ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1459 - (show annotations)
Tue Mar 4 10:45:15 2014 UTC (5 years, 11 months ago) by ph10
File size: 263558 byte(s)
Error occurred while calculating annotation data.
Preparations for next release.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.35-RC1 04-March-2014
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.
103 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
104 not take account if existing stack usage. There is now a new global
105 variable called pcre_stack_guard that can be set to point to an external
106 function to check stack availability. It is called at the start of
107 processing every parenthesized group.
110 Version 8.34 15-December-2013
111 -----------------------------
113 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
114 executable memory. Patch inspired by Carsten Klein.
116 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
117 this macro is never tested and has no effect, because the work to support
118 coverage involves only compiling and linking options and special targets in
119 the Makefile. The comment in config.h implied that defining the macro would
120 enable coverage support, which is totally false. There was also support for
121 setting this macro in the CMake files (my fault, I just copied it from
122 configure). SUPPORT_GCOV has now been removed.
124 3. Make a small performance improvement in strlen16() and strlen32() in
125 pcretest.
127 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
128 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
129 these cases). There was also one in pcretest.c.
131 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
133 6. In UTF mode, the code for checking whether a group could match an empty
134 string (which is used for indefinitely repeated groups to allow for
135 breaking an infinite loop) was broken when the group contained a repeated
136 negated single-character class with a character that occupied more than one
137 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
138 UTF-8 mode). The effect was undefined: the group might or might not be
139 deemed as matching an empty string, or the program might have crashed.
141 7. The code for checking whether a group could match an empty string was not
142 recognizing that \h, \H, \v, \V, and \R must match a character.
144 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
145 an empty string. If it can, pcretest shows this in its information output.
147 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
148 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
149 matched by pcre_exec() without using JIT:
151 (a) If the rest of the pattern did not match after a maximal run of
152 grapheme clusters, the code for backing up to try with fewer of them
153 did not always back up over a full grapheme when characters that do not
154 have the modifier quality were involved, e.g. Hangul syllables.
156 (b) If the match point in a subject started with modifier character, and
157 there was no match, the code could incorrectly back up beyond the match
158 point, and potentially beyond the first character in the subject,
159 leading to a segfault or an incorrect match result.
161 10. A conditional group with an assertion condition could lead to PCRE
162 recording an incorrect first data item for a match if no other first data
163 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
164 first data item, and therefore matched "ca" after "c" instead of at the
165 start.
167 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
168 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
170 12. The source of pcregrep now includes z/OS-specific code so that it can be
171 compiled for z/OS as part of the special z/OS distribution.
173 13. Added the -T and -TM options to pcretest.
175 14. The code in pcre_compile.c for creating the table of named capturing groups
176 has been refactored. Instead of creating the table dynamically during the
177 actual compiling pass, the information is remembered during the pre-compile
178 pass (on the stack unless there are more than 20 named groups, in which
179 case malloc() is used) and the whole table is created before the actual
180 compile happens. This has simplified the code (it is now nearly 150 lines
181 shorter) and prepared the way for better handling of references to groups
182 with duplicate names.
184 15. A back reference to a named subpattern when there is more than one of the
185 same name now checks them in the order in which they appear in the pattern.
186 The first one that is set is used for the reference. Previously only the
187 first one was inspected. This change makes PCRE more compatible with Perl.
189 16. Unicode character properties were updated from Unicode 6.3.0.
191 17. The compile-time code for auto-possessification has been refactored, based
192 on a patch by Zoltan Herczeg. It now happens after instead of during
193 compilation. The code is cleaner, and more cases are handled. The option
194 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
195 options in pcretest are provided to set it. It can also be set by
196 (*NO_AUTO_POSSESS) at the start of a pattern.
198 18. The character VT has been added to the default ("C" locale) set of
199 characters that match \s and are generally treated as white space,
200 following this same change in Perl 5.18. There is now no difference between
201 "Perl space" and "POSIX space". Whether VT is treated as white space in
202 other locales depends on the locale.
204 19. The code for checking named groups as conditions, either for being set or
205 for being recursed, has been refactored (this is related to 14 and 15
206 above). Processing unduplicated named groups should now be as fast at
207 numerical groups, and processing duplicated groups should be faster than
208 before.
210 20. Two patches to the CMake build system, by Alexander Barkov:
212 (1) Replace the "source" command by "." in CMakeLists.txt because
213 "source" is a bash-ism.
215 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
216 without these the CMake build does not work on Solaris.
218 21. Perl has changed its handling of \8 and \9. If there is no previously
219 encountered capturing group of those numbers, they are treated as the
220 literal characters 8 and 9 instead of a binary zero followed by the
221 literals. PCRE now does the same.
223 22. Following Perl, added \o{} to specify codepoints in octal, making it
224 possible to specify values greater than 0777 and also making them
225 unambiguous.
227 23. Perl now gives an error for missing closing braces after \x{... instead of
228 treating the string as literal. PCRE now does the same.
230 24. RunTest used to grumble if an inappropriate test was selected explicitly,
231 but just skip it when running all tests. This make it awkward to run ranges
232 of tests when one of them was inappropriate. Now it just skips any
233 inappropriate tests, as it always did when running all tests.
235 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
236 character types such as \d or \w, too many callouts were inserted, and the
237 data that they returned was rubbish.
239 26. In UCP mode, \s was not matching two of the characters that Perl matches,
240 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
241 were matched by \h. The code has now been refactored so that the lists of
242 the horizontal and vertical whitespace characters used for \h and \v (which
243 are defined only in one place) are now also used for \s.
245 27. Add JIT support for the 64 bit TileGX architecture.
246 Patch by Jiong Wang (Tilera Corporation).
248 28. Possessive quantifiers for classes (both explicit and automatically
249 generated) now use special opcodes instead of wrapping in ONCE brackets.
251 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
252 (because it's meaningless), this was not happening when PCRE_CASELESS was
253 set. Not wrong, but inefficient.
255 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
256 properties for \w, \d, etc) is present in a test regex. Otherwise if the
257 test contains no characters greater than 255, Perl doesn't realise it
258 should be using Unicode semantics.
260 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
261 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
262 does in Unicode mode.
264 32. Added the "forbid" facility to pcretest so that putting tests into the
265 wrong test files can sometimes be quickly detected.
267 33. There is now a limit (default 250) on the depth of nesting of parentheses.
268 This limit is imposed to control the amount of system stack used at compile
269 time. It can be changed at build time by --with-parens-nest-limit=xxx or
270 the equivalent in CMake.
272 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
273 errors. Perl warns for these when in warning mode, but PCRE has no facility
274 for giving warnings.
276 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
277 However, this was not working for (?!) because it is optimized to (*FAIL),
278 for which PCRE does not allow quantifiers. The optimization is now disabled
279 when a quantifier follows (?!). I can't see any use for this, but it makes
280 things uniform.
282 36. Perl no longer allows group names to start with digits, so I have made this
283 change also in PCRE. It simplifies the code a bit.
285 37. In extended mode, Perl ignores spaces before a + that indicates a
286 possessive quantifier. PCRE allowed a space before the quantifier, but not
287 before the possessive +. It now does.
289 38. The use of \K (reset reported match start) within a repeated possessive
290 group such as (a\Kb)*+ was not working.
292 40. Document that the same character tables must be used at compile time and
293 run time, and that the facility to pass tables to pcre_exec() and
294 pcre_dfa_exec() is for use only with saved/restored patterns.
296 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
297 features for Builds with MSVC:
299 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
300 builds). The .rc files can be used to set FileDescription and many other
301 attributes.
303 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
304 This allows higher-level build scripts which want .pdb files to avoid
305 hard-coding the exact files needed."
307 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
308 mean "start of word" and "end of word", respectively, as a transition aid.
310 43. A minimizing repeat of a class containing codepoints greater than 255 in
311 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
312 compiled to use the heap for recursion.
314 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
315 is configured.
318 Version 8.33 28-May-2013
319 ------------------------
321 1. Added 'U' to some constants that are compared to unsigned integers, to
322 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
323 variables that are added to signed variables, to ensure the result is
324 signed and can be negated.
326 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
327 CMake config files.
329 3. Revise the creation of config.h.generic so that all boolean macros are
330 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
331 overriding via -D on the command line possible.
333 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
334 to unsigned int is reported to make a quite noticeable speed difference in
335 a specific Windows environment. Testing on Linux did also appear to show
336 some benefit (and it is clearly not harmful). Also fixed the definition of
337 Xop which should be unsigned.
339 5. Related to (4), changing the definition of the intermediate variable cc
340 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
341 improvements.
343 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
344 unnecessary spaces.
346 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
347 and later.
349 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
351 9. Optimizing fast_forward_start_bits in JIT.
353 10. Adding support for callouts in JIT, and fixing some issues revealed
354 during this work. Namely:
356 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
358 (b) Minimum length was not checked before the matching is started.
360 11. The value of capture_last that is passed to callouts was incorrect in some
361 cases when there was a capture on one path that was subsequently abandoned
362 after a backtrack. Also, the capture_last value is now reset after a
363 recursion, since all captures are also reset in this case.
365 12. The interpreter no longer returns the "too many substrings" error in the
366 case when an overflowing capture is in a branch that is subsequently
367 abandoned after a backtrack.
369 13. In the pathological case when an offset vector of size 2 is used, pcretest
370 now prints out the matched string after a yield of 0 or 1.
372 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
373 Only supported by the JIT compiler at the moment.
375 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
377 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
378 offset of the starting point of the matching process, provided the offsets
379 vector is large enough.
381 17. The \A escape now records a lookbehind value of 1, though its execution
382 does not actually inspect the previous character. This is to ensure that,
383 in partial multi-segment matching, at least one character from the old
384 segment is retained when a new segment is processed. Otherwise, if there
385 are no lookbehinds in the pattern, \A might match incorrectly at the start
386 of a new segment.
388 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
390 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
391 gives some modest performance improvement in 8-bit mode.
393 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
394 be expressed in certain programming languages using Universal Character
395 Names.
397 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
398 which points out that "non characters" are not "characters that may not
399 appear in Unicode strings" but rather "characters that are reserved for
400 internal use and have only local meaning".
402 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
403 there was a conditional group that depended on an assertion, if the
404 assertion was false, the callout that immediately followed the alternation
405 in the condition was skipped when pcre_exec() was used for matching.
407 23. Allow an explicit callout to be inserted before an assertion that is the
408 condition for a conditional group, for compatibility with automatic
409 callouts, which always insert a callout at this point.
411 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
412 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
414 25. (*PRUNE) is now supported by the JIT compiler.
416 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
418 27. Fix the case where there are two or more SKIPs with arguments that may be
419 ignored.
421 28. (*SKIP) is now supported by the JIT compiler.
423 29. (*THEN) is now supported by the JIT compiler.
425 30. Update RunTest with additional test selector options.
427 31. The way PCRE handles backtracking verbs has been changed in two ways.
429 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
430 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
431 backtracking. In some cases this makes it more Perl-compatible, but Perl's
432 rather obscure rules do not always do the same thing.
434 (2) Previously, backtracking verbs were confined within assertions. This is
435 no longer the case for positive assertions, except for (*ACCEPT). Again,
436 this sometimes improves Perl compatibility, and sometimes does not.
438 32. A number of tests that were in test 2 because Perl did things differently
439 have been moved to test 1, because either Perl or PCRE has changed, and
440 these tests are now compatible.
442 32. Backtracking control verbs are now handled in the same way in JIT and
443 interpreter.
445 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
446 contained a forward subroutine reference caused a compile error.
448 34. Auto-detect and optimize limited repetitions in JIT.
450 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
451 blocking (*UTF) etc.
453 36. In the interpreter, maximizing pattern repetitions for characters and
454 character types now use tail recursion, which reduces stack usage.
456 37. The value of the max lookbehind was not correctly preserved if a compiled
457 and saved regex was reloaded on a host of different endianness.
459 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
460 of the compiled pattern block, expand the flags field from 16 to 32 bits
461 because it was almost full.
463 39. Try madvise first before posix_madvise.
465 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
466 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
467 original change locked out all matches of empty strings. This has been
468 changed so that one match of an empty string per line is recognized.
469 Subsequent searches on the same line (for colouring or for --only-matching,
470 for example) do not recognize empty strings.
472 41. Applied a user patch to fix a number of spelling mistakes in comments.
474 42. Data lines longer than 65536 caused pcretest to crash.
476 43. Clarified the data type for length and startoffset arguments for pcre_exec
477 and pcre_dfa_exec in the function-specific man pages, where they were
478 explicitly stated to be in bytes, never having been updated. I also added
479 some clarification to the pcreapi man page.
481 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
482 a segmentation fault.
485 Version 8.32 30-November-2012
486 -----------------------------
488 1. Improved JIT compiler optimizations for first character search and single
489 character iterators.
491 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
492 Patch by Daniel Richard G.
494 3. Single character iterator optimizations in the JIT compiler.
496 4. Improved JIT compiler optimizations for character ranges.
498 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
499 Reported by Giuseppe D'Angelo.
501 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
502 of a line, was being set incorrectly in cases where .* appeared inside
503 atomic brackets at the start of a pattern, or where there was a subsequent
504 *PRUNE or *SKIP.
506 7. Improved instruction cache flush for POWER/PowerPC.
507 Patch by Daniel Richard G.
509 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
510 grep:
512 (a) There is now no limit to the number of patterns to be matched.
514 (b) An error is given if a pattern is too long.
516 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
517 are now supported.
519 (d) --exclude-from and --include-from (multiple use) have been added.
521 (e) Exclusions and inclusions now apply to all files and directories, not
522 just to those obtained from scanning a directory recursively.
524 (f) Multiple uses of -f and --file-list are now supported.
526 (g) In a Windows environment, the default for -d has been changed from
527 "read" (the GNU grep default) to "skip", because otherwise the presence
528 of a directory in the file list provokes an error.
530 (h) The documentation has been revised and clarified in places.
532 9. Improve the matching speed of capturing brackets.
534 10. Changed the meaning of \X so that it now matches a Unicode extended
535 grapheme cluster.
537 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
538 out POSIX threads when JIT support is configured.
540 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
542 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
543 structure could go wrong in environments where size_t is not the same size
544 as int.
546 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
547 set.
549 15. The EBCDIC support had decayed; later updates to the code had included
550 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
551 general tidy up of EBCDIC-related issues, and the documentation was also
552 not quite right. There is now a test that can be run on ASCII systems to
553 check some of the EBCDIC-related things (but is it not a full test).
555 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
556 in a small tidy to the code.
558 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
560 18. If the --only-matching (-o) option in pcregrep is specified multiple
561 times, each one causes appropriate output. For example, -o1 -o2 outputs the
562 substrings matched by the 1st and 2nd capturing parentheses. A separating
563 string can be specified by --om-separator (default empty).
565 19. Improving the first n character searches.
567 20. Turn case lists for horizontal and vertical white space into macros so that
568 they are defined only once.
570 21. This set of changes together give more compatible Unicode case-folding
571 behaviour for characters that have more than one other case when UCP
572 support is available.
574 (a) The Unicode property table now has offsets into a new table of sets of
575 three or more characters that are case-equivalent. The MultiStage2.py
576 script that generates these tables (the pcre_ucd.c file) now scans
577 CaseFolding.txt instead of UnicodeData.txt for character case
578 information.
580 (b) The code for adding characters or ranges of characters to a character
581 class has been abstracted into a generalized function that also handles
582 case-independence. In UTF-mode with UCP support, this uses the new data
583 to handle characters with more than one other case.
585 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
586 whose other case is greater than 256 are now correctly matched
587 caselessly. Previously, the high codepoint matched the low one, but not
588 vice versa.
590 (d) The processing of \h, \H, \v, and \ in character classes now makes use
591 of the new class addition function, using character lists defined as
592 macros alongside the case definitions of 20 above.
594 (e) Caseless back references now work with characters that have more than
595 one other case.
597 (f) General caseless matching of characters with more than one other case
598 is supported.
600 22. Unicode character properties were updated from Unicode 6.2.0
602 23. Improved CMake support under Windows. Patch by Daniel Richard G.
604 24. Add support for 32-bit character strings, and UTF-32
606 25. Major JIT compiler update (code refactoring and bugfixing).
607 Experimental Sparc 32 support is added.
609 26. Applied a modified version of Daniel Richard G's patch to create
610 pcre.h.generic and config.h.generic by "make" instead of in the
611 PrepareRelease script.
613 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
614 pcre_compile.c when checking for a zero character.
616 28. Introducing a native interface for JIT. Through this interface, the compiled
617 machine code can be directly executed. The purpose of this interface is to
618 provide fast pattern matching, so several sanity checks are not performed.
619 However, feature tests are still performed. The new interface provides
620 1.4x speedup compared to the old one.
622 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
623 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
624 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
626 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
627 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
628 the "old" RFC 2279). Instead, it ended up passing a negative length to
629 pcre_exec().
631 31. Add support for GCC's visibility feature to hide internal functions.
633 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
634 "unknown -C option" after outputting 0 or 1.
636 33. There is now support for generating a code coverage report for the test
637 suite in environments where gcc is the compiler and lcov is installed. This
638 is mainly for the benefit of the developers.
640 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
641 unaddressable using valgrind annotations, allowing valgrind to detect
642 invalid memory accesses. This is mainly for the benefit of the developers.
644 25. (*UTF) can now be used to start a pattern in any of the three libraries.
646 26. Give configure error if --enable-cpp but no C++ compiler found.
649 Version 8.31 06-July-2012
650 -------------------------
652 1. Fixing a wrong JIT test case and some compiler warnings.
654 2. Removed a bashism from the RunTest script.
656 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
657 to unsigned type, result still unsigned" that was given by an MS compiler
658 on encountering the code "-sizeof(xxx)".
660 4. Partial matching support is added to the JIT compiler.
662 5. Fixed several bugs concerned with partial matching of items that consist
663 of more than one character:
665 (a) /^(..)\1/ did not partially match "aba" because checking references was
666 done on an "all or nothing" basis. This also applied to repeated
667 references.
669 (b) \R did not give a hard partial match if \r was found at the end of the
670 subject.
672 (c) \X did not give a hard partial match after matching one or more
673 characters at the end of the subject.
675 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
676 a partial match for the string "\r".
678 (e) When newline was set to CRLF, the metacharacter "." did not recognize
679 a partial match for a CR character at the end of the subject string.
681 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
682 running pcretest, the text "(JIT)" added to the output whenever JIT is
683 actually used to run the match.
685 7. Individual JIT compile options can be set in pcretest by following -s+[+]
686 or /S+[+] with a digit between 1 and 7.
688 8. OP_NOT now supports any UTF character not just single-byte ones.
690 9. (*MARK) control verb is now supported by the JIT compiler.
692 10. The command "./RunTest list" lists the available tests without actually
693 running any of them. (Because I keep forgetting what they all are.)
697 12. Applied a (slightly modified) user-supplied patch that improves performance
698 when the heap is used for recursion (compiled with --disable-stack-for-
699 recursion). Instead of malloc and free for each heap frame each time a
700 logical recursion happens, frames are retained on a chain and re-used where
701 possible. This sometimes gives as much as 30% improvement.
703 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
704 call.
706 14. As documented, (*COMMIT) is now confined to within a positive assertion.
708 15. It is now possible to link pcretest with libedit as an alternative to
709 libreadline.
711 16. (*COMMIT) control verb is now supported by the JIT compiler.
713 17. The Unicode data tables have been updated to Unicode 6.1.0.
715 18. Added --file-list option to pcregrep.
717 19. Added binary file support to pcregrep, including the -a, --binary-files,
718 -I, and --text options.
720 20. The madvise function is renamed for posix_madvise for QNX compatibility
721 reasons. Fixed by Giuseppe D'Angelo.
723 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
725 22. Changed the option for creating symbolic links for 16-bit man pages from
726 -s to -sf so that re-installing does not cause issues.
728 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
730 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
731 matches in certain environments (the workspace was not being correctly
732 retained). Also added to pcre_dfa_exec() a simple plausibility check on
733 some of the workspace data at the beginning of a restart.
735 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
736 was not doing so when it should - probably a typo introduced by SVN 528
737 (change 8.10/14).
739 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
740 \w+ when the character tables indicated that \x{c4} was a word character.
741 There were several related cases, all because the tests for doing a table
742 lookup were testing for characters less than 127 instead of 255.
744 27. If a pattern contains capturing parentheses that are not used in a match,
745 their slots in the ovector are set to -1. For those that are higher than
746 any matched groups, this happens at the end of processing. In the case when
747 there were back references that the ovector was too small to contain
748 (causing temporary malloc'd memory to be used during matching), and the
749 highest capturing number was not used, memory off the end of the ovector
750 was incorrectly being set to -1. (It was using the size of the temporary
751 memory instead of the true size.)
753 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
754 ovector size, it uses memory at the end of the block that it has got.
756 29. Check for an overlong MARK name and give an error at compile time. The
757 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
759 30. JIT compiler update.
761 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
762 Rill for the patch.
764 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
766 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
768 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
769 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
770 was enabled.
772 35. Improve JIT code generation for greedy plus quantifier.
774 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
775 1 to "aa" instead of to an empty string. The bug affected repeated groups
776 that could potentially match an empty string.
778 37. Optimizing single character iterators in JIT.
780 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
781 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
782 codepoints that are too big for the mode are faulted, and in a UTF mode,
783 disallowed codepoints are also faulted.
785 39. If PCRE was compiled with UTF support, in three places in the DFA
786 matcher there was code that should only have been obeyed in UTF mode, but
787 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
788 processing when bytes with values greater than 127 were present. In 16-bit
789 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
790 both cases the values are those that cannot be the first data item in a UTF
791 character. The three items that might have provoked this were recursions,
792 possessively repeated groups, and atomic groups.
794 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
795 and pcregrep, because some OS require shared objects to be explicitly
796 passed to ld, causing the link step to fail if they are not.
798 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
799 mode, patterns that started with \h* or \R* might be incorrectly matched.
802 Version 8.30 04-February-2012
803 -----------------------------
805 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
806 name is defined in ctype.h.
808 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
809 only in quite long subpatterns.
811 3. Removed the function pcre_info(), which has been obsolete and deprecated
812 since it was replaced by pcre_fullinfo() in February 2000.
814 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
815 match a (*MARK), and the match failed at the start of the subject, a
816 reference to memory before the start of the subject could occur. This bug
817 was introduced by fix 17 of release 8.21.
819 5. A reference to an unset group with zero minimum repetition was giving
820 totally wrong answers (in non-JavaScript-compatibility mode). For example,
821 /(another)?(\1?)test/ matched against "hello world test". This bug was
822 introduced in release 8.13.
824 6. Add support for 16-bit character strings (a large amount of work involving
825 many changes and refactorings).
827 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
828 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
829 from a file.
831 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
832 rounding is not applied in this particular case).
834 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
835 if they appear, or are escaped, in patterns.
837 10. Get rid of a number of -Wunused-but-set-variable warnings.
839 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
840 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
841 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
842 also returns the mark "x". This bug applied to capturing parentheses,
843 non-capturing parentheses, and atomic parentheses. It also applied to some
844 assertions.
846 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
847 information out of configure.ac instead of relying on pcre.h.generic, which
848 is not stored in the repository.
850 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
851 -lreadline.
853 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
855 15. Applied Graycode's patch to put the top-level frame on the stack rather
856 than the heap when not using the stack for recursion. This gives a
857 performance improvement in many cases when recursion is not deep.
859 16. Experimental code added to "pcretest -C" to output the stack frame size.
862 Version 8.21 12-Dec-2011
863 ------------------------
865 1. Updating the JIT compiler.
867 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
868 are added as well.
870 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
871 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
872 calling _pcre_jit_exec. Some extra comments are added.
874 4. (*MARK) settings inside atomic groups that do not contain any capturing
875 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
876 was introduced by change 18 for 8.20.
878 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
879 ECMA-262 standard.
881 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
882 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
883 This bug was probably introduced by change 9 of 8.13.
885 7. While fixing 6 above, I noticed that a number of other items were being
886 incorrectly rejected as "not fixed length". This arose partly because newer
887 opcodes had not been added to the fixed-length checking code. I have (a)
888 corrected the bug and added tests for these items, and (b) arranged for an
889 error to occur if an unknown opcode is encountered while checking for fixed
890 length instead of just assuming "not fixed length". The items that were
891 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
892 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
893 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
895 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
896 being incorrectly compiled and would have given unpredicatble results.
898 9. A possessively repeated subpattern with minimum repeat count greater than
899 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
900 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
901 the first (A) could occur when it should not.
903 10. Add a cast and remove a redundant test from the code.
905 11. JIT should use pcre_malloc/pcre_free for allocation.
907 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
908 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
909 is anything other than /usr, -L is still shown).
911 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
913 14. Perl does not support \N without a following name in a [] class; PCRE now
914 also gives an error.
916 15. If a forward reference was repeated with an upper limit of around 2000,
917 it caused the error "internal error: overran compiling workspace". The
918 maximum number of forward references (including repeats) was limited by the
919 internal workspace, and dependent on the LINK_SIZE. The code has been
920 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
921 the default depends on LINK_SIZE. There is a new upper limit (for safety)
922 of around 200,000 forward references. While doing this, I also speeded up
923 the filling in of repeated forward references.
925 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
926 incorrectly expecting the subject to contain another "a" after the start.
928 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
929 in the match, the SKIP should be ignored. This was not happening; instead
930 the SKIP was being treated as NOMATCH. For patterns such as
931 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
932 tested.
934 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
935 now much more compatible with Perl, in particular in cases where the result
936 is a non-match for a non-anchored pattern. For example, if
937 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
938 "m", where previously it did not return a name. A side effect of this
939 change is that for partial matches, the last encountered mark name is
940 returned, as for non matches. A number of tests that were previously not
941 Perl-compatible have been moved into the Perl-compatible test files. The
942 refactoring has had the pleasing side effect of removing one argument from
943 the match() function, thus reducing its stack requirements.
945 19. If the /S+ option was used in pcretest to study a pattern using JIT,
946 subsequent uses of /S (without +) incorrectly behaved like /S+.
948 21. Retrieve executable code size support for the JIT compiler and fixing
949 some warnings.
951 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
952 not work when the shorter character appeared right at the end of the
953 subject string.
955 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
956 systems.
958 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
959 output it when the /M option is used in pcretest.
961 25. The CheckMan script was not being included in the distribution. Also, added
962 an explicit "perl" to run Perl scripts from the PrepareRelease script
963 because this is reportedly needed in Windows.
965 26. If study data was being save in a file and studying had not found a set of
966 "starts with" bytes for the pattern, the data written to the file (though
967 never used) was taken from uninitialized memory and so caused valgrind to
968 complain.
970 27. Updated RunTest.bat as provided by Sheri Pierce.
972 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
974 29. Computation of memory usage for the table of capturing group names was
975 giving an unnecessarily large value.
978 Version 8.20 21-Oct-2011
979 ------------------------
981 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
982 a POSIX class. After further experiments with Perl, which convinced me that
983 Perl has bugs and confusions, a closing square bracket is no longer allowed
984 in a POSIX name. This bug also affected patterns with classes that started
985 with full stops.
987 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
988 captured substring, but while checking the failing first alternative,
989 substring 1 is temporarily captured. If the output vector supplied to
990 pcre_exec() was not big enough for this capture, the yield of the function
991 was still zero ("insufficient space for captured substrings"). This cannot
992 be totally fixed without adding another stack variable, which seems a lot
993 of expense for a edge case. However, I have improved the situation in cases
994 such as /(a)(b)x|abc/ matched against "abc", where the return code
995 indicates that fewer than the maximum number of slots in the ovector have
996 been set.
998 3. Related to (2) above: when there are more back references in a pattern than
999 slots in the output vector, pcre_exec() uses temporary memory during
1000 matching, and copies in the captures as far as possible afterwards. It was
1001 using the entire output vector, but this conflicts with the specification
1002 that only 2/3 is used for passing back captured substrings. Now it uses
1003 only the first 2/3, for compatibility. This is, of course, another edge
1004 case.
1006 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1007 main code base, and can be used by building with --enable-jit. When this is
1008 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1009 runtime --no-jit option is given.
1011 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1012 ovector, the return from the function was zero, implying that there were
1013 other matches that did not fit. The correct "exactly full" value is now
1014 returned.
1016 6. If a subpattern that was called recursively or as a subroutine contained
1017 (*PRUNE) or any other control that caused it to give a non-standard return,
1018 invalid errors such as "Error -26 (nested recursion at the same subject
1019 position)" or even infinite loops could occur.
1021 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1022 computing the minimum length on reaching *ACCEPT, and so ended up with the
1023 wrong value of 1 rather than 0. Further investigation indicates that
1024 computing a minimum subject length in the presence of *ACCEPT is difficult
1025 (think back references, subroutine calls), and so I have changed the code
1026 so that no minimum is registered for a pattern that contains *ACCEPT.
1028 8. If (*THEN) was present in the first (true) branch of a conditional group,
1029 it was not handled as intended. [But see 16 below.]
1031 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1032 Sheri Pierce.
1034 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1035 the first byte in a match must be "a".
1037 11. Change 17 for 8.13 increased the recursion depth for patterns like
1038 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1039 pattern contains any instances of (*THEN). If it does not, the old
1040 optimizations are restored. It would be nice to do this on a per-group
1041 basis, but at the moment that is not feasible.
1043 12. In some environments, the output of pcretest -C is CRLF terminated. This
1044 broke RunTest's code that checks for the link size. A single white space
1045 character after the value is now allowed for.
1047 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1048 For "fr", it uses the Windows-specific input and output files.
1050 14. If (*THEN) appeared in a group that was called recursively or as a
1051 subroutine, it did not work as intended. [But see next item.]
1053 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1054 pattern fragments (but not containing any | characters). If A and B are
1055 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1056 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1057 D. In other words, Perl considers parentheses that do not contain any |
1058 characters to be part of a surrounding alternative, whereas PCRE was
1059 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1060 differently. PCRE now behaves in the same way as Perl, except in the case
1061 of subroutine/recursion calls such as (?1) which have in any case always
1062 been different (but PCRE had them first :-).
1064 16. Related to 15 above: Perl does not treat the | in a conditional group as
1065 creating alternatives. Such a group is treated in the same way as an
1066 ordinary group without any | characters when processing (*THEN). PCRE has
1067 been changed to match Perl's behaviour.
1069 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1070 RunGrepTest script failed.
1072 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1073 inevitable for groups that contain captures, but it can lead to a lot of
1074 stack use in large patterns. The old behaviour has been restored for atomic
1075 groups that do not contain any capturing parentheses.
1077 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1078 suppress the check for a minimum subject length at run time. (If it was
1079 given to pcre_exec() or pcre_dfa_exec() it did work.)
1081 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1082 fail to work when decoding hex characters in data strings in EBCDIC
1083 environments.
1085 21. It appears that in at least one Mac OS environment, the isxdigit() function
1086 is implemented as a macro that evaluates to its argument more than once,
1087 contravening the C 90 Standard (I haven't checked a later standard). There
1088 was an instance in pcretest which caused it to go wrong when processing
1089 \x{...} escapes in subject strings. The has been rewritten to avoid using
1090 things like p++ in the argument of isxdigit().
1093 Version 8.13 16-Aug-2011
1094 ------------------------
1096 1. The Unicode data tables have been updated to Unicode 6.0.0.
1098 2. Two minor typos in pcre_internal.h have been fixed.
1100 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1101 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1102 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1104 4. There were a number of related bugs in the code for matching backrefences
1105 caselessly in UTF-8 mode when codes for the characters concerned were
1106 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1107 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1108 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1109 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1110 2-byte code at the end of the subject (it thought there wasn't enough data
1111 left).
1113 5. Comprehensive information about what went wrong is now returned by
1114 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1115 as the output vector has at least 2 elements. The offset of the start of
1116 the failing character and a reason code are placed in the vector.
1118 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1119 now returned is for the first byte of the failing character, instead of the
1120 last byte inspected. This is an incompatible change, but I hope it is small
1121 enough not to be a problem. It makes the returned offset consistent with
1122 pcre_exec() and pcre_dfa_exec().
1124 7. pcretest now gives a text phrase as well as the error number when
1125 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1126 failure, the offset and reason code are output.
1128 8. When \R was used with a maximizing quantifier it failed to skip backwards
1129 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1130 back over a single character (\n). This seems wrong (because it treated the
1131 two characters as a single entity when going forwards), conflicts with the
1132 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1133 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1134 has been changed.
1136 9. Some internal refactoring has changed the processing so that the handling
1137 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1138 time (the PCRE_DOTALL option was changed this way some time ago: version
1139 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1140 which was always a bit of a fudge. It also means that there is one less
1141 argument for the match() function, which reduces its stack requirements
1142 slightly. This change also fixes an incompatibility with Perl: the pattern
1143 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1145 10. More internal refactoring has drastically reduced the number of recursive
1146 calls to match() for possessively repeated groups such as (abc)++ when
1147 using pcre_exec().
1149 11. While implementing 10, a number of bugs in the handling of groups were
1150 discovered and fixed:
1152 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1153 (a|)*(?1) gave a compile-time internal error.
1154 ((a|)+)+ did not notice that the outer group could match an empty string.
1155 (^a|^)+ was not marked as anchored.
1156 (.*a|.*)+ was not marked as matching at start or after a newline.
1158 12. Yet more internal refactoring has removed another argument from the match()
1159 function. Special calls to this function are now indicated by setting a
1160 value in a variable in the "match data" data block.
1162 13. Be more explicit in pcre_study() instead of relying on "default" for
1163 opcodes that mean there is no starting character; this means that when new
1164 ones are added and accidentally left out of pcre_study(), testing should
1165 pick them up.
1167 14. The -s option of pcretest has been documented for ages as being an old
1168 synonym of -m (show memory usage). I have changed it to mean "force study
1169 for every regex", that is, assume /S for every regex. This is similar to -i
1170 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1171 using it. It makes it easier to run collections of tests with and without
1172 study enabled, and thereby test pcre_study() more easily. All the standard
1173 tests are now run with and without -s (but some patterns can be marked as
1174 "never study" - see 20 below).
1176 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1177 restoration of the capturing data to the outer values was not happening
1178 correctly.
1180 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1181 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1182 pattern had matched an empty string, and so incorrectly returned a no
1183 match.
1185 17. There was optimizing code for the last branch of non-capturing parentheses,
1186 and also for the obeyed branch of a conditional subexpression, which used
1187 tail recursion to cut down on stack usage. Unfortunately, now that there is
1188 the possibility of (*THEN) occurring in these branches, tail recursion is
1189 no longer possible because the return has to be checked for (*THEN). These
1190 two optimizations have therefore been removed. [But see 8.20/11 above.]
1192 18. If a pattern containing \R was studied, it was assumed that \R always
1193 matched two bytes, thus causing the minimum subject length to be
1194 incorrectly computed because \R can also match just one byte.
1196 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1197 was incorrectly computed.
1199 20. If /S is present twice on a test pattern in pcretest input, it now
1200 *disables* studying, thereby overriding the use of -s on the command line
1201 (see 14 above). This is necessary for one or two tests to keep the output
1202 identical in both cases.
1204 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1205 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1207 22. When an atomic group that contained a capturing parenthesis was
1208 successfully matched, but the branch in which it appeared failed, the
1209 capturing was not being forgotten if a higher numbered group was later
1210 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1211 group 1 to "a", when in fact it should be unset. This applied to multi-
1212 branched capturing and non-capturing groups, repeated or not, and also to
1213 positive assertions (capturing in negative assertions does not happen
1214 in PCRE) and also to nested atomic groups.
1216 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1217 subject after a captured substring, to make it easier to tell which of a
1218 number of identical substrings has been captured.
1220 24. The way atomic groups are processed by pcre_exec() has been changed so that
1221 if they are repeated, backtracking one repetition now resets captured
1222 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1223 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1224 "aaa". Previously, it would have been "a". As part of this code
1225 refactoring, the way recursive calls are handled has also been changed.
1227 25. If an assertion condition captured any substrings, they were not passed
1228 back unless some other capturing happened later. For example, if
1229 (?(?=(a))a) was matched against "a", no capturing was returned.
1231 26. When studying a pattern that contained subroutine calls or assertions,
1232 the code for finding the minimum length of a possible match was handling
1233 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1234 group 1 called group 2 while simultaneously a separate group 2 called group
1235 1). A stack overflow occurred in this case. I have fixed this by limiting
1236 the recursion depth to 10.
1238 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1239 Fortmann. This supports explicit test numbers on the command line, and has
1240 argument validation and error reporting.
1242 28. An instance of \X with an unlimited repeat could fail if at any point the
1243 first character it looked at was a mark character.
1245 29. Some minor code refactoring concerning Unicode properties and scripts
1246 should reduce the stack requirement of match() slightly.
1248 30. Added the '=' option to pcretest to check the setting of unused capturing
1249 slots at the end of the pattern, which are documented as being -1, but are
1250 not included in the return count.
1252 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1253 compiled something random. Now it gives a compile-time error (as does
1254 Perl).
1256 32. A *MARK encountered during the processing of a positive assertion is now
1257 recorded and passed back (compatible with Perl).
1259 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1260 had alternative anchored branches, the search for a second match in a line
1261 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1262 matched the line "0102" twice. The same bug affected patterns that started
1263 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1264 twice.
1266 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1267 does, and because of capturing effects, quantifying parenthesized
1268 assertions may at times be useful. Quantifiers are now allowed for
1269 parenthesized assertions.
1271 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1273 36. \g was being checked for fancy things in a character class, when it should
1274 just be a literal "g".
1276 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1277 appearance of a nested POSIX class supersedes an apparent external class.
1278 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1279 unescaped square brackets may also appear as part of class names. For
1280 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1281 more like Perl. (But see 8.20/1 above.)
1283 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1284 was because it thought it was \N{name}, which is not supported).
1286 39. Add minix to OS list not supporting the -S option in pcretest.
1288 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1289 cannot analyze patterns in sufficient detail to catch mutual recursions
1290 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1291 subgroup is called recursively as a subpattern for a second time at the
1292 same position in the subject string. In previous releases this might have
1293 been caught by the recursion limit, or it might have run out of stack.
1295 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1296 happen only once. PCRE was, however incorrectly giving a compile time error
1297 "recursive call could loop indefinitely" because it cannot analyze the
1298 pattern in sufficient detail. The compile time test no longer happens when
1299 PCRE is compiling a conditional subpattern, but actual runaway loops are
1300 now caught at runtime (see 40 above).
1302 42. It seems that Perl allows any characters other than a closing parenthesis
1303 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1304 has been changed to be the same.
1306 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1307 as not to get warnings when autogen.sh is called. Also changed
1308 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1310 44. To help people who use pcregrep to scan files containing exceedingly long
1311 lines, the following changes have been made:
1313 (a) The default value of the buffer size parameter has been increased from
1314 8K to 20K. (The actual buffer used is three times this size.)
1316 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1317 PCRE is built.
1319 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1320 to be set at run time.
1322 (d) Numerical values in pcregrep options can be followed by K or M, for
1323 example --buffer-size=50K.
1325 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1326 given and the return code is set to 2.
1328 45. Add a pointer to the latest mark to the callout data block.
1330 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1331 partial match of an empty string instead of no match. This was specific to
1332 the use of ".".
1334 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1335 complete match instead of a partial match. This bug was dependent on both
1336 the PCRE_UTF8 and PCRE_DOTALL options being set.
1338 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1339 starting byte set, because \b was not being ignored.
1342 Version 8.12 15-Jan-2011
1343 ------------------------
1345 1. Fixed some typos in the markup of the man pages, and wrote a script that
1346 checks for such things as part of the documentation building process.
1348 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1349 --match-limit and --recursion-limit options (added for 8.11). In
1350 particular, this made one of the standard tests fail. (The integer value
1351 went into the wrong half of a long int.)
1353 3. If the --colour option was given to pcregrep with -v (invert match), it
1354 did strange things, either producing crazy output, or crashing. It should,
1355 of course, ignore a request for colour when reporting lines that do not
1356 match.
1358 4. Another pcregrep bug caused similar problems if --colour was specified with
1359 -M (multiline) and the pattern match finished with a line ending.
1361 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1362 matched in multiline mode, the following line was shown as part of the
1363 match. This seems wrong, so I have changed it.
1365 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1366 the check for further matches in the same line (so they could be coloured)
1367 to overrun the end of the current line. If another match was found, it was
1368 incorrectly shown (and then shown again when found in the next line).
1370 7. If pcregrep was compiled under Windows, there was a reference to the
1371 function pcregrep_exit() before it was defined. I am assuming this was
1372 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1373 reported by a user. I've moved the definition above the reference.
1376 Version 8.11 10-Dec-2010
1377 ------------------------
1379 1. (*THEN) was not working properly if there were untried alternatives prior
1380 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1381 backtracked to try for "b" instead of moving to the next alternative branch
1382 at the same level (in this case, to look for "c"). The Perl documentation
1383 is clear that when (*THEN) is backtracked onto, it goes to the "next
1384 alternative in the innermost enclosing group".
1386 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1387 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1388 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1389 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1390 (*THEN).
1392 3. If \s appeared in a character class, it removed the VT character from
1393 the class, even if it had been included by some previous item, for example
1394 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1395 of \s, but is part of the POSIX "space" class.)
1397 4. A partial match never returns an empty string (because you can always
1398 match an empty string at the end of the subject); however the checking for
1399 an empty string was starting at the "start of match" point. This has been
1400 changed to the "earliest inspected character" point, because the returned
1401 data for a partial match starts at this character. This means that, for
1402 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1403 (previously it gave "no match").
1405 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1406 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1407 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1408 has an implication that the given string is incomplete (because a partial
1409 match is preferred over a full match). For this reason, these items now
1410 give a partial match in this situation. [Aside: previously, the one case
1411 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1412 match rather than a full match, which was wrong by the old rules, but is
1413 now correct.]
1415 6. There was a bug in the handling of #-introduced comments, recognized when
1416 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1417 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1418 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1419 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1420 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1421 places in pcre_compile().
1423 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1424 comments when looking ahead for named forward references to subpatterns,
1425 the only newline sequence it recognized was NL. It now handles newlines
1426 according to the set newline convention.
1428 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1429 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1430 cater for a lack of strerror(). These oversights have been fixed.
1432 9. Added --match-limit and --recursion-limit to pcregrep.
1434 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1436 11. When the -o option was used, pcregrep was setting a return code of 1, even
1437 when matches were found, and --line-buffered was not being honoured.
1439 12. Added an optional parentheses number to the -o and --only-matching options
1440 of pcregrep.
1442 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1443 can match an empty string. The code to do it in pcretest and pcredemo
1444 needed fixing:
1446 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1447 only one byte after an empty string match just before CRLF (this case
1448 just got forgotten; "any" and "anycrlf" were OK).
1450 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1451 mode when an empty string match preceded an ASCII character followed by
1452 a non-ASCII character. (The code for advancing by one character rather
1453 than one byte was nonsense.)
1455 (c) The pcredemo.c sample program did not have any code at all to handle
1456 the cases when CRLF is a valid newline sequence.
1458 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1459 as a starting offset was within the subject string. There is now a new
1460 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1461 negative or greater than the length of the string. In order to test this,
1462 pcretest is extended to allow the setting of negative starting offsets.
1464 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1465 starting offset points to the beginning of a UTF-8 character was
1466 unnecessarily clumsy. I tidied it up.
1468 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1469 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1471 17. Nobody had reported that the --include_dir option, which was added in
1472 release 7.7 should have been called --include-dir (hyphen, not underscore)
1473 for compatibility with GNU grep. I have changed it to --include-dir, but
1474 left --include_dir as an undocumented synonym, and the same for
1475 --exclude-dir, though that is not available in GNU grep, at least as of
1476 release 2.5.4.
1478 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1479 characters from a string of bytes) have been redefined so as not to use
1480 loops, in order to improve performance in some environments. At the same
1481 time, I abstracted some of the common code into auxiliary macros to save
1482 repetition (this should not affect the compiled code).
1484 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1485 compile-time error is now given if \c is not followed by an ASCII
1486 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1487 different, and any byte value is allowed.)
1489 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1490 START_OPTIMIZE option, which is now allowed at compile time - but just
1491 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1492 to pcregrep and other applications that have no direct access to PCRE
1493 options. The new /Y option in pcretest sets this option when calling
1494 pcre_compile().
1496 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1497 back references. Groups containing recursive back references were forced to
1498 be atomic by that change, but in the case of named groups, the amount of
1499 memory required was incorrectly computed, leading to "Failed: internal
1500 error: code overflow". This has been fixed.
1502 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1503 pcretest.c, to avoid build problems in some Borland environments.
1506 Version 8.10 25-Jun-2010
1507 ------------------------
1509 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1510 THEN.
1512 2. (*ACCEPT) was not working when inside an atomic group.
1514 3. Inside a character class, \B is treated as a literal by default, but
1515 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1516 causes the error). The code is unchanged, but I tidied the documentation.
1518 4. Inside a character class, PCRE always treated \R and \X as literals,
1519 whereas Perl faults them if its -w option is set. I have changed PCRE so
1520 that it faults them when PCRE_EXTRA is set.
1522 5. Added support for \N, which always matches any character other than
1523 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1525 6. When compiling pcregrep with newer versions of gcc which may have
1526 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1527 declared with attribute warn_unused_result" were given. Just casting the
1528 result to (void) does not stop the warnings; a more elaborate fudge is
1529 needed. I've used a macro to implement this.
1531 7. Minor change to pcretest.c to avoid a compiler warning.
1533 8. Added four artifical Unicode properties to help with an option to make
1534 \s etc use properties (see next item). The new properties are: Xan
1535 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1537 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1538 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1539 this option. Modified pcretest to add /W to test this facility. Added
1540 REG_UCP to make it available via the POSIX interface.
1542 10. Added --line-buffered to pcregrep.
1544 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1545 studied, and the match started with a letter with a code point greater than
1546 127 whose first byte was different to the first byte of the other case of
1547 the letter, the other case of this starting letter was not recognized
1548 (#976).
1550 12. If a pattern that was studied started with a repeated Unicode property
1551 test, for example, \p{Nd}+, there was the theoretical possibility of
1552 setting up an incorrect bitmap of starting bytes, but fortunately it could
1553 not have actually happened in practice until change 8 above was made (it
1554 added property types that matched character-matching opcodes).
1556 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1557 possible starting bytes for non-anchored patterns.
1559 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1560 \R, and also a number of cases that involve Unicode properties, both
1561 explicit and implicit when PCRE_UCP is set.
1563 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1564 input, it could crash or give wrong results if characters with values
1565 greater than 0xc0 were present in the subject string. (Detail: it assumed
1566 UTF-8 input when processing these items.)
1568 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1569 size_t is 64-bit (#991).
1571 17. Added a check for running out of memory when PCRE is compiled with
1572 --disable-stack-for-recursion (#990).
1574 18. If the last data line in a file for pcretest does not have a newline on
1575 the end, a newline was missing in the output.
1577 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1578 less than 128) in its various bitmaps. However, there is a facility for
1579 generating tables according to the current locale when PCRE is compiled. It
1580 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1581 characters, are recognized by isspace() and therefore were getting set in
1582 these tables, and indeed these tables seem to approximate to ISO 8859. This
1583 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1584 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1585 which of course cannot start UTF-8 characters. I have changed the code so
1586 that only real ASCII characters (less than 128) and the correct starting
1587 bytes for UTF-8 encodings are set for characters greater than 127 when in
1588 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1589 altogether.)
1591 20. Added the /T option to pcretest so as to be able to run tests with non-
1592 standard character tables, thus making it possible to include the tests
1593 used for 19 above in the standard set of tests.
1595 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1596 reference to a subpattern the other side of a comment that contains an
1597 opening parenthesis caused either an internal compiling error, or a
1598 reference to the wrong subpattern.
1601 Version 8.02 19-Mar-2010
1602 ------------------------
1604 1. The Unicode data tables have been updated to Unicode 5.2.0.
1606 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1607 configured.
1609 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1610 original author of that file, following a query about its status.
1612 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1613 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1615 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1616 quantifier applied to a forward-referencing subroutine call, could compile
1617 incorrect code or give the error "internal error: previously-checked
1618 referenced subpattern not found".
1620 6. Both MS Visual Studio and Symbian OS have problems with initializing
1621 variables to point to external functions. For these systems, therefore,
1622 pcre_malloc etc. are now initialized to local functions that call the
1623 relevant global functions.
1625 7. There were two entries missing in the vectors called coptable and poptable
1626 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1627 I've fixed the data, and added a kludgy way of testing at compile time that
1628 the lengths are correct (equal to the number of opcodes).
1630 8. Following on from 7, I added a similar kludge to check the length of the
1631 eint vector in pcreposix.c.
1633 9. Error texts for pcre_compile() are held as one long string to avoid too
1634 much relocation at load time. To find a text, the string is searched,
1635 counting zeros. There was no check for running off the end of the string,
1636 which could happen if a new error number was added without updating the
1637 string.
1639 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1641 11. \K was not working if it appeared in an atomic group or in a group that
1642 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1643 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1644 the assertion is positive, but not if it is negative.
1646 12. Change 11 fortuitously reduced the size of the stack frame used in the
1647 "match()" function of pcre_exec.c by one pointer. Forthcoming
1648 implementation of support for (*MARK) will need an extra pointer on the
1649 stack; I have reserved it now, so that the stack frame size does not
1650 decrease.
1652 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1653 item in branch that calls a recursion is a subroutine call - as in the
1654 second branch in the above example - was incorrectly given the compile-
1655 time error "recursive call could loop indefinitely" because pcre_compile()
1656 was not correctly checking the subroutine for matching a non-empty string.
1658 14. The checks for overrunning compiling workspace could trigger after an
1659 overrun had occurred. This is a "should never occur" error, but it can be
1660 triggered by pathological patterns such as hundreds of nested parentheses.
1661 The checks now trigger 100 bytes before the end of the workspace.
1663 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1666 Version 8.01 19-Jan-2010
1667 ------------------------
1669 1. If a pattern contained a conditional subpattern with only one branch (in
1670 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1671 computed the wrong minimum data length (which is of course zero for such
1672 subpatterns). This could cause incorrect "no match" results.
1674 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1675 the pattern is reset in the first branch, pcre_compile() failed with
1676 "internal error: code overflow at offset...". This happened only when
1677 the reset was to the original external option setting. (An optimization
1678 abstracts leading options settings into an external setting, which was the
1679 cause of this.)
1681 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1682 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1683 assertion pattern did not match (meaning that the assertion was true), it
1684 was incorrectly treated as false if the SKIP had been reached during the
1685 matching. This also applied to assertions used as conditions.
1687 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1688 assertion subpattern, including such a pattern used as a condition,
1689 unpredictable results occurred, instead of the error return
1692 5. The C++ GlobalReplace function was not working like Perl for the special
1693 situation when an empty string is matched. It now does the fancy magic
1694 stuff that is necessary.
1696 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1697 removed. (These were left over from very, very early versions of PCRE.)
1699 7. Some cosmetic changes to the code to make life easier when compiling it
1700 as part of something else:
1702 (a) Change DEBUG to PCRE_DEBUG.
1704 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1705 called "current" as "current_branch", to prevent a collision with the
1706 Linux macro when compiled as a kernel module.
1708 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1709 prevent a collision with the Linux macro when compiled as a kernel
1710 module.
1712 8. In pcre_compile() there are some checks for integer overflows that used to
1713 cast potentially large values to (double). This has been changed to that
1714 when building, a check for int64_t is made, and if it is found, it is used
1715 instead, thus avoiding the use of floating point arithmetic. (There is no
1716 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1717 double.
1719 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1720 2005 (difference between two addresses compared to an unsigned value).
1722 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1723 custom one, because of the following reported problem in Windows:
1725 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1726 under Win32.
1727 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1728 therefore missing the function definition.
1729 - The compiler thus generates a "C" signature for the test function.
1730 - The linker fails to find the "C" function.
1731 - PCRE fails to configure if asked to do so against libbz2.
1733 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1734 messages were output:
1736 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1737 rerunning libtoolize, to keep the correct libtool macros in-tree.
1738 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1740 I have done both of these things.
1742 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1743 most of the time, it *can* run out if it is given a pattern that contains a
1744 runaway infinite recursion. I updated the discussion in the pcrestack man
1745 page.
1747 13. Now that we have gone to the x.xx style of version numbers, the minor
1748 version may start with zero. Using 08 or 09 is a bad idea because users
1749 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1750 interpreted as invalid octal numbers. I've updated the previous comment in
1751 configure.ac, and also added a check that gives an error if 08 or 09 are
1752 used.
1754 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1755 causing partial matching to fail when the end of the subject matched \W
1756 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1758 15. There were some discrepancies between the declarations in pcre_internal.h
1759 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1760 their definitions. The declarations used "const uschar *" and the
1761 definitions used USPTR. Even though USPTR is normally defined as "const
1762 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1763 reported that: "This difference in casting confuses some C++ compilers, for
1764 example, SunCC recognizes above declarations as different functions and
1765 generates broken code for hbpcre." I have changed the declarations to use
1766 USPTR.
1768 16. GNU libtool is named differently on some systems. The autogen.sh script now
1769 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1770 (FreeBSD).
1772 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1773 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1774 comment: "Figure out how to create a longlong from a string: strtoll and
1775 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1776 instance, but it only takes 2 args instead of 3!"
1778 18. A subtle bug concerned with back references has been fixed by a change of
1779 specification, with a corresponding code fix. A pattern such as
1780 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1781 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1782 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1783 same bug. Such groups have to be quantified to be useful, or contained
1784 inside another quantified group. (If there's no repetition, the reference
1785 can never match.) The problem arises because, having left the group and
1786 moved on to the rest of the pattern, a later failure that backtracks into
1787 the group uses the captured value from the final iteration of the group
1788 rather than the correct earlier one. I have fixed this in PCRE by forcing
1789 any group that contains a reference to itself to be an atomic group; that
1790 is, there cannot be any backtracking into it once it has completed. This is
1791 similar to recursive and subroutine calls.
1794 Version 8.00 19-Oct-09
1795 ----------------------
1797 1. The table for translating pcre_compile() error codes into POSIX error codes
1798 was out-of-date, and there was no check on the pcre_compile() error code
1799 being within the table. This could lead to an OK return being given in
1800 error.
1802 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1803 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1804 in a Windows environment.
1806 3. The pcregrep --count option prints the count for each file even when it is
1807 zero, as does GNU grep. However, pcregrep was also printing all files when
1808 --files-with-matches was added. Now, when both options are given, it prints
1809 counts only for those files that have at least one match. (GNU grep just
1810 prints the file name in this circumstance, but including the count seems
1811 more useful - otherwise, why use --count?) Also ensured that the
1812 combination -clh just lists non-zero counts, with no names.
1814 4. The long form of the pcregrep -F option was incorrectly implemented as
1815 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1816 but it seems right to fix it, and I didn't think it was worth preserving
1817 the old behaviour.
1819 5. The command line items --regex=pattern and --regexp=pattern were not
1820 recognized by pcregrep, which required --regex pattern or --regexp pattern
1821 (with a space rather than an '='). The man page documented the '=' forms,
1822 which are compatible with GNU grep; these now work.
1824 6. No libpcreposix.pc file was created for pkg-config; there was just
1825 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1827 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1828 when UCP support is not needed, by modifying the Python script that
1829 generates it from Unicode data files. This should not matter if the module
1830 is correctly used as a library, but I received one complaint about 50K of
1831 unwanted data. My guess is that the person linked everything into his
1832 program rather than using a library. Anyway, it does no harm.
1834 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1835 was a minimum greater than 1 for a wide character in a possessive
1836 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1837 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1838 character. Chaos in the form of incorrect output or a compiling loop could
1839 result.
1841 9. The restrictions on what a pattern can contain when partial matching is
1842 requested for pcre_exec() have been removed. All patterns can now be
1843 partially matched by this function. In addition, if there are at least two
1844 slots in the offset vector, the offset of the earliest inspected character
1845 for the match and the offset of the end of the subject are set in them when
1846 PCRE_ERROR_PARTIAL is returned.
1848 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1849 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1850 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1851 and may be more useful for multi-segment matching.
1853 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1854 used to be given if ever the end of the subject was reached; now it is
1855 given only if matching could not proceed because another character was
1856 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1857 string "Z", which now yields "no match" instead of "partial match". In the
1858 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1859 final character ended with (*FAIL).
1861 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1862 if the pattern had a "must contain" character that was already found in the
1863 earlier partial match, unless partial matching was again requested. For
1864 example, with the pattern /dog.(body)?/, the "must contain" character is
1865 "g". If the first part-match was for the string "dog", restarting with
1866 "sbody" failed. This bug has been fixed.
1868 13. The string returned by pcre_dfa_exec() after a partial match has been
1869 changed so that it starts at the first inspected character rather than the
1870 first character of the match. This makes a difference only if the pattern
1871 starts with a lookbehind assertion or \b or \B (\K is not supported by
1872 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1873 matching functions compatible, and I think it's the right thing to do.
1875 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1876 so that the demonstration program is easily available in environments where
1877 PCRE has not been installed from source.
1879 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1880 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1881 library.
1883 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1884 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1885 is not the first non-POSIX option to be added. Clearly some people find
1886 these options useful.
1888 17. If a caller to the POSIX matching function regexec() passes a non-zero
1889 value for nmatch with a NULL value for pmatch, the value of
1890 nmatch is forced to zero.
1892 18. RunGrepTest did not have a test for the availability of the -u option of
1893 the diff command, as RunTest does. It now checks in the same way as
1894 RunTest, and also checks for the -b option.
1896 19. If an odd number of negated classes containing just a single character
1897 interposed, within parentheses, between a forward reference to a named
1898 subpattern and the definition of the subpattern, compilation crashed with
1899 an internal error, complaining that it could not find the referenced
1900 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1901 [The bug was that it was starting one character too far in when skipping
1902 over the character class, thus treating the ] as data rather than
1903 terminating the class. This meant it could skip too much.]
1905 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1906 /g option in pcretest when the pattern contains \K, which makes it possible
1907 to have an empty string match not at the start, even when the pattern is
1908 anchored. Updated pcretest and pcredemo to use this option.
1910 21. If the maximum number of capturing subpatterns in a recursion was greater
1911 than the maximum at the outer level, the higher number was returned, but
1912 with unset values at the outer level. The correct (outer level) value is
1913 now given.
1915 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1916 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1917 make it do so. The string so far is captured, making this feature
1918 compatible with Perl.
1920 23. The tests have been re-organized, adding tests 11 and 12, to make it
1921 possible to check the Perl 5.10 features against Perl 5.10.
1923 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1924 pattern matches a fixed length string. PCRE did not allow this; now it
1925 does. Neither allows recursion.
1927 25. I finally figured out how to implement a request to provide the minimum
1928 length of subject string that was needed in order to match a given pattern.
1929 (It was back references and recursion that I had previously got hung up
1930 on.) This code has now been added to pcre_study(); it finds a lower bound
1931 to the length of subject needed. It is not necessarily the greatest lower
1932 bound, but using it to avoid searching strings that are too short does give
1933 some useful speed-ups. The value is available to calling programs via
1934 pcre_fullinfo().
1936 26. While implementing 25, I discovered to my embarrassment that pcretest had
1937 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1938 study optimizations had never been tested with that matching function.
1939 Oops. What is worse, even when it was passed study data, there was a bug in
1940 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1941 were also very few tests of studied patterns with pcre_dfa_exec().
1943 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1944 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1945 on the other side of the coin, they are no longer allowed to have different
1946 names, because these cannot be distinguished in PCRE, and this has caused
1947 confusion. (This is a difference from Perl.)
1949 28. When duplicate subpattern names are present (necessarily with different
1950 numbers, as required by 27 above), and a test is made by name in a
1951 conditional pattern, either for a subpattern having been matched, or for
1952 recursion in such a pattern, all the associated numbered subpatterns are
1953 tested, and the overall condition is true if the condition is true for any
1954 one of them. This is the way Perl works, and is also more like the way
1955 testing by number works.
1958 Version 7.9 11-Apr-09
1959 ---------------------
1961 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1962 (pcretest), all targets were linked against these libraries. This included
1963 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1964 libraries. This caused unwanted dependencies to be created. This problem
1965 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1966 pcretest is linked with readline.
1968 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1969 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1970 moved outside it again, because FALSE and TRUE are already defined in AIX,
1971 but BOOL is not.
1973 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1974 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1976 4. The pcregrep documentation said spaces were inserted as well as colons (or
1977 hyphens) following file names and line numbers when outputting matching
1978 lines. This is not true; no spaces are inserted. I have also clarified the
1979 wording for the --colour (or --color) option.
1981 5. In pcregrep, when --colour was used with -o, the list of matching strings
1982 was not coloured; this is different to GNU grep, so I have changed it to be
1983 the same.
1985 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1986 each matching line was coloured. Now it goes on to look for further matches
1987 of any of the test patterns, which is the same behaviour as GNU grep.
1989 7. A pattern that could match an empty string could cause pcregrep to loop; it
1990 doesn't make sense to accept an empty string match in pcregrep, so I have
1991 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1992 seems to be how GNU grep behaves. [But see later change 40 for release
1993 8.33.]
1995 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1996 start or after a newline", because the conditional assertion was not being
1997 correctly handled. The rule now is that both the assertion and what follows
1998 in the first alternative must satisfy the test.
2000 9. If auto-callout was enabled in a pattern with a conditional group whose
2001 condition was an assertion, PCRE could crash during matching, both with
2002 pcre_exec() and pcre_dfa_exec().
2004 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2005 used for matching.
2007 11. Unicode property support in character classes was not working for
2008 characters (bytes) greater than 127 when not in UTF-8 mode.
2010 12. Added the -M command line option to pcretest.
2012 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2014 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2016 16. Added comments and documentation about mis-use of no_arg in the C++
2017 wrapper.
2019 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2020 from Martin Jerabek that uses macro names for all relevant character and
2021 string constants.
2023 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2024 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2025 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2026 these, but not everybody uses configure.
2028 19. A conditional group that had only one branch was not being correctly
2029 recognized as an item that could match an empty string. This meant that an
2030 enclosing group might also not be so recognized, causing infinite looping
2031 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2032 with the subject "ab", where knowledge that the repeated group can match
2033 nothing is needed in order to break the loop.
2035 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2036 exec(), but without supplying a callout function, matching went wrong.
2038 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2039 leak if the size of the offset vector was greater than 30. When the vector
2040 is smaller, the saved offsets during recursion go onto a local stack
2041 vector, but for larger vectors malloc() is used. It was failing to free
2042 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2043 error, in fact).
2045 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2046 heapframe that is used only when UTF-8 support is enabled. This caused no
2047 problem, but was untidy.
2049 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2050 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2051 included within another project.
2053 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2054 slightly modified by me:
2056 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2057 not building pcregrep.
2059 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2060 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2062 25. Forward references, both numeric and by name, in patterns that made use of
2063 duplicate group numbers, could behave incorrectly or give incorrect errors,
2064 because when scanning forward to find the reference group, PCRE was not
2065 taking into account the duplicate group numbers. A pattern such as
2066 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2068 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2069 the feature of a custom pointer more persuasive (as requested by a user).
2071 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2072 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2073 pre-defined.
2075 28. Added support for (*UTF8) at the start of a pattern.
2077 29. Arrange for flags added by the "release type" setting in CMake to be shown
2078 in the configuration summary.
2081 Version 7.8 05-Sep-08
2082 ---------------------
2084 1. Replaced UCP searching code with optimized version as implemented for Ad
2085 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2086 stage table and inline lookup instead of a function, giving speed ups of 2
2087 to 5 times on some simple patterns that I tested. Permission was given to
2088 distribute the MultiStage2.py script that generates the tables (it's not in
2089 the tarball, but is in the Subversion repository).
2091 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2092 scripts.
2094 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2095 a group with a zero qualifier. The result of the study could be incorrect,
2096 or the function might crash, depending on the pattern.
2098 4. Caseless matching was not working for non-ASCII characters in back
2099 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2100 It now works when Unicode Property Support is available.
2102 5. In pcretest, an escape such as \x{de} in the data was always generating
2103 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2104 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2105 truncation.
2107 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2109 7. Added two (int) casts to pcregrep when printing the difference of two
2110 pointers, in case they are 64-bit values.
2112 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2113 test 2 if it fails.
2115 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2116 and a #define of that name to empty if it is not externally set. This is to
2117 allow users of MSVC to set it if necessary.
2119 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2120 the convenience functions in the pcre_get.c source file.
2122 11. An option change at the start of a pattern that had top-level alternatives
2123 could cause overwriting and/or a crash. This command provoked a crash in
2124 some environments:
2126 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2128 This potential security problem was recorded as CVE-2008-2371.
2130 12. For a pattern where the match had to start at the beginning or immediately
2131 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2132 pcre_dfa_exec() could read past the end of the passed subject if there was
2133 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2134 pcretest so that it places the subject at the end of its malloc-ed buffer.
2136 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2137 exec() might read past the end of the data buffer in UTF-8 mode.
2139 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2140 the data contained the byte 0x85 as part of a UTF-8 character within its
2141 first line. This applied both to normal and DFA matching.
2143 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2144 /^[^d]*?$/8 failed to match "abc".
2146 16. Added a missing copyright notice to pcrecpp_internal.h.
2148 17. Make it more clear in the documentation that values returned from
2149 pcre_exec() in ovector are byte offsets, not character counts.
2151 18. Tidied a few places to stop certain compilers from issuing warnings.
2153 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2154 supplied by Stefan Weber. I made a further small update for 7.8 because
2155 there is a change of source arrangements: the pcre_searchfuncs.c module is
2156 replaced by pcre_ucd.c.
2159 Version 7.7 07-May-08
2160 ---------------------
2162 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2163 a string to a long long, pretend we don't even have a long long." This is
2164 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2166 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2167 pre-7.6 versions, which defined a global no_arg variable instead of putting
2168 it in the RE class. (See also #8 below.)
2170 3. Remove a line of dead code, identified by coverity and reported by Nuno
2171 Lopes.
2173 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2175 (1) The include/exclude patterns were being applied to the whole pathnames
2176 of files, instead of just to the final components.
2178 (2) If there was more than one level of directory, the subdirectories were
2179 skipped unless they satisfied the include/exclude conditions. This is
2180 inconsistent with GNU grep (and could even be seen as contrary to the
2181 pcregrep specification - which I improved to make it absolutely clear).
2182 The action now is always to scan all levels of directory, and just
2183 apply the include/exclude patterns to regular files.
2185 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2186 --exclude_dir in the tests to avoid scanning .svn directories.
2188 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2189 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2190 doesn't support NULs in patterns.
2192 7. Added some missing "const"s to declarations of static tables in
2193 pcre_compile.c and pcre_dfa_exec.c.
2195 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2196 caused by fix #2 above. (Subsequently also a second patch to fix the
2197 first patch. And a third patch - this was a messy problem.)
2199 9. Applied Craig's patch to remove the use of push_back().
2201 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2202 matching function regexec().
2204 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2205 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2206 references. PCRE supports relative numbers with this syntax (I don't think
2207 Oniguruma does).
2209 12. Previously, a group with a zero repeat such as (...){0} was completely
2210 omitted from the compiled regex. However, this means that if the group
2211 was called as a subroutine from elsewhere in the pattern, things went wrong
2212 (an internal error was given). Such groups are now left in the compiled
2213 pattern, with a new opcode that causes them to be skipped at execution
2214 time.
2216 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2217 to the way PCRE behaves:
2219 (a) A lone ] character is dis-allowed (Perl treats it as data).
2221 (b) A back reference to an unmatched subpattern matches an empty string
2222 (Perl fails the current match path).
2224 (c) A data ] in a character class must be notated as \] because if the
2225 first data character in a class is ], it defines an empty class. (In
2226 Perl it is not possible to have an empty class.) The empty class []
2227 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2228 The negative empty class [^] matches any one character, independently
2229 of the DOTALL setting.
2231 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2232 non-existent subpattern following a character class starting with ']' and
2233 containing () gave an internal compiling error instead of "reference to
2234 non-existent subpattern". Fortunately, when the pattern did exist, the
2235 compiled code was correct. (When scanning forwards to check for the
2236 existence of the subpattern, it was treating the data ']' as terminating
2237 the class, so got the count wrong. When actually compiling, the reference
2238 was subsequently set up correctly.)
2240 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2241 it was being rejected as not supported by pcre_dfa_exec(), even though
2242 other assertions are supported. I have made pcre_dfa_exec() support
2243 (*FAIL).
2245 16. The implementation of 13c above involved the invention of a new opcode,
2246 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2247 cannot be changed at match time, I realized I could make a small
2248 improvement to matching performance by compiling OP_ALLANY instead of
2249 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2250 on the OP_ANY path.
2252 17. Compiling pcretest on Windows with readline support failed without the
2253 following two fixes: (1) Make the unistd.h include conditional on
2254 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2256 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2257 ncurses library to be included for pcretest when ReadLine support is
2258 requested, but also to allow for it to be overridden. This patch came from
2259 Daniel Bergström.
2261 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2262 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2263 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2264 spotting this.
2267 Version 7.6 28-Jan-08
2268 ---------------------
2270 1. A character class containing a very large number of characters with
2271 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2272 overflow.
2274 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2275 HAVE_LONG_LONG is not defined.
2277 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2278 bring them up to date and include new features. This patch includes:
2280 - Fixed PH's badly added libz and libbz2 support.
2281 - Fixed a problem with static linking.
2282 - Added pcredemo. [But later removed - see 7 below.]
2283 - Fixed dftables problem and added an option.
2284 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2286 - Added readline support for pcretest.
2287 - Added an listing of the option settings after cmake has run.
2289 4. A user submitted a patch to Makefile that makes it easy to create
2290 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2291 Makefile.am that cause it to include this special target, without
2292 affecting anything else. Note that the same mingw target plus all
2293 the other distribution libraries and programs are now supported
2294 when configuring with CMake (see 6 below) instead of with
2295 Configure/Make.
2297 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2298 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2299 exported in the Windows port". It has not yet been confirmed that the patch
2300 solves the problem, but it does no harm.
2302 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2303 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2304 with CMake, and also correct the comment about stack recursion.
2306 7. Remove the automatic building of pcredemo from the ./configure system and
2307 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2308 of a program that users should build themselves after PCRE is installed, so
2309 building it automatically is not really right. What is more, it gave
2310 trouble in some build environments.
2312 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2315 Version 7.5 10-Jan-08
2316 ---------------------
2318 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2319 values in parens when parsing an RE using the C++ wrapper."
2321 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2322 Characters greater than 255 were excluded from the class instead of being
2323 included.
2325 3. The same bug as (2) above applied to negated POSIX classes such as
2326 [:^space:].
2328 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2329 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2330 I have changed it.
2332 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2333 first named subpattern) and a construct such as (?&a) would reference the
2334 first named subpattern whose name started with "a" (in other words, the
2335 length check was missing). Both these problems are fixed. "Subpattern name
2336 expected" is now given for (?&) (a zero-length name), and this patch also
2337 makes it give the same error for \k'' (previously it complained that that
2338 was a reference to a non-existent subpattern).
2340 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2341 this is right because (?- can be followed by option settings as well as by
2342 digits. I have, however, made the messages clearer.
2344 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2345 than the number used in the conditional) now cause a compile-time error.
2346 This is actually not compatible with Perl, which accepts such patterns, but
2347 treats the conditional as always being FALSE (as PCRE used to), but it
2348 seems to me that giving a diagnostic is better.
2350 8. Change "alphameric" to the more common word "alphanumeric" in comments
2351 and messages.
2353 9. Fix two occurrences of "backslash" in comments that should have been
2354 "backspace".
2356 10. Remove two redundant lines of code that can never be obeyed (their function
2357 was moved elsewhere).
2359 11. The program that makes PCRE's Unicode character property table had a bug
2360 which caused it to generate incorrect table entries for sequences of
2361 characters that have the same character type, but are in different scripts.
2362 It amalgamated them into a single range, with the script of the first of
2363 them. In other words, some characters were in the wrong script. There were
2364 thirteen such cases, affecting characters in the following ranges:
2366 U+002b0 - U+002c1
2367 U+0060c - U+0060d
2368 U+0061e - U+00612
2369 U+0064b - U+0065e
2370 U+0074d - U+0076d
2371 U+01800 - U+01805
2372 U+01d00 - U+01d77
2373 U+01d9b - U+01dbf
2374 U+0200b - U+0200f
2375 U+030fc - U+030fe
2376 U+03260 - U+0327f
2377 U+0fb46 - U+0fbb1
2378 U+10450 - U+1049d
2380 12. The -o option (show only the matching part of a line) for pcregrep was not
2381 compatible with GNU grep in that, if there was more than one match in a
2382 line, it showed only the first of them. It now behaves in the same way as
2383 GNU grep.
2385 13. If the -o and -v options were combined for pcregrep, it printed a blank
2386 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2387 does the same. The return code can be used to tell if there were any
2388 non-matching lines.
2390 14. Added --file-offsets and --line-offsets to pcregrep.
2392 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2393 infinitely looping recursion. The bug was that positive lookaheads were not
2394 being skipped when checking for a possible empty match (negative lookaheads
2395 and both kinds of lookbehind were skipped).
2397 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2398 inclusion of <windows.h> to before rather than after the definition of
2399 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2401 17. Specifying a possessive quantifier with a specific limit for a Unicode
2402 character property caused pcre_compile() to compile bad code, which led at
2403 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2404 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2405 caused the error; without that there was no problem.
2407 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2409 19. Added --enable-pcretest-libreadline.
2411 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2412 RE::GlobalReplace(). As a result, the number of replacements returned was
2413 double what it should be. I removed one of the increments, but Craig sent a
2414 later patch that removed the other one (the right fix) and added unit tests
2415 that check the return values (which was not done before).
2417 21. Several CMake things:
2419 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2420 the names libpcre and libpcreposix, not just pcre and pcreposix.
2422 (2) The above change means that pcretest and pcregrep are now correctly
2423 linked with the newly-built libraries, not previously installed ones.
2427 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2428 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2429 UTF-8 newline character). The key issue is that the pattern starts .*;
2430 this means that the match must be either at the beginning, or after a
2431 newline. The bug was in the code for advancing after a failed match and
2432 checking that the new position followed a newline. It was not taking
2433 account of UTF-8 characters correctly.
2435 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2436 character classes. PCRE was not treating the sequence [:...:] as a
2437 character class unless the ... were all letters. Perl, however, seems to
2438 allow any characters between [: and :], though of course it rejects as
2439 unknown any "names" that contain non-letters, because all the known class
2440 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2441 for example, whereas PCRE did not - it did not recognize a POSIX character
2442 class. This seemed a bit dangerous, so the code has been changed to be
2443 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2444 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2445 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2446 Perl does, and where it didn't before.
2448 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2449 Windows environments %n is disabled by default.
2452 Version 7.4 21-Sep-07
2453 ---------------------
2455 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2456 means that a class such as [\s] counted as "explicit reference to CR or
2457 LF". That isn't really right - the whole point of the change was to try to
2458 help when there was an actual mention of one of the two characters. So now
2459 the change happens only if \r or \n (or a literal CR or LF) character is
2460 encountered.
2462 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2463 of both had grown to the point where there were only 3 bits left.
2464 Fortunately, there was spare space in the data structure, and so I have
2465 moved the internal flags into a new 16-bit field to free up more option
2466 bits.
2468 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2469 but did not set the internal JCHANGED flag - either of these is enough to
2470 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2471 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2472 start sets both bits.
2474 4. Added options (at build time, compile time, exec time) to change \R from
2475 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2477 5. doc/pcresyntax.html was missing from the distribution.
2479 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2480 compatibility, even though it is no longer used.
2482 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2483 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2484 windows.h file is present (where different names are used). [This was
2485 reversed later after testing - see 16 below.]
2487 8. Changed all #include <config.h> to #include "config.h". There were also
2488 some further <pcre.h> cases that I changed to "pcre.h".
2490 9. When pcregrep was used with the --colour option, it missed the line ending
2491 sequence off the lines that it output.
2493 10. It was pointed out to me that arrays of string pointers cause lots of
2494 relocations when a shared library is dynamically loaded. A technique of
2495 using a single long string with a table of offsets can drastically reduce
2496 these. I have refactored PCRE in four places to do this. The result is
2497 dramatic:
2499 Originally: 290
2500 After changing UCP table: 187
2501 After changing error message table: 43
2502 After changing table of "verbs" 36
2503 After changing table of Posix names 22
2505 Thanks to the folks working on Gregex for glib for this insight.
2507 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2508 unicode-properties was also set.
2510 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2512 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2513 checked only for CRLF.
2515 14. Added casts to pcretest.c to avoid compiler warnings.
2517 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2519 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2520 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2521 entirely. This removes changes made in 7 above.
2523 17. The CMake files have been updated, and there is now more information about
2524 building with CMake in the NON-UNIX-USE document.
2527 Version 7.3 28-Aug-07
2528 ---------------------
2530 1. In the rejigging of the build system that eventually resulted in 7.1, the
2531 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2532 brackets there is not right, since it causes compilers to look for an
2533 installed pcre.h, not the version that is in the source that is being
2534 compiled (which of course may be different). I have changed it back to:
2536 #include "pcre.h"
2538 I have a vague recollection that the change was concerned with compiling in
2539 different directories, but in the new build system, that is taken care of
2540 by the VPATH setting the Makefile.
2542 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2543 when the subject happened to end in the byte 0x85 (e.g. if the last
2544 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2545 characters but of course it shouldn't be taken as a newline when it is part
2546 of another character. The bug was that, for an unlimited repeat of . in
2547 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2548 characters when looking for a newline.
2550 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2552 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2553 in debug output.
2555 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2556 long printing in the pcrecpp unittest when running under MinGW.
2558 6. ESC_K was left out of the EBCDIC table.
2560 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2561 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2562 limit also applies to "virtual nesting" when a pattern is recursive, and in
2563 this case 1000 isn't so big. I have been able to remove this limit at the
2564 expense of backing off one optimization in certain circumstances. Normally,
2565 when pcre_exec() would call its internal match() function recursively and
2566 immediately return the result unconditionally, it uses a "tail recursion"
2567 feature to save stack. However, when a subpattern that can match an empty
2568 string has an unlimited repetition quantifier, it no longer makes this
2569 optimization. That gives it a stack frame in which to save the data for
2570 checking that an empty string has been matched. Previously this was taken
2571 from the 1000-entry workspace that had been reserved. So now there is no
2572 explicit limit, but more stack is used.
2574 8. Applied Daniel's patches to solve problems with the import/export magic
2575 syntax that is required for Windows, and which was going wrong for the
2576 pcreposix and pcrecpp parts of the library. These were overlooked when this
2577 problem was solved for the main library.
2579 9. There were some crude static tests to avoid integer overflow when computing
2580 the size of patterns that contain repeated groups with explicit upper
2581 limits. As the maximum quantifier is 65535, the maximum group length was
2582 set at 30,000 so that the product of these two numbers did not overflow a
2583 32-bit integer. However, it turns out that people want to use groups that
2584 are longer than 30,000 bytes (though not repeat them that many times).
2585 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2586 made it possible to implement the integer overflow checks in a much more
2587 dynamic way, which I have now done. The artificial limitation on group
2588 length has been removed - we now have only the limit on the total length of
2589 the compiled pattern, which depends on the LINK_SIZE setting.
2591 10. Fixed a bug in the documentation for get/copy named substring when
2592 duplicate names are permitted. If none of the named substrings are set, the
2593 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2594 empty string.
2596 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2597 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2598 because the ] is interpreted as the first data character and the
2599 terminating ] is not found. PCRE has been made compatible with Perl in this
2600 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2601 cause memory overwriting.
2603 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2604 string has been matched (to stop an infinite loop). It was not recognizing
2605 a conditional subpattern that could match an empty string if that
2606 subpattern was within another subpattern. For example, it looped when
2607 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2608 condition was not nested. This bug has been fixed.
2610 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2611 past the start of the subject in the presence of bytes with the top bit
2612 set, for example "\x8aBCD".
2614 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2615 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2617 14. Optimized (?!) to (*FAIL).
2619 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2620 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2621 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2622 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2623 does: it's just the validity check that is more restrictive.
2625 16. Inserted checks for integer overflows during escape sequence (backslash)
2626 processing, and also fixed erroneous offset values for syntax errors during
2627 backslash processing.
2629 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2630 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2632 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2633 caused an overrun.
2635 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2636 something other than just ASCII characters) inside a group that had an
2637 unlimited repeat caused a loop at compile time (while checking to see
2638 whether the group could match an empty string).
2640 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2641 [\P{Any}] did so. (Error in the code for printing property names.)
2643 21. An orphan \E inside a character class could cause a crash.
2645 22. A repeated capturing bracket such as (A)? could cause a wild memory
2646 reference during compilation.
2648 23. There are several functions in pcre_compile() that scan along a compiled
2649 expression for various reasons (e.g. to see if it's fixed length for look
2650 behind). There were bugs in these functions when a repeated \p or \P was
2651 present in the pattern. These operators have additional parameters compared
2652 with \d, etc, and these were not being taken into account when moving along
2653 the compiled data. Specifically:
2655 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2656 length.
2658 (b) An item such as \pL+ within a repeated group could cause crashes or
2659 loops.
2661 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2662 "reference to non-existent subpattern" error.
2664 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2666 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2667 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2669 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2671 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2672 character were causing crashes (broken optimization).
2674 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2675 \p or \P) caused a compile-time loop.
2677 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2678 break. For example, the unstudied pattern [\r\n]A does not match the string
2679 "\r\nA" because change 7.0/46 below moves the current point on by two
2680 characters after failing to match at the start. However, the pattern \nA
2681 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2682 the same is true. There doesn't seem any very clean way out of this, but
2683 what I have chosen to do makes the common cases work: PCRE now takes note
2684 of whether there can be an explicit match for \r or \n anywhere in the
2685 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2686 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2687 pattern has explicit CR or LF references.
2689 29. Added (*CR) etc for changing newline setting at start of pattern.
2692 Version 7.2 19-Jun-07
2693 ---------------------
2695 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2696 which is apparently normally available under Windows.
2698 2. Re-jig the pcregrep tests with different newline settings in an attempt
2699 to make them independent of the local environment's newline setting.
2701 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2703 4. Some of the "internals" tests were previously cut out when the link size
2704 was not 2, because the output contained actual offsets. The recent new
2705 "Z" feature of pcretest means that these can be cut out, making the tests
2706 usable with all link sizes.
2708 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2709 stack recursion. This gives a massive performance boost under BSD, but just
2710 a small improvement under Linux. However, it saves one field in the frame
2711 in all cases.
2713 6. Added more features from the forthcoming Perl 5.10:
2715 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2716 recursion call. It refers to the nth most recently opened parentheses.
2718 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2719 to be opened parentheses.
2721 (c) Conditions that refer to capturing parentheses can be specified
2722 relatively, for example, (?(-2)... or (?(+3)...
2724 (d) \K resets the start of the current match so that everything before
2725 is not part of it.
2727 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2729 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2730 reference syntax.
2732 (g) (?| introduces a group in which the numbering of parentheses in each
2733 alternative starts with the same number.
2735 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2737 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2740 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2741 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2742 for detecting groups that can match an empty string.
2744 9. A pattern with a very large number of alternatives (more than several
2745 hundred) was running out of internal workspace during the pre-compile
2746 phase, where pcre_compile() figures out how much memory will be needed. A
2747 bit of new cunning has reduced the workspace needed for groups with
2748 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2749 workspace instead of running out of the 4096 that are available.
2751 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2753 11. Applied patch from Google to remove an optimization that didn't quite work.
2754 The report of the bug said:
2756 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2757 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2758 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2760 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2761 it matched the wrong number of bytes.
2764 Version 7.1 24-Apr-07
2765 ---------------------
2767 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2768 that is more "standard", making use of automake and other Autotools. There
2769 is some re-arrangement of the files and adjustment of comments consequent
2770 on this.
2772 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2773 for recursive directory scanning broke on some systems because the files
2774 are not scanned in any specific order and on different systems the order
2775 was different. A call to "sort" has been inserted into RunGrepTest for the
2776 approprate test as a short-term fix. In the longer term there may be an
2777 alternative.
2779 3. I had an email from Eric Raymond about problems translating some of PCRE's
2780 man pages to HTML (despite the fact that I distribute HTML pages, some
2781 people do their own conversions for various reasons). The problems
2782 concerned the use of low-level troff macros .br and .in. I have therefore
2783 removed all such uses from the man pages (some were redundant, some could
2784 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2785 HTML has been updated to handle .nf/.fi and to complain if it encounters
2786 .br or .in.
2788 4. Updated comments in configure.ac that get placed in config.h.in and also
2789 arranged for config.h to be included in the distribution, with the name
2790 config.h.generic, for the benefit of those who have to compile without
2791 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2793 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2794 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2795 makevp.bat for the current PCRE, using the additional files
2796 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2798 6. A Windows user reported a minor discrepancy with test 2, which turned out
2799 to be caused by a trailing space on an input line that had got lost in his
2800 copy. The trailing space was an accident, so I've just removed it.
2802 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2803 that is needed.
2805 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2806 as "const" (a) because they are and (b) because it helps the PHP
2807 maintainers who have recently made a script to detect big data structures
2808 in the php code that should be moved to the .rodata section. I remembered
2809 to update Builducptable as well, so it won't revert if ucptable.h is ever
2810 re-created.
2812 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2813 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2814 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2815 support is not required. This saves 1.5-2K of code, which is important in
2816 some applications.
2818 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2819 so as not to refer to the tables, even though these functions will never be
2820 called when UTF-8 support is disabled. Otherwise there are problems with a
2821 shared library.
2823 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2825 (a) It was defining its arguments as char * instead of void *.
2827 (b) It was assuming that all moves were upwards in memory; this was true
2828 a long time ago when I wrote it, but is no longer the case.
2830 The emulated memove() is provided for those environments that have neither
2831 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2832 is clearly not the case, as these two bugs were recently reported.
2834 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2835 and Detrail to create the HTML documentation, the .txt form of the man
2836 pages, and it removes trailing spaces from listed files. It also creates
2837 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2838 case, it wraps all the #defines with #ifndefs. This script should be run
2839 before "make dist".
2841 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2842 with Unicode property support.
2844 (a) For a maximizing quantifier, if the two different cases of the
2845 character were of different lengths in their UTF-8 codings (there are
2846 some cases like this - I found 11), and the matching function had to
2847 back up over a mixture of the two cases, it incorrectly assumed they
2848 were both the same length.
2850 (b) When PCRE was configured to use the heap rather than the stack for
2851 recursion during matching, it was not correctly preserving the data for
2852 the other case of a UTF-8 character when checking ahead for a match
2853 while processing a minimizing repeat. If the check also involved
2854 matching a wide character, but failed, corruption could cause an
2855 erroneous result when trying to check for a repeat of the original
2856 character.
2858 13. Some tidying changes to the testing mechanism:
2860 (a) The RunTest script now detects the internal link size and whether there
2861 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2862 values substituted by "configure". (The RunGrepTest script already did
2863 this for UTF-8.) The configure.ac script no longer substitutes the
2864 relevant variables.
2866 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2867 with length and offset values. This means that the output is different
2868 for different internal link sizes. Test 2 is skipped for link sizes
2869 other than 2 because of this, bypassing the problem. Unfortunately,
2870 there was also a test in test 3 (the locale tests) that used /B and
2871 failed for link sizes other than 2. Rather than cut the whole test out,
2872 I have added a new /Z option to pcretest that replaces the length and
2873 offset values with spaces. This is now used to make test 3 independent
2874 of link size. (Test 2 will be tidied up later.)
2876 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2877 segmentation fault instead of returning the appropriate error message.
2879 15. In multiline mode when the newline sequence was set to "any", the pattern
2880 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2881 This doesn't seem right; it now treats the CRLF combination as the line
2882 ending, and so does not match in that case. It's only a pattern such as ^$
2883 that would hit this one: something like ^ABC$ would have failed after \r
2884 and then tried again after \r\n.
2886 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2887 in an attempt to make files that differ only in their line terminators
2888 compare equal. This works on Linux.
2890 17. Under certain error circumstances pcregrep might try to free random memory
2891 as it exited. This is now fixed, thanks to valgrind.
2893 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2894 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2895 was because its rules for how to advance for /g after matching an empty
2896 string at the end of a line did not allow for this case. They now check for
2897 it specially.
2899 20. pcretest is supposed to handle patterns and data of any length, by
2900 extending its buffers when necessary. It was getting this wrong when the
2901 buffer for a data line had to be extended.
2903 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2904 CRLF as a newline sequence.
2906 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2907 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2908 I have nevertheless tidied it up.
2910 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2912 24. Added a man page for pcre-config.
2915 Version 7.0 19-Dec-06
2916 ---------------------
2918 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2919 moving to gcc 4.1.1.
2921 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2922 sys/time.h, which is documented as needed for this function. It doesn't
2923 seem to matter on Linux, but it showed up on some releases of OS X.
2925 3. It seems that there are systems where bytes whose values are greater than
2926 127 match isprint() in the "C" locale. The "C" locale should be the
2927 default when a C program starts up. In most systems, only ASCII printing
2928 characters match isprint(). This difference caused the output from pcretest
2929 to vary, making some of the tests fail. I have changed pcretest so that:
2931 (a) When it is outputting text in the compiled version of a pattern, bytes
2932 other than 32-126 are always shown as hex escapes.
2934 (b) When it is outputting text that is a matched part of a subject string,
2935 it does the same, unless a different locale has been set for the match
2936 (using the /L modifier). In this case, it uses isprint() to decide.
2938 4. Fixed a major bug that caused incorrect computation of the amount of memory
2939 required for a compiled pattern when options that changed within the
2940 pattern affected the logic of the preliminary scan that determines the
2941 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2942 that the computed length was too small. The symptoms of this bug were
2943 either the PCRE error "internal error: code overflow" from pcre_compile(),
2944 or a glibc crash with a message such as "pcretest: free(): invalid next
2945 size (fast)". Examples of patterns that provoked this bug (shown in
2946 pcretest format) are:
2948 /(?-x: )/x
2949 /(?x)(?-x: \s*#\s*)/
2950 /((?i)[\x{c0}])/8
2951 /(?i:[\x{c0}])/8
2953 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2954 is now done differently.
2956 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2957 wrapper classes; (b) implement a new function in the C++ scanner that is
2958 more efficient than the old way of doing things because it avoids levels of
2959 recursion in the regex matching; (c) add a paragraph to the documentation
2960 for the FullMatch() function.
2962 6. The escape sequence \n was being treated as whatever was defined as
2963 "newline". Not only was this contrary to the documentation, which states
2964 that \n is character 10 (hex 0A), but it also went horribly wrong when
2965 "newline" was defined as CRLF. This has been fixed.
2967 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2968 was being set to -1 for the "end of line" case (supposedly a value that no
2969 character can have). Though this value is never used (the check for end of
2970 line is "zero bytes in current character"), it caused compiler complaints.
2971 I've changed it to 0xffffffff.
2973 8. In pcre_version.c, the version string was being built by a sequence of
2974 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2975 string (as it is for production releases) called a macro with an empty
2976 argument. The C standard says the result of this is undefined. The gcc
2977 compiler treats it as an empty string (which was what was wanted) but it is
2978 reported that Visual C gives an error. The source has been hacked around to
2979 avoid this problem.
2981 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2982 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2983 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2984 of them did).
2986 10. Originally, pcretest opened its input and output without "b"; then I was
2987 told that "b" was needed in some environments, so it was added for release
2988 5.0 to both the input and output. (It makes no difference on Unix-like
2989 systems.) Later I was told that it is wrong for the input on Windows. I've
2990 now abstracted the modes into two macros, to make it easier to fiddle with
2991 them, and removed "b" from the input mode under Windows.
2993 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2995 12. Added -help and --help to pcretest as an official way of being reminded
2996 of the options.
2998 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2999 and pcrecpp.cc because they annoy compilers at high warning levels.
3001 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3003 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3004 scripts are not C programs :-) and which was not noticed because it works
3005 on Linux.
3007 16. pcretest is supposed to handle any length of pattern and data line (as one
3008 line or as a continued sequence of lines) by extending its input buffer if
3009 necessary. This feature was broken for very long pattern lines, leading to
3010 a string of junk being passed to pcre_compile() if the pattern was longer
3011 than about 50K.
3013 17. I have done a major re-factoring of the way pcre_compile() computes the
3014 amount of memory needed for a compiled pattern. Previously, there was code
3015 that made a preliminary scan of the pattern in order to do this. That was
3016 OK when PCRE was new, but as the facilities have expanded, it has become
3017 harder and harder to keep it in step with the real compile phase, and there
3018 have been a number of bugs (see for example, 4 above). I have now found a
3019 cunning way of running the real compile function in a "fake" mode that
3020 enables it to compute how much memory it would need, while actually only
3021 ever using a few hundred bytes of working memory and without too many
3022 tests of the mode. This should make future maintenance and development
3023 easier. A side effect of this work is that the limit of 200 on the nesting
3024 depth of parentheses has been removed (though this was never a serious
3025 limitation, I suspect). However, there is a downside: pcre_compile() now
3026 runs more slowly than before (30% or more, depending on the pattern). I
3027 hope this isn't a big issue. There is no effect on runtime performance.
3029 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3030 newline (only possible for the last line of a file) and it was a
3031 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3033 19. Added additional timing features to pcretest. (1) The -tm option now times
3034 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3035 separate command line item, by a number that specifies the number of
3036 repeats to use when timing. The default is 50000; this gives better
3037 precision, but takes uncomfortably long for very large patterns.
3039 20. Extended pcre_study() to be more clever in cases where a branch of a
3040 subpattern has no definite first character. For example, (a*|b*)[cd] would
3041 previously give no result from pcre_study(). Now it recognizes that the
3042 first character must be a, b, c, or d.
3044 21. There was an incorrect error "recursive call could loop indefinitely" if
3045 a subpattern (or the entire pattern) that was being tested for matching an
3046 empty string contained only one non-empty item after a nested subpattern.
3047 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3048 incorrectly, because the \d was being skipped in the check.
3050 22. The pcretest program now has a new pattern option /B and a command line
3051 option -b, which is equivalent to adding /B to every pattern. This causes
3052 it to show the compiled bytecode, without the additional information that
3053 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3054 is the same as /B/I).
3056 23. A new optimization is now able automatically to treat some sequences such
3057 as a*b as a*+b. More specifically, if something simple (such as a character
3058 or a simple class like \d) has an unlimited quantifier, and is followed by
3059 something that cannot possibly match the quantified thing, the quantifier
3060 is automatically "possessified".
3062 24. A recursive reference to a subpattern whose number was greater than 39
3063 went wrong under certain circumstances in UTF-8 mode. This bug could also
3064 have affected the operation of pcre_study().
3066 25. Realized that a little bit of performance could be had by replacing
3067 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3069 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3071 27. Possessive quantifiers such as a++ were previously implemented by turning
3072 them into atomic groups such as ($>a+). Now they have their own opcodes,
3073 which improves performance. This includes the automatically created ones
3074 from 23 above.
3076 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3077 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3078 the first matched character to be a colon. This applied both to named and
3079 numbered groups.
3081 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3083 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3084 building PCRE on Windows easier, so I have included it in the distribution.
3086 31. There is now a check in pcretest against a ridiculously large number being
3087 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3088 loop, the loop is abandoned.
3090 32. Forward references to subpatterns in conditions such as (?(2)...) where
3091 subpattern 2 is defined later cause pcre_compile() to search forwards in
3092 the pattern for the relevant set of parentheses. This search went wrong
3093 when there were unescaped parentheses in a character class, parentheses
3094 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3096 33. "Subroutine" calls and backreferences were previously restricted to
3097 referencing subpatterns earlier in the regex. This restriction has now
3098 been removed.
3100 34. Added a number of extra features that are going to be in Perl 5.10. On the
3101 whole, these are just syntactic alternatives for features that PCRE had
3102 previously implemented using the Python syntax or my own invention. The
3103 other formats are all retained for compatibility.
3105 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3106 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3107 also .NET compatible.
3109 (b) A recursion or subroutine call to a named group can now be defined as
3110 (?&name) as well as (?P>name).
3112 (c) A backreference to a named group can now be defined as \k<name> or
3113 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3114 5.10, are also .NET compatible.
3116 (d) A conditional reference to a named group can now use the syntax
3117 (?(<name>) or (?('name') as well as (?(name).
3119 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3120 groups (named and numbered) that are never evaluated inline, but can be
3121 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3122 is always false. There may be only one alternative in such a group.
3124 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3125 as the simple (?(R). The condition is true only if the most recent
3126 recursion is that of the given number or name. It does not search out
3127 through the entire recursion stack.
3129 (g) The escape \gN or \g{N} has been added, where N is a positive or
3130 negative number, specifying an absolute or relative reference.
3132 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3133 some "unreachable code" warnings.
3135 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3136 things, this adds five new scripts.
3138 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3139 There were also incompatibilities regarding the handling of \Q..\E inside
3140 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3141 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3143 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3144 matches an empty string, and forcibly breaks the loop. There were bugs in
3145 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3146 against aaaa the result was just "a" rather than "aaaa", for example. Two
3147 separate and independent bugs (that affected different cases) have been
3148 fixed.
3150 39. Refactored the code to abolish the use of different opcodes for small
3151 capturing bracket numbers. This is a tidy that I avoided doing when I
3152 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3153 The new approach is not only tidier, it makes it possible to reduce the
3154 memory needed to fix the previous bug (38).
3156 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3157 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3158 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3159 mode.
3161 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3162 report.
3164 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3165 copy construction and assignment in the C++ wrapper.
3167 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3168 couple of bugs that could have given wrong results in the "--newline=crlf"
3169 case.
3171 44. Added a number of casts and did some reorganization of signed/unsigned int
3172 variables following suggestions from Dair Grant. Also renamed the variable
3173 "this" as "item" because it is a C++ keyword.
3175 45. Arranged for dftables to add
3177 #include "pcre_internal.h"
3179 to pcre_chartables.c because without it, gcc 4.x may remove the array
3180 definition from the final binary if PCRE is built into a static library and
3181 dead code stripping is activated.
3183 46. For an unanchored pattern, if a match attempt fails at the start of a
3184 newline sequence, and the newline setting is CRLF or ANY, and the next two
3185 characters are CRLF, advance by two characters instead of one.
3188 Version 6.7 04-Jul-06
3189 ---------------------
3191 1. In order to handle tests when input lines are enormously long, pcretest has
3192 been re-factored so that it automatically extends its buffers when
3193 necessary. The code is crude, but this _is_ just a test program. The
3194 default size has been increased from 32K to 50K.
3196 2. The code in pcre_study() was using the value of the re argument before
3197 testing it for NULL. (Of course, in any sensible call of the function, it
3198 won't be NULL.)
3200 3. The memmove() emulation function in pcre_internal.h, which is used on
3201 systems that lack both memmove() and bcopy() - that is, hardly ever -
3202 was missing a "static" storage class specifier.
3204 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3205 containing an extended class (one that cannot be represented by a bitmap
3206 because it contains high-valued characters or Unicode property items, e.g.
3207 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3208 pattern, but PCRE should not loop if you do not (it no longer does).
3209 [Detail: two cases were found: (a) a repeated subpattern containing an
3210 extended class; (b) a recursive reference to a subpattern that followed a
3211 previous extended class. It wasn't skipping over the extended class
3212 correctly when UTF-8 mode was not set.]
3214 5. A negated single-character class was not being recognized as fixed-length
3215 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3216 compile error "lookbehind assertion is not fixed length".
3218 6. The RunPerlTest auxiliary script was showing an unexpected difference
3219 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3220 write a Perl script that can interpret lines of an input file either as
3221 byte characters or as UTF-8, which is what "perltest" was being required to
3222 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3223 can't do is switch easily at run time between having the "use utf8;" pragma
3224 or not. In the end, I fudged it by using the RunPerlTest script to insert
3225 "use utf8;" explicitly for the UTF-8 tests.
3227 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3228 the end of the subject string, contrary to the documentation and to what
3229 Perl does. This was true of both matching functions. Now it matches only at
3230 the start of the subject and immediately after *internal* newlines.
3232 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3233 a pointer to an int instead of a pointer to an unsigned long int. This
3234 caused problems on 64-bit systems.
3236 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3237 instance of the 'standard' template library not being so standard".
3239 10. There was no check on the number of named subpatterns nor the maximum
3240 length of a subpattern name. The product of these values is used to compute
3241 the size of the memory block for a compiled pattern. By supplying a very
3242 long subpattern name and a large number of named subpatterns, the size
3243 computation could be caused to overflow. This is now prevented by limiting
3244 the length of names to 32 characters, and the number of named subpatterns
3245 to 10,000.
3247 11. Subpatterns that are repeated with specific counts have to be replicated in
3248 the compiled pattern. The size of memory for this was computed from the
3249 length of the subpattern and the repeat count. The latter is limited to
3250 65535, but there was no limit on the former, meaning that integer overflow
3251 could in principle occur. The compiled length of a repeated subpattern is
3252 now limited to 30,000 bytes in order to prevent this.
3254 12. Added the optional facility to have named substrings with the same name.
3256 13. Added the ability to use a named substring as a condition, using the
3257 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3258 are numbers (not recommended). Forward references are permitted.
3260 14. Added forward references in named backreferences (if you see what I mean).
3262 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3263 pattern could run off the end of the subject. For example, the pattern
3264 "(?s)(.{1,5})"8 did this with the subject "ab".
3266 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3267 PCRE_CASELESS was set when matching characters that were quantified with ?
3268 or *.
3270 17. A character class other than a single negated character that had a minimum
3271 but no maximum quantifier - for example [ab]{6,} - was not handled
3272 correctly by pce_dfa_exec(). It would match only one character.
3274 18. A valid (though odd) pattern that looked like a POSIX character
3275 class but used an invalid character after [ (for example [[,abc,]]) caused
3276 pcre_compile() to give the error "Failed: internal error: code overflow" or
3277 in some cases to crash with a glibc free() error. This could even happen if
3278 the pattern terminated after [[ but there just happened to be a sequence of
3279 letters, a binary zero, and a closing ] in the memory that followed.
3281 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3282 over the years. Originally (before any Unicode support), just the bottom 8
3283 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3284 output from "man perlunicode" includes this:
3286 The regular expression compiler produces polymorphic opcodes. That
3287 is, the pattern adapts to the data and automatically switches to
3288 the Unicode character scheme when presented with Unicode data--or
3289 instead uses a traditional byte scheme when presented with byte
3290 data.
3292 Sadly, a wide octal escape does not cause a switch, and in a string with
3293 no other multibyte characters, these octal escapes are treated as before.
3294 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3295 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3296 Unicode string.
3298 I have not perpetrated such confusion in PCRE. Up till now, it took just
3299 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3300 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3301 translate to the appropriate multibyte character.
3303 29. Applied some refactoring to reduce the number of warnings from Microsoft
3304 and Borland compilers. This has included removing the fudge introduced
3305 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3306 a warning about an unused variable.
3308 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3309 characters since release 4.0, because Perl (from release 5.004) does not.
3310 [Or at least, is documented not to: some releases seem to be in conflict
3311 with the documentation.] However, when a pattern was studied with
3312 pcre_study() and all its branches started with \s, PCRE still included VT
3313 as a possible starting character. Of course, this did no harm; it just
3314 caused an unnecessary match attempt.
3316 22. Removed a now-redundant internal flag bit that recorded the fact that case
3317 dependency changed within the pattern. This was once needed for "required
3318 byte" processing, but is no longer used. This recovers a now-scarce options
3319 bit. Also moved the least significant internal flag bit to the most-
3320 significant bit of the word, which was not previously used (hangover from
3321 the days when it was an int rather than a uint) to free up another bit for
3322 the future.
3324 23. Added support for CRLF line endings as well as CR and LF. As well as the
3325 default being selectable at build time, it can now be changed at runtime
3326 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3327 specify that it is scanning data with non-default line endings.
3329 24. Changed the definition of CXXLINK to make it agree with the definition of
3330 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3332 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3333 recursions. This makes a big different to stack usage for some patterns.
3335 26. If a subpattern containing a named recursion or subroutine reference such
3336 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3337 the space required for the compiled pattern went wrong and gave too small a
3338 value. Depending on the environment, this could lead to "Failed: internal
3339 error: code overflow at offset 49" or "glibc detected double free or
3340 corruption" errors.
3342 27. Applied patches from Google (a) to support the new newline modes and (b) to
3343 advance over multibyte UTF-8 characters in GlobalReplace.
3345 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3346 difference for some implementation of PCRE in some Windows version.
3348 29. Added some extra testing facilities to pcretest:
3350 \q<number> in a data line sets the "match limit" value
3351 \Q<number> in a data line sets the "match recursion limt" value
3352 -S <number> sets the stack size, where <number> is in megabytes
3354 The -S option isn't available for Windows.
3357 Version 6.6 06-Feb-06
3358 ---------------------
3360 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3361 in pcreposix.h. I have copied the definition from pcre.h.
3363 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3364 because pcre.h is no longer a built file.
3366 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3367 not normally included in the compiled code.
3370 Version 6.5 01-Feb-06
3371 ---------------------
3373 1. When using the partial match feature with pcre_dfa_exec(), it was not
3374 anchoring the second and subsequent partial matches at the new starting
3375 point. This could lead to incorrect results. For example, with the pattern
3376 /1234/, partially matching against "123" and then "a4" gave a match.
3378 2. Changes to pcregrep:
3380 (a) All non-match returns from pcre_exec() were being treated as failures
3381 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3382 error message is output. Some extra information is given for the
3384 probably the only errors that are likely to be caused by users (by
3385 specifying a regex that has nested indefinite repeats, for instance).
3386 If there are more than 20 of these errors, pcregrep is abandoned.
3388 (b) A binary zero was treated as data while matching, but terminated the
3389 output line if it was written out. This has been fixed: binary zeroes
3390 are now no different to any other data bytes.
3392 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3393 used to set a locale for matching. The --locale=xxxx long option has
3394 been added (no short equivalent) to specify a locale explicitly on the
3395 pcregrep command, overriding the environment variables.
3397 (d) When -B was used with -n, some line numbers in the output were one less
3398 than they should have been.
3400 (e) Added the -o (--only-matching) option.
3402 (f) If -A or -C was used with -c (count only), some lines of context were
3403 accidentally printed for the final match.
3405 (g) Added the -H (--with-filename) option.
3407 (h) The combination of options -rh failed to suppress file names for files
3408 that were found from directory arguments.
3410 (i) Added the -D (--devices) and -d (--directories) options.
3412 (j) Added the -F (--fixed-strings) option.
3414 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3416 (l) Added the --colo(u)r option.
3418 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3419 is not present by default.
3421 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3422 items such as (?R) or (?1), when the recursion could match a number of
3423 alternatives. If it matched one of the alternatives, but subsequently,
3424 outside the recursion, there was a failure, the code tried to back up into
3425 the recursion. However, because of the way PCRE is implemented, this is not
3426 possible, and the result was an incorrect result from the match.
3428 In order to prevent this happening, the specification of recursion has
3429 been changed so that all such subpatterns are automatically treated as
3430 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3432 4. I had overlooked the fact that, in some locales, there are characters for
3433 which isalpha() is true but neither isupper() nor islower() are true. In
3434 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3435 and ordfeminine) are like this. This affected the treatment of \w and \W
3436 when they appeared in character classes, but not when they appeared outside
3437 a character class. The bit map for "word" characters is now created
3438 separately from the results of isalnum() instead of just taking it from the
3439 upper, lower, and digit maps. (Plus the underscore character, of course.)
3441 5. The above bug also affected the handling of POSIX character classes such as
3442 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3443 permanent tables. Instead, the bit maps for such a class were previously
3444 created as the appropriate unions of the upper, lower, and digit bitmaps.
3445 Now they are created by subtraction from the [[:word:]] class, which has
3446 its own bitmap.
3448 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3449 It is created by subtracting the vertical space characters (\x09, \x0a,
3450 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3451 subtraction was done in the overall bitmap for a character class, meaning
3452 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3453 be recognized. This bug has been fixed.
3455 7. Patches from the folks at Google:
3457 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3458 real life, but is still worth protecting against".
3460 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3461 regular expressions".
3463 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3464 have it.
3466 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3467 "configure" and the latter not, in order to fix a problem somebody had
3468 with compiling the Arg class on HP-UX.
3470 (e) Improve the error-handling of the C++ wrapper a little bit.
3472 (f) New tests for checking recursion limiting.
3474 8. The pcre_memmove() function, which is used only if the environment does not
3475 have a standard memmove() function (and is therefore rarely compiled),
3476 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3477 returning a result (though PCRE never actually uses the result).
3479 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3480 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3481 returned instead of calling malloc() with an overflowing number that would
3482 most likely cause subsequent chaos.
3484 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3486 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3487 with this option is matched, the nmatch and pmatch options of regexec() are
3488 ignored.
3490 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3491 provided in case anyone wants to the the POSIX interface with UTF-8
3492 strings.
3494 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3495 C++ linking (needed for some HP-UX environments).
3497 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3498 (unused parameter) and in the pcre_printint() function (omitted "default"
3499 switch label when the default is to do nothing).
3501 15. Added some code to make it possible, when PCRE is compiled as a C++
3502 library, to replace subject pointers for pcre_exec() with a smart pointer
3503 class, thus making it possible to process discontinuous strings.
3505 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3506 much the same function. They were added by different people who were trying
3507 to make PCRE easy to compile on non-Unix systems. It has been suggested
3508 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3509 for compiling on Windows systems. I have therefore replaced it with
3510 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3511 defaults to "extern" for C or "extern C" for C++, which works fine on
3512 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3513 SCOPE with something explicit in config.h. In addition:
3515 (a) pcreposix.h still had just "extern" instead of either of these macros;
3516 I have replaced it with PCRE_DATA_SCOPE.
3518 (b) Functions such as _pcre_xclass(), which are internal to the library,
3519 but external in the C sense, all had PCRE_EXPORT in their definitions.
3520 This is apparently wrong for the Windows case, so I have removed it.
3521 (It makes no difference on Unix-like systems.)
3523 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3524 of recursive calls to match(). This is different to MATCH_LIMIT because
3525 that limits the total number of calls to match(), not all of which increase
3526 the depth of recursion. Limiting the recursion depth limits the amount of
3527 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3528 when PCRE is compiled, and changed at run time. A patch from Google adds
3529 this functionality to the C++ interface.
3531 18. Changes to the handling of Unicode character properties:
3533 (a) Updated the table to Unicode 4.1.0.
3535 (b) Recognize characters that are not in the table as "Cn" (undefined).
3537 (c) I revised the way the table is implemented to a much improved format
3538 which includes recognition of ranges. It now supports the ranges that
3539 are defined in UnicodeData.txt, and it also amalgamates other
3540 characters into ranges. This has reduced the number of entries in the
3541 table from around 16,000 to around 3,000, thus reducing its size
3542 considerably. I realized I did not need to use a tree structure after
3543 all - a binary chop search is just as efficient. Having reduced the
3544 number of entries, I extended their size from 6 bytes to 8 bytes to
3545 allow for more data.
3547 (d) Added support for Unicode script names via properties such as \p{Han}.
3549 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3550 matching that character.
3552 20. When matching a repeated Unicode property with a minimum greater than zero,
3553 (for example \pL{2,}), PCRE could look past the end of the subject if it
3554 reached it while seeking the minimum number of characters. This could
3555 happen only if some of the characters were more than one byte long, because
3556 there is a check for at least the minimum number of bytes.
3558 21. Refactored the implementation of \p and \P so as to be more general, to
3559 allow for more different types of property in future. This has changed the
3560 compiled form incompatibly. Anybody with saved compiled patterns that use
3561 \p or \P will have to recompile them.
3563 22. Added "Any" and "L&" to the supported property types.
3565 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3566 but give a compile time error if the value is greater than 0xff.
3568 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3569 accidentally not being installed or uninstalled.
3571 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3572 made were to insert the current release number. This seemed silly, because
3573 it made things harder for people building PCRE on systems that don't run
3574 "configure". I have turned pcre.h into a distributed file, no longer built
3575 by "configure", with the version identification directly included. There is
3576 no longer a pcre.h.in file.
3578 However, this change necessitated a change to the pcre-config script as
3579 well. It is built from pcre-config.in, and one of the substitutions was the
3580 release number. I have updated configure.ac so that ./configure now finds
3581 the release number by grepping pcre.h.
3583 26. Added the ability to run the tests under valgrind.
3586 Version 6.4 05-Sep-05
3587 ---------------------
3589 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3590 "--" to be printed when multiple files were scanned, even when none of the
3591 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3592 consider it to be a bug, and have restored the previous behaviour.
3594 2. A couple of code tidies to get rid of compiler warnings.
3596 3. The pcretest program used to cheat by referring to symbols in the library
3597 whose names begin with _pcre_. These are internal symbols that are not
3598 really supposed to be visible externally, and in some environments it is
3599 possible to suppress them. The cheating is now confined to including
3600 certain files from the library's source, which is a bit cleaner.
3602 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3603 file's purpose clearer.
3605 5. Reorganized pcre_ucp_findchar().
3608 Version 6.3 15-Aug-05
3609 ---------------------
3611 1. The file libpcre.pc.in did not have general read permission in the tarball.
3613 2. There were some problems when building without C++ support:
3615 (a) If C++ support was not built, "make install" and "make test" still
3616 tried to test it.
3618 (b) There were problems when the value of CXX was explicitly set. Some
3619 changes have been made to try to fix these, and ...
3621 (c) --disable-cpp can now be used to explicitly disable C++ support.
3623 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3624 backslash in a target when C++ was disabled. This confuses some
3625 versions of "make", apparently. Using an intermediate variable solves
3626 this. (Same for CPP_LOBJ.)
3628 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3629 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3630 necessary on certain architectures.
3632 4. Added a setting of -export-symbols-regex to the link command to remove
3633 those symbols that are exported in the C sense, but actually are local
3634 within the library, and not documented. Their names all begin with
3635 "_pcre_". This is not a perfect job, because (a) we have to except some
3636 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3637 available (and never for static libraries). I have made a note to try to
3638 find a way round (a) in the future.
3641 Version 6.2 01-Aug-05
3642 ---------------------
3644 1. There was no test for integer overflow of quantifier values. A construction
3645 such as {1111111111111111} would give undefined results. What is worse, if
3646 a minimum quantifier for a parenthesized subpattern overflowed and became
3647 negative, the calculation of the memory size went wrong. This could have
3648 led to memory overwriting.
3650 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3652 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3653 operating environments where this matters.
3655 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3656 PCRE options from within the C++ wrapper.
3658 5. Named capturing subpatterns were not being correctly counted when a pattern
3659 was compiled. This caused two problems: (a) If there were more than 100
3660 such subpatterns, the calculation of the memory needed for the whole
3661 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3662 back references of the form \12, where the number was greater than 9, were
3663 not recognized as back references, even though there were sufficient
3664 previous subpatterns.
3666 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3667 versions of gcc, e.g. 2.95.4.
3670 Version 6.1 21-Jun-05
3671 ---------------------
3673 1. There was one reference to the variable "posix" in pcretest.c that was not
3674 surrounded by "#if !defined NOPOSIX".
3676 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3677 the cross-check on the old pcre_info() function, for the benefit of the
3678 cut-down version of PCRE that is currently imported into Exim.
3680 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3681 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3682 patterns that start (?i) but I don't think that matters. The use of (?i) is
3683 just an example; this all applies to the other options as well.
3685 4. Since libtool seems to echo the compile commands it is issuing, the output
3686 from "make" can be reduced a bit by putting "@" in front of each libtool
3687 compile command.
3689 5. Patch from the folks at Google for configure.in to be a bit more thorough
3690 in checking for a suitable C++ installation before trying to compile the
3691 C++ stuff. This should fix a reported problem when a compiler was present,
3692 but no suitable headers.
3694 6. The man pages all had just "PCRE" as their title. I have changed them to
3695 be the relevant file name. I have also arranged that these names are
3696 retained in the file doc/pcre.txt, which is a concatenation in text format
3697 of all the man pages except the little individual ones for each function.
3699 7. The NON-UNIX-USE file had not been updated for the different set of source
3700 files that come with release 6. I also added a few comments about the C++
3701 wrapper.
3704 Version 6.0 07-Jun-05
3705 ---------------------
3707 1. Some minor internal re-organization to help with my DFA experiments.
3709 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3710 didn't matter for the library itself when fully configured, but did matter
3711 when compiling without UCP support, or within Exim, where the ucp files are
3712 not imported.
3714 3. Refactoring of the library code to split up the various functions into
3715 different source modules. The addition of the new DFA matching code (see
3716 below) to a single monolithic source would have made it really too
3717 unwieldy, quite apart from causing all the code to be include in a
3718 statically linked application, when only some functions are used. This is
3719 relevant even without the DFA addition now that patterns can be compiled in
3720 one application and matched in another.
3722 The downside of splitting up is that there have to be some external
3723 functions and data tables that are used internally in different modules of
3724 the library but which are not part of the API. These have all had their
3725 names changed to start with "_pcre_" so that they are unlikely to clash
3726 with other external names.
3728 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3729 a different (DFA) algorithm. Although it is slower than the original
3730 function, it does have some advantages for certain types of matching
3731 problem.
3733 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3734 including restarting after a partial match.
3736 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3737 defined when compiling for Windows was sent to me. I have put it into the
3738 code, though I have no means of testing or verifying it.
3740 7. Added the pcre_refcount() auxiliary function.
3742 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3743 match before or at the first newline in the subject string. In pcretest,
3744 the /f option on a pattern can be used to set this.
3746 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3747 would behave wrongly. This has been present in PCRE since release 4.0.
3749 10. A number of changes to the pcregrep command:
3751 (a) Refactored how -x works; insert ^(...)$ instead of setting
3752 PCRE_ANCHORED and checking the length, in preparation for adding
3753 something similar for -w.
3755 (b) Added the -w (match as a word) option.
3757 (c) Refactored the way lines are read and buffered so as to have more
3758 than one at a time available.
3760 (d) Implemented a pcregrep test script.
3762 (e) Added the -M (multiline match) option. This allows patterns to match
3763 over several lines of the subject. The buffering ensures that at least
3764 8K, or the rest of the document (whichever is the shorter) is available
3765 for matching (and similarly the previous 8K for lookbehind assertions).
3767 (f) Changed the --help output so that it now says
3769 -w, --word-regex(p)
3771 instead of two lines, one with "regex" and the other with "regexp"
3772 because that confused at least one person since the short forms are the
3773 same. (This required a bit of code, as the output is generated
3774 automatically from a table. It wasn't just a text change.)
3776 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3777 option but starts with a hyphen. Could be a pattern or a path name
3778 starting with a hyphen, for instance.
3780 (h) "-" can be given as a file name to represent stdin.
3782 (i) When file names are being printed, "(standard input)" is used for
3783 the standard input, for compatibility with GNU grep. Previously
3784 "<stdin>" was used.
3786 (j) The option --label=xxx can be used to supply a name to be used for
3787 stdin when file names are being printed. There is no short form.
3789 (k) Re-factored the options decoding logic because we are going to add
3790 two more options that take data. Such options can now be given in four
3791 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3793 (l) Added the -A, -B, and -C options for requesting that lines of context
3794 around matches be printed.
3796 (m) Added the -L option to print the names of files that do not contain
3797 any matching lines, that is, the complement of -l.
3799 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3800 continue to scan other files.
3802 (o) The -s option was incorrectly implemented. For compatibility with other
3803 greps, it now suppresses the error message for a non-existent or non-
3804 accessible file (but not the return code). There is a new option called
3805 -q that suppresses the output of matching lines, which was what -s was
3806 previously doing.
3808 (p) Added --include and --exclude options to specify files for inclusion
3809 and exclusion when recursing.
3811 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3812 Hopefully, it now does.
3814 12. Missing cast in pcre_study().
3816 13. Added an "uninstall" target to the makefile.
3818 14. Replaced "extern" in the function prototypes in Makefile.in with
3819 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3820 world, but is set differently for Windows.
3822 15. Added a second compiling function called pcre_compile2(). The only
3823 difference is that it has an extra argument, which is a pointer to an
3824 integer error code. When there is a compile-time failure, this is set
3825 non-zero, in addition to the error test pointer being set to point to an
3826 error message. The new argument may be NULL if no error number is required
3827 (but then you may as well call pcre_compile(), which is now just a
3828 wrapper). This facility is provided because some applications need a
3829 numeric error indication, but it has also enabled me to tidy up the way
3830 compile-time errors are handled in the POSIX wrapper.
3832 16. Added VPATH=.libs to the makefile; this should help when building with one
3833 prefix path and installing with another. (Or so I'm told by someone who
3834 knows more about this stuff than I do.)
3836 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3837 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3838 match everything, including newlines. This is not POSIX-compatible, but
3839 somebody wanted the feature. From pcretest it can be activated by using
3840 both the P and the s flags.
3842 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3844 19. libpcre.pc was being incorrectly installed as executable.
3846 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3847 it now also looks for '\r' so that it will work unmodified on Windows.
3849 21. Added Google's contributed C++ wrapper to the distribution.
3851 22. Added some untidy missing memory free() calls in pcretest, to keep
3852 Electric Fence happy when testing.
3856 Version 5.0 13-Sep-04
3857 ---------------------
3859 1. Internal change: literal characters are no longer packed up into items
3860 containing multiple characters in a single byte-string. Each character
3861 is now matched using a separate opcode. However, there may be more than one
3862 byte in the character in UTF-8 mode.
3864 2. The pcre_callout_block structure has two new fields: pattern_position and
3865 next_item_length. These contain the offset in the pattern to the next match
3866 item, and its length, respectively.
3868 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3869 insertion of callouts before each pattern item. Added the /C option to
3870 pcretest to make use of this.
3872 4. On the advice of a Windows user, the lines
3874 #if defined(_WIN32) || defined(WIN32)
3875 _setmode( _fileno( stdout ), 0x8000 );
3876 #endif /* defined(_WIN32) || defined(WIN32) */
3878 have been added to the source of pcretest. This apparently does useful
3879 magic in relation to line terminators.
3881 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3882 for the benefit of those environments where the "b" makes a difference.
3884 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3885 to know about it. I have put a hack into configure.in that adds in code
3886 to set GCC=yes if CC=icc. This seems to end up at a point in the
3887 generated configure script that is early enough to affect the setting of
3888 compiler options, which is what is needed, but I have no means of testing
3889 whether it really works. (The user who reported this had patched the
3890 generated configure script, which of course I cannot do.)
3892 LATER: After change 22 below (new libtool files), the configure script
3893 seems to know about icc (and also ecc). Therefore, I have commented out
3894 this hack in configure.in.
3896 7. Added support for pkg-config (2 patches were sent in).
3898 8. Negated POSIX character classes that used a combination of internal tables
3899 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3900 [[:^ascii]]. Typically, they would match almost any characters. The other
3901 POSIX classes were not broken in this way.
3903 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3904 to find the match, as PCRE was deluded into thinking that the match had to
3905 start at the start point or following a newline. The same bug applied to
3906 patterns with negative forward assertions or any backward assertions
3907 preceding ".*" at the start, unless the pattern required a fixed first
3908 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3910 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3911 starting at the last subject character, bytes beyond the end of the subject
3912 string were read.
3914 11. Renamed the variable "class" as "classbits" to make life easier for C++
3915 users. (Previously there was a macro definition, but it apparently wasn't
3916 enough.)
3918 12. Added the new field "tables" to the extra data so that tables can be passed
3919 in at exec time, or the internal tables can be re-selected. This allows
3920 a compiled regex to be saved and re-used at a later time by a different
3921 program that might have everything at different addresses.
3923 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3924 -R library as well as a -L library.
3926 14. The debugging options of pcretest (-d on the command line or D on a
3927 pattern) showed incorrect output for anything following an extended class
3928 that contained multibyte characters and which was followed by a quantifier.
3930 15. Added optional support for general category Unicode character properties
3931 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3932 support. It adds about 90K to the size of the library. The meanings of the
3933 inbuilt class escapes such as \d and \s have NOT been changed.
3935 16. Updated pcredemo.c to include calls to free() to release the memory for the
3936 compiled pattern.
3938 17. The generated file chartables.c was being created in the source directory
3939 instead of in the building directory. This caused the build to fail if the
3940 source directory was different from the building directory, and was
3941 read-only.
3943 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3944 file. No doubt somebody will tell me if they don't make sense... Also added
3945 Dan Mooney's comments about building on OpenVMS.
3947 19. Added support for partial matching via the PCRE_PARTIAL option for
3948 pcre_exec() and the \P data escape in pcretest.
3950 20. Extended pcretest with 3 new pattern features:
3952 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3953 write the compiled pattern to the file whose name is "rest-of-line".
3954 This is a straight binary dump of the data, with the saved pointer to
3955 the character tables forced to be NULL. The study data, if any, is
3956 written too. After writing, pcretest reads a new pattern.
3958 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3959 compiled pattern from the given file. There must not be any
3960 occurrences of "<" in the file name (pretty unlikely); if there are,
3961 pcretest will instead treat the initial "<" as a pattern delimiter.
3962 After reading in the pattern, pcretest goes on to read data lines as
3963 usual.
3965 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3966 and 16-bit fields in a compiled pattern, to simulate a pattern that
3967 was compiled on a host of opposite endianness.
3969 21. The pcre-exec() function can now cope with patterns that were compiled on
3970 hosts of opposite endianness, with this restriction:
3972 As for any compiled expression that is saved and used later, the tables
3973 pointer field cannot be preserved; the extra_data field in the arguments
3974 to pcre_exec() should be used to pass in a tables address if a value
3975 other than the default internal tables were used at compile time.
3977 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3978 now diagnosed as an error. Previously, most of the time, a negative number
3979 would have been treated as zero, but if in addition "ovector" was passed as
3980 NULL, a crash could occur.
3982 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3983 new versions from the libtool 1.5 distribution (the last one is a copy of
3984 a file called libtool.m4). This seems to have fixed the need to patch
3985 "configure" to support Darwin 1.3 (which I used to do). However, I still
3986 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3987 workstation).
3989 24. Changed the PCRE licence to be the more standard "BSD" licence.
3992 Version 4.5 01-Dec-03
3993 ---------------------
3995 1. There has been some re-arrangement of the code for the match() function so
3996 that it can be compiled in a version that does not call itself recursively.
3997 Instead, it keeps those local variables that need separate instances for
3998 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3999 needs to "recurse". Keeping track of where control must go is done by means
4000 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4001 hide most of the details from the main code, and operates only if
4002 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4003 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4004 operating.
4006 To make it easier for callers to provide specially tailored get/free
4007 functions for this usage, two new functions, pcre_stack_malloc, and
4008 pcre_stack_free, are used. They are always called in strict stacking order,
4009 and the size of block requested is always the same.
4011 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4012 PCRE has been compiled to use the stack or the heap for recursion. The
4013 -C option of pcretest uses this to show which version is compiled.
4015 A new data escape \S, is added to pcretest; it causes the amounts of store
4016 obtained and freed by both kinds of malloc/free at match time to be added
4017 to the output.
4019 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4020 what's available on my current Linux desktop machine.
4022 3. When matching a UTF-8 string, the test for a valid string at the start has
4023 been extended. If start_offset is not zero, PCRE now checks that it points
4024 to a byte that is the start of a UTF-8 character. If not, it returns
4025 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4026 this is necessary because there may be backward assertions in the pattern.
4027 When matching the same subject several times, it may save resources to use
4028 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4030 4. The code for checking the validity of UTF-8 strings has been tightened so
4031 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4032 containing "overlong sequences".
4034 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4035 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4036 should have been "&", but it just so happened that all the cases this let
4037 through by mistake were picked up later in the function.
4039 6. I had used a variable called "isblank" - this is a C99 function, causing
4040 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4042 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4043 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4044 so the version is shown; (c) stop "make test" echoing "./RunTest".
4046 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4048 9. The prototype for memmove() for systems that don't have it was using
4049 size_t, but the inclusion of the header that defines size_t was later. I've
4050 moved the #includes for the C headers earlier to avoid this.
4052 10. Added some adjustments to the code to make it easier to compiler on certain
4053 special systems:
4055 (a) Some "const" qualifiers were missing.
4056 (b) Added the macro EXPORT before all exported functions; by default this
4057 is defined to be empty.
4058 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4059 that it reads its output file name as an argument instead of writing
4060 to the standard output and assuming this can be redirected.
4062 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4063 class containing characters with values greater than 255, PCRE compilation
4064 went into a loop.
4066 12. A recursive reference to a subpattern that was within another subpattern
4067 that had a minimum quantifier of zero caused PCRE to crash. For example,
4068 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4069 recursion. If the recursively-called subpattern itself had a zero repeat,
4070 that was OK.
4072 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4073 buffer into which it was copied (for escape processing) was still set at
4074 1024, so long lines caused crashes.
4076 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4077 "internal error: code overflow...". This applied to any character class
4078 that was followed by a possessive quantifier.
4080 15. Modified the Makefile to add libpcre.la as a prerequisite for
4081 libpcreposix.la because I was told this is needed for a parallel build to
4082 work.
4084 16. If a pattern that contained .* following optional items at the start was
4085 studied, the wrong optimizing data was generated, leading to matching
4086 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4087 matching string must start with a or b or c. The correct conclusion for
4088 this pattern is that a match can start with any character.
4091 Version 4.4 13-Aug-03
4092 ---------------------
4094 1. In UTF-8 mode, a character class containing characters with values between
4095 127 and 255 was not handled correctly if the compiled pattern was studied.
4096 In fixing this, I have also improved the studying algorithm for such
4097 classes (slightly).
4099 2. Three internal functions had redundant arguments passed to them. Removal
4100 might give a very teeny performance improvement.
4102 3. Documentation bug: the value of the capture_top field in a callout is *one
4103 more than* the number of the hightest numbered captured substring.
4105 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4106 in incorrectly linking with a previously installed version. They now link
4107 explicitly with libpcre.la.
4109 5. configure.in no longer needs to recognize Cygwin specially.
4111 6. A problem in pcre.in for Windows platforms is fixed.
4113 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4114 pcretest, it used to include the size of the study block as part of its
4115 output. Unfortunately, the structure contains a field that has a different
4116 size on different hardware architectures. This meant that the tests that
4117 showed this size failed. As the block is currently always of a fixed size,
4118 this information isn't actually particularly useful in pcretest output, so
4119 I have just removed it.
4121 8. Three pre-processor statements accidentally did not start in column 1.
4122 Sadly, there are *still* compilers around that complain, even though
4123 standard C has not required this for well over a decade. Sigh.
4125 9. In pcretest, the code for checking callouts passed small integers in the
4126 callout_data field, which is a void * field. However, some picky compilers
4127 complained about the casts involved for this on 64-bit systems. Now
4128 pcretest passes the address of the small integer instead, which should get
4129 rid of the warnings.
4131 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4132 both compile and run time, and gives an error if an invalid UTF-8 sequence
4133 is found. There is a option for disabling this check in cases where the
4134 string is known to be correct and/or the maximum performance is wanted.
4136 11. In response to a bug report, I changed one line in Makefile.in from
4138 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4139 to
4140 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4142 to look similar to other lines, but I have no way of telling whether this
4143 is the right thing to do, as I do not use Windows. No doubt I'll get told
4144 if it's wrong...
4147 Version 4.3 21-May-03
4148 ---------------------
4150 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4151 Makefile.
4153 2. Some refactoring to improve the quality of the code:
4155 (i) The utf8_table... variables are now declared "const".
4157 (ii) The code for \cx, which used the "case flipping" table to upper case
4158 lower case letters, now just substracts 32. This is ASCII-specific,
4159 but the whole concept of \cx is ASCII-specific, so it seems
4160 reasonable.
4162 (iii) PCRE was using its character types table to recognize decimal and
4163 hexadecimal digits in the pattern. This is silly, because it handles
4164 only 0-9, a-f, and A-F, but the character types table is locale-
4165 specific, which means strange things might happen. A private
4166 table is now used for this - though it costs 256 bytes, a table is
4167 much faster than multiple explicit tests. Of course, the standard
4168 character types table is still used for matching digits in subject
4169 strings against \d.
4171 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4172 ending in _t are). So I've renamed it as ESC_tee.
4174 3. The first argument for regexec() in the POSIX wrapper should have been
4175 defined as "const".
4177 4. Changed pcretest to use malloc() for its buffers so that they can be
4178 Electric Fenced for debugging.
4180 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4181 to read one or more bytes before the start of the subject string. Often this
4182 had no effect on PCRE's behaviour, but in some circumstances it could
4183 provoke a segmentation fault.
4185 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4186 to try to read one or more bytes before the start of the subject string.
4188 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4189 UTF-8 support could misbehave in various ways if the subject string
4190 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4191 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4192 back over UTF-8 characters.)
4195 Version 4.2 14-Apr-03
4196 ---------------------
4198 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4200 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4201 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4202 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4203 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4204 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4206 Note: automatic setting of the BUILD variables is not yet working
4207 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4208 compile-time but not at link-time
4209 [LINK]: use for linking executables only
4210 make different versions for Windows and non-Windows
4211 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4212 libraries
4213 [LINK_FOR_BUILD]: new variable
4214 [OBJEXT]: use throughout
4215 [EXEEXT]: use throughout
4216 <winshared>: new target
4217 <wininstall>: new target
4218 <dftables.o>: use native compiler
4219 <dftables>: use native linker
4220 <install>: handle Windows platform correctly
4221 <clean>: ditto
4222 <check>: ditto
4223 copy DLL to top builddir before testing
4225 As part of these changes, -no-undefined was removed again. This was reported
4226 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4227 in any case.
4229 3. Some tidies to get rid of compiler warnings:
4231 . In the match_data structure, match_limit was an unsigned long int, whereas
4232 match_call_count was an int. I've made them both unsigned long ints.
4234 . In pcretest the fact that a const uschar * doesn't automatically cast to
4235 a void * provoked a warning.
4237 . Turning on some more compiler warnings threw up some "shadow" variables
4238 and a few more missing casts.
4240 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4241 option, a class that contained a single character with a value between 128
4242 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4244 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4245 option, a class that contained several characters, but with at least one
4246 whose value was between 128 and 255 caused PCRE to crash.
4249 Version 4.1 12-Mar-03
4250 ---------------------
4252 1. Compiling with gcc -pedantic found a couple of places where casts were
4253 needed, and a string in dftables.c that was longer than standard compilers are
4254 required to support.
4256 2. Compiling with Sun's compiler found a few more places where the code could
4257 be tidied up in order to avoid warnings.
4259 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4260 first of these names is deprecated in the latest Autoconf in favour of the name
4261 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4262 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4263 analogy I have changed it to CFLAGS_FOR_BUILD.
4265 4. Added -no-undefined to the linking command in the Makefile, because this is
4266 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4267 linking step for the pcreposix library.
4269 5. PCRE was failing to diagnose the case of two named groups with the same
4270 name.
4272 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4273 literal character that is needed in the subject for a match, and scans along to
4274 ensure that it is present before embarking on the full matching process. This
4275 saves time in cases of nested unlimited repeats that are never going to match.
4276 Problem: the scan can take a lot of time if the subject is very long (e.g.
4277 megabytes), thus penalizing straightforward matches. It is now done only if the
4278 amount of subject to be scanned is less than 1000 bytes.
4280 7. A lesser problem with the same optimization is that it was recording the
4281 first character of an anchored pattern as "needed", thus provoking a search
4282 right along the subject, even when the first match of the pattern was going to
4283 fail. The "needed" character is now not set for anchored patterns, unless it
4284 follows something in the pattern that is of non-fixed length. Thus, it still
4285 fulfils its original purpose of finding quick non-matches in cases of nested
4286 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4289 Version 4.0 17-Feb-03
4290 ---------------------
4292 1. If a comment in an extended regex that started immediately after a meta-item
4293 extended to the end of string, PCRE compiled incorrect data. This could lead to
4294 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4296 2. Moved to autoconf 2.53 and libtool 1.4.2.
4298 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4299 the special perltest8 script is no longer needed - all the tests can be run
4300 from a single perltest script.
4302 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4303 by \s. It has now been removed in PCRE. This means it isn't recognized as
4304 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4305 class [:space:] *does* include VT, thereby creating a mess.
4307 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4308 space and tab.
4310 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4311 its new features into the main test script, reducing the number of scripts.
4313 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4314 were backward compatible, and made the (?i) apply to the whole pattern, as if
4315 /i were given. Now it behaves more logically, and applies the option setting
4316 only to what follows. PCRE has been changed to follow suit. However, if it
4317 finds options settings right at the start of the pattern, it extracts them into
4318 the global options, as before. Thus, they show up in the info data.
4320 8. Added support for the \Q...\E escape sequence. Characters in between are
4321 treated as literals. This is slightly different from Perl in that $ and @ are
4322 also handled as literals inside the quotes. In Perl, they will cause variable
4323 interpolation. Note the following examples:
4325 Pattern PCRE matches Perl matches
4327 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4328 \Qabc\$xyz\E abc\$xyz abc\$xyz
4329 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4331 For compatibility with Perl, \Q...\E sequences are recognized inside character
4332 classes as well as outside them.
4334 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4335 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4336 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4337 signed/unsigned warnings.
4339 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4340 option for pcretest, so I've replaced it by a simple function that does just
4341 that job.
4343 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4344 "pcregrep -".
4346 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4347 Java package. This provides some syntactic sugar for simple cases of what my
4348 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4349 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4350 item, you can use this simplified notation. Note that only makes sense with
4351 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4352 greediness, whatever the setting of the PCRE_UNGREEDY option.
4354 13. A change of greediness default within a pattern was not taking effect at
4355 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4356 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4357 was abstracted outside.
4359 14. PCRE now supports the \G assertion. It is true when the current matching
4360 position is at the start point of the match. This differs from \A when the
4361 starting offset is non-zero. Used with the /g option of pcretest (or similar
4362 code), it works in the same way as it does for Perl's /g option. If all
4363 alternatives of a regex begin with \G, the expression is anchored to the start
4364 match position, and the "anchored" flag is set in the compiled expression.
4366 15. Some bugs concerning the handling of certain option changes within patterns
4367 have been fixed. These applied to options other than (?ims). For example,
4368 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4369 been the other way round. Some of this was related to change 7 above.
4371 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4372 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4373 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4374 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4376 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4377 mode. Unlike ".", it always matches newline, whatever the setting of
4378 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4379 assertions. Perl allows it, but it doesn't (in general) work because it can't
4380 calculate the length of the lookbehind. At least, that's the case for Perl
4381 5.8.0 - I've been told they are going to document that it doesn't work in
4382 future.
4384 18. Added an error diagnosis for escapes that PCRE does not support: these are
4385 \L, \l, \N, \P, \p, \U, \u, and \X.
4387 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4388 reading past the end of the pattern in cases such as /[abcd/.
4390 20. PCRE was getting more memory than necessary for patterns with classes that
4391 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4393 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4394 compiling PCRE for use with Virtual Pascal.
4396 22. Small fix to the Makefile to make it work properly if the build is done
4397 outside the source tree.
4399 23. Added a new extension: a condition to go with recursion. If a conditional
4400 subpattern starts with (?(R) the "true" branch is used if recursion has
4401 happened, whereas the "false" branch is used only at the top level.
4403 24. When there was a very long string of literal characters (over 255 bytes
4404 without UTF support, over 250 bytes with UTF support), the computation of how
4405 much memory was required could be incorrect, leading to segfaults or other
4406 strange effects.
4408 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4409 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4410 there was a subsequent back reference to those brackets. This meant that, for
4411 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4412 possible to check for precisely this case. All we can do is abandon the
4413 optimization if .* occurs inside capturing brackets when there are any back
4414 references whatsoever. (See below for a better fix that came later.)
4416 26. The handling of the optimization for finding the first character of a
4417 non-anchored pattern, and for finding a character that is required later in the
4418 match were failing in some cases. This didn't break the matching; it just
4419 failed to optimize when it could. The way this is done has been re-implemented.
4421 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4423 28. Added a new feature that provides some of the functionality that Perl
4424 provides with (?{...}). The facility is termed a "callout". The way it is done
4425 in PCRE is for the caller to provide an optional function, by setting
4426 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4427 global variable. By default it is unset, which disables all calling out. To get
4428 the function called, the regex must include (?C) at appropriate points. This
4429 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4430 This provides a means of identifying different callout points. When PCRE
4431 reaches such a point in the regex, if pcre_callout has been set, the external
4432 function is called. It is provided with data in a structure called
4433 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4434 matching continues; if it returns a non-zero value, the match at the current
4435 point fails. However, backtracking will occur if possible. [This was changed
4436 later and other features added - see item 49 below.]
4438 29. pcretest is upgraded to test the callout functionality. It provides a
4439 callout function that displays information. By default, it shows the start of
4440 the match and the current position in the text. There are some new data escapes
4441 to vary what happens:
4443 \C+ in addition, show current contents of captured substrings
4444 \C- do not supply a callout function
4445 \C!n return 1 when callout number n is reached
4446 \C!n!m return 1 when callout number n is reached for the mth time
4448 30. If pcregrep was called with the -l option and just a single file name, it
4449 output "<stdin>" if a match was found, instead of the file name.
4451 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4452 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4453 pcre_exec(). This saves a malloc/free per call. The default value of
4454 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4455 when configuring.
4457 32. The default maximum size of a compiled pattern is 64K. There have been a
4458 few cases of people hitting this limit. The code now uses macros to handle the
4459 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4460 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4461 configuring. Tests 2 and 5 work only with 2-byte links because they output
4462 debugging information about compiled patterns.
4464 33. Internal code re-arrangements:
4466 (a) Moved the debugging function for printing out a compiled regex into
4467 its own source file (printint.c) and used #include to pull it into
4468 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4469 separate copies.
4471 (b) Defined the list of op-code names for debugging as a macro in
4472 internal.h so that it is next to the definition of the opcodes.
4474 (c) Defined a table of op-code lengths for simpler skipping along compiled
4475 code. This is again a macro in internal.h so that it is next to the
4476 definition of the opcodes.
4478 34. Added support for recursive calls to individual subpatterns, along the
4479 lines of Robin Houston's patch (but implemented somewhat differently).
4481 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4482 allow it to read and process whole directories in Win32. This code was
4483 contributed by Lionel Fourquaux; it has not been tested by me.
4485 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4486 used to name a group. Names consist of alphanumerics and underscores, and must
4487 be unique. Back references use the syntax (?P=name) and recursive calls use
4488 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4489 numbers. The function pcre_fullinfo() can be used after compilation to extract
4490 a name/number map. There are three relevant calls:
4492 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4493 PCRE_INFO_NAMECOUNT yields the number of entries
4494 PCRE_INFO_NAMETABLE yields a pointer to the map.
4496 The map is a vector of fixed-size entries. The size of each entry depends on
4497 the length of the longest name used. The first two bytes of each entry are the
4498 group number, most significant byte first. There follows the corresponding
4499 name, zero terminated. The names are in alphabetical order.
4501 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4502 case instead of 255. Making it the same both with and without UTF-8 support
4503 means that the same test output works with both.
4505 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4506 calling malloc() with a zero argument.
4508 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4509 optimization. I've improved things by keeping a bitmap of backreferences with
4510 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4511 fact referenced, the optimization can be applied. It is unlikely that a
4512 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4513 the match to follow \n) will appear inside brackets with a number greater than
4514 31, but if it does, any back reference > 31 suppresses the optimization.
4516 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4517 of disabling numbered capturing parentheses. Any opening parenthesis that is
4518 not followed by ? behaves as if it were followed by ?: but named parentheses
4519 can still be used for capturing (and they will acquire numbers in the usual
4520 way).
4522 41. Redesigned the return codes from the match() function into yes/no/error so
4523 that errors can be passed back from deep inside the nested calls. A malloc
4524 failure while inside a recursive subpattern call now causes the
4525 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4527 42. It is now possible to set a limit on the number of times the match()
4528 function is called in a call to pcre_exec(). This facility makes it possible to
4529 limit the amount of recursion and backtracking, though not in a directly
4530 obvious way, because the match() function is used in a number of different
4531 circumstances. The count starts from zero for each position in the subject
4532 string (for non-anchored patterns). The default limit is, for compatibility, a
4533 large number, namely 10 000 000. You can change this in two ways:
4535 (a) When configuring PCRE before making, you can use --with-match-limit=n
4536 to set a default value for the compiled library.
4538 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4539 a different value is set. See 45 below.
4541 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4543 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4544 of things that can be changed at compile time. The first argument specifies
4545 what is wanted and the second points to where the information is to be placed.
4546 The current list of available information is:
4550 The output is an integer that is set to one if UTF-8 support is available;
4551 otherwise it is set to zero.
4555 The output is an integer that it set to the value of the code that is used for
4556 newline. It is either LF (10) or CR (13).
4560 The output is an integer that contains the number of bytes used for internal
4561 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4565 The output is an integer that contains the threshold above which the POSIX
4566 interface uses malloc() for output vectors. See item 31 above.
4570 The output is an unsigned integer that contains the default limit of the number
4571 of match() calls in a pcre_exec() execution. See 42 above.
4573 44. pcretest has been upgraded by the addition of the -C option. This causes it
4574 to extract all the available output from the new pcre_config() function, and to
4575 output it. The program then exits immediately.
4577 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4578 order to support additional features. One way would have been to define
4579 pcre_exec2() (for example) with extra arguments, but this would not have been
4580 extensible, and would also have required all calls to the original function to
4581 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4582 is used for passing in "extra" data from pcre_study().
4584 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4585 contains the following fields:
4587 flags a bitmap indicating which of the following fields are set
4588 study_data opaque data from pcre_study()
4589 match_limit a way of specifying a limit on match() calls for a specific
4590 call to pcre_exec()
4591 callout_data data for callouts (see 49 below)
4593 The flag bits are also defined in pcre.h, and are
4599 The pcre_study() function now returns one of these new pcre_extra blocks, with
4600 the actual study data pointed to by the study_data field, and the
4601 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4602 before. That is, this change is entirely upwards-compatible and requires no
4603 change to existing code.
4605 If you want to pass in additional data to pcre_exec(), you can either place it
4606 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4607 block.
4609 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4610 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4611 times with different match limits, until it finds the minimum value needed for
4612 pcre_exec() to complete. The value is then output. This can be instructive; for
4613 most simple matches the number is quite small, but for pathological cases it
4614 gets very large very quickly.
4616 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4617 returns the size of the data block pointed to by the study_data field in a
4618 pcre_extra block, that is, the value that was passed as the argument to
4619 pcre_malloc() when PCRE was getting memory in which to place the information
4620 created by pcre_study(). The fourth argument should point to a size_t variable.
4621 pcretest has been extended so that this information is shown after a successful
4622 pcre_study() call when information about the compiled regex is being displayed.
4624 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4625 because what follows is always an absolute path. (Later: it turns out that this
4626 is more than cosmetic for MinGW, because it doesn't like empty path
4627 components.)
4629 49. Some changes have been made to the callout feature (see 28 above):
4631 (i) A callout function now has three choices for what it returns:
4633 0 => success, carry on matching
4634 > 0 => failure at this point, but backtrack if possible
4635 < 0 => serious error, return this value from pcre_exec()
4637 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4638 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4639 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4640 use by callout functions. It will never be used by PCRE itself.
4642 (ii) The pcre_extra structure (see 45 above) has a void * field called
4643 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4644 pcre_callout_block structure has a field of the same name. The contents of
4645 the field passed in the pcre_extra structure are passed to the callout
4646 function in the corresponding field in the callout block. This makes it
4647 easier to use the same callout-containing regex from multiple threads. For
4648 testing, the pcretest program has a new data escape
4650 \C*n pass the number n (may be negative) as callout_data
4652 If the callout function in pcretest receives a non-zero value as
4653 callout_data, it returns that value.
4655 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4656 there were some redundant $(CFLAGS) in commands that are now specified as
4657 $(LINK), which already includes $(CFLAGS).
4659 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4660 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4661 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4662 one-byte characters throughout. Note that case-insensitive matching applies
4663 only to characters whose values are less than 256. PCRE doesn't support the
4664 notion of cases for higher-valued characters.
4666 (i) A character class whose characters are all within 0-255 is handled as
4667 a bit map, and the map is inverted for negative classes. Previously, a
4668 character > 255 always failed to match such a class; however it should
4669 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4671 (ii) A negated character class with a single character < 255 is coded as
4672 "not this character" (OP_NOT). This wasn't working properly when the test
4673 character was multibyte, either singly or repeated.
4675 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4676 mode, for example: \x{100}{2,3}.
4678 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4679 singly or repeated) now correctly test multibyte characters. However,
4680 PCRE doesn't recognize any characters with values greater than 255 as
4681 digits, spaces, or word characters. Such characters always match \D, \S,
4682 and \W, and never match \d, \s, or \w.
4684 (v) Classes may now contain characters and character ranges with values
4685 greater than 255. For example: [ab\x{100}-\x{400}].
4687 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4688 PCRE in UTF-8 mode.
4690 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4691 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4692 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4693 value.)
4695 53. The single man page has become too large. I have therefore split it up into
4696 a number of separate man pages. These also give rise to individual HTML pages;
4697 these are now put in a separate directory, and there is an index.html page that
4698 lists them all. Some hyperlinking between the pages has been installed.
4700 54. Added convenience functions for handling named capturing parentheses.
4702 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4703 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4704 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4705 are faulted.
4707 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4708 calling configure. These values are used when compiling the dftables.c program
4709 which is run to generate the source of the default character tables. They
4710 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4711 you will need to set these values.
4713 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4716 Version 3.9 02-Jan-02
4717 ---------------------
4719 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4721 2. If --disable-static was given, the building process failed when trying to
4722 build pcretest and pcregrep. (For some reason it was using libtool to compile
4723 them, which is not right, as they aren't part of the library.)
4726 Version 3.8 18-Dec-01
4727 ---------------------
4729 1. The experimental UTF-8 code was completely screwed up. It was packing the
4730 bytes in the wrong order. How dumb can you get?
4733 Version 3.7 29-Oct-01
4734 ---------------------
4736 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4737 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4738 this didn't happen under Solaris 8, where I normally test things.
4740 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4741 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4742 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4743 this fix an hour or so after the initial 3.7 release.)
4746 Version 3.6 23-Oct-01
4747 ---------------------
4749 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4750 offsets passed as NULL with zero offset count.
4752 2. The config.guess and config.sub files had not been updated when I moved to
4753 the latest autoconf.
4756 Version 3.5 15-Aug-01
4757 ---------------------
4759 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4760 had been forgotten.
4762 2. By using declared but undefined structures, we can avoid using "void"
4763 definitions in pcre.h while keeping the internal definitions of the structures
4764 private.
4766 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4767 user point of view, this means that both static and shared libraries are built
4768 by default, but this can be individually controlled. More of the work of
4769 handling this static/shared cases is now inside libtool instead of PCRE's make
4770 file.
4772 4. The pcretest utility is now installed along with pcregrep because it is
4773 useful for users (to test regexs) and by doing this, it automatically gets
4774 relinked by libtool. The documentation has been turned into a man page, so
4775 there are now .1, .txt, and .html versions in /doc.
4777 5. Upgrades to pcregrep:
4778 (i) Added long-form option names like gnu grep.
4779 (ii) Added --help to list all options with an explanatory phrase.
4780 (iii) Added -r, --recursive to recurse into sub-directories.
4781 (iv) Added -f, --file to read patterns from a file.
4783 6. pcre_exec() was referring to its "code" argument before testing that
4784 argument for NULL (and giving an error if it was NULL).
4786 7. Upgraded Makefile.in to allow for compiling in a different directory from
4787 the source directory.
4789 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4790 options bits, the pointer it was passed was to an int instead of to an unsigned
4791 long int. This mattered only on 64-bit systems.
4793 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4794 generated) instead of pcre.in, which it its source. Also made the same change
4795 in several of the .c files.
4797 10. A new release of gcc defines printf() as a macro, which broke pcretest
4798 because it had an ifdef in the middle of a string argument for printf(). Fixed
4799 by using separate calls to printf().
4801 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4802 script, to force use of CR or LF instead of \n in the source. On non-Unix
4803 systems, the value can be set in config.h.
4805 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4806 absolute limit. Changed the text of the error message to make this clear, and
4807 likewise updated the man page.
4809 13. The limit of 99 on the number of capturing subpatterns has been removed.
4810 The new limit is 65535, which I hope will not be a "real" limit.
4813 Version 3.4 22-Aug-00
4814 ---------------------
4816 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4818 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4821 Version 3.3 01-Aug-00
4822 ---------------------
4824 1. If an octal character was given, but the value was greater than \377, it
4825 was not getting masked to the least significant bits, as documented. This could
4826 lead to crashes in some systems.
4828 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4829 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4831 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4832 These just pass their arguments on to (pcre_free)(), but they are provided
4833 because some uses of PCRE bind it to non-C systems that can call its functions,
4834 but cannot call free() or pcre_free() directly.
4836 4. Add "make test" as a synonym for "make check". Corrected some comments in
4837 the Makefile.
4839 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4840 Makefile.
4842 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4843 command called pgrep for grepping around the active processes.
4845 7. Added the beginnings of support for UTF-8 character strings.
4847 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4848 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4849 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4850 out for the ar command.)
4853 Version 3.2 12-May-00
4854 ---------------------
4856 This is purely a bug fixing release.
4858 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4859 of ZA. This was just one example of several cases that could provoke this bug,
4860 which was introduced by change 9 of version 2.00. The code for breaking
4861 infinite loops after an iteration that matches an empty string was't working
4862 correctly.
4864 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4865 when matched against abbab (for example). After matching an empty string, it
4866 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4867 caused it to match further down the string than it should.
4869 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4870 was there because it doesn't seem to be needed, and it causes trouble on some
4871 systems, as it is not a Standard C header. It has been removed.
4873 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4874 were reported on the Macintosh. The changes were from
4876 while ((c = *(++ptr)) != 0 && c != '\n');
4877 to
4878 while ((c = *(++ptr)) != 0 && c != '\n') ;
4880 Totally extraordinary, but if that's what it takes...
4882 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4883 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4884 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4885 assumes the way PCRE uses memmove() (always moving upwards).
4887 6. PCRE is being used in one environment where strchr() is not available. There
4888 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4889 faster code anyway.
4892 Version 3.1 09-Feb-00
4893 ---------------------
4895 The only change in this release is the fixing of some bugs in Makefile.in for
4896 the "install" target:
4898 (1) It was failing to install pcreposix.h.
4900 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4903 Version 3.0 01-Feb-00
4904 ---------------------
4906 1. Add support for the /+ modifier to perltest (to output $` like it does in
4907 pcretest).
4909 2. Add support for the /g modifier to perltest.
4911 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4912 matches null strings.
4914 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4915 pattern. Perl treats empty patterns specially - it reuses the most recent
4916 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4917 effect.
4919 5. The POSIX interface was broken in that it was just handing over the POSIX
4920 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4921 required a bigger vector, with some working space on the end. This means that
4922 the POSIX wrapper now has to get and free some memory, and copy the results.
4924 6. Added some simple autoconf support, placing the test data and the
4925 documentation in separate directories, re-organizing some of the
4926 information files, and making it build pcre-config (a GNU standard). Also added
4927 libtool support for building PCRE as a shared library, which is now the
4928 default.
4930 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4931 09 are not valid octal constants. Single digits will be used for minor values
4932 less than 10.
4934 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4935 existing programs that set these in the POSIX interface can use PCRE without
4936 modification.
4938 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4939 return all that pcre_info() returns, plus additional data. The pcre_info()
4940 function is retained for compatibility, but is considered to be obsolete.
4942 10. Added experimental recursion feature (?R) to handle one common case that
4943 Perl 5.6 will be able to do with (?p{...}).
4945 11. Added support for POSIX character classes like [:alpha:], which Perl is
4946 adopting.
4949 Version 2.08 31-Aug-99
4950 ----------------------
4952 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4953 trying to match at the startoffset position, but instead was moving forward to
4954 the next newline as if a previous match had failed.
4956 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4957 and could get into a loop if a null string was matched other than at the start
4958 of the subject.
4960 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4961 be distinguished at compile time, and for completeness also added PCRE_DATE.
4963 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4964 in GnuWin32 environments.
4967 Version 2.07 29-Jul-99
4968 ----------------------
4970 1. The documentation is now supplied in plain text form and HTML as well as in
4971 the form of man page sources.
4973 2. C++ compilers don't like assigning (void *) values to other pointer types.
4974 In particular this affects malloc(). Although there is no problem in Standard
4975 C, I've put in casts to keep C++ compilers happy.
4977 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4978 should be (const char *).
4980 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4981 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4982 However, I haven't made this a standard facility. The documentation doesn't
4983 mention it, and the Makefile doesn't support it.
4985 5. The Makefile now contains an "install" target, with editable destinations at
4986 the top of the file. The pcretest program is not installed.
4988 6. pgrep -V now gives the PCRE version number and date.
4990 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4991 causing the entire string to be ignored, instead of just the last character.
4993 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4994 non-matching string, it can take a very, very long time, even for strings of
4995 quite modest length, because of the nested recursion. PCRE now does better in
4996 some of these cases. It does this by remembering the last required literal
4997 character in the pattern, and pre-searching the subject to ensure it is present
4998 before running the real match. In other words, it applies a heuristic to detect
4999 some types of certain failure quickly, and in the above example, if presented
5000 with a string that has no trailing " it gives "no match" very quickly.
5002 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5003 other alternatives are tried instead.
5006 Version 2.06 09-Jun-99
5007 ----------------------
5009 1. Change pcretest's output for amount of store used to show just the code
5010 space, because the remainder (the data block) varies in size between 32-bit and
5011 64-bit systems.
5013 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5014 start matching at. This allows lookbehinds to work when searching for multiple
5015 occurrences in a string.
5017 3. Added additional options to pcretest for testing multiple occurrences:
5019 /+ outputs the rest of the string that follows a match
5020 /g loops for multiple occurrences, using the new startoffset argument
5021 /G loops for multiple occurrences by passing an incremented pointer
5023 4. PCRE wasn't doing the "first character" optimization for patterns starting
5024 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5025 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5026 the letter 'x'. On long subject strings, this gives a significant speed-up.
5029 Version 2.05 21-Apr-99
5030 ----------------------
5032 1. Changed the type of magic_number from int to long int so that it works
5033 properly on 16-bit systems.
5035 2. Fixed a bug which caused patterns starting with .* not to work correctly
5036 when the subject string contained newline characters. PCRE was assuming
5037 anchoring for such patterns in all cases, which is not correct because .* will
5038 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5039 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5040 must be retried after every newline in the subject.
5043 Version 2.04 18-Feb-99
5044 ----------------------
5046 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5047 computation of the store needed to hold the pattern was incorrect (too large).
5048 If such patterns were nested a few deep, this could multiply and become a real
5049 problem.
5051 2. Added /M option to pcretest to show the memory requirement of a specific
5052 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5054 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5055 compiled in such a way that the backtracking after subsequent failure was
5056 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5057 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5060 Version 2.03 02-Feb-99
5061 ----------------------
5063 1. Fixed typo and small mistake in man page.
5065 2. Added 4th condition (GPL supersedes if conflict) and created separate
5066 LICENCE file containing the conditions.
5068 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5069 Perl, that is the internal \ allows the delimiter to be included in the
5070 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5071 the final delimiter, add \ to the end of the pattern (to test the error).
5073 4. Added the convenience functions for extracting substrings after a successful
5074 match. Updated pcretest to make it able to test these functions.
5077 Version 2.02 14-Jan-99
5078 ----------------------
5080 1. Initialized the working variables associated with each extraction so that
5081 their saving and restoring doesn't refer to uninitialized store.
5083 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5084 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5085 fix the problem.
5087 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5088 calls, and wasn't printing the correct value for compiling calls. Increased the
5089 default value of LOOPREPEAT, and the number of significant figures in the
5090 times.
5092 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5094 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5095 a building problem on Windows NT with a FAT file system.
5098 Version 2.01 21-Oct-98
5099 ----------------------
5101 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5102 to character tables built by pcre_maketables() in the current locale. If NULL
5103 is passed, the default tables are used.
5106 Version 2.00 24-Sep-98
5107 ----------------------
5109 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5110 it any more.
5112 2. Allow quantification of (?>) groups, and make it work correctly.
5114 3. The first character computation wasn't working for (?>) groups.
5116 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5117 end of the subject) and add 5.005's \z, which really does match only at the
5118 very end of the subject.
5120 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5122 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5123 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5124 localized options. All options to pcre_study() were also removed.
5126 7. Add other new features from 5.005:
5128 $(?<= positive lookbehind
5129 $(?<! negative lookbehind
5130 (?imsx-imsx) added the unsetting capability
5131 such a setting is global if at outer level; local otherwise
5132 (?imsx-imsx:) non-capturing groups with option setting
5133 (?(cond)re|re) conditional pattern matching
5135 A backreference to itself in a repeated group matches the previous
5136 captured string.
5138 8. General tidying up of studying (both automatic and via "study")
5139 consequential on the addition of new assertions.
5141 9. As in 5.005, unlimited repeated groups that could match an empty substring
5142 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5143 runtime if any iteration does actually match an empty substring.
5145 10. Include the RunTest script in the distribution.
5147 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5148 discrepancies, some of which were old and were also with respect to 5.004. They
5149 have now been fixed.
5152 Version 1.09 28-Apr-98
5153 ----------------------
5155 1. A negated single character class followed by a quantifier with a minimum
5156 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5157 program crashes, or just wrong answers. This did not apply to negated classes
5158 containing more than one character, or to minima other than one.
5161 Version 1.08 27-Mar-98
5162 ----------------------
5164 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5166 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5167 latter must appear before anything that relies on it in the pattern.
5170 Version 1.07 16-Feb-98
5171 ----------------------
5173 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5174 repeat of a potentially empty string).
5177 Version 1.06 23-Jan-98
5178 ----------------------
5180 1. Added Markus Oberhumer's little patches for C++.
5182 2. Literal strings longer than 255 characters were broken.
5185 Version 1.05 23-Dec-97
5186 ----------------------
5188 1. Negated character classes containing more than one character were failing if
5189 PCRE_CASELESS was set at run time.
5192 Version 1.04 19-Dec-97
5193 ----------------------
5195 1. Corrected the man page, where some "const" qualifiers had been omitted.
5197 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5198 input syntax.
5200 3. Fixed memory leak which occurred when a regex with back references was
5201 matched with an offsets vector that wasn't big enough. The temporary memory
5202 that is used in this case wasn't being freed if the match failed.
5204 4. Tidied pcretest to ensure it frees memory that it gets.
5206 5. Temporary memory was being obtained in the case where the passed offsets
5207 vector was exactly big enough.
5209 6. Corrected definition of offsetof() from change 5 below.
5211 7. I had screwed up change 6 below and broken the rules for the use of
5212 setjmp(). Now fixed.
5215 Version 1.03 18-Dec-97
5216 ----------------------
5218 1. A erroneous regex with a missing opening parenthesis was correctly
5219 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5220 on some systems.
5222 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5223 it was reported that one broken compiler failed on the former because "code" is
5224 also an independent variable.
5226 3. The erroneous regex a[]b caused an array overrun reference.
5228 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5229 fail on data ending with that character. (It was going on too far, and checking
5230 the next character, typically a binary zero.) This was specific to the
5231 optimized code for single-character negative classes.
5233 5. Added a contributed patch from the TIN world which does the following:
5235 + Add an undef for memmove, in case the the system defines a macro for it.
5237 + Add a definition of offsetof(), in case there isn't one. (I don't know
5238 the reason behind this - offsetof() is part of the ANSI standard - but
5239 it does no harm).
5241 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5242 most of the places where whitespace preceded '#'. I have given up and
5243 allowed the remaining 2 cases to be at the margin.
5245 + Rename some variables in pcre to eliminate shadowing. This seems very
5246 pedantic, but does no harm, of course.
5248 6. Moved the call to setjmp() into its own function, to get rid of warnings
5249 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5251 7. Constructs such as \d{8,} were compiling into the equivalent of
5252 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5253 outcome, but in this particular case used more store than had been allocated,
5254 which caused the bug to be discovered because it threw up an internal error.
5256 8. The debugging code in both pcre and pcretest for outputting the compiled
5257 form of a regex was going wrong in the case of back references followed by
5258 curly-bracketed repeats.
5261 Version 1.02 12-Dec-97
5262 ----------------------
5264 1. Typos in pcre.3 and comments in the source fixed.
5266 2. Applied a contributed patch to get rid of places where it used to remove
5267 'const' from variables, and fixed some signed/unsigned and uninitialized
5268 variable warnings.
5270 3. Added the "runtest" target to Makefile.
5272 4. Set default compiler flag to -O2 rather than just -O.
5275 Version 1.01 19-Nov-97
5276 ----------------------
5278 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5279 like /([ab]*)*/, that is, for classes with more than one character in them.
5281 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5282 as /((?>a*))*/ (a PCRE_EXTRA facility).
5285 Version 1.00 18-Nov-97
5286 ----------------------
5288 1. Added compile-time macros to support systems such as SunOS4 which don't have
5289 memmove() or strerror() but have other things that can be used instead.
5291 2. Arranged that "make clean" removes the executables.
5294 Version 0.99 27-Oct-97
5295 ----------------------
5297 1. Fixed bug in code for optimizing classes with only one character. It was
5298 initializing a 32-byte map regardless, which could cause it to run off the end
5299 of the memory it had got.
5301 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5304 Version 0.98 22-Oct-97
5305 ----------------------
5307 1. Fixed bug in code for handling temporary memory usage when there are more
5308 back references than supplied space in the ovector. This could cause segfaults.
5311 Version 0.97 21-Oct-97
5312 ----------------------
5314 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5316 2. Optimized negated single characters not to use a bit map.
5318 3. Brought error texts together as macro definitions; clarified some of them;
5319 fixed one that was wrong - it said "range out of order" when it meant "invalid
5320 escape sequence".
5322 4. Changed some char * arguments to const char *.
5324 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5326 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5327 pcretest.
5330 Version 0.96 16-Oct-97
5331 ----------------------
5333 1. Added a simple "pgrep" utility to the distribution.
5335 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5336 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5337 where "ddd" means "one or more decimal digits".
5339 3. Fixed serious bug. If a pattern had a back reference, but the call to
5340 pcre_exec() didn't supply a large enough ovector to record the related
5341 identifying subpattern, the match always failed. PCRE now remembers the number
5342 of the largest back reference, and gets some temporary memory in which to save
5343 the offsets during matching if necessary, in order to ensure that
5344 backreferences always work.
5346 4. Increased the compatibility with Perl in a number of ways:
5348 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5349 to request this handling. The option can be set at compile or exec time.
5351 (b) $ matches before a terminating newline by default; an option
5352 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5353 mode). The option can be set at compile or exec time.
5355 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5356 the same as Perl's. If the decimal number it represents is less than 10
5357 or there aren't that many previous left capturing parentheses, an octal
5358 escape is read. Inside a character class, it's always an octal escape,
5359 even if it is a single digit.
5361 (d) An escaped but undefined alphabetic character is taken as a literal,
5362 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5363 escapes.
5365 (e) {0} is now permitted. (The previous item is removed from the compiled
5366 pattern).
5368 5. Changed all the names of code files so that the basic parts are no longer
5369 than 10 characters, and abolished the teeny "globals.c" file.
5371 6. Changed the handling of character classes; they are now done with a 32-byte
5372 bit map always.
5374 7. Added the -d and /D options to pcretest to make it possible to look at the
5375 internals of compilation without having to recompile pcre.
5378 Version 0.95 23-Sep-97
5379 ----------------------
5381 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5382 \x20 at the start of a run of normal characters. These were being treated as
5383 real characters, instead of the source characters being re-checked.
5386 Version 0.94 18-Sep-97
5387 ----------------------
5389 1. The functions are now thread-safe, with the caveat that the global variables
5390 containing pointers to malloc() and free() or alternative functions are the
5391 same for all threads.
5393 2. Get pcre_study() to generate a bitmap of initial characters for non-
5394 anchored patterns when this is possible, and use it if passed to pcre_exec().
5397 Version 0.93 15-Sep-97
5398 ----------------------
5400 1. /(b)|(:+)/ was computing an incorrect first character.
5402 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5403 but not actually doing anything yet.
5405 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5406 as Perl does (e.g. [-az] or [az-]).
5408 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5409 all possible positions.
5411 5. Split up into different modules to avoid including unneeded functions in a
5412 compiled binary. However, compile and exec are still in one module. The "study"
5413 function is split off.
5415 6. The character tables are now in a separate module whose source is generated
5416 by an auxiliary program - but can then be edited by hand if required. There are
5417 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5418 toupper() in the code.
5420 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5421 make them global. Abolish the function for setting them, as the caller can now
5422 set them directly.
5425 Version 0.92 11-Sep-97
5426 ----------------------
5428 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5429 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5431 2. Caseless matching was not working in character classes if the characters in
5432 the pattern were in upper case.
5434 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5436 4. Make PCRE_ANCHORED public and accept as a compile option.
5438 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5439 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5440 pass them.
5442 6. Give an error if bad option bits passed at compile or run time.
5444 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5445 pcretest to cause it to pass that flag.
5447 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5448 options, and the first character, if set.
5450 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5453 Version 0.91 10-Sep-97
5454 ----------------------
5456 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5457 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5459 2. PCRE was looping on encountering an indefinitely repeated back reference to
5460 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5461 Perl does - treats the match as successful.
5463 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5