of pcretest has been documented for ages as being an old
1162 synonym of -m (show memory usage). I have changed it to mean "force study
1163 for every regex", that is, assume /S for every regex. This is similar to -i
1164 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1165 using it. It makes it easier to run collections of tests with and without
1166 study enabled, and thereby test pcre_study() more easily. All the standard
1167 tests are now run with and without -s (but some patterns can be marked as
1168 "never study" - see 20 below).
1170 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1171 restoration of the capturing data to the outer values was not happening
1172 correctly.
1174 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1175 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1176 pattern had matched an empty string, and so incorrectly returned a no
1177 match.
1179 17. There was optimizing code for the last branch of non-capturing parentheses,
1180 and also for the obeyed branch of a conditional subexpression, which used
1181 tail recursion to cut down on stack usage. Unfortunately, now that there is
1182 the possibility of (*THEN) occurring in these branches, tail recursion is
1183 no longer possible because the return has to be checked for (*THEN). These
1184 two optimizations have therefore been removed. [But see 8.20/11 above.]
1186 18. If a pattern containing \R was studied, it was assumed that \R always
1187 matched two bytes, thus causing the minimum subject length to be
1188 incorrectly computed because \R can also match just one byte.
1190 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1191 was incorrectly computed.
1193 20. If /S is present twice on a test pattern in pcretest input, it now
1194 *disables* studying, thereby overriding the use of -s on the command line
1195 (see 14 above). This is necessary for one or two tests to keep the output
1196 identical in both cases.
1198 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1199 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1201 22. When an atomic group that contained a capturing parenthesis was
1202 successfully matched, but the branch in which it appeared failed, the
1203 capturing was not being forgotten if a higher numbered group was later
1204 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1205 group 1 to "a", when in fact it should be unset. This applied to multi-
1206 branched capturing and non-capturing groups, repeated or not, and also to
1207 positive assertions (capturing in negative assertions does not happen
1208 in PCRE) and also to nested atomic groups.
1210 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1211 subject after a captured substring, to make it easier to tell which of a
1212 number of identical substrings has been captured.
1214 24. The way atomic groups are processed by pcre_exec() has been changed so that
1215 if they are repeated, backtracking one repetition now resets captured
1216 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1217 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1218 "aaa". Previously, it would have been "a". As part of this code
1219 refactoring, the way recursive calls are handled has also been changed.
1221 25. If an assertion condition captured any substrings, they were not passed
1222 back unless some other capturing happened later. For example, if
1223 (?(?=(a))a) was matched against "a", no capturing was returned.
1225 26. When studying a pattern that contained subroutine calls or assertions,
1226 the code for finding the minimum length of a possible match was handling
1227 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1228 group 1 called group 2 while simultaneously a separate group 2 called group
1229 1). A stack overflow occurred in this case. I have fixed this by limiting
1230 the recursion depth to 10.
1232 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1233 Fortmann. This supports explicit test numbers on the command line, and has
1234 argument validation and error reporting.
1236 28. An instance of \X with an unlimited repeat could fail if at any point the
1237 first character it looked at was a mark character.
1239 29. Some minor code refactoring concerning Unicode properties and scripts
1240 should reduce the stack requirement of match() slightly.
1242 30. Added the '=' option to pcretest to check the setting of unused capturing
1243 slots at the end of the pattern, which are documented as being -1, but are
1244 not included in the return count.
1246 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1247 compiled something random. Now it gives a compile-time error (as does
1248 Perl).
1250 32. A *MARK encountered during the processing of a positive assertion is now
1251 recorded and passed back (compatible with Perl).
1253 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1254 had alternative anchored branches, the search for a second match in a line
1255 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1256 matched the line "0102" twice. The same bug affected patterns that started
1257 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1258 twice.
1260 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1261 does, and because of capturing effects, quantifying parenthesized
1262 assertions may at times be useful. Quantifiers are now allowed for
1263 parenthesized assertions.
1265 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1267 36. \g was being checked for fancy things in a character class, when it should
1268 just be a literal "g".
1270 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1271 appearance of a nested POSIX class supersedes an apparent external class.
1272 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1273 unescaped square brackets may also appear as part of class names. For
1274 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1275 more like Perl. (But see 8.20/1 above.)
1277 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1278 was because it thought it was \N{name}, which is not supported).
1280 39. Add minix to OS list not supporting the -S option in pcretest.
1282 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1283 cannot analyze patterns in sufficient detail to catch mutual recursions
1284 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1285 subgroup is called recursively as a subpattern for a second time at the
1286 same position in the subject string. In previous releases this might have
1287 been caught by the recursion limit, or it might have run out of stack.
1289 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1290 happen only once. PCRE was, however incorrectly giving a compile time error
1291 "recursive call could loop indefinitely" because it cannot analyze the
1292 pattern in sufficient detail. The compile time test no longer happens when
1293 PCRE is compiling a conditional subpattern, but actual runaway loops are
1294 now caught at runtime (see 40 above).
1296 42. It seems that Perl allows any characters other than a closing parenthesis
1297 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1298 has been changed to be the same.
1300 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1301 as not to get warnings when autogen.sh is called. Also changed
1302 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1304 44. To help people who use pcregrep to scan files containing exceedingly long
1305 lines, the following changes have been made:
1307 (a) The default value of the buffer size parameter has been increased from
1308 8K to 20K. (The actual buffer used is three times this size.)
1310 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1311 PCRE is built.
1313 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1314 to be set at run time.
1316 (d) Numerical values in pcregrep options can be followed by K or M, for
1317 example --buffer-size=50K.
1319 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1320 given and the return code is set to 2.
1322 45. Add a pointer to the latest mark to the callout data block.
1324 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1325 partial match of an empty string instead of no match. This was specific to
1326 the use of ".".
1328 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1329 complete match instead of a partial match. This bug was dependent on both
1330 the PCRE_UTF8 and PCRE_DOTALL options being set.
1332 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1333 starting byte set, because \b was not being ignored.
1336 Version 8.12 15-Jan-2011
1337 ------------------------
1339 1. Fixed some typos in the markup of the man pages, and wrote a script that
1340 checks for such things as part of the documentation building process.
1342 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1343 --match-limit and --recursion-limit options (added for 8.11). In
1344 particular, this made one of the standard tests fail. (The integer value
1345 went into the wrong half of a long int.)
1347 3. If the --colour option was given to pcregrep with -v (invert match), it
1348 did strange things, either producing crazy output, or crashing. It should,
1349 of course, ignore a request for colour when reporting lines that do not
1350 match.
1352 4. Another pcregrep bug caused similar problems if --colour was specified with
1353 -M (multiline) and the pattern match finished with a line ending.
1355 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1356 matched in multiline mode, the following line was shown as part of the
1357 match. This seems wrong, so I have changed it.
1359 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1360 the check for further matches in the same line (so they could be coloured)
1361 to overrun the end of the current line. If another match was found, it was
1362 incorrectly shown (and then shown again when found in the next line).
1364 7. If pcregrep was compiled under Windows, there was a reference to the
1365 function pcregrep_exit() before it was defined. I am assuming this was
1366 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1367 reported by a user. I've moved the definition above the reference.
1370 Version 8.11 10-Dec-2010
1371 ------------------------
1373 1. (*THEN) was not working properly if there were untried alternatives prior
1374 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1375 backtracked to try for "b" instead of moving to the next alternative branch
1376 at the same level (in this case, to look for "c"). The Perl documentation
1377 is clear that when (*THEN) is backtracked onto, it goes to the "next
1378 alternative in the innermost enclosing group".
1380 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1381 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1382 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1383 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1384 (*THEN).
1386 3. If \s appeared in a character class, it removed the VT character from
1387 the class, even if it had been included by some previous item, for example
1388 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1389 of \s, but is part of the POSIX "space" class.)
1391 4. A partial match never returns an empty string (because you can always
1392 match an empty string at the end of the subject); however the checking for
1393 an empty string was starting at the "start of match" point. This has been
1394 changed to the "earliest inspected character" point, because the returned
1395 data for a partial match starts at this character. This means that, for
1396 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1397 (previously it gave "no match").
1399 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1400 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1401 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1402 has an implication that the given string is incomplete (because a partial
1403 match is preferred over a full match). For this reason, these items now
1404 give a partial match in this situation. [Aside: previously, the one case
1405 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1406 match rather than a full match, which was wrong by the old rules, but is
1407 now correct.]
1409 6. There was a bug in the handling of #-introduced comments, recognized when
1410 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1411 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1412 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1413 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1414 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1415 places in pcre_compile().
1417 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1418 comments when looking ahead for named forward references to subpatterns,
1419 the only newline sequence it recognized was NL. It now handles newlines
1420 according to the set newline convention.
1422 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1423 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1424 cater for a lack of strerror(). These oversights have been fixed.
1426 9. Added --match-limit and --recursion-limit to pcregrep.
1428 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1430 11. When the -o option was used, pcregrep was setting a return code of 1, even
1431 when matches were found, and --line-buffered was not being honoured.
1433 12. Added an optional parentheses number to the -o and --only-matching options
1434 of pcregrep.
1436 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1437 can match an empty string. The code to do it in pcretest and pcredemo
1438 needed fixing:
1440 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1441 only one byte after an empty string match just before CRLF (this case
1442 just got forgotten; "any" and "anycrlf" were OK).
1444 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1445 mode when an empty string match preceded an ASCII character followed by
1446 a non-ASCII character. (The code for advancing by one character rather
1447 than one byte was nonsense.)
1449 (c) The pcredemo.c sample program did not have any code at all to handle
1450 the cases when CRLF is a valid newline sequence.
1452 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1453 as a starting offset was within the subject string. There is now a new
1454 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1455 negative or greater than the length of the string. In order to test this,
1456 pcretest is extended to allow the setting of negative starting offsets.
1458 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1459 starting offset points to the beginning of a UTF-8 character was
1460 unnecessarily clumsy. I tidied it up.
1462 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1463 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1465 17. Nobody had reported that the --include_dir option, which was added in
1466 release 7.7 should have been called --include-dir (hyphen, not underscore)
1467 for compatibility with GNU grep. I have changed it to --include-dir, but
1468 left --include_dir as an undocumented synonym, and the same for
1469 --exclude-dir, though that is not available in GNU grep, at least as of
1470 release 2.5.4.
1472 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1473 characters from a string of bytes) have been redefined so as not to use
1474 loops, in order to improve performance in some environments. At the same
1475 time, I abstracted some of the common code into auxiliary macros to save
1476 repetition (this should not affect the compiled code).
1478 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1479 compile-time error is now given if \c is not followed by an ASCII
1480 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1481 different, and any byte value is allowed.)
1483 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1484 START_OPTIMIZE option, which is now allowed at compile time - but just
1485 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1486 to pcregrep and other applications that have no direct access to PCRE
1487 options. The new /Y option in pcretest sets this option when calling
1488 pcre_compile().
1490 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1491 back references. Groups containing recursive back references were forced to
1492 be atomic by that change, but in the case of named groups, the amount of
1493 memory required was incorrectly computed, leading to "Failed: internal
1494 error: code overflow". This has been fixed.
1496 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1497 pcretest.c, to avoid build problems in some Borland environments.
1500 Version 8.10 25-Jun-2010
1501 ------------------------
1503 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1504 THEN.
1506 2. (*ACCEPT) was not working when inside an atomic group.
1508 3. Inside a character class, \B is treated as a literal by default, but
1509 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1510 causes the error). The code is unchanged, but I tidied the documentation.
1512 4. Inside a character class, PCRE always treated \R and \X as literals,
1513 whereas Perl faults them if its -w option is set. I have changed PCRE so
1514 that it faults them when PCRE_EXTRA is set.
1516 5. Added support for \N, which always matches any character other than
1517 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1519 6. When compiling pcregrep with newer versions of gcc which may have
1520 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1521 declared with attribute warn_unused_result" were given. Just casting the
1522 result to (void) does not stop the warnings; a more elaborate fudge is
1523 needed. I've used a macro to implement this.
1525 7. Minor change to pcretest.c to avoid a compiler warning.
1527 8. Added four artifical Unicode properties to help with an option to make
1528 \s etc use properties (see next item). The new properties are: Xan
1529 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1531 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1532 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1533 this option. Modified pcretest to add /W to test this facility. Added
1534 REG_UCP to make it available via the POSIX interface.
1536 10. Added --line-buffered to pcregrep.
1538 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1539 studied, and the match started with a letter with a code point greater than
1540 127 whose first byte was different to the first byte of the other case of
1541 the letter, the other case of this starting letter was not recognized
1542 (#976).
1544 12. If a pattern that was studied started with a repeated Unicode property
1545 test, for example, \p{Nd}+, there was the theoretical possibility of
1546 setting up an incorrect bitmap of starting bytes, but fortunately it could
1547 not have actually happened in practice until change 8 above was made (it
1548 added property types that matched character-matching opcodes).
1550 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1551 possible starting bytes for non-anchored patterns.
1553 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1554 \R, and also a number of cases that involve Unicode properties, both
1555 explicit and implicit when PCRE_UCP is set.
1557 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1558 input, it could crash or give wrong results if characters with values
1559 greater than 0xc0 were present in the subject string. (Detail: it assumed
1560 UTF-8 input when processing these items.)
1562 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1563 size_t is 64-bit (#991).
1565 17. Added a check for running out of memory when PCRE is compiled with
1566 --disable-stack-for-recursion (#990).
1568 18. If the last data line in a file for pcretest does not have a newline on
1569 the end, a newline was missing in the output.
1571 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1572 less than 128) in its various bitmaps. However, there is a facility for
1573 generating tables according to the current locale when PCRE is compiled. It
1574 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1575 characters, are recognized by isspace() and therefore were getting set in
1576 these tables, and indeed these tables seem to approximate to ISO 8859. This
1577 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1578 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1579 which of course cannot start UTF-8 characters. I have changed the code so
1580 that only real ASCII characters (less than 128) and the correct starting
1581 bytes for UTF-8 encodings are set for characters greater than 127 when in
1582 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1583 altogether.)
1585 20. Added the /T option to pcretest so as to be able to run tests with non-
1586 standard character tables, thus making it possible to include the tests
1587 used for 19 above in the standard set of tests.
1589 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1590 reference to a subpattern the other side of a comment that contains an
1591 opening parenthesis caused either an internal compiling error, or a
1592 reference to the wrong subpattern.
1595 Version 8.02 19-Mar-2010
1596 ------------------------
1598 1. The Unicode data tables have been updated to Unicode 5.2.0.
1600 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1601 configured.
1603 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1604 original author of that file, following a query about its status.
1606 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1607 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1609 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1610 quantifier applied to a forward-referencing subroutine call, could compile
1611 incorrect code or give the error "internal error: previously-checked
1612 referenced subpattern not found".
1614 6. Both MS Visual Studio and Symbian OS have problems with initializing
1615 variables to point to external functions. For these systems, therefore,
1616 pcre_malloc etc. are now initialized to local functions that call the
1617 relevant global functions.
1619 7. There were two entries missing in the vectors called coptable and poptable
1620 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1621 I've fixed the data, and added a kludgy way of testing at compile time that
1622 the lengths are correct (equal to the number of opcodes).
1624 8. Following on from 7, I added a similar kludge to check the length of the
1625 eint vector in pcreposix.c.
1627 9. Error texts for pcre_compile() are held as one long string to avoid too
1628 much relocation at load time. To find a text, the string is searched,
1629 counting zeros. There was no check for running off the end of the string,
1630 which could happen if a new error number was added without updating the
1631 string.
1633 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1635 11. \K was not working if it appeared in an atomic group or in a group that
1636 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1637 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1638 the assertion is positive, but not if it is negative.
1640 12. Change 11 fortuitously reduced the size of the stack frame used in the
1641 "match()" function of pcre_exec.c by one pointer. Forthcoming
1642 implementation of support for (*MARK) will need an extra pointer on the
1643 stack; I have reserved it now, so that the stack frame size does not
1644 decrease.
1646 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1647 item in branch that calls a recursion is a subroutine call - as in the
1648 second branch in the above example - was incorrectly given the compile-
1649 time error "recursive call could loop indefinitely" because pcre_compile()
1650 was not correctly checking the subroutine for matching a non-empty string.
1652 14. The checks for overrunning compiling workspace could trigger after an
1653 overrun had occurred. This is a "should never occur" error, but it can be
1654 triggered by pathological patterns such as hundreds of nested parentheses.
1655 The checks now trigger 100 bytes before the end of the workspace.
1657 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1660 Version 8.01 19-Jan-2010
1661 ------------------------
1663 1. If a pattern contained a conditional subpattern with only one branch (in
1664 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1665 computed the wrong minimum data length (which is of course zero for such
1666 subpatterns). This could cause incorrect "no match" results.
1668 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1669 the pattern is reset in the first branch, pcre_compile() failed with
1670 "internal error: code overflow at offset...". This happened only when
1671 the reset was to the original external option setting. (An optimization
1672 abstracts leading options settings into an external setting, which was the
1673 cause of this.)
1675 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1676 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1677 assertion pattern did not match (meaning that the assertion was true), it
1678 was incorrectly treated as false if the SKIP had been reached during the
1679 matching. This also applied to assertions used as conditions.
1681 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1682 assertion subpattern, including such a pattern used as a condition,
1683 unpredictable results occurred, instead of the error return
1686 5. The C++ GlobalReplace function was not working like Perl for the special
1687 situation when an empty string is matched. It now does the fancy magic
1688 stuff that is necessary.
1690 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1691 removed. (These were left over from very, very early versions of PCRE.)
1693 7. Some cosmetic changes to the code to make life easier when compiling it
1694 as part of something else:
1696 (a) Change DEBUG to PCRE_DEBUG.
1698 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1699 called "current" as "current_branch", to prevent a collision with the
1700 Linux macro when compiled as a kernel module.
1702 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1703 prevent a collision with the Linux macro when compiled as a kernel
1704 module.
1706 8. In pcre_compile() there are some checks for integer overflows that used to
1707 cast potentially large values to (double). This has been changed to that
1708 when building, a check for int64_t is made, and if it is found, it is used
1709 instead, thus avoiding the use of floating point arithmetic. (There is no
1710 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1711 double.
1713 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1714 2005 (difference between two addresses compared to an unsigned value).
1716 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1717 custom one, because of the following reported problem in Windows:
1719 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1720 under Win32.
1721 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1722 therefore missing the function definition.
1723 - The compiler thus generates a "C" signature for the test function.
1724 - The linker fails to find the "C" function.
1725 - PCRE fails to configure if asked to do so against libbz2.
1727 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1728 messages were output:
1730 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1731 rerunning libtoolize, to keep the correct libtool macros in-tree.
1732 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1734 I have done both of these things.
1736 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1737 most of the time, it *can* run out if it is given a pattern that contains a
1738 runaway infinite recursion. I updated the discussion in the pcrestack man
1739 page.
1741 13. Now that we have gone to the x.xx style of version numbers, the minor
1742 version may start with zero. Using 08 or 09 is a bad idea because users
1743 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1744 interpreted as invalid octal numbers. I've updated the previous comment in
1745 configure.ac, and also added a check that gives an error if 08 or 09 are
1746 used.
1748 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1749 causing partial matching to fail when the end of the subject matched \W
1750 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1752 15. There were some discrepancies between the declarations in pcre_internal.h
1753 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1754 their definitions. The declarations used "const uschar *" and the
1755 definitions used USPTR. Even though USPTR is normally defined as "const
1756 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1757 reported that: "This difference in casting confuses some C++ compilers, for
1758 example, SunCC recognizes above declarations as different functions and
1759 generates broken code for hbpcre." I have changed the declarations to use
1760 USPTR.
1762 16. GNU libtool is named differently on some systems. The autogen.sh script now
1763 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1764 (FreeBSD).
1766 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1767 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1768 comment: "Figure out how to create a longlong from a string: strtoll and
1769 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1770 instance, but it only takes 2 args instead of 3!"
1772 18. A subtle bug concerned with back references has been fixed by a change of
1773 specification, with a corresponding code fix. A pattern such as
1774 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1775 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1776 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1777 same bug. Such groups have to be quantified to be useful, or contained
1778 inside another quantified group. (If there's no repetition, the reference
1779 can never match.) The problem arises because, having left the group and
1780 moved on to the rest of the pattern, a later failure that backtracks into
1781 the group uses the captured value from the final iteration of the group
1782 rather than the correct earlier one. I have fixed this in PCRE by forcing
1783 any group that contains a reference to itself to be an atomic group; that
1784 is, there cannot be any backtracking into it once it has completed. This is
1785 similar to recursive and subroutine calls.
1788 Version 8.00 19-Oct-09
1789 ----------------------
1791 1. The table for translating pcre_compile() error codes into POSIX error codes
1792 was out-of-date, and there was no check on the pcre_compile() error code
1793 being within the table. This could lead to an OK return being given in
1794 error.
1796 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1797 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1798 in a Windows environment.
1800 3. The pcregrep --count option prints the count for each file even when it is
1801 zero, as does GNU grep. However, pcregrep was also printing all files when
1802 --files-with-matches was added. Now, when both options are given, it prints
1803 counts only for those files that have at least one match. (GNU grep just
1804 prints the file name in this circumstance, but including the count seems
1805 more useful - otherwise, why use --count?) Also ensured that the
1806 combination -clh just lists non-zero counts, with no names.
1808 4. The long form of the pcregrep -F option was incorrectly implemented as
1809 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1810 but it seems right to fix it, and I didn't think it was worth preserving
1811 the old behaviour.
1813 5. The command line items --regex=pattern and --regexp=pattern were not
1814 recognized by pcregrep, which required --regex pattern or --regexp pattern
1815 (with a space rather than an '='). The man page documented the '=' forms,
1816 which are compatible with GNU grep; these now work.
1818 6. No libpcreposix.pc file was created for pkg-config; there was just
1819 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1821 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1822 when UCP support is not needed, by modifying the Python script that
1823 generates it from Unicode data files. This should not matter if the module
1824 is correctly used as a library, but I received one complaint about 50K of
1825 unwanted data. My guess is that the person linked everything into his
1826 program rather than using a library. Anyway, it does no harm.
1828 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1829 was a minimum greater than 1 for a wide character in a possessive
1830 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1831 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1832 character. Chaos in the form of incorrect output or a compiling loop could
1833 result.
1835 9. The restrictions on what a pattern can contain when partial matching is
1836 requested for pcre_exec() have been removed. All patterns can now be
1837 partially matched by this function. In addition, if there are at least two
1838 slots in the offset vector, the offset of the earliest inspected character
1839 for the match and the offset of the end of the subject are set in them when
1840 PCRE_ERROR_PARTIAL is returned.
1842 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1843 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1844 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1845 and may be more useful for multi-segment matching.
1847 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1848 used to be given if ever the end of the subject was reached; now it is
1849 given only if matching could not proceed because another character was
1850 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1851 string "Z", which now yields "no match" instead of "partial match". In the
1852 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1853 final character ended with (*FAIL).
1855 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1856 if the pattern had a "must contain" character that was already found in the
1857 earlier partial match, unless partial matching was again requested. For
1858 example, with the pattern /dog.(body)?/, the "must contain" character is
1859 "g". If the first part-match was for the string "dog", restarting with
1860 "sbody" failed. This bug has been fixed.
1862 13. The string returned by pcre_dfa_exec() after a partial match has been
1863 changed so that it starts at the first inspected character rather than the
1864 first character of the match. This makes a difference only if the pattern
1865 starts with a lookbehind assertion or \b or \B (\K is not supported by
1866 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1867 matching functions compatible, and I think it's the right thing to do.
1869 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1870 so that the demonstration program is easily available in environments where
1871 PCRE has not been installed from source.
1873 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1874 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1875 library.
1877 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1878 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1879 is not the first non-POSIX option to be added. Clearly some people find
1880 these options useful.
1882 17. If a caller to the POSIX matching function regexec() passes a non-zero
1883 value for nmatch with a NULL value for pmatch, the value of
1884 nmatch is forced to zero.
1886 18. RunGrepTest did not have a test for the availability of the -u option of
1887 the diff command, as RunTest does. It now checks in the same way as
1888 RunTest, and also checks for the -b option.
1890 19. If an odd number of negated classes containing just a single character
1891 interposed, within parentheses, between a forward reference to a named
1892 subpattern and the definition of the subpattern, compilation crashed with
1893 an internal error, complaining that it could not find the referenced
1894 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1895 [The bug was that it was starting one character too far in when skipping
1896 over the character class, thus treating the ] as data rather than
1897 terminating the class. This meant it could skip too much.]
1899 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1900 /g option in pcretest when the pattern contains \K, which makes it possible
1901 to have an empty string match not at the start, even when the pattern is
1902 anchored. Updated pcretest and pcredemo to use this option.
1904 21. If the maximum number of capturing subpatterns in a recursion was greater
1905 than the maximum at the outer level, the higher number was returned, but
1906 with unset values at the outer level. The correct (outer level) value is
1907 now given.
1909 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1910 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1911 make it do so. The string so far is captured, making this feature
1912 compatible with Perl.
1914 23. The tests have been re-organized, adding tests 11 and 12, to make it
1915 possible to check the Perl 5.10 features against Perl 5.10.
1917 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1918 pattern matches a fixed length string. PCRE did not allow this; now it
1919 does. Neither allows recursion.
1921 25. I finally figured out how to implement a request to provide the minimum
1922 length of subject string that was needed in order to match a given pattern.
1923 (It was back references and recursion that I had previously got hung up
1924 on.) This code has now been added to pcre_study(); it finds a lower bound
1925 to the length of subject needed. It is not necessarily the greatest lower
1926 bound, but using it to avoid searching strings that are too short does give
1927 some useful speed-ups. The value is available to calling programs via
1928 pcre_fullinfo().
1930 26. While implementing 25, I discovered to my embarrassment that pcretest had
1931 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1932 study optimizations had never been tested with that matching function.
1933 Oops. What is worse, even when it was passed study data, there was a bug in
1934 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1935 were also very few tests of studied patterns with pcre_dfa_exec().
1937 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1938 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1939 on the other side of the coin, they are no longer allowed to have different
1940 names, because these cannot be distinguished in PCRE, and this has caused
1941 confusion. (This is a difference from Perl.)
1943 28. When duplicate subpattern names are present (necessarily with different
1944 numbers, as required by 27 above), and a test is made by name in a
1945 conditional pattern, either for a subpattern having been matched, or for
1946 recursion in such a pattern, all the associated numbered subpatterns are
1947 tested, and the overall condition is true if the condition is true for any
1948 one of them. This is the way Perl works, and is also more like the way
1949 testing by number works.
1952 Version 7.9 11-Apr-09
1953 ---------------------
1955 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1956 (pcretest), all targets were linked against these libraries. This included
1957 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1958 libraries. This caused unwanted dependencies to be created. This problem
1959 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1960 pcretest is linked with readline.
1962 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1963 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1964 moved outside it again, because FALSE and TRUE are already defined in AIX,
1965 but BOOL is not.
1967 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1968 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1970 4. The pcregrep documentation said spaces were inserted as well as colons (or
1971 hyphens) following file names and line numbers when outputting matching
1972 lines. This is not true; no spaces are inserted. I have also clarified the
1973 wording for the --colour (or --color) option.
1975 5. In pcregrep, when --colour was used with -o, the list of matching strings
1976 was not coloured; this is different to GNU grep, so I have changed it to be
1977 the same.
1979 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1980 each matching line was coloured. Now it goes on to look for further matches
1981 of any of the test patterns, which is the same behaviour as GNU grep.
1983 7. A pattern that could match an empty string could cause pcregrep to loop; it
1984 doesn't make sense to accept an empty string match in pcregrep, so I have
1985 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1986 seems to be how GNU grep behaves. [But see later change 40 for release
1987 8.33.]
1989 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1990 start or after a newline", because the conditional assertion was not being
1991 correctly handled. The rule now is that both the assertion and what follows
1992 in the first alternative must satisfy the test.
1994 9. If auto-callout was enabled in a pattern with a conditional group whose
1995 condition was an assertion, PCRE could crash during matching, both with
1996 pcre_exec() and pcre_dfa_exec().
1998 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1999 used for matching.
2001 11. Unicode property support in character classes was not working for
2002 characters (bytes) greater than 127 when not in UTF-8 mode.
2004 12. Added the -M command line option to pcretest.
2006 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2008 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2010 16. Added comments and documentation about mis-use of no_arg in the C++
2011 wrapper.
2013 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2014 from Martin Jerabek that uses macro names for all relevant character and
2015 string constants.
2017 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2018 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2019 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2020 these, but not everybody uses configure.
2022 19. A conditional group that had only one branch was not being correctly
2023 recognized as an item that could match an empty string. This meant that an
2024 enclosing group might also not be so recognized, causing infinite looping
2025 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2026 with the subject "ab", where knowledge that the repeated group can match
2027 nothing is needed in order to break the loop.
2029 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2030 exec(), but without supplying a callout function, matching went wrong.
2032 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2033 leak if the size of the offset vector was greater than 30. When the vector
2034 is smaller, the saved offsets during recursion go onto a local stack
2035 vector, but for larger vectors malloc() is used. It was failing to free
2036 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2037 error, in fact).
2039 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2040 heapframe that is used only when UTF-8 support is enabled. This caused no
2041 problem, but was untidy.
2043 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2044 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2045 included within another project.
2047 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2048 slightly modified by me:
2050 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2051 not building pcregrep.
2053 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2054 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2056 25. Forward references, both numeric and by name, in patterns that made use of
2057 duplicate group numbers, could behave incorrectly or give incorrect errors,
2058 because when scanning forward to find the reference group, PCRE was not
2059 taking into account the duplicate group numbers. A pattern such as
2060 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2062 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2063 the feature of a custom pointer more persuasive (as requested by a user).
2065 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2066 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2067 pre-defined.
2069 28. Added support for (*UTF8) at the start of a pattern.
2071 29. Arrange for flags added by the "release type" setting in CMake to be shown
2072 in the configuration summary.
2075 Version 7.8 05-Sep-08
2076 ---------------------
2078 1. Replaced UCP searching code with optimized version as implemented for Ad
2079 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2080 stage table and inline lookup instead of a function, giving speed ups of 2
2081 to 5 times on some simple patterns that I tested. Permission was given to
2082 distribute the MultiStage2.py script that generates the tables (it's not in
2083 the tarball, but is in the Subversion repository).
2085 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2086 scripts.
2088 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2089 a group with a zero qualifier. The result of the study could be incorrect,
2090 or the function might crash, depending on the pattern.
2092 4. Caseless matching was not working for non-ASCII characters in back
2093 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2094 It now works when Unicode Property Support is available.
2096 5. In pcretest, an escape such as \x{de} in the data was always generating
2097 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2098 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2099 truncation.
2101 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2103 7. Added two (int) casts to pcregrep when printing the difference of two
2104 pointers, in case they are 64-bit values.
2106 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2107 test 2 if it fails.
2109 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2110 and a #define of that name to empty if it is not externally set. This is to
2111 allow users of MSVC to set it if necessary.
2113 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2114 the convenience functions in the pcre_get.c source file.
2116 11. An option change at the start of a pattern that had top-level alternatives
2117 could cause overwriting and/or a crash. This command provoked a crash in
2118 some environments:
2120 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2122 This potential security problem was recorded as CVE-2008-2371.
2124 12. For a pattern where the match had to start at the beginning or immediately
2125 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2126 pcre_dfa_exec() could read past the end of the passed subject if there was
2127 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2128 pcretest so that it places the subject at the end of its malloc-ed buffer.
2130 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2131 exec() might read past the end of the data buffer in UTF-8 mode.
2133 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2134 the data contained the byte 0x85 as part of a UTF-8 character within its
2135 first line. This applied both to normal and DFA matching.
2137 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2138 /^[^d]*?$/8 failed to match "abc".
2140 16. Added a missing copyright notice to pcrecpp_internal.h.
2142 17. Make it more clear in the documentation that values returned from
2143 pcre_exec() in ovector are byte offsets, not character counts.
2145 18. Tidied a few places to stop certain compilers from issuing warnings.
2147 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2148 supplied by Stefan Weber. I made a further small update for 7.8 because
2149 there is a change of source arrangements: the pcre_searchfuncs.c module is
2150 replaced by pcre_ucd.c.
2153 Version 7.7 07-May-08
2154 ---------------------
2156 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2157 a string to a long long, pretend we don't even have a long long." This is
2158 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2160 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2161 pre-7.6 versions, which defined a global no_arg variable instead of putting
2162 it in the RE class. (See also #8 below.)
2164 3. Remove a line of dead code, identified by coverity and reported by Nuno
2165 Lopes.
2167 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2169 (1) The include/exclude patterns were being applied to the whole pathnames
2170 of files, instead of just to the final components.
2172 (2) If there was more than one level of directory, the subdirectories were
2173 skipped unless they satisfied the include/exclude conditions. This is
2174 inconsistent with GNU grep (and could even be seen as contrary to the
2175 pcregrep specification - which I improved to make it absolutely clear).
2176 The action now is always to scan all levels of directory, and just
2177 apply the include/exclude patterns to regular files.
2179 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2180 --exclude_dir in the tests to avoid scanning .svn directories.
2182 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2183 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2184 doesn't support NULs in patterns.
2186 7. Added some missing "const"s to declarations of static tables in
2187 pcre_compile.c and pcre_dfa_exec.c.
2189 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2190 caused by fix #2 above. (Subsequently also a second patch to fix the
2191 first patch. And a third patch - this was a messy problem.)
2193 9. Applied Craig's patch to remove the use of push_back().
2195 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2196 matching function regexec().
2198 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2199 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2200 references. PCRE supports relative numbers with this syntax (I don't think
2201 Oniguruma does).
2203 12. Previously, a group with a zero repeat such as (...){0} was completely
2204 omitted from the compiled regex. However, this means that if the group
2205 was called as a subroutine from elsewhere in the pattern, things went wrong
2206 (an internal error was given). Such groups are now left in the compiled
2207 pattern, with a new opcode that causes them to be skipped at execution
2208 time.
2210 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2211 to the way PCRE behaves:
2213 (a) A lone ] character is dis-allowed (Perl treats it as data).
2215 (b) A back reference to an unmatched subpattern matches an empty string
2216 (Perl fails the current match path).
2218 (c) A data ] in a character class must be notated as \] because if the
2219 first data character in a class is ], it defines an empty class. (In
2220 Perl it is not possible to have an empty class.) The empty class []
2221 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2222 The negative empty class [^] matches any one character, independently
2223 of the DOTALL setting.
2225 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2226 non-existent subpattern following a character class starting with ']' and
2227 containing () gave an internal compiling error instead of "reference to
2228 non-existent subpattern". Fortunately, when the pattern did exist, the
2229 compiled code was correct. (When scanning forwards to check for the
2230 existence of the subpattern, it was treating the data ']' as terminating
2231 the class, so got the count wrong. When actually compiling, the reference
2232 was subsequently set up correctly.)
2234 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2235 it was being rejected as not supported by pcre_dfa_exec(), even though
2236 other assertions are supported. I have made pcre_dfa_exec() support
2237 (*FAIL).
2239 16. The implementation of 13c above involved the invention of a new opcode,
2240 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2241 cannot be changed at match time, I realized I could make a small
2242 improvement to matching performance by compiling OP_ALLANY instead of
2243 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2244 on the OP_ANY path.
2246 17. Compiling pcretest on Windows with readline support failed without the
2247 following two fixes: (1) Make the unistd.h include conditional on
2248 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2250 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2251 ncurses library to be included for pcretest when ReadLine support is
2252 requested, but also to allow for it to be overridden. This patch came from
2253 Daniel Bergström.
2255 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2256 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2257 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2258 spotting this.
2261 Version 7.6 28-Jan-08
2262 ---------------------
2264 1. A character class containing a very large number of characters with
2265 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2266 overflow.
2268 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2269 HAVE_LONG_LONG is not defined.
2271 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2272 bring them up to date and include new features. This patch includes:
2274 - Fixed PH's badly added libz and libbz2 support.
2275 - Fixed a problem with static linking.
2276 - Added pcredemo. [But later removed - see 7 below.]
2277 - Fixed dftables problem and added an option.
2278 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2280 - Added readline support for pcretest.
2281 - Added an listing of the option settings after cmake has run.
2283 4. A user submitted a patch to Makefile that makes it easy to create
2284 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2285 Makefile.am that cause it to include this special target, without
2286 affecting anything else. Note that the same mingw target plus all
2287 the other distribution libraries and programs are now supported
2288 when configuring with CMake (see 6 below) instead of with
2289 Configure/Make.
2291 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2292 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2293 exported in the Windows port". It has not yet been confirmed that the patch
2294 solves the problem, but it does no harm.
2296 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2297 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2298 with CMake, and also correct the comment about stack recursion.
2300 7. Remove the automatic building of pcredemo from the ./configure system and
2301 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2302 of a program that users should build themselves after PCRE is installed, so
2303 building it automatically is not really right. What is more, it gave
2304 trouble in some build environments.
2306 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2309 Version 7.5 10-Jan-08
2310 ---------------------
2312 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2313 values in parens when parsing an RE using the C++ wrapper."
2315 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2316 Characters greater than 255 were excluded from the class instead of being
2317 included.
2319 3. The same bug as (2) above applied to negated POSIX classes such as
2320 [:^space:].
2322 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2323 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2324 I have changed it.
2326 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2327 first named subpattern) and a construct such as (?&a) would reference the
2328 first named subpattern whose name started with "a" (in other words, the
2329 length check was missing). Both these problems are fixed. "Subpattern name
2330 expected" is now given for (?&) (a zero-length name), and this patch also
2331 makes it give the same error for \k'' (previously it complained that that
2332 was a reference to a non-existent subpattern).
2334 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2335 this is right because (?- can be followed by option settings as well as by
2336 digits. I have, however, made the messages clearer.
2338 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2339 than the number used in the conditional) now cause a compile-time error.
2340 This is actually not compatible with Perl, which accepts such patterns, but
2341 treats the conditional as always being FALSE (as PCRE used to), but it
2342 seems to me that giving a diagnostic is better.
2344 8. Change "alphameric" to the more common word "alphanumeric" in comments
2345 and messages.
2347 9. Fix two occurrences of "backslash" in comments that should have been
2348 "backspace".
2350 10. Remove two redundant lines of code that can never be obeyed (their function
2351 was moved elsewhere).
2353 11. The program that makes PCRE's Unicode character property table had a bug
2354 which caused it to generate incorrect table entries for sequences of
2355 characters that have the same character type, but are in different scripts.
2356 It amalgamated them into a single range, with the script of the first of
2357 them. In other words, some characters were in the wrong script. There were
2358 thirteen such cases, affecting characters in the following ranges:
2360 U+002b0 - U+002c1
2361 U+0060c - U+0060d
2362 U+0061e - U+00612
2363 U+0064b - U+0065e
2364 U+0074d - U+0076d
2365 U+01800 - U+01805
2366 U+01d00 - U+01d77
2367 U+01d9b - U+01dbf
2368 U+0200b - U+0200f
2369 U+030fc - U+030fe
2370 U+03260 - U+0327f
2371 U+0fb46 - U+0fbb1
2372 U+10450 - U+1049d
2374 12. The -o option (show only the matching part of a line) for pcregrep was not
2375 compatible with GNU grep in that, if there was more than one match in a
2376 line, it showed only the first of them. It now behaves in the same way as
2377 GNU grep.
2379 13. If the -o and -v options were combined for pcregrep, it printed a blank
2380 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2381 does the same. The return code can be used to tell if there were any
2382 non-matching lines.
2384 14. Added --file-offsets and --line-offsets to pcregrep.
2386 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2387 infinitely looping recursion. The bug was that positive lookaheads were not
2388 being skipped when checking for a possible empty match (negative lookaheads
2389 and both kinds of lookbehind were skipped).
2391 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2392 inclusion of <windows.h> to before rather than after the definition of
2393 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2395 17. Specifying a possessive quantifier with a specific limit for a Unicode
2396 character property caused pcre_compile() to compile bad code, which led at
2397 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2398 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2399 caused the error; without that there was no problem.
2401 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2403 19. Added --enable-pcretest-libreadline.
2405 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2406 RE::GlobalReplace(). As a result, the number of replacements returned was
2407 double what it should be. I removed one of the increments, but Craig sent a
2408 later patch that removed the other one (the right fix) and added unit tests
2409 that check the return values (which was not done before).
2411 21. Several CMake things:
2413 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2414 the names libpcre and libpcreposix, not just pcre and pcreposix.
2416 (2) The above change means that pcretest and pcregrep are now correctly
2417 linked with the newly-built libraries, not previously installed ones.
2421 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2422 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2423 UTF-8 newline character). The key issue is that the pattern starts .*;
2424 this means that the match must be either at the beginning, or after a
2425 newline. The bug was in the code for advancing after a failed match and
2426 checking that the new position followed a newline. It was not taking
2427 account of UTF-8 characters correctly.
2429 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2430 character classes. PCRE was not treating the sequence [:...:] as a
2431 character class unless the ... were all letters. Perl, however, seems to
2432 allow any characters between [: and :], though of course it rejects as
2433 unknown any "names" that contain non-letters, because all the known class
2434 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2435 for example, whereas PCRE did not - it did not recognize a POSIX character
2436 class. This seemed a bit dangerous, so the code has been changed to be
2437 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2438 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2439 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2440 Perl does, and where it didn't before.
2442 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2443 Windows environments %n is disabled by default.
2446 Version 7.4 21-Sep-07
2447 ---------------------
2449 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2450 means that a class such as [\s] counted as "explicit reference to CR or
2451 LF". That isn't really right - the whole point of the change was to try to
2452 help when there was an actual mention of one of the two characters. So now
2453 the change happens only if \r or \n (or a literal CR or LF) character is
2454 encountered.
2456 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2457 of both had grown to the point where there were only 3 bits left.
2458 Fortunately, there was spare space in the data structure, and so I have
2459 moved the internal flags into a new 16-bit field to free up more option
2460 bits.
2462 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2463 but did not set the internal JCHANGED flag - either of these is enough to
2464 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2465 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2466 start sets both bits.
2468 4. Added options (at build time, compile time, exec time) to change \R from
2469 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2471 5. doc/pcresyntax.html was missing from the distribution.
2473 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2474 compatibility, even though it is no longer used.
2476 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2477 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2478 windows.h file is present (where different names are used). [This was
2479 reversed later after testing - see 16 below.]
2481 8. Changed all #include <config.h> to #include "config.h". There were also
2482 some further <pcre.h> cases that I changed to "pcre.h".
2484 9. When pcregrep was used with the --colour option, it missed the line ending
2485 sequence off the lines that it output.
2487 10. It was pointed out to me that arrays of string pointers cause lots of
2488 relocations when a shared library is dynamically loaded. A technique of
2489 using a single long string with a table of offsets can drastically reduce
2490 these. I have refactored PCRE in four places to do this. The result is
2491 dramatic:
2493 Originally: 290
2494 After changing UCP table: 187
2495 After changing error message table: 43
2496 After changing table of "verbs" 36
2497 After changing table of Posix names 22
2499 Thanks to the folks working on Gregex for glib for this insight.
2501 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2502 unicode-properties was also set.
2504 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2506 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2507 checked only for CRLF.
2509 14. Added casts to pcretest.c to avoid compiler warnings.
2511 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2513 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2514 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2515 entirely. This removes changes made in 7 above.
2517 17. The CMake files have been updated, and there is now more information about
2518 building with CMake in the NON-UNIX-USE document.
2521 Version 7.3 28-Aug-07
2522 ---------------------
2524 1. In the rejigging of the build system that eventually resulted in 7.1, the
2525 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2526 brackets there is not right, since it causes compilers to look for an
2527 installed pcre.h, not the version that is in the source that is being
2528 compiled (which of course may be different). I have changed it back to:
2530 #include "pcre.h"
2532 I have a vague recollection that the change was concerned with compiling in
2533 different directories, but in the new build system, that is taken care of
2534 by the VPATH setting the Makefile.
2536 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2537 when the subject happened to end in the byte 0x85 (e.g. if the last
2538 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2539 characters but of course it shouldn't be taken as a newline when it is part
2540 of another character. The bug was that, for an unlimited repeat of . in
2541 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2542 characters when looking for a newline.
2544 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2546 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2547 in debug output.
2549 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2550 long printing in the pcrecpp unittest when running under MinGW.
2552 6. ESC_K was left out of the EBCDIC table.
2554 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2555 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2556 limit also applies to "virtual nesting" when a pattern is recursive, and in
2557 this case 1000 isn't so big. I have been able to remove this limit at the
2558 expense of backing off one optimization in certain circumstances. Normally,
2559 when pcre_exec() would call its internal match() function recursively and
2560 immediately return the result unconditionally, it uses a "tail recursion"
2561 feature to save stack. However, when a subpattern that can match an empty
2562 string has an unlimited repetition quantifier, it no longer makes this
2563 optimization. That gives it a stack frame in which to save the data for
2564 checking that an empty string has been matched. Previously this was taken
2565 from the 1000-entry workspace that had been reserved. So now there is no
2566 explicit limit, but more stack is used.
2568 8. Applied Daniel's patches to solve problems with the import/export magic
2569 syntax that is required for Windows, and which was going wrong for the
2570 pcreposix and pcrecpp parts of the library. These were overlooked when this
2571 problem was solved for the main library.
2573 9. There were some crude static tests to avoid integer overflow when computing
2574 the size of patterns that contain repeated groups with explicit upper
2575 limits. As the maximum quantifier is 65535, the maximum group length was
2576 set at 30,000 so that the product of these two numbers did not overflow a
2577 32-bit integer. However, it turns out that people want to use groups that
2578 are longer than 30,000 bytes (though not repeat them that many times).
2579 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2580 made it possible to implement the integer overflow checks in a much more
2581 dynamic way, which I have now done. The artificial limitation on group
2582 length has been removed - we now have only the limit on the total length of
2583 the compiled pattern, which depends on the LINK_SIZE setting.
2585 10. Fixed a bug in the documentation for get/copy named substring when
2586 duplicate names are permitted. If none of the named substrings are set, the
2587 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2588 empty string.
2590 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2591 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2592 because the ] is interpreted as the first data character and the
2593 terminating ] is not found. PCRE has been made compatible with Perl in this
2594 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2595 cause memory overwriting.
2597 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2598 string has been matched (to stop an infinite loop). It was not recognizing
2599 a conditional subpattern that could match an empty string if that
2600 subpattern was within another subpattern. For example, it looped when
2601 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2602 condition was not nested. This bug has been fixed.
2604 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2605 past the start of the subject in the presence of bytes with the top bit
2606 set, for example "\x8aBCD".
2608 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2609 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2611 14. Optimized (?!) to (*FAIL).
2613 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2614 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2615 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2616 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2617 does: it's just the validity check that is more restrictive.
2619 16. Inserted checks for integer overflows during escape sequence (backslash)
2620 processing, and also fixed erroneous offset values for syntax errors during
2621 backslash processing.
2623 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2624 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2626 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2627 caused an overrun.
2629 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2630 something other than just ASCII characters) inside a group that had an
2631 unlimited repeat caused a loop at compile time (while checking to see
2632 whether the group could match an empty string).
2634 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2635 [\P{Any}] did so. (Error in the code for printing property names.)
2637 21. An orphan \E inside a character class could cause a crash.
2639 22. A repeated capturing bracket such as (A)? could cause a wild memory
2640 reference during compilation.
2642 23. There are several functions in pcre_compile() that scan along a compiled
2643 expression for various reasons (e.g. to see if it's fixed length for look
2644 behind). There were bugs in these functions when a repeated \p or \P was
2645 present in the pattern. These operators have additional parameters compared
2646 with \d, etc, and these were not being taken into account when moving along
2647 the compiled data. Specifically:
2649 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2650 length.
2652 (b) An item such as \pL+ within a repeated group could cause crashes or
2653 loops.
2655 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2656 "reference to non-existent subpattern" error.
2658 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2660 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2661 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2663 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2665 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2666 character were causing crashes (broken optimization).
2668 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2669 \p or \P) caused a compile-time loop.
2671 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2672 break. For example, the unstudied pattern [\r\n]A does not match the string
2673 "\r\nA" because change 7.0/46 below moves the current point on by two
2674 characters after failing to match at the start. However, the pattern \nA
2675 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2676 the same is true. There doesn't seem any very clean way out of this, but
2677 what I have chosen to do makes the common cases work: PCRE now takes note
2678 of whether there can be an explicit match for \r or \n anywhere in the
2679 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2680 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2681 pattern has explicit CR or LF references.
2683 29. Added (*CR) etc for changing newline setting at start of pattern.
2686 Version 7.2 19-Jun-07
2687 ---------------------
2689 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2690 which is apparently normally available under Windows.
2692 2. Re-jig the pcregrep tests with different newline settings in an attempt
2693 to make them independent of the local environment's newline setting.
2695 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2697 4. Some of the "internals" tests were previously cut out when the link size
2698 was not 2, because the output contained actual offsets. The recent new
2699 "Z" feature of pcretest means that these can be cut out, making the tests
2700 usable with all link sizes.
2702 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2703 stack recursion. This gives a massive performance boost under BSD, but just
2704 a small improvement under Linux. However, it saves one field in the frame
2705 in all cases.
2707 6. Added more features from the forthcoming Perl 5.10:
2709 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2710 recursion call. It refers to the nth most recently opened parentheses.
2712 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2713 to be opened parentheses.
2715 (c) Conditions that refer to capturing parentheses can be specified
2716 relatively, for example, (?(-2)... or (?(+3)...
2718 (d) \K resets the start of the current match so that everything before
2719 is not part of it.
2721 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2723 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2724 reference syntax.
2726 (g) (?| introduces a group in which the numbering of parentheses in each
2727 alternative starts with the same number.
2729 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2731 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2734 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2735 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2736 for detecting groups that can match an empty string.
2738 9. A pattern with a very large number of alternatives (more than several
2739 hundred) was running out of internal workspace during the pre-compile
2740 phase, where pcre_compile() figures out how much memory will be needed. A
2741 bit of new cunning has reduced the workspace needed for groups with
2742 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2743 workspace instead of running out of the 4096 that are available.
2745 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2747 11. Applied patch from Google to remove an optimization that didn't quite work.
2748 The report of the bug said:
2750 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2751 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2752 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2754 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2755 it matched the wrong number of bytes.
2758 Version 7.1 24-Apr-07
2759 ---------------------
2761 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2762 that is more "standard", making use of automake and other Autotools. There
2763 is some re-arrangement of the files and adjustment of comments consequent
2764 on this.
2766 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2767 for recursive directory scanning broke on some systems because the files
2768 are not scanned in any specific order and on different systems the order
2769 was different. A call to "sort" has been inserted into RunGrepTest for the
2770 approprate test as a short-term fix. In the longer term there may be an
2771 alternative.
2773 3. I had an email from Eric Raymond about problems translating some of PCRE's
2774 man pages to HTML (despite the fact that I distribute HTML pages, some
2775 people do their own conversions for various reasons). The problems
2776 concerned the use of low-level troff macros .br and .in. I have therefore
2777 removed all such uses from the man pages (some were redundant, some could
2778 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2779 HTML has been updated to handle .nf/.fi and to complain if it encounters
2780 .br or .in.
2782 4. Updated comments in configure.ac that get placed in config.h.in and also
2783 arranged for config.h to be included in the distribution, with the name
2784 config.h.generic, for the benefit of those who have to compile without
2785 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2787 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2788 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2789 makevp.bat for the current PCRE, using the additional files
2790 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2792 6. A Windows user reported a minor discrepancy with test 2, which turned out
2793 to be caused by a trailing space on an input line that had got lost in his
2794 copy. The trailing space was an accident, so I've just removed it.
2796 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2797 that is needed.
2799 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2800 as "const" (a) because they are and (b) because it helps the PHP
2801 maintainers who have recently made a script to detect big data structures
2802 in the php code that should be moved to the .rodata section. I remembered
2803 to update Builducptable as well, so it won't revert if ucptable.h is ever
2804 re-created.
2806 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2807 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2808 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2809 support is not required. This saves 1.5-2K of code, which is important in
2810 some applications.
2812 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2813 so as not to refer to the tables, even though these functions will never be
2814 called when UTF-8 support is disabled. Otherwise there are problems with a
2815 shared library.
2817 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2819 (a) It was defining its arguments as char * instead of void *.
2821 (b) It was assuming that all moves were upwards in memory; this was true
2822 a long time ago when I wrote it, but is no longer the case.
2824 The emulated memove() is provided for those environments that have neither
2825 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2826 is clearly not the case, as these two bugs were recently reported.
2828 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2829 and Detrail to create the HTML documentation, the .txt form of the man
2830 pages, and it removes trailing spaces from listed files. It also creates
2831 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2832 case, it wraps all the #defines with #ifndefs. This script should be run
2833 before "make dist".
2835 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2836 with Unicode property support.
2838 (a) For a maximizing quantifier, if the two different cases of the
2839 character were of different lengths in their UTF-8 codings (there are
2840 some cases like this - I found 11), and the matching function had to
2841 back up over a mixture of the two cases, it incorrectly assumed they
2842 were both the same length.
2844 (b) When PCRE was configured to use the heap rather than the stack for
2845 recursion during matching, it was not correctly preserving the data for
2846 the other case of a UTF-8 character when checking ahead for a match
2847 while processing a minimizing repeat. If the check also involved
2848 matching a wide character, but failed, corruption could cause an
2849 erroneous result when trying to check for a repeat of the original
2850 character.
2852 13. Some tidying changes to the testing mechanism:
2854 (a) The RunTest script now detects the internal link size and whether there
2855 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2856 values substituted by "configure". (The RunGrepTest script already did
2857 this for UTF-8.) The configure.ac script no longer substitutes the
2858 relevant variables.
2860 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2861 with length and offset values. This means that the output is different
2862 for different internal link sizes. Test 2 is skipped for link sizes
2863 other than 2 because of this, bypassing the problem. Unfortunately,
2864 there was also a test in test 3 (the locale tests) that used /B and
2865 failed for link sizes other than 2. Rather than cut the whole test out,
2866 I have added a new /Z option to pcretest that replaces the length and
2867 offset values with spaces. This is now used to make test 3 independent
2868 of link size. (Test 2 will be tidied up later.)
2870 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2871 segmentation fault instead of returning the appropriate error message.
2873 15. In multiline mode when the newline sequence was set to "any", the pattern
2874 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2875 This doesn't seem right; it now treats the CRLF combination as the line
2876 ending, and so does not match in that case. It's only a pattern such as ^$
2877 that would hit this one: something like ^ABC$ would have failed after \r
2878 and then tried again after \r\n.
2880 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2881 in an attempt to make files that differ only in their line terminators
2882 compare equal. This works on Linux.
2884 17. Under certain error circumstances pcregrep might try to free random memory
2885 as it exited. This is now fixed, thanks to valgrind.
2887 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2888 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2889 was because its rules for how to advance for /g after matching an empty
2890 string at the end of a line did not allow for this case. They now check for
2891 it specially.
2893 20. pcretest is supposed to handle patterns and data of any length, by
2894 extending its buffers when necessary. It was getting this wrong when the
2895 buffer for a data line had to be extended.
2897 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2898 CRLF as a newline sequence.
2900 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2901 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2902 I have nevertheless tidied it up.
2904 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2906 24. Added a man page for pcre-config.
2909 Version 7.0 19-Dec-06
2910 ---------------------
2912 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2913 moving to gcc 4.1.1.
2915 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2916 sys/time.h, which is documented as needed for this function. It doesn't
2917 seem to matter on Linux, but it showed up on some releases of OS X.
2919 3. It seems that there are systems where bytes whose values are greater than
2920 127 match isprint() in the "C" locale. The "C" locale should be the
2921 default when a C program starts up. In most systems, only ASCII printing
2922 characters match isprint(). This difference caused the output from pcretest
2923 to vary, making some of the tests fail. I have changed pcretest so that:
2925 (a) When it is outputting text in the compiled version of a pattern, bytes
2926 other than 32-126 are always shown as hex escapes.
2928 (b) When it is outputting text that is a matched part of a subject string,
2929 it does the same, unless a different locale has been set for the match
2930 (using the /L modifier). In this case, it uses isprint() to decide.
2932 4. Fixed a major bug that caused incorrect computation of the amount of memory
2933 required for a compiled pattern when options that changed within the
2934 pattern affected the logic of the preliminary scan that determines the
2935 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2936 that the computed length was too small. The symptoms of this bug were
2937 either the PCRE error "internal error: code overflow" from pcre_compile(),
2938 or a glibc crash with a message such as "pcretest: free(): invalid next
2939 size (fast)". Examples of patterns that provoked this bug (shown in
2940 pcretest format) are:
2942 /(?-x: )/x
2943 /(?x)(?-x: \s*#\s*)/
2944 /((?i)[\x{c0}])/8
2945 /(?i:[\x{c0}])/8
2947 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2948 is now done differently.
2950 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2951 wrapper classes; (b) implement a new function in the C++ scanner that is
2952 more efficient than the old way of doing things because it avoids levels of
2953 recursion in the regex matching; (c) add a paragraph to the documentation
2954 for the FullMatch() function.
2956 6. The escape sequence \n was being treated as whatever was defined as
2957 "newline". Not only was this contrary to the documentation, which states
2958 that \n is character 10 (hex 0A), but it also went horribly wrong when
2959 "newline" was defined as CRLF. This has been fixed.
2961 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2962 was being set to -1 for the "end of line" case (supposedly a value that no
2963 character can have). Though this value is never used (the check for end of
2964 line is "zero bytes in current character"), it caused compiler complaints.
2965 I've changed it to 0xffffffff.
2967 8. In pcre_version.c, the version string was being built by a sequence of
2968 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2969 string (as it is for production releases) called a macro with an empty
2970 argument. The C standard says the result of this is undefined. The gcc
2971 compiler treats it as an empty string (which was what was wanted) but it is
2972 reported that Visual C gives an error. The source has been hacked around to
2973 avoid this problem.
2975 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2976 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2977 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2978 of them did).
2980 10. Originally, pcretest opened its input and output without "b"; then I was
2981 told that "b" was needed in some environments, so it was added for release
2982 5.0 to both the input and output. (It makes no difference on Unix-like
2983 systems.) Later I was told that it is wrong for the input on Windows. I've
2984 now abstracted the modes into two macros, to make it easier to fiddle with
2985 them, and removed "b" from the input mode under Windows.
2987 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2989 12. Added -help and --help to pcretest as an official way of being reminded
2990 of the options.
2992 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2993 and pcrecpp.cc because they annoy compilers at high warning levels.
2995 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2997 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2998 scripts are not C programs :-) and which was not noticed because it works
2999 on Linux.
3001 16. pcretest is supposed to handle any length of pattern and data line (as one
3002 line or as a continued sequence of lines) by extending its input buffer if
3003 necessary. This feature was broken for very long pattern lines, leading to
3004 a string of junk being passed to pcre_compile() if the pattern was longer
3005 than about 50K.
3007 17. I have done a major re-factoring of the way pcre_compile() computes the
3008 amount of memory needed for a compiled pattern. Previously, there was code
3009 that made a preliminary scan of the pattern in order to do this. That was
3010 OK when PCRE was new, but as the facilities have expanded, it has become
3011 harder and harder to keep it in step with the real compile phase, and there
3012 have been a number of bugs (see for example, 4 above). I have now found a
3013 cunning way of running the real compile function in a "fake" mode that
3014 enables it to compute how much memory it would need, while actually only
3015 ever using a few hundred bytes of working memory and without too many
3016 tests of the mode. This should make future maintenance and development
3017 easier. A side effect of this work is that the limit of 200 on the nesting
3018 depth of parentheses has been removed (though this was never a serious
3019 limitation, I suspect). However, there is a downside: pcre_compile() now
3020 runs more slowly than before (30% or more, depending on the pattern). I
3021 hope this isn't a big issue. There is no effect on runtime performance.
3023 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3024 newline (only possible for the last line of a file) and it was a
3025 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3027 19. Added additional timing features to pcretest. (1) The -tm option now times
3028 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3029 separate command line item, by a number that specifies the number of
3030 repeats to use when timing. The default is 50000; this gives better
3031 precision, but takes uncomfortably long for very large patterns.
3033 20. Extended pcre_study() to be more clever in cases where a branch of a
3034 subpattern has no definite first character. For example, (a*|b*)[cd] would
3035 previously give no result from pcre_study(). Now it recognizes that the
3036 first character must be a, b, c, or d.
3038 21. There was an incorrect error "recursive call could loop indefinitely" if
3039 a subpattern (or the entire pattern) that was being tested for matching an
3040 empty string contained only one non-empty item after a nested subpattern.
3041 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3042 incorrectly, because the \d was being skipped in the check.
3044 22. The pcretest program now has a new pattern option /B and a command line
3045 option -b, which is equivalent to adding /B to every pattern. This causes
3046 it to show the compiled bytecode, without the additional information that
3047 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3048 is the same as /B/I).
3050 23. A new optimization is now able automatically to treat some sequences such
3051 as a*b as a*+b. More specifically, if something simple (such as a character
3052 or a simple class like \d) has an unlimited quantifier, and is followed by
3053 something that cannot possibly match the quantified thing, the quantifier
3054 is automatically "possessified".
3056 24. A recursive reference to a subpattern whose number was greater than 39
3057 went wrong under certain circumstances in UTF-8 mode. This bug could also
3058 have affected the operation of pcre_study().
3060 25. Realized that a little bit of performance could be had by replacing
3061 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3063 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3065 27. Possessive quantifiers such as a++ were previously implemented by turning
3066 them into atomic groups such as ($>a+). Now they have their own opcodes,
3067 which improves performance. This includes the automatically created ones
3068 from 23 above.
3070 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3071 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3072 the first matched character to be a colon. This applied both to named and
3073 numbered groups.
3075 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3077 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3078 building PCRE on Windows easier, so I have included it in the distribution.
3080 31. There is now a check in pcretest against a ridiculously large number being
3081 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3082 loop, the loop is abandoned.
3084 32. Forward references to subpatterns in conditions such as (?(2)...) where
3085 subpattern 2 is defined later cause pcre_compile() to search forwards in
3086 the pattern for the relevant set of parentheses. This search went wrong
3087 when there were unescaped parentheses in a character class, parentheses
3088 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3090 33. "Subroutine" calls and backreferences were previously restricted to
3091 referencing subpatterns earlier in the regex. This restriction has now
3092 been removed.
3094 34. Added a number of extra features that are going to be in Perl 5.10. On the
3095 whole, these are just syntactic alternatives for features that PCRE had
3096 previously implemented using the Python syntax or my own invention. The
3097 other formats are all retained for compatibility.
3099 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3100 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3101 also .NET compatible.
3103 (b) A recursion or subroutine call to a named group can now be defined as
3104 (?&name) as well as (?P>name).
3106 (c) A backreference to a named group can now be defined as \k<name> or
3107 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3108 5.10, are also .NET compatible.
3110 (d) A conditional reference to a named group can now use the syntax
3111 (?(<name>) or (?('name') as well as (?(name).
3113 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3114 groups (named and numbered) that are never evaluated inline, but can be
3115 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3116 is always false. There may be only one alternative in such a group.
3118 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3119 as the simple (?(R). The condition is true only if the most recent
3120 recursion is that of the given number or name. It does not search out
3121 through the entire recursion stack.
3123 (g) The escape \gN or \g{N} has been added, where N is a positive or
3124 negative number, specifying an absolute or relative reference.
3126 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3127 some "unreachable code" warnings.
3129 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3130 things, this adds five new scripts.
3132 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3133 There were also incompatibilities regarding the handling of \Q..\E inside
3134 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3135 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3137 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3138 matches an empty string, and forcibly breaks the loop. There were bugs in
3139 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3140 against aaaa the result was just "a" rather than "aaaa", for example. Two
3141 separate and independent bugs (that affected different cases) have been
3142 fixed.
3144 39. Refactored the code to abolish the use of different opcodes for small
3145 capturing bracket numbers. This is a tidy that I avoided doing when I
3146 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3147 The new approach is not only tidier, it makes it possible to reduce the
3148 memory needed to fix the previous bug (38).
3150 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3151 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3152 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3153 mode.
3155 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3156 report.
3158 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3159 copy construction and assignment in the C++ wrapper.
3161 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3162 couple of bugs that could have given wrong results in the "--newline=crlf"
3163 case.
3165 44. Added a number of casts and did some reorganization of signed/unsigned int
3166 variables following suggestions from Dair Grant. Also renamed the variable
3167 "this" as "item" because it is a C++ keyword.
3169 45. Arranged for dftables to add
3171 #include "pcre_internal.h"
3173 to pcre_chartables.c because without it, gcc 4.x may remove the array
3174 definition from the final binary if PCRE is built into a static library and
3175 dead code stripping is activated.
3177 46. For an unanchored pattern, if a match attempt fails at the start of a
3178 newline sequence, and the newline setting is CRLF or ANY, and the next two
3179 characters are CRLF, advance by two characters instead of one.
3182 Version 6.7 04-Jul-06
3183 ---------------------
3185 1. In order to handle tests when input lines are enormously long, pcretest has
3186 been re-factored so that it automatically extends its buffers when
3187 necessary. The code is crude, but this _is_ just a test program. The
3188 default size has been increased from 32K to 50K.
3190 2. The code in pcre_study() was using the value of the re argument before
3191 testing it for NULL. (Of course, in any sensible call of the function, it
3192 won't be NULL.)
3194 3. The memmove() emulation function in pcre_internal.h, which is used on
3195 systems that lack both memmove() and bcopy() - that is, hardly ever -
3196 was missing a "static" storage class specifier.
3198 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3199 containing an extended class (one that cannot be represented by a bitmap
3200 because it contains high-valued characters or Unicode property items, e.g.
3201 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3202 pattern, but PCRE should not loop if you do not (it no longer does).
3203 [Detail: two cases were found: (a) a repeated subpattern containing an
3204 extended class; (b) a recursive reference to a subpattern that followed a
3205 previous extended class. It wasn't skipping over the extended class
3206 correctly when UTF-8 mode was not set.]
3208 5. A negated single-character class was not being recognized as fixed-length
3209 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3210 compile error "lookbehind assertion is not fixed length".
3212 6. The RunPerlTest auxiliary script was showing an unexpected difference
3213 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3214 write a Perl script that can interpret lines of an input file either as
3215 byte characters or as UTF-8, which is what "perltest" was being required to
3216 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3217 can't do is switch easily at run time between having the "use utf8;" pragma
3218 or not. In the end, I fudged it by using the RunPerlTest script to insert
3219 "use utf8;" explicitly for the UTF-8 tests.
3221 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3222 the end of the subject string, contrary to the documentation and to what
3223 Perl does. This was true of both matching functions. Now it matches only at
3224 the start of the subject and immediately after *internal* newlines.
3226 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3227 a pointer to an int instead of a pointer to an unsigned long int. This
3228 caused problems on 64-bit systems.
3230 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3231 instance of the 'standard' template library not being so standard".
3233 10. There was no check on the number of named subpatterns nor the maximum
3234 length of a subpattern name. The product of these values is used to compute
3235 the size of the memory block for a compiled pattern. By supplying a very
3236 long subpattern name and a large number of named subpatterns, the size
3237 computation could be caused to overflow. This is now prevented by limiting
3238 the length of names to 32 characters, and the number of named subpatterns
3239 to 10,000.
3241 11. Subpatterns that are repeated with specific counts have to be replicated in
3242 the compiled pattern. The size of memory for this was computed from the
3243 length of the subpattern and the repeat count. The latter is limited to
3244 65535, but there was no limit on the former, meaning that integer overflow
3245 could in principle occur. The compiled length of a repeated subpattern is
3246 now limited to 30,000 bytes in order to prevent this.
3248 12. Added the optional facility to have named substrings with the same name.
3250 13. Added the ability to use a named substring as a condition, using the
3251 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3252 are numbers (not recommended). Forward references are permitted.
3254 14. Added forward references in named backreferences (if you see what I mean).
3256 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3257 pattern could run off the end of the subject. For example, the pattern
3258 "(?s)(.{1,5})"8 did this with the subject "ab".
3260 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3261 PCRE_CASELESS was set when matching characters that were quantified with ?
3262 or *.
3264 17. A character class other than a single negated character that had a minimum
3265 but no maximum quantifier - for example [ab]{6,} - was not handled
3266 correctly by pce_dfa_exec(). It would match only one character.
3268 18. A valid (though odd) pattern that looked like a POSIX character
3269 class but used an invalid character after [ (for example [[,abc,]]) caused
3270 pcre_compile() to give the error "Failed: internal error: code overflow" or
3271 in some cases to crash with a glibc free() error. This could even happen if
3272 the pattern terminated after [[ but there just happened to be a sequence of
3273 letters, a binary zero, and a closing ] in the memory that followed.
3275 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3276 over the years. Originally (before any Unicode support), just the bottom 8
3277 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3278 output from "man perlunicode" includes this:
3280 The regular expression compiler produces polymorphic opcodes. That
3281 is, the pattern adapts to the data and automatically switches to
3282 the Unicode character scheme when presented with Unicode data--or
3283 instead uses a traditional byte scheme when presented with byte
3284 data.
3286 Sadly, a wide octal escape does not cause a switch, and in a string with
3287 no other multibyte characters, these octal escapes are treated as before.
3288 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3289 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3290 Unicode string.
3292 I have not perpetrated such confusion in PCRE. Up till now, it took just
3293 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3294 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3295 translate to the appropriate multibyte character.
3297 29. Applied some refactoring to reduce the number of warnings from Microsoft
3298 and Borland compilers. This has included removing the fudge introduced
3299 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3300 a warning about an unused variable.
3302 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3303 characters since release 4.0, because Perl (from release 5.004) does not.
3304 [Or at least, is documented not to: some releases seem to be in conflict
3305 with the documentation.] However, when a pattern was studied with
3306 pcre_study() and all its branches started with \s, PCRE still included VT
3307 as a possible starting character. Of course, this did no harm; it just
3308 caused an unnecessary match attempt.
3310 22. Removed a now-redundant internal flag bit that recorded the fact that case
3311 dependency changed within the pattern. This was once needed for "required
3312 byte" processing, but is no longer used. This recovers a now-scarce options
3313 bit. Also moved the least significant internal flag bit to the most-
3314 significant bit of the word, which was not previously used (hangover from
3315 the days when it was an int rather than a uint) to free up another bit for
3316 the future.
3318 23. Added support for CRLF line endings as well as CR and LF. As well as the
3319 default being selectable at build time, it can now be changed at runtime
3320 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3321 specify that it is scanning data with non-default line endings.
3323 24. Changed the definition of CXXLINK to make it agree with the definition of
3324 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3326 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3327 recursions. This makes a big different to stack usage for some patterns.
3329 26. If a subpattern containing a named recursion or subroutine reference such
3330 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3331 the space required for the compiled pattern went wrong and gave too small a
3332 value. Depending on the environment, this could lead to "Failed: internal
3333 error: code overflow at offset 49" or "glibc detected double free or
3334 corruption" errors.
3336 27. Applied patches from Google (a) to support the new newline modes and (b) to
3337 advance over multibyte UTF-8 characters in GlobalReplace.
3339 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3340 difference for some implementation of PCRE in some Windows version.
3342 29. Added some extra testing facilities to pcretest:
3344 \q<number> in a data line sets the "match limit" value
3345 \Q<number> in a data line sets the "match recursion limt" value
3346 -S <number> sets the stack size, where <number> is in megabytes
3348 The -S option isn't available for Windows.
3351 Version 6.6 06-Feb-06
3352 ---------------------
3354 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3355 in pcreposix.h. I have copied the definition from pcre.h.
3357 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3358 because pcre.h is no longer a built file.
3360 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3361 not normally included in the compiled code.
3364 Version 6.5 01-Feb-06
3365 ---------------------
3367 1. When using the partial match feature with pcre_dfa_exec(), it was not
3368 anchoring the second and subsequent partial matches at the new starting
3369 point. This could lead to incorrect results. For example, with the pattern
3370 /1234/, partially matching against "123" and then "a4" gave a match.
3372 2. Changes to pcregrep:
3374 (a) All non-match returns from pcre_exec() were being treated as failures
3375 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3376 error message is output. Some extra information is given for the
3378 probably the only errors that are likely to be caused by users (by
3379 specifying a regex that has nested indefinite repeats, for instance).
3380 If there are more than 20 of these errors, pcregrep is abandoned.
3382 (b) A binary zero was treated as data while matching, but terminated the
3383 output line if it was written out. This has been fixed: binary zeroes
3384 are now no different to any other data bytes.
3386 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3387 used to set a locale for matching. The --locale=xxxx long option has
3388 been added (no short equivalent) to specify a locale explicitly on the
3389 pcregrep command, overriding the environment variables.
3391 (d) When -B was used with -n, some line numbers in the output were one less
3392 than they should have been.
3394 (e) Added the -o (--only-matching) option.
3396 (f) If -A or -C was used with -c (count only), some lines of context were
3397 accidentally printed for the final match.
3399 (g) Added the -H (--with-filename) option.
3401 (h) The combination of options -rh failed to suppress file names for files
3402 that were found from directory arguments.
3404 (i) Added the -D (--devices) and -d (--directories) options.
3406 (j) Added the -F (--fixed-strings) option.
3408 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3410 (l) Added the --colo(u)r option.
3412 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3413 is not present by default.
3415 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3416 items such as (?R) or (?1), when the recursion could match a number of
3417 alternatives. If it matched one of the alternatives, but subsequently,
3418 outside the recursion, there was a failure, the code tried to back up into
3419 the recursion. However, because of the way PCRE is implemented, this is not
3420 possible, and the result was an incorrect result from the match.
3422 In order to prevent this happening, the specification of recursion has
3423 been changed so that all such subpatterns are automatically treated as
3424 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3426 4. I had overlooked the fact that, in some locales, there are characters for
3427 which isalpha() is true but neither isupper() nor islower() are true. In
3428 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3429 and ordfeminine) are like this. This affected the treatment of \w and \W
3430 when they appeared in character classes, but not when they appeared outside
3431 a character class. The bit map for "word" characters is now created
3432 separately from the results of isalnum() instead of just taking it from the
3433 upper, lower, and digit maps. (Plus the underscore character, of course.)
3435 5. The above bug also affected the handling of POSIX character classes such as
3436 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3437 permanent tables. Instead, the bit maps for such a class were previously
3438 created as the appropriate unions of the upper, lower, and digit bitmaps.
3439 Now they are created by subtraction from the [[:word:]] class, which has
3440 its own bitmap.
3442 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3443 It is created by subtracting the vertical space characters (\x09, \x0a,
3444 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3445 subtraction was done in the overall bitmap for a character class, meaning
3446 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3447 be recognized. This bug has been fixed.
3449 7. Patches from the folks at Google:
3451 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3452 real life, but is still worth protecting against".
3454 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3455 regular expressions".
3457 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3458 have it.
3460 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3461 "configure" and the latter not, in order to fix a problem somebody had
3462 with compiling the Arg class on HP-UX.
3464 (e) Improve the error-handling of the C++ wrapper a little bit.
3466 (f) New tests for checking recursion limiting.
3468 8. The pcre_memmove() function, which is used only if the environment does not
3469 have a standard memmove() function (and is therefore rarely compiled),
3470 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3471 returning a result (though PCRE never actually uses the result).
3473 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3474 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3475 returned instead of calling malloc() with an overflowing number that would
3476 most likely cause subsequent chaos.
3478 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3480 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3481 with this option is matched, the nmatch and pmatch options of regexec() are
3482 ignored.
3484 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3485 provided in case anyone wants to the the POSIX interface with UTF-8
3486 strings.
3488 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3489 C++ linking (needed for some HP-UX environments).
3491 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3492 (unused parameter) and in the pcre_printint() function (omitted "default"
3493 switch label when the default is to do nothing).
3495 15. Added some code to make it possible, when PCRE is compiled as a C++
3496 library, to replace subject pointers for pcre_exec() with a smart pointer
3497 class, thus making it possible to process discontinuous strings.
3499 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3500 much the same function. They were added by different people who were trying
3501 to make PCRE easy to compile on non-Unix systems. It has been suggested
3502 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3503 for compiling on Windows systems. I have therefore replaced it with
3504 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3505 defaults to "extern" for C or "extern C" for C++, which works fine on
3506 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3507 SCOPE with something explicit in config.h. In addition:
3509 (a) pcreposix.h still had just "extern" instead of either of these macros;
3510 I have replaced it with PCRE_DATA_SCOPE.
3512 (b) Functions such as _pcre_xclass(), which are internal to the library,
3513 but external in the C sense, all had PCRE_EXPORT in their definitions.
3514 This is apparently wrong for the Windows case, so I have removed it.
3515 (It makes no difference on Unix-like systems.)
3517 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3518 of recursive calls to match(). This is different to MATCH_LIMIT because
3519 that limits the total number of calls to match(), not all of which increase
3520 the depth of recursion. Limiting the recursion depth limits the amount of
3521 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3522 when PCRE is compiled, and changed at run time. A patch from Google adds
3523 this functionality to the C++ interface.
3525 18. Changes to the handling of Unicode character properties:
3527 (a) Updated the table to Unicode 4.1.0.
3529 (b) Recognize characters that are not in the table as "Cn" (undefined).
3531 (c) I revised the way the table is implemented to a much improved format
3532 which includes recognition of ranges. It now supports the ranges that
3533 are defined in UnicodeData.txt, and it also amalgamates other
3534 characters into ranges. This has reduced the number of entries in the
3535 table from around 16,000 to around 3,000, thus reducing its size
3536 considerably. I realized I did not need to use a tree structure after
3537 all - a binary chop search is just as efficient. Having reduced the
3538 number of entries, I extended their size from 6 bytes to 8 bytes to
3539 allow for more data.
3541 (d) Added support for Unicode script names via properties such as \p{Han}.
3543 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3544 matching that character.
3546 20. When matching a repeated Unicode property with a minimum greater than zero,
3547 (for example \pL{2,}), PCRE could look past the end of the subject if it
3548 reached it while seeking the minimum number of characters. This could
3549 happen only if some of the characters were more than one byte long, because
3550 there is a check for at least the minimum number of bytes.
3552 21. Refactored the implementation of \p and \P so as to be more general, to
3553 allow for more different types of property in future. This has changed the
3554 compiled form incompatibly. Anybody with saved compiled patterns that use
3555 \p or \P will have to recompile them.
3557 22. Added "Any" and "L&" to the supported property types.
3559 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3560 but give a compile time error if the value is greater than 0xff.
3562 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3563 accidentally not being installed or uninstalled.
3565 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3566 made were to insert the current release number. This seemed silly, because
3567 it made things harder for people building PCRE on systems that don't run
3568 "configure". I have turned pcre.h into a distributed file, no longer built
3569 by "configure", with the version identification directly included. There is
3570 no longer a pcre.h.in file.
3572 However, this change necessitated a change to the pcre-config script as
3573 well. It is built from pcre-config.in, and one of the substitutions was the
3574 release number. I have updated configure.ac so that ./configure now finds
3575 the release number by grepping pcre.h.
3577 26. Added the ability to run the tests under valgrind.
3580 Version 6.4 05-Sep-05
3581 ---------------------
3583 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3584 "--" to be printed when multiple files were scanned, even when none of the
3585 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3586 consider it to be a bug, and have restored the previous behaviour.
3588 2. A couple of code tidies to get rid of compiler warnings.
3590 3. The pcretest program used to cheat by referring to symbols in the library
3591 whose names begin with _pcre_. These are internal symbols that are not
3592 really supposed to be visible externally, and in some environments it is
3593 possible to suppress them. The cheating is now confined to including
3594 certain files from the library's source, which is a bit cleaner.
3596 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3597 file's purpose clearer.
3599 5. Reorganized pcre_ucp_findchar().
3602 Version 6.3 15-Aug-05
3603 ---------------------
3605 1. The file libpcre.pc.in did not have general read permission in the tarball.
3607 2. There were some problems when building without C++ support:
3609 (a) If C++ support was not built, "make install" and "make test" still
3610 tried to test it.
3612 (b) There were problems when the value of CXX was explicitly set. Some
3613 changes have been made to try to fix these, and ...
3615 (c) --disable-cpp can now be used to explicitly disable C++ support.
3617 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3618 backslash in a target when C++ was disabled. This confuses some
3619 versions of "make", apparently. Using an intermediate variable solves
3620 this. (Same for CPP_LOBJ.)
3622 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3623 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3624 necessary on certain architectures.
3626 4. Added a setting of -export-symbols-regex to the link command to remove
3627 those symbols that are exported in the C sense, but actually are local
3628 within the library, and not documented. Their names all begin with
3629 "_pcre_". This is not a perfect job, because (a) we have to except some
3630 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3631 available (and never for static libraries). I have made a note to try to
3632 find a way round (a) in the future.
3635 Version 6.2 01-Aug-05
3636 ---------------------
3638 1. There was no test for integer overflow of quantifier values. A construction
3639 such as {1111111111111111} would give undefined results. What is worse, if
3640 a minimum quantifier for a parenthesized subpattern overflowed and became
3641 negative, the calculation of the memory size went wrong. This could have
3642 led to memory overwriting.
3644 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3646 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3647 operating environments where this matters.
3649 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3650 PCRE options from within the C++ wrapper.
3652 5. Named capturing subpatterns were not being correctly counted when a pattern
3653 was compiled. This caused two problems: (a) If there were more than 100
3654 such subpatterns, the calculation of the memory needed for the whole
3655 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3656 back references of the form \12, where the number was greater than 9, were
3657 not recognized as back references, even though there were sufficient
3658 previous subpatterns.
3660 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3661 versions of gcc, e.g. 2.95.4.
3664 Version 6.1 21-Jun-05
3665 ---------------------
3667 1. There was one reference to the variable "posix" in pcretest.c that was not
3668 surrounded by "#if !defined NOPOSIX".
3670 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3671 the cross-check on the old pcre_info() function, for the benefit of the
3672 cut-down version of PCRE that is currently imported into Exim.
3674 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3675 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3676 patterns that start (?i) but I don't think that matters. The use of (?i) is
3677 just an example; this all applies to the other options as well.
3679 4. Since libtool seems to echo the compile commands it is issuing, the output
3680 from "make" can be reduced a bit by putting "@" in front of each libtool
3681 compile command.
3683 5. Patch from the folks at Google for configure.in to be a bit more thorough
3684 in checking for a suitable C++ installation before trying to compile the
3685 C++ stuff. This should fix a reported problem when a compiler was present,
3686 but no suitable headers.
3688 6. The man pages all had just "PCRE" as their title. I have changed them to
3689 be the relevant file name. I have also arranged that these names are
3690 retained in the file doc/pcre.txt, which is a concatenation in text format
3691 of all the man pages except the little individual ones for each function.
3693 7. The NON-UNIX-USE file had not been updated for the different set of source
3694 files that come with release 6. I also added a few comments about the C++
3695 wrapper.
3698 Version 6.0 07-Jun-05
3699 ---------------------
3701 1. Some minor internal re-organization to help with my DFA experiments.
3703 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3704 didn't matter for the library itself when fully configured, but did matter
3705 when compiling without UCP support, or within Exim, where the ucp files are
3706 not imported.
3708 3. Refactoring of the library code to split up the various functions into
3709 different source modules. The addition of the new DFA matching code (see
3710 below) to a single monolithic source would have made it really too
3711 unwieldy, quite apart from causing all the code to be include in a
3712 statically linked application, when only some functions are used. This is
3713 relevant even without the DFA addition now that patterns can be compiled in
3714 one application and matched in another.
3716 The downside of splitting up is that there have to be some external
3717 functions and data tables that are used internally in different modules of
3718 the library but which are not part of the API. These have all had their
3719 names changed to start with "_pcre_" so that they are unlikely to clash
3720 with other external names.
3722 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3723 a different (DFA) algorithm. Although it is slower than the original
3724 function, it does have some advantages for certain types of matching
3725 problem.
3727 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3728 including restarting after a partial match.
3730 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3731 defined when compiling for Windows was sent to me. I have put it into the
3732 code, though I have no means of testing or verifying it.
3734 7. Added the pcre_refcount() auxiliary function.
3736 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3737 match before or at the first newline in the subject string. In pcretest,
3738 the /f option on a pattern can be used to set this.
3740 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3741 would behave wrongly. This has been present in PCRE since release 4.0.
3743 10. A number of changes to the pcregrep command:
3745 (a) Refactored how -x works; insert ^(...)$ instead of setting
3746 PCRE_ANCHORED and checking the length, in preparation for adding
3747 something similar for -w.
3749 (b) Added the -w (match as a word) option.
3751 (c) Refactored the way lines are read and buffered so as to have more
3752 than one at a time available.
3754 (d) Implemented a pcregrep test script.
3756 (e) Added the -M (multiline match) option. This allows patterns to match
3757 over several lines of the subject. The buffering ensures that at least
3758 8K, or the rest of the document (whichever is the shorter) is available
3759 for matching (and similarly the previous 8K for lookbehind assertions).
3761 (f) Changed the --help output so that it now says
3763 -w, --word-regex(p)
3765 instead of two lines, one with "regex" and the other with "regexp"
3766 because that confused at least one person since the short forms are the
3767 same. (This required a bit of code, as the output is generated
3768 automatically from a table. It wasn't just a text change.)
3770 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3771 option but starts with a hyphen. Could be a pattern or a path name
3772 starting with a hyphen, for instance.
3774 (h) "-" can be given as a file name to represent stdin.
3776 (i) When file names are being printed, "(standard input)" is used for
3777 the standard input, for compatibility with GNU grep. Previously
3778 "<stdin>" was used.
3780 (j) The option --label=xxx can be used to supply a name to be used for
3781 stdin when file names are being printed. There is no short form.
3783 (k) Re-factored the options decoding logic because we are going to add
3784 two more options that take data. Such options can now be given in four
3785 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3787 (l) Added the -A, -B, and -C options for requesting that lines of context
3788 around matches be printed.
3790 (m) Added the -L option to print the names of files that do not contain
3791 any matching lines, that is, the complement of -l.
3793 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3794 continue to scan other files.
3796 (o) The -s option was incorrectly implemented. For compatibility with other
3797 greps, it now suppresses the error message for a non-existent or non-
3798 accessible file (but not the return code). There is a new option called
3799 -q that suppresses the output of matching lines, which was what -s was
3800 previously doing.
3802 (p) Added --include and --exclude options to specify files for inclusion
3803 and exclusion when recursing.
3805 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3806 Hopefully, it now does.
3808 12. Missing cast in pcre_study().
3810 13. Added an "uninstall" target to the makefile.
3812 14. Replaced "extern" in the function prototypes in Makefile.in with
3813 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3814 world, but is set differently for Windows.
3816 15. Added a second compiling function called pcre_compile2(). The only
3817 difference is that it has an extra argument, which is a pointer to an
3818 integer error code. When there is a compile-time failure, this is set
3819 non-zero, in addition to the error test pointer being set to point to an
3820 error message. The new argument may be NULL if no error number is required
3821 (but then you may as well call pcre_compile(), which is now just a
3822 wrapper). This facility is provided because some applications need a
3823 numeric error indication, but it has also enabled me to tidy up the way
3824 compile-time errors are handled in the POSIX wrapper.
3826 16. Added VPATH=.libs to the makefile; this should help when building with one
3827 prefix path and installing with another. (Or so I'm told by someone who
3828 knows more about this stuff than I do.)
3830 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3831 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3832 match everything, including newlines. This is not POSIX-compatible, but
3833 somebody wanted the feature. From pcretest it can be activated by using
3834 both the P and the s flags.
3836 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3838 19. libpcre.pc was being incorrectly installed as executable.
3840 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3841 it now also looks for '\r' so that it will work unmodified on Windows.
3843 21. Added Google's contributed C++ wrapper to the distribution.
3845 22. Added some untidy missing memory free() calls in pcretest, to keep
3846 Electric Fence happy when testing.
3850 Version 5.0 13-Sep-04
3851 ---------------------
3853 1. Internal change: literal characters are no longer packed up into items
3854 containing multiple characters in a single byte-string. Each character
3855 is now matched using a separate opcode. However, there may be more than one
3856 byte in the character in UTF-8 mode.
3858 2. The pcre_callout_block structure has two new fields: pattern_position and
3859 next_item_length. These contain the offset in the pattern to the next match
3860 item, and its length, respectively.
3862 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3863 insertion of callouts before each pattern item. Added the /C option to
3864 pcretest to make use of this.
3866 4. On the advice of a Windows user, the lines
3868 #if defined(_WIN32) || defined(WIN32)
3869 _setmode( _fileno( stdout ), 0x8000 );
3870 #endif /* defined(_WIN32) || defined(WIN32) */
3872 have been added to the source of pcretest. This apparently does useful
3873 magic in relation to line terminators.
3875 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3876 for the benefit of those environments where the "b" makes a difference.
3878 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3879 to know about it. I have put a hack into configure.in that adds in code
3880 to set GCC=yes if CC=icc. This seems to end up at a point in the
3881 generated configure script that is early enough to affect the setting of
3882 compiler options, which is what is needed, but I have no means of testing
3883 whether it really works. (The user who reported this had patched the
3884 generated configure script, which of course I cannot do.)
3886 LATER: After change 22 below (new libtool files), the configure script
3887 seems to know about icc (and also ecc). Therefore, I have commented out
3888 this hack in configure.in.
3890 7. Added support for pkg-config (2 patches were sent in).
3892 8. Negated POSIX character classes that used a combination of internal tables
3893 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3894 [[:^ascii]]. Typically, they would match almost any characters. The other
3895 POSIX classes were not broken in this way.
3897 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3898 to find the match, as PCRE was deluded into thinking that the match had to
3899 start at the start point or following a newline. The same bug applied to
3900 patterns with negative forward assertions or any backward assertions
3901 preceding ".*" at the start, unless the pattern required a fixed first
3902 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3904 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3905 starting at the last subject character, bytes beyond the end of the subject
3906 string were read.
3908 11. Renamed the variable "class" as "classbits" to make life easier for C++
3909 users. (Previously there was a macro definition, but it apparently wasn't
3910 enough.)
3912 12. Added the new field "tables" to the extra data so that tables can be passed
3913 in at exec time, or the internal tables can be re-selected. This allows
3914 a compiled regex to be saved and re-used at a later time by a different
3915 program that might have everything at different addresses.
3917 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3918 -R library as well as a -L library.
3920 14. The debugging options of pcretest (-d on the command line or D on a
3921 pattern) showed incorrect output for anything following an extended class
3922 that contained multibyte characters and which was followed by a quantifier.
3924 15. Added optional support for general category Unicode character properties
3925 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3926 support. It adds about 90K to the size of the library. The meanings of the
3927 inbuilt class escapes such as \d and \s have NOT been changed.
3929 16. Updated pcredemo.c to include calls to free() to release the memory for the
3930 compiled pattern.
3932 17. The generated file chartables.c was being created in the source directory
3933 instead of in the building directory. This caused the build to fail if the
3934 source directory was different from the building directory, and was
3935 read-only.
3937 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3938 file. No doubt somebody will tell me if they don't make sense... Also added
3939 Dan Mooney's comments about building on OpenVMS.
3941 19. Added support for partial matching via the PCRE_PARTIAL option for
3942 pcre_exec() and the \P data escape in pcretest.
3944 20. Extended pcretest with 3 new pattern features:
3946 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3947 write the compiled pattern to the file whose name is "rest-of-line".
3948 This is a straight binary dump of the data, with the saved pointer to
3949 the character tables forced to be NULL. The study data, if any, is
3950 written too. After writing, pcretest reads a new pattern.
3952 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3953 compiled pattern from the given file. There must not be any
3954 occurrences of "<" in the file name (pretty unlikely); if there are,
3955 pcretest will instead treat the initial "<" as a pattern delimiter.
3956 After reading in the pattern, pcretest goes on to read data lines as
3957 usual.
3959 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3960 and 16-bit fields in a compiled pattern, to simulate a pattern that
3961 was compiled on a host of opposite endianness.
3963 21. The pcre-exec() function can now cope with patterns that were compiled on
3964 hosts of opposite endianness, with this restriction:
3966 As for any compiled expression that is saved and used later, the tables
3967 pointer field cannot be preserved; the extra_data field in the arguments
3968 to pcre_exec() should be used to pass in a tables address if a value
3969 other than the default internal tables were used at compile time.
3971 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3972 now diagnosed as an error. Previously, most of the time, a negative number
3973 would have been treated as zero, but if in addition "ovector" was passed as
3974 NULL, a crash could occur.
3976 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3977 new versions from the libtool 1.5 distribution (the last one is a copy of
3978 a file called libtool.m4). This seems to have fixed the need to patch
3979 "configure" to support Darwin 1.3 (which I used to do). However, I still
3980 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3981 workstation).
3983 24. Changed the PCRE licence to be the more standard "BSD" licence.
3986 Version 4.5 01-Dec-03
3987 ---------------------
3989 1. There has been some re-arrangement of the code for the match() function so
3990 that it can be compiled in a version that does not call itself recursively.
3991 Instead, it keeps those local variables that need separate instances for
3992 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3993 needs to "recurse". Keeping track of where control must go is done by means
3994 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3995 hide most of the details from the main code, and operates only if
3996 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3997 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3998 operating.
4000 To make it easier for callers to provide specially tailored get/free
4001 functions for this usage, two new functions, pcre_stack_malloc, and
4002 pcre_stack_free, are used. They are always called in strict stacking order,
4003 and the size of block requested is always the same.
4005 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4006 PCRE has been compiled to use the stack or the heap for recursion. The
4007 -C option of pcretest uses this to show which version is compiled.
4009 A new data escape \S, is added to pcretest; it causes the amounts of store
4010 obtained and freed by both kinds of malloc/free at match time to be added
4011 to the output.
4013 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4014 what's available on my current Linux desktop machine.
4016 3. When matching a UTF-8 string, the test for a valid string at the start has
4017 been extended. If start_offset is not zero, PCRE now checks that it points
4018 to a byte that is the start of a UTF-8 character. If not, it returns
4019 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4020 this is necessary because there may be backward assertions in the pattern.
4021 When matching the same subject several times, it may save resources to use
4022 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4024 4. The code for checking the validity of UTF-8 strings has been tightened so
4025 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4026 containing "overlong sequences".
4028 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4029 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4030 should have been "&", but it just so happened that all the cases this let
4031 through by mistake were picked up later in the function.
4033 6. I had used a variable called "isblank" - this is a C99 function, causing
4034 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4036 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4037 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4038 so the version is shown; (c) stop "make test" echoing "./RunTest".
4040 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4042 9. The prototype for memmove() for systems that don't have it was using
4043 size_t, but the inclusion of the header that defines size_t was later. I've
4044 moved the #includes for the C headers earlier to avoid this.
4046 10. Added some adjustments to the code to make it easier to compiler on certain
4047 special systems:
4049 (a) Some "const" qualifiers were missing.
4050 (b) Added the macro EXPORT before all exported functions; by default this
4051 is defined to be empty.
4052 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4053 that it reads its output file name as an argument instead of writing
4054 to the standard output and assuming this can be redirected.
4056 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4057 class containing characters with values greater than 255, PCRE compilation
4058 went into a loop.
4060 12. A recursive reference to a subpattern that was within another subpattern
4061 that had a minimum quantifier of zero caused PCRE to crash. For example,
4062 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4063 recursion. If the recursively-called subpattern itself had a zero repeat,
4064 that was OK.
4066 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4067 buffer into which it was copied (for escape processing) was still set at
4068 1024, so long lines caused crashes.
4070 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4071 "internal error: code overflow...". This applied to any character class
4072 that was followed by a possessive quantifier.
4074 15. Modified the Makefile to add libpcre.la as a prerequisite for
4075 libpcreposix.la because I was told this is needed for a parallel build to
4076 work.
4078 16. If a pattern that contained .* following optional items at the start was
4079 studied, the wrong optimizing data was generated, leading to matching
4080 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4081 matching string must start with a or b or c. The correct conclusion for
4082 this pattern is that a match can start with any character.
4085 Version 4.4 13-Aug-03
4086 ---------------------
4088 1. In UTF-8 mode, a character class containing characters with values between
4089 127 and 255 was not handled correctly if the compiled pattern was studied.
4090 In fixing this, I have also improved the studying algorithm for such
4091 classes (slightly).
4093 2. Three internal functions had redundant arguments passed to them. Removal
4094 might give a very teeny performance improvement.
4096 3. Documentation bug: the value of the capture_top field in a callout is *one
4097 more than* the number of the hightest numbered captured substring.
4099 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4100 in incorrectly linking with a previously installed version. They now link
4101 explicitly with libpcre.la.
4103 5. configure.in no longer needs to recognize Cygwin specially.
4105 6. A problem in pcre.in for Windows platforms is fixed.
4107 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4108 pcretest, it used to include the size of the study block as part of its
4109 output. Unfortunately, the structure contains a field that has a different
4110 size on different hardware architectures. This meant that the tests that
4111 showed this size failed. As the block is currently always of a fixed size,
4112 this information isn't actually particularly useful in pcretest output, so
4113 I have just removed it.
4115 8. Three pre-processor statements accidentally did not start in column 1.
4116 Sadly, there are *still* compilers around that complain, even though
4117 standard C has not required this for well over a decade. Sigh.
4119 9. In pcretest, the code for checking callouts passed small integers in the
4120 callout_data field, which is a void * field. However, some picky compilers
4121 complained about the casts involved for this on 64-bit systems. Now
4122 pcretest passes the address of the small integer instead, which should get
4123 rid of the warnings.
4125 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4126 both compile and run time, and gives an error if an invalid UTF-8 sequence
4127 is found. There is a option for disabling this check in cases where the
4128 string is known to be correct and/or the maximum performance is wanted.
4130 11. In response to a bug report, I changed one line in Makefile.in from
4132 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4133 to
4134 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4136 to look similar to other lines, but I have no way of telling whether this
4137 is the right thing to do, as I do not use Windows. No doubt I'll get told
4138 if it's wrong...
4141 Version 4.3 21-May-03
4142 ---------------------
4144 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4145 Makefile.
4147 2. Some refactoring to improve the quality of the code:
4149 (i) The utf8_table... variables are now declared "const".
4151 (ii) The code for \cx, which used the "case flipping" table to upper case
4152 lower case letters, now just substracts 32. This is ASCII-specific,
4153 but the whole concept of \cx is ASCII-specific, so it seems
4154 reasonable.
4156 (iii) PCRE was using its character types table to recognize decimal and
4157 hexadecimal digits in the pattern. This is silly, because it handles
4158 only 0-9, a-f, and A-F, but the character types table is locale-
4159 specific, which means strange things might happen. A private
4160 table is now used for this - though it costs 256 bytes, a table is
4161 much faster than multiple explicit tests. Of course, the standard
4162 character types table is still used for matching digits in subject
4163 strings against \d.
4165 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4166 ending in _t are). So I've renamed it as ESC_tee.
4168 3. The first argument for regexec() in the POSIX wrapper should have been
4169 defined as "const".
4171 4. Changed pcretest to use malloc() for its buffers so that they can be
4172 Electric Fenced for debugging.
4174 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4175 to read one or more bytes before the start of the subject string. Often this
4176 had no effect on PCRE's behaviour, but in some circumstances it could
4177 provoke a segmentation fault.
4179 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4180 to try to read one or more bytes before the start of the subject string.
4182 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4183 UTF-8 support could misbehave in various ways if the subject string
4184 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4185 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4186 back over UTF-8 characters.)
4189 Version 4.2 14-Apr-03
4190 ---------------------
4192 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4194 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4195 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4196 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4197 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4198 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4200 Note: automatic setting of the BUILD variables is not yet working
4201 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4202 compile-time but not at link-time
4203 [LINK]: use for linking executables only
4204 make different versions for Windows and non-Windows
4205 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4206 libraries
4207 [LINK_FOR_BUILD]: new variable
4208 [OBJEXT]: use throughout
4209 [EXEEXT]: use throughout
4210 <winshared>: new target
4211 <wininstall>: new target
4212 <dftables.o>: use native compiler
4213 <dftables>: use native linker
4214 <install>: handle Windows platform correctly
4215 <clean>: ditto
4216 <check>: ditto
4217 copy DLL to top builddir before testing
4219 As part of these changes, -no-undefined was removed again. This was reported
4220 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4221 in any case.
4223 3. Some tidies to get rid of compiler warnings:
4225 . In the match_data structure, match_limit was an unsigned long int, whereas
4226 match_call_count was an int. I've made them both unsigned long ints.
4228 . In pcretest the fact that a const uschar * doesn't automatically cast to
4229 a void * provoked a warning.
4231 . Turning on some more compiler warnings threw up some "shadow" variables
4232 and a few more missing casts.
4234 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4235 option, a class that contained a single character with a value between 128
4236 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4238 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4239 option, a class that contained several characters, but with at least one
4240 whose value was between 128 and 255 caused PCRE to crash.
4243 Version 4.1 12-Mar-03
4244 ---------------------
4246 1. Compiling with gcc -pedantic found a couple of places where casts were
4247 needed, and a string in dftables.c that was longer than standard compilers are
4248 required to support.
4250 2. Compiling with Sun's compiler found a few more places where the code could
4251 be tidied up in order to avoid warnings.
4253 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4254 first of these names is deprecated in the latest Autoconf in favour of the name
4255 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4256 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4257 analogy I have changed it to CFLAGS_FOR_BUILD.
4259 4. Added -no-undefined to the linking command in the Makefile, because this is
4260 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4261 linking step for the pcreposix library.
4263 5. PCRE was failing to diagnose the case of two named groups with the same
4264 name.
4266 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4267 literal character that is needed in the subject for a match, and scans along to
4268 ensure that it is present before embarking on the full matching process. This
4269 saves time in cases of nested unlimited repeats that are never going to match.
4270 Problem: the scan can take a lot of time if the subject is very long (e.g.
4271 megabytes), thus penalizing straightforward matches. It is now done only if the
4272 amount of subject to be scanned is less than 1000 bytes.
4274 7. A lesser problem with the same optimization is that it was recording the
4275 first character of an anchored pattern as "needed", thus provoking a search
4276 right along the subject, even when the first match of the pattern was going to
4277 fail. The "needed" character is now not set for anchored patterns, unless it
4278 follows something in the pattern that is of non-fixed length. Thus, it still
4279 fulfils its original purpose of finding quick non-matches in cases of nested
4280 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4283 Version 4.0 17-Feb-03
4284 ---------------------
4286 1. If a comment in an extended regex that started immediately after a meta-item
4287 extended to the end of string, PCRE compiled incorrect data. This could lead to
4288 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4290 2. Moved to autoconf 2.53 and libtool 1.4.2.
4292 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4293 the special perltest8 script is no longer needed - all the tests can be run
4294 from a single perltest script.
4296 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4297 by \s. It has now been removed in PCRE. This means it isn't recognized as
4298 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4299 class [:space:] *does* include VT, thereby creating a mess.
4301 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4302 space and tab.
4304 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4305 its new features into the main test script, reducing the number of scripts.
4307 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4308 were backward compatible, and made the (?i) apply to the whole pattern, as if
4309 /i were given. Now it behaves more logically, and applies the option setting
4310 only to what follows. PCRE has been changed to follow suit. However, if it
4311 finds options settings right at the start of the pattern, it extracts them into
4312 the global options, as before. Thus, they show up in the info data.
4314 8. Added support for the \Q...\E escape sequence. Characters in between are
4315 treated as literals. This is slightly different from Perl in that $ and @ are
4316 also handled as literals inside the quotes. In Perl, they will cause variable
4317 interpolation. Note the following examples:
4319 Pattern PCRE matches Perl matches
4321 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4322 \Qabc\$xyz\E abc\$xyz abc\$xyz
4323 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4325 For compatibility with Perl, \Q...\E sequences are recognized inside character
4326 classes as well as outside them.
4328 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4329 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4330 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4331 signed/unsigned warnings.
4333 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4334 option for pcretest, so I've replaced it by a simple function that does just
4335 that job.
4337 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4338 "pcregrep -".
4340 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4341 Java package. This provides some syntactic sugar for simple cases of what my
4342 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4343 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4344 item, you can use this simplified notation. Note that only makes sense with
4345 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4346 greediness, whatever the setting of the PCRE_UNGREEDY option.
4348 13. A change of greediness default within a pattern was not taking effect at
4349 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4350 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4351 was abstracted outside.
4353 14. PCRE now supports the \G assertion. It is true when the current matching
4354 position is at the start point of the match. This differs from \A when the
4355 starting offset is non-zero. Used with the /g option of pcretest (or similar
4356 code), it works in the same way as it does for Perl's /g option. If all
4357 alternatives of a regex begin with \G, the expression is anchored to the start
4358 match position, and the "anchored" flag is set in the compiled expression.
4360 15. Some bugs concerning the handling of certain option changes within patterns
4361 have been fixed. These applied to options other than (?ims). For example,
4362 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4363 been the other way round. Some of this was related to change 7 above.
4365 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4366 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4367 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4368 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4370 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4371 mode. Unlike ".", it always matches newline, whatever the setting of
4372 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4373 assertions. Perl allows it, but it doesn't (in general) work because it can't
4374 calculate the length of the lookbehind. At least, that's the case for Perl
4375 5.8.0 - I've been told they are going to document that it doesn't work in
4376 future.
4378 18. Added an error diagnosis for escapes that PCRE does not support: these are
4379 \L, \l, \N, \P, \p, \U, \u, and \X.
4381 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4382 reading past the end of the pattern in cases such as /[abcd/.
4384 20. PCRE was getting more memory than necessary for patterns with classes that
4385 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4387 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4388 compiling PCRE for use with Virtual Pascal.
4390 22. Small fix to the Makefile to make it work properly if the build is done
4391 outside the source tree.
4393 23. Added a new extension: a condition to go with recursion. If a conditional
4394 subpattern starts with (?(R) the "true" branch is used if recursion has
4395 happened, whereas the "false" branch is used only at the top level.
4397 24. When there was a very long string of literal characters (over 255 bytes
4398 without UTF support, over 250 bytes with UTF support), the computation of how
4399 much memory was required could be incorrect, leading to segfaults or other
4400 strange effects.
4402 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4403 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4404 there was a subsequent back reference to those brackets. This meant that, for
4405 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4406 possible to check for precisely this case. All we can do is abandon the
4407 optimization if .* occurs inside capturing brackets when there are any back
4408 references whatsoever. (See below for a better fix that came later.)
4410 26. The handling of the optimization for finding the first character of a
4411 non-anchored pattern, and for finding a character that is required later in the
4412 match were failing in some cases. This didn't break the matching; it just
4413 failed to optimize when it could. The way this is done has been re-implemented.
4415 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4417 28. Added a new feature that provides some of the functionality that Perl
4418 provides with (?{...}). The facility is termed a "callout". The way it is done
4419 in PCRE is for the caller to provide an optional function, by setting
4420 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4421 global variable. By default it is unset, which disables all calling out. To get
4422 the function called, the regex must include (?C) at appropriate points. This
4423 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4424 This provides a means of identifying different callout points. When PCRE
4425 reaches such a point in the regex, if pcre_callout has been set, the external
4426 function is called. It is provided with data in a structure called
4427 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4428 matching continues; if it returns a non-zero value, the match at the current
4429 point fails. However, backtracking will occur if possible. [This was changed
4430 later and other features added - see item 49 below.]
4432 29. pcretest is upgraded to test the callout functionality. It provides a
4433 callout function that displays information. By default, it shows the start of
4434 the match and the current position in the text. There are some new data escapes
4435 to vary what happens:
4437 \C+ in addition, show current contents of captured substrings
4438 \C- do not supply a callout function
4439 \C!n return 1 when callout number n is reached
4440 \C!n!m return 1 when callout number n is reached for the mth time
4442 30. If pcregrep was called with the -l option and just a single file name, it
4443 output "<stdin>" if a match was found, instead of the file name.
4445 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4446 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4447 pcre_exec(). This saves a malloc/free per call. The default value of
4448 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4449 when configuring.
4451 32. The default maximum size of a compiled pattern is 64K. There have been a
4452 few cases of people hitting this limit. The code now uses macros to handle the
4453 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4454 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4455 configuring. Tests 2 and 5 work only with 2-byte links because they output
4456 debugging information about compiled patterns.
4458 33. Internal code re-arrangements:
4460 (a) Moved the debugging function for printing out a compiled regex into
4461 its own source file (printint.c) and used #include to pull it into
4462 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4463 separate copies.
4465 (b) Defined the list of op-code names for debugging as a macro in
4466 internal.h so that it is next to the definition of the opcodes.
4468 (c) Defined a table of op-code lengths for simpler skipping along compiled
4469 code. This is again a macro in internal.h so that it is next to the
4470 definition of the opcodes.
4472 34. Added support for recursive calls to individual subpatterns, along the
4473 lines of Robin Houston's patch (but implemented somewhat differently).
4475 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4476 allow it to read and process whole directories in Win32. This code was
4477 contributed by Lionel Fourquaux; it has not been tested by me.
4479 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4480 used to name a group. Names consist of alphanumerics and underscores, and must
4481 be unique. Back references use the syntax (?P=name) and recursive calls use
4482 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4483 numbers. The function pcre_fullinfo() can be used after compilation to extract
4484 a name/number map. There are three relevant calls:
4486 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4487 PCRE_INFO_NAMECOUNT yields the number of entries
4488 PCRE_INFO_NAMETABLE yields a pointer to the map.
4490 The map is a vector of fixed-size entries. The size of each entry depends on
4491 the length of the longest name used. The first two bytes of each entry are the
4492 group number, most significant byte first. There follows the corresponding
4493 name, zero terminated. The names are in alphabetical order.
4495 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4496 case instead of 255. Making it the same both with and without UTF-8 support
4497 means that the same test output works with both.
4499 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4500 calling malloc() with a zero argument.
4502 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4503 optimization. I've improved things by keeping a bitmap of backreferences with
4504 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4505 fact referenced, the optimization can be applied. It is unlikely that a
4506 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4507 the match to follow \n) will appear inside brackets with a number greater than
4508 31, but if it does, any back reference > 31 suppresses the optimization.
4510 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4511 of disabling numbered capturing parentheses. Any opening parenthesis that is
4512 not followed by ? behaves as if it were followed by ?: but named parentheses
4513 can still be used for capturing (and they will acquire numbers in the usual
4514 way).
4516 41. Redesigned the return codes from the match() function into yes/no/error so
4517 that errors can be passed back from deep inside the nested calls. A malloc
4518 failure while inside a recursive subpattern call now causes the
4519 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4521 42. It is now possible to set a limit on the number of times the match()
4522 function is called in a call to pcre_exec(). This facility makes it possible to
4523 limit the amount of recursion and backtracking, though not in a directly
4524 obvious way, because the match() function is used in a number of different
4525 circumstances. The count starts from zero for each position in the subject
4526 string (for non-anchored patterns). The default limit is, for compatibility, a
4527 large number, namely 10 000 000. You can change this in two ways:
4529 (a) When configuring PCRE before making, you can use --with-match-limit=n
4530 to set a default value for the compiled library.
4532 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4533 a different value is set. See 45 below.
4535 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4537 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4538 of things that can be changed at compile time. The first argument specifies
4539 what is wanted and the second points to where the information is to be placed.
4540 The current list of available information is:
4544 The output is an integer that is set to one if UTF-8 support is available;
4545 otherwise it is set to zero.
4549 The output is an integer that it set to the value of the code that is used for
4550 newline. It is either LF (10) or CR (13).
4554 The output is an integer that contains the number of bytes used for internal
4555 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4559 The output is an integer that contains the threshold above which the POSIX
4560 interface uses malloc() for output vectors. See item 31 above.
4564 The output is an unsigned integer that contains the default limit of the number
4565 of match() calls in a pcre_exec() execution. See 42 above.
4567 44. pcretest has been upgraded by the addition of the -C option. This causes it
4568 to extract all the available output from the new pcre_config() function, and to
4569 output it. The program then exits immediately.
4571 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4572 order to support additional features. One way would have been to define
4573 pcre_exec2() (for example) with extra arguments, but this would not have been
4574 extensible, and would also have required all calls to the original function to
4575 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4576 is used for passing in "extra" data from pcre_study().
4578 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4579 contains the following fields:
4581 flags a bitmap indicating which of the following fields are set
4582 study_data opaque data from pcre_study()
4583 match_limit a way of specifying a limit on match() calls for a specific
4584 call to pcre_exec()
4585 callout_data data for callouts (see 49 below)
4587 The flag bits are also defined in pcre.h, and are
4593 The pcre_study() function now returns one of these new pcre_extra blocks, with
4594 the actual study data pointed to by the study_data field, and the
4595 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4596 before. That is, this change is entirely upwards-compatible and requires no
4597 change to existing code.
4599 If you want to pass in additional data to pcre_exec(), you can either place it
4600 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4601 block.
4603 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4604 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4605 times with different match limits, until it finds the minimum value needed for
4606 pcre_exec() to complete. The value is then output. This can be instructive; for
4607 most simple matches the number is quite small, but for pathological cases it
4608 gets very large very quickly.
4610 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4611 returns the size of the data block pointed to by the study_data field in a
4612 pcre_extra block, that is, the value that was passed as the argument to
4613 pcre_malloc() when PCRE was getting memory in which to place the information
4614 created by pcre_study(). The fourth argument should point to a size_t variable.
4615 pcretest has been extended so that this information is shown after a successful
4616 pcre_study() call when information about the compiled regex is being displayed.
4618 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4619 because what follows is always an absolute path. (Later: it turns out that this
4620 is more than cosmetic for MinGW, because it doesn't like empty path
4621 components.)
4623 49. Some changes have been made to the callout feature (see 28 above):
4625 (i) A callout function now has three choices for what it returns:
4627 0 => success, carry on matching
4628 > 0 => failure at this point, but backtrack if possible
4629 < 0 => serious error, return this value from pcre_exec()
4631 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4632 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4633 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4634 use by callout functions. It will never be used by PCRE itself.
4636 (ii) The pcre_extra structure (see 45 above) has a void * field called
4637 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4638 pcre_callout_block structure has a field of the same name. The contents of
4639 the field passed in the pcre_extra structure are passed to the callout
4640 function in the corresponding field in the callout block. This makes it
4641 easier to use the same callout-containing regex from multiple threads. For
4642 testing, the pcretest program has a new data escape
4644 \C*n pass the number n (may be negative) as callout_data
4646 If the callout function in pcretest receives a non-zero value as
4647 callout_data, it returns that value.
4649 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4650 there were some redundant $(CFLAGS) in commands that are now specified as
4651 $(LINK), which already includes $(CFLAGS).
4653 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4654 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4655 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4656 one-byte characters throughout. Note that case-insensitive matching applies
4657 only to characters whose values are less than 256. PCRE doesn't support the
4658 notion of cases for higher-valued characters.
4660 (i) A character class whose characters are all within 0-255 is handled as
4661 a bit map, and the map is inverted for negative classes. Previously, a
4662 character > 255 always failed to match such a class; however it should
4663 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4665 (ii) A negated character class with a single character < 255 is coded as
4666 "not this character" (OP_NOT). This wasn't working properly when the test
4667 character was multibyte, either singly or repeated.
4669 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4670 mode, for example: \x{100}{2,3}.
4672 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4673 singly or repeated) now correctly test multibyte characters. However,
4674 PCRE doesn't recognize any characters with values greater than 255 as
4675 digits, spaces, or word characters. Such characters always match \D, \S,
4676 and \W, and never match \d, \s, or \w.
4678 (v) Classes may now contain characters and character ranges with values
4679 greater than 255. For example: [ab\x{100}-\x{400}].
4681 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4682 PCRE in UTF-8 mode.
4684 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4685 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4686 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4687 value.)
4689 53. The single man page has become too large. I have therefore split it up into
4690 a number of separate man pages. These also give rise to individual HTML pages;
4691 these are now put in a separate directory, and there is an index.html page that
4692 lists them all. Some hyperlinking between the pages has been installed.
4694 54. Added convenience functions for handling named capturing parentheses.
4696 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4697 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4698 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4699 are faulted.
4701 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4702 calling configure. These values are used when compiling the dftables.c program
4703 which is run to generate the source of the default character tables. They
4704 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4705 you will need to set these values.
4707 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4710 Version 3.9 02-Jan-02
4711 ---------------------
4713 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4715 2. If --disable-static was given, the building process failed when trying to
4716 build pcretest and pcregrep. (For some reason it was using libtool to compile
4717 them, which is not right, as they aren't part of the library.)
4720 Version 3.8 18-Dec-01
4721 ---------------------
4723 1. The experimental UTF-8 code was completely screwed up. It was packing the
4724 bytes in the wrong order. How dumb can you get?
4727 Version 3.7 29-Oct-01
4728 ---------------------
4730 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4731 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4732 this didn't happen under Solaris 8, where I normally test things.
4734 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4735 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4736 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4737 this fix an hour or so after the initial 3.7 release.)
4740 Version 3.6 23-Oct-01
4741 ---------------------
4743 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4744 offsets passed as NULL with zero offset count.
4746 2. The config.guess and config.sub files had not been updated when I moved to
4747 the latest autoconf.
4750 Version 3.5 15-Aug-01
4751 ---------------------
4753 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4754 had been forgotten.
4756 2. By using declared but undefined structures, we can avoid using "void"
4757 definitions in pcre.h while keeping the internal definitions of the structures
4758 private.
4760 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4761 user point of view, this means that both static and shared libraries are built
4762 by default, but this can be individually controlled. More of the work of
4763 handling this static/shared cases is now inside libtool instead of PCRE's make
4764 file.
4766 4. The pcretest utility is now installed along with pcregrep because it is
4767 useful for users (to test regexs) and by doing this, it automatically gets
4768 relinked by libtool. The documentation has been turned into a man page, so
4769 there are now .1, .txt, and .html versions in /doc.
4771 5. Upgrades to pcregrep:
4772 (i) Added long-form option names like gnu grep.
4773 (ii) Added --help to list all options with an explanatory phrase.
4774 (iii) Added -r, --recursive to recurse into sub-directories.
4775 (iv) Added -f, --file to read patterns from a file.
4777 6. pcre_exec() was referring to its "code" argument before testing that
4778 argument for NULL (and giving an error if it was NULL).
4780 7. Upgraded Makefile.in to allow for compiling in a different directory from
4781 the source directory.
4783 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4784 options bits, the pointer it was passed was to an int instead of to an unsigned
4785 long int. This mattered only on 64-bit systems.
4787 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4788 generated) instead of pcre.in, which it its source. Also made the same change
4789 in several of the .c files.
4791 10. A new release of gcc defines printf() as a macro, which broke pcretest
4792 because it had an ifdef in the middle of a string argument for printf(). Fixed
4793 by using separate calls to printf().
4795 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4796 script, to force use of CR or LF instead of \n in the source. On non-Unix
4797 systems, the value can be set in config.h.
4799 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4800 absolute limit. Changed the text of the error message to make this clear, and
4801 likewise updated the man page.
4803 13. The limit of 99 on the number of capturing subpatterns has been removed.
4804 The new limit is 65535, which I hope will not be a "real" limit.
4807 Version 3.4 22-Aug-00
4808 ---------------------
4810 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4812 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4815 Version 3.3 01-Aug-00
4816 ---------------------
4818 1. If an octal character was given, but the value was greater than \377, it
4819 was not getting masked to the least significant bits, as documented. This could
4820 lead to crashes in some systems.
4822 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4823 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4825 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4826 These just pass their arguments on to (pcre_free)(), but they are provided
4827 because some uses of PCRE bind it to non-C systems that can call its functions,
4828 but cannot call free() or pcre_free() directly.
4830 4. Add "make test" as a synonym for "make check". Corrected some comments in
4831 the Makefile.
4833 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4834 Makefile.
4836 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4837 command called pgrep for grepping around the active processes.
4839 7. Added the beginnings of support for UTF-8 character strings.
4841 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4842 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4843 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4844 out for the ar command.)
4847 Version 3.2 12-May-00
4848 ---------------------
4850 This is purely a bug fixing release.
4852 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4853 of ZA. This was just one example of several cases that could provoke this bug,
4854 which was introduced by change 9 of version 2.00. The code for breaking
4855 infinite loops after an iteration that matches an empty string was't working
4856 correctly.
4858 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4859 when matched against abbab (for example). After matching an empty string, it
4860 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4861 caused it to match further down the string than it should.
4863 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4864 was there because it doesn't seem to be needed, and it causes trouble on some
4865 systems, as it is not a Standard C header. It has been removed.
4867 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4868 were reported on the Macintosh. The changes were from
4870 while ((c = *(++ptr)) != 0 && c != '\n');
4871 to
4872 while ((c = *(++ptr)) != 0 && c != '\n') ;
4874 Totally extraordinary, but if that's what it takes...
4876 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4877 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4878 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4879 assumes the way PCRE uses memmove() (always moving upwards).
4881 6. PCRE is being used in one environment where strchr() is not available. There
4882 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4883 faster code anyway.
4886 Version 3.1 09-Feb-00
4887 ---------------------
4889 The only change in this release is the fixing of some bugs in Makefile.in for
4890 the "install" target:
4892 (1) It was failing to install pcreposix.h.
4894 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4897 Version 3.0 01-Feb-00
4898 ---------------------
4900 1. Add support for the /+ modifier to perltest (to output $` like it does in
4901 pcretest).
4903 2. Add support for the /g modifier to perltest.
4905 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4906 matches null strings.
4908 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4909 pattern. Perl treats empty patterns specially - it reuses the most recent
4910 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4911 effect.
4913 5. The POSIX interface was broken in that it was just handing over the POSIX
4914 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4915 required a bigger vector, with some working space on the end. This means that
4916 the POSIX wrapper now has to get and free some memory, and copy the results.
4918 6. Added some simple autoconf support, placing the test data and the
4919 documentation in separate directories, re-organizing some of the
4920 information files, and making it build pcre-config (a GNU standard). Also added
4921 libtool support for building PCRE as a shared library, which is now the
4922 default.
4924 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4925 09 are not valid octal constants. Single digits will be used for minor values
4926 less than 10.
4928 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4929 existing programs that set these in the POSIX interface can use PCRE without
4930 modification.
4932 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4933 return all that pcre_info() returns, plus additional data. The pcre_info()
4934 function is retained for compatibility, but is considered to be obsolete.
4936 10. Added experimental recursion feature (?R) to handle one common case that
4937 Perl 5.6 will be able to do with (?p{...}).
4939 11. Added support for POSIX character classes like [:alpha:], which Perl is
4940 adopting.
4943 Version 2.08 31-Aug-99
4944 ----------------------
4946 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4947 trying to match at the startoffset position, but instead was moving forward to
4948 the next newline as if a previous match had failed.
4950 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4951 and could get into a loop if a null string was matched other than at the start
4952 of the subject.
4954 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4955 be distinguished at compile time, and for completeness also added PCRE_DATE.
4957 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4958 in GnuWin32 environments.
4961 Version 2.07 29-Jul-99
4962 ----------------------
4964 1. The documentation is now supplied in plain text form and HTML as well as in
4965 the form of man page sources.
4967 2. C++ compilers don't like assigning (void *) values to other pointer types.
4968 In particular this affects malloc(). Although there is no problem in Standard
4969 C, I've put in casts to keep C++ compilers happy.
4971 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4972 should be (const char *).
4974 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4975 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4976 However, I haven't made this a standard facility. The documentation doesn't
4977 mention it, and the Makefile doesn't support it.
4979 5. The Makefile now contains an "install" target, with editable destinations at
4980 the top of the file. The pcretest program is not installed.
4982 6. pgrep -V now gives the PCRE version number and date.
4984 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4985 causing the entire string to be ignored, instead of just the last character.
4987 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4988 non-matching string, it can take a very, very long time, even for strings of
4989 quite modest length, because of the nested recursion. PCRE now does better in
4990 some of these cases. It does this by remembering the last required literal
4991 character in the pattern, and pre-searching the subject to ensure it is present
4992 before running the real match. In other words, it applies a heuristic to detect
4993 some types of certain failure quickly, and in the above example, if presented
4994 with a string that has no trailing " it gives "no match" very quickly.
4996 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4997 other alternatives are tried instead.
5000 Version 2.06 09-Jun-99
5001 ----------------------
5003 1. Change pcretest's output for amount of store used to show just the code
5004 space, because the remainder (the data block) varies in size between 32-bit and
5005 64-bit systems.
5007 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5008 start matching at. This allows lookbehinds to work when searching for multiple
5009 occurrences in a string.
5011 3. Added additional options to pcretest for testing multiple occurrences:
5013 /+ outputs the rest of the string that follows a match
5014 /g loops for multiple occurrences, using the new startoffset argument
5015 /G loops for multiple occurrences by passing an incremented pointer
5017 4. PCRE wasn't doing the "first character" optimization for patterns starting
5018 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5019 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5020 the letter 'x'. On long subject strings, this gives a significant speed-up.
5023 Version 2.05 21-Apr-99
5024 ----------------------
5026 1. Changed the type of magic_number from int to long int so that it works
5027 properly on 16-bit systems.
5029 2. Fixed a bug which caused patterns starting with .* not to work correctly
5030 when the subject string contained newline characters. PCRE was assuming
5031 anchoring for such patterns in all cases, which is not correct because .* will
5032 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5033 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5034 must be retried after every newline in the subject.
5037 Version 2.04 18-Feb-99
5038 ----------------------
5040 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5041 computation of the store needed to hold the pattern was incorrect (too large).
5042 If such patterns were nested a few deep, this could multiply and become a real
5043 problem.
5045 2. Added /M option to pcretest to show the memory requirement of a specific
5046 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5048 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5049 compiled in such a way that the backtracking after subsequent failure was
5050 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5051 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5054 Version 2.03 02-Feb-99
5055 ----------------------
5057 1. Fixed typo and small mistake in man page.
5059 2. Added 4th condition (GPL supersedes if conflict) and created separate
5060 LICENCE file containing the conditions.
5062 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5063 Perl, that is the internal \ allows the delimiter to be included in the
5064 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5065 the final delimiter, add \ to the end of the pattern (to test the error).
5067 4. Added the convenience functions for extracting substrings after a successful
5068 match. Updated pcretest to make it able to test these functions.
5071 Version 2.02 14-Jan-99
5072 ----------------------
5074 1. Initialized the working variables associated with each extraction so that
5075 their saving and restoring doesn't refer to uninitialized store.
5077 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5078 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5079 fix the problem.
5081 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5082 calls, and wasn't printing the correct value for compiling calls. Increased the
5083 default value of LOOPREPEAT, and the number of significant figures in the
5084 times.
5086 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5088 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5089 a building problem on Windows NT with a FAT file system.
5092 Version 2.01 21-Oct-98
5093 ----------------------
5095 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5096 to character tables built by pcre_maketables() in the current locale. If NULL
5097 is passed, the default tables are used.
5100 Version 2.00 24-Sep-98
5101 ----------------------
5103 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5104 it any more.
5106 2. Allow quantification of (?>) groups, and make it work correctly.
5108 3. The first character computation wasn't working for (?>) groups.
5110 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5111 end of the subject) and add 5.005's \z, which really does match only at the
5112 very end of the subject.
5114 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5116 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5117 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5118 localized options. All options to pcre_study() were also removed.
5120 7. Add other new features from 5.005:
5122 $(?<= positive lookbehind
5123 $(?<! negative lookbehind
5124 (?imsx-imsx) added the unsetting capability
5125 such a setting is global if at outer level; local otherwise
5126 (?imsx-imsx:) non-capturing groups with option setting
5127 (?(cond)re|re) conditional pattern matching
5129 A backreference to itself in a repeated group matches the previous
5130 captured string.
5132 8. General tidying up of studying (both automatic and via "study")
5133 consequential on the addition of new assertions.
5135 9. As in 5.005, unlimited repeated groups that could match an empty substring
5136 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5137 runtime if any iteration does actually match an empty substring.
5139 10. Include the RunTest script in the distribution.
5141 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5142 discrepancies, some of which were old and were also with respect to 5.004. They
5143 have now been fixed.
5146 Version 1.09 28-Apr-98
5147 ----------------------
5149 1. A negated single character class followed by a quantifier with a minimum
5150 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5151 program crashes, or just wrong answers. This did not apply to negated classes
5152 containing more than one character, or to minima other than one.
5155 Version 1.08 27-Mar-98
5156 ----------------------
5158 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5160 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5161 latter must appear before anything that relies on it in the pattern.
5164 Version 1.07 16-Feb-98
5165 ----------------------
5167 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5168 repeat of a potentially empty string).
5171 Version 1.06 23-Jan-98
5172 ----------------------
5174 1. Added Markus Oberhumer's little patches for C++.
5176 2. Literal strings longer than 255 characters were broken.
5179 Version 1.05 23-Dec-97
5180 ----------------------
5182 1. Negated character classes containing more than one character were failing if
5183 PCRE_CASELESS was set at run time.
5186 Version 1.04 19-Dec-97
5187 ----------------------
5189 1. Corrected the man page, where some "const" qualifiers had been omitted.
5191 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5192 input syntax.
5194 3. Fixed memory leak which occurred when a regex with back references was
5195 matched with an offsets vector that wasn't big enough. The temporary memory
5196 that is used in this case wasn't being freed if the match failed.
5198 4. Tidied pcretest to ensure it frees memory that it gets.
5200 5. Temporary memory was being obtained in the case where the passed offsets
5201 vector was exactly big enough.
5203 6. Corrected definition of offsetof() from change 5 below.
5205 7. I had screwed up change 6 below and broken the rules for the use of
5206 setjmp(). Now fixed.
5209 Version 1.03 18-Dec-97
5210 ----------------------
5212 1. A erroneous regex with a missing opening parenthesis was correctly
5213 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5214 on some systems.
5216 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5217 it was reported that one broken compiler failed on the former because "code" is
5218 also an independent variable.
5220 3. The erroneous regex a[]b caused an array overrun reference.
5222 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5223 fail on data ending with that character. (It was going on too far, and checking
5224 the next character, typically a binary zero.) This was specific to the
5225 optimized code for single-character negative classes.
5227 5. Added a contributed patch from the TIN world which does the following:
5229 + Add an undef for memmove, in case the the system defines a macro for it.
5231 + Add a definition of offsetof(), in case there isn't one. (I don't know
5232 the reason behind this - offsetof() is part of the ANSI standard - but
5233 it does no harm).
5235 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5236 most of the places where whitespace preceded '#'. I have given up and
5237 allowed the remaining 2 cases to be at the margin.
5239 + Rename some variables in pcre to eliminate shadowing. This seems very
5240 pedantic, but does no harm, of course.
5242 6. Moved the call to setjmp() into its own function, to get rid of warnings
5243 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5245 7. Constructs such as \d{8,} were compiling into the equivalent of
5246 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5247 outcome, but in this particular case used more store than had been allocated,
5248 which caused the bug to be discovered because it threw up an internal error.
5250 8. The debugging code in both pcre and pcretest for outputting the compiled
5251 form of a regex was going wrong in the case of back references followed by
5252 curly-bracketed repeats.
5255 Version 1.02 12-Dec-97
5256 ----------------------
5258 1. Typos in pcre.3 and comments in the source fixed.
5260 2. Applied a contributed patch to get rid of places where it used to remove
5261 'const' from variables, and fixed some signed/unsigned and uninitialized
5262 variable warnings.
5264 3. Added the "runtest" target to Makefile.
5266 4. Set default compiler flag to -O2 rather than just -O.
5269 Version 1.01 19-Nov-97
5270 ----------------------
5272 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5273 like /([ab]*)*/, that is, for classes with more than one character in them.
5275 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5276 as /((?>a*))*/ (a PCRE_EXTRA facility).
5279 Version 1.00 18-Nov-97
5280 ----------------------
5282 1. Added compile-time macros to support systems such as SunOS4 which don't have
5283 memmove() or strerror() but have other things that can be used instead.
5285 2. Arranged that "make clean" removes the executables.
5288 Version 0.99 27-Oct-97
5289 ----------------------
5291 1. Fixed bug in code for optimizing classes with only one character. It was
5292 initializing a 32-byte map regardless, which could cause it to run off the end
5293 of the memory it had got.
5295 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5298 Version 0.98 22-Oct-97
5299 ----------------------
5301 1. Fixed bug in code for handling temporary memory usage when there are more
5302 back references than supplied space in the ovector. This could cause segfaults.
5305 Version 0.97 21-Oct-97
5306 ----------------------
5308 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5310 2. Optimized negated single characters not to use a bit map.
5312 3. Brought error texts together as macro definitions; clarified some of them;
5313 fixed one that was wrong - it said "range out of order" when it meant "invalid
5314 escape sequence".
5316 4. Changed some char * arguments to const char *.
5318 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5320 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5321 pcretest.
5324 Version 0.96 16-Oct-97
5325 ----------------------
5327 1. Added a simple "pgrep" utility to the distribution.
5329 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5330 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5331 where "ddd" means "one or more decimal digits".
5333 3. Fixed serious bug. If a pattern had a back reference, but the call to
5334 pcre_exec() didn't supply a large enough ovector to record the related
5335 identifying subpattern, the match always failed. PCRE now remembers the number
5336 of the largest back reference, and gets some temporary memory in which to save
5337 the offsets during matching if necessary, in order to ensure that
5338 backreferences always work.
5340 4. Increased the compatibility with Perl in a number of ways:
5342 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5343 to request this handling. The option can be set at compile or exec time.
5345 (b) $ matches before a terminating newline by default; an option
5346 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5347 mode). The option can be set at compile or exec time.
5349 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5350 the same as Perl's. If the decimal number it represents is less than 10
5351 or there aren't that many previous left capturing parentheses, an octal
5352 escape is read. Inside a character class, it's always an octal escape,
5353 even if it is a single digit.
5355 (d) An escaped but undefined alphabetic character is taken as a literal,
5356 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5357 escapes.
5359 (e) {0} is now permitted. (The previous item is removed from the compiled
5360 pattern).
5362 5. Changed all the names of code files so that the basic parts are no longer
5363 than 10 characters, and abolished the teeny "globals.c" file.
5365 6. Changed the handling of character classes; they are now done with a 32-byte
5366 bit map always.
5368 7. Added the -d and /D options to pcretest to make it possible to look at the
5369 internals of compilation without having to recompile pcre.
5372 Version 0.95 23-Sep-97
5373 ----------------------
5375 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5376 \x20 at the start of a run of normal characters. These were being treated as
5377 real characters, instead of the source characters being re-checked.
5380 Version 0.94 18-Sep-97
5381 ----------------------
5383 1. The functions are now thread-safe, with the caveat that the global variables
5384 containing pointers to malloc() and free() or alternative functions are the
5385 same for all threads.
5387 2. Get pcre_study() to generate a bitmap of initial characters for non-
5388 anchored patterns when this is possible, and use it if passed to pcre_exec().
5391 Version 0.93 15-Sep-97
5392 ----------------------
5394 1. /(b)|(:+)/ was computing an incorrect first character.
5396 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5397 but not actually doing anything yet.
5399 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5400 as Perl does (e.g. [-az] or [az-]).
5402 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5403 all possible positions.
5405 5. Split up into different modules to avoid including unneeded functions in a
5406 compiled binary. However, compile and exec are still in one module. The "study"
5407 function is split off.
5409 6. The character tables are now in a separate module whose source is generated
5410 by an auxiliary program - but can then be edited by hand if required. There are
5411 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5412 toupper() in the code.
5414 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5415 make them global. Abolish the function for setting them, as the caller can now
5416 set them directly.
5419 Version 0.92 11-Sep-97
5420 ----------------------
5422 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5423 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5425 2. Caseless matching was not working in character classes if the characters in
5426 the pattern were in upper case.
5428 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5430 4. Make PCRE_ANCHORED public and accept as a compile option.
5432 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5433 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5434 pass them.
5436 6. Give an error if bad option bits passed at compile or run time.
5438 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5439 pcretest to cause it to pass that flag.
5441 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5442 options, and the first character, if set.
5444 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5447 Version 0.91 10-Sep-97
5448 ----------------------
5450 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5451 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5453 2. PCRE was looping on encountering an indefinitely repeated back reference to
5454 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5455 Perl does - treats the match as successful.
5457 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5