7 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
8 which is apparently normally available under Windows.
10 2. Re-jig the pcregrep tests with different newline settings in an attempt
11 to make them independent of the local environment's newline setting.
13 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
15 4. Some of the "internals" tests were previously cut out when the link size
16 was not 2, because the output contained actual offsets. The recent new
17 "Z" feature of pcretest means that these can be cut out, making the tests
18 usable with all link sizes.
20 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
21 stack recursion. This gives a massive performance boost under BSD, but just
22 a small improvement under Linux. However, it saves one field in the frame
25 6. Added more features from the forthcoming Perl 5.10:
27 (a) (?-n) (where n is a string of digits) is a relative subroutine or
28 recursion call. It refers to the nth most recently opened parentheses.
30 (b) (?+n) is also a relative subroutine call; it refers to the nth next
31 to be opened parentheses.
33 (c) Conditions that refer to capturing parentheses can be specified
34 relatively, for example, (?(-2)... or (?(+3)...
36 (d) \K resets the start of the current match so that everything before
39 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
41 (f) \g{name} is another synonym - part of Perl 5.10's unification of
44 (g) (?| introduces a group in which the numbering of parentheses in each
45 alternative starts with the same number.
47 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
49 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
52 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
53 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
54 for detecting groups that can match an empty string.
56 9. A pattern with a very large number of alternatives (more than several
57 hundred) was running out of internal workspace during the pre-compile
58 phase, where pcre_compile() figures out how much memory will be needed. A
59 bit of new cunning has reduced the workspace needed for groups with
60 alternatives. The 1000-alternative test pattern now uses 12 bytes of
61 workspace instead of running out of the 4096 that are available.
63 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
65 11. Applied patch from Google to remove an optimization that didn't quite work.
66 The report of the bug said:
68 pcrecpp::RE("a*").FullMatch("aaa") matches, while
69 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
70 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
72 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
73 it matched the wrong number of bytes.
79 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
80 that is more "standard", making use of automake and other Autotools. There
81 is some re-arrangement of the files and adjustment of comments consequent
84 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
85 for recursive directory scanning broke on some systems because the files
86 are not scanned in any specific order and on different systems the order
87 was different. A call to "sort" has been inserted into RunGrepTest for the
88 approprate test as a short-term fix. In the longer term there may be an
91 3. I had an email from Eric Raymond about problems translating some of PCRE's
92 man pages to HTML (despite the fact that I distribute HTML pages, some
93 people do their own conversions for various reasons). The problems
94 concerned the use of low-level troff macros .br and .in. I have therefore
95 removed all such uses from the man pages (some were redundant, some could
96 be replaced by .nf/.fi pairs). The 132html script that I use to generate
97 HTML has been updated to handle .nf/.fi and to complain if it encounters
100 4. Updated comments in configure.ac that get placed in config.h.in and also
101 arranged for config.h to be included in the distribution, with the name
102 config.h.generic, for the benefit of those who have to compile without
103 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
105 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
106 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
107 makevp.bat for the current PCRE, using the additional files
108 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
110 6. A Windows user reported a minor discrepancy with test 2, which turned out
111 to be caused by a trailing space on an input line that had got lost in his
112 copy. The trailing space was an accident, so I've just removed it.
114 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
117 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
118 as "const" (a) because they are and (b) because it helps the PHP
119 maintainers who have recently made a script to detect big data structures
120 in the php code that should be moved to the .rodata section. I remembered
121 to update Builducptable as well, so it won't revert if ucptable.h is ever
124 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
125 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
126 order to be able to cut out the UTF-8 tables in the latter when UTF-8
127 support is not required. This saves 1.5-2K of code, which is important in
130 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
131 so as not to refer to the tables, even though these functions will never be
132 called when UTF-8 support is disabled. Otherwise there are problems with a
135 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
137 (a) It was defining its arguments as char * instead of void *.
139 (b) It was assuming that all moves were upwards in memory; this was true
140 a long time ago when I wrote it, but is no longer the case.
142 The emulated memove() is provided for those environments that have neither
143 memmove() nor bcopy(). I didn't think anyone used it these days, but that
144 is clearly not the case, as these two bugs were recently reported.
146 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
147 and Detrail to create the HTML documentation, the .txt form of the man
148 pages, and it removes trailing spaces from listed files. It also creates
149 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
150 case, it wraps all the #defines with #ifndefs. This script should be run
153 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
154 with Unicode property support.
156 (a) For a maximizing quantifier, if the two different cases of the
157 character were of different lengths in their UTF-8 codings (there are
158 some cases like this - I found 11), and the matching function had to
159 back up over a mixture of the two cases, it incorrectly assumed they
160 were both the same length.
162 (b) When PCRE was configured to use the heap rather than the stack for
163 recursion during matching, it was not correctly preserving the data for
164 the other case of a UTF-8 character when checking ahead for a match
165 while processing a minimizing repeat. If the check also involved
166 matching a wide character, but failed, corruption could cause an
167 erroneous result when trying to check for a repeat of the original
170 13. Some tidying changes to the testing mechanism:
172 (a) The RunTest script now detects the internal link size and whether there
173 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
174 values substituted by "configure". (The RunGrepTest script already did
175 this for UTF-8.) The configure.ac script no longer substitutes the
178 (b) The debugging options /B and /D in pcretest show the compiled bytecode
179 with length and offset values. This means that the output is different
180 for different internal link sizes. Test 2 is skipped for link sizes
181 other than 2 because of this, bypassing the problem. Unfortunately,
182 there was also a test in test 3 (the locale tests) that used /B and
183 failed for link sizes other than 2. Rather than cut the whole test out,
184 I have added a new /Z option to pcretest that replaces the length and
185 offset values with spaces. This is now used to make test 3 independent
186 of link size. (Test 2 will be tidied up later.)
188 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
189 segmentation fault instead of returning the appropriate error message.
191 15. In multiline mode when the newline sequence was set to "any", the pattern
192 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
193 This doesn't seem right; it now treats the CRLF combination as the line
194 ending, and so does not match in that case. It's only a pattern such as ^$
195 that would hit this one: something like ^ABC$ would have failed after \r
196 and then tried again after \r\n.
198 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
199 in an attempt to make files that differ only in their line terminators
200 compare equal. This works on Linux.
202 17. Under certain error circumstances pcregrep might try to free random memory
203 as it exited. This is now fixed, thanks to valgrind.
205 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
206 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
207 was because its rules for how to advance for /g after matching an empty
208 string at the end of a line did not allow for this case. They now check for
211 20. pcretest is supposed to handle patterns and data of any length, by
212 extending its buffers when necessary. It was getting this wrong when the
213 buffer for a data line had to be extended.
215 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
216 CRLF as a newline sequence.
218 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
219 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
220 I have nevertheless tidied it up.
222 23. Added some casts to kill warnings from HP-UX ia64 compiler.
224 24. Added a man page for pcre-config.
227 Version 7.0 19-Dec-06
228 ---------------------
230 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
233 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
234 sys/time.h, which is documented as needed for this function. It doesn't
235 seem to matter on Linux, but it showed up on some releases of OS X.
237 3. It seems that there are systems where bytes whose values are greater than
238 127 match isprint() in the "C" locale. The "C" locale should be the
239 default when a C program starts up. In most systems, only ASCII printing
240 characters match isprint(). This difference caused the output from pcretest
241 to vary, making some of the tests fail. I have changed pcretest so that:
243 (a) When it is outputting text in the compiled version of a pattern, bytes
244 other than 32-126 are always shown as hex escapes.
246 (b) When it is outputting text that is a matched part of a subject string,
247 it does the same, unless a different locale has been set for the match
248 (using the /L modifier). In this case, it uses isprint() to decide.
250 4. Fixed a major bug that caused incorrect computation of the amount of memory
251 required for a compiled pattern when options that changed within the
252 pattern affected the logic of the preliminary scan that determines the
253 length. The relevant options are -x, and -i in UTF-8 mode. The result was
254 that the computed length was too small. The symptoms of this bug were
255 either the PCRE error "internal error: code overflow" from pcre_compile(),
256 or a glibc crash with a message such as "pcretest: free(): invalid next
257 size (fast)". Examples of patterns that provoked this bug (shown in
258 pcretest format) are:
265 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
266 is now done differently.
268 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
269 wrapper classes; (b) implement a new function in the C++ scanner that is
270 more efficient than the old way of doing things because it avoids levels of
271 recursion in the regex matching; (c) add a paragraph to the documentation
272 for the FullMatch() function.
274 6. The escape sequence \n was being treated as whatever was defined as
275 "newline". Not only was this contrary to the documentation, which states
276 that \n is character 10 (hex 0A), but it also went horribly wrong when
277 "newline" was defined as CRLF. This has been fixed.
279 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
280 was being set to -1 for the "end of line" case (supposedly a value that no
281 character can have). Though this value is never used (the check for end of
282 line is "zero bytes in current character"), it caused compiler complaints.
283 I've changed it to 0xffffffff.
285 8. In pcre_version.c, the version string was being built by a sequence of
286 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
287 string (as it is for production releases) called a macro with an empty
288 argument. The C standard says the result of this is undefined. The gcc
289 compiler treats it as an empty string (which was what was wanted) but it is
290 reported that Visual C gives an error. The source has been hacked around to
293 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
294 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
295 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
298 10. Originally, pcretest opened its input and output without "b"; then I was
299 told that "b" was needed in some environments, so it was added for release
300 5.0 to both the input and output. (It makes no difference on Unix-like
301 systems.) Later I was told that it is wrong for the input on Windows. I've
302 now abstracted the modes into two macros, to make it easier to fiddle with
303 them, and removed "b" from the input mode under Windows.
305 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
307 12. Added -help and --help to pcretest as an official way of being reminded
310 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
311 and pcrecpp.cc because they annoy compilers at high warning levels.
313 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
315 15. Fixed an occurrence of == in configure.ac that should have been = (shell
316 scripts are not C programs :-) and which was not noticed because it works
319 16. pcretest is supposed to handle any length of pattern and data line (as one
320 line or as a continued sequence of lines) by extending its input buffer if
321 necessary. This feature was broken for very long pattern lines, leading to
322 a string of junk being passed to pcre_compile() if the pattern was longer
325 17. I have done a major re-factoring of the way pcre_compile() computes the
326 amount of memory needed for a compiled pattern. Previously, there was code
327 that made a preliminary scan of the pattern in order to do this. That was
328 OK when PCRE was new, but as the facilities have expanded, it has become
329 harder and harder to keep it in step with the real compile phase, and there
330 have been a number of bugs (see for example, 4 above). I have now found a
331 cunning way of running the real compile function in a "fake" mode that
332 enables it to compute how much memory it would need, while actually only
333 ever using a few hundred bytes of working memory and without too many
334 tests of the mode. This should make future maintenance and development
335 easier. A side effect of this work is that the limit of 200 on the nesting
336 depth of parentheses has been removed (though this was never a serious
337 limitation, I suspect). However, there is a downside: pcre_compile() now
338 runs more slowly than before (30% or more, depending on the pattern). I
339 hope this isn't a big issue. There is no effect on runtime performance.
341 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
342 newline (only possible for the last line of a file) and it was a
343 pattern that set a locale (followed by /Lsomething), pcretest crashed.
345 19. Added additional timing features to pcretest. (1) The -tm option now times
346 matching only, not compiling. (2) Both -t and -tm can be followed, as a
347 separate command line item, by a number that specifies the number of
348 repeats to use when timing. The default is 50000; this gives better
349 precision, but takes uncomfortably long for very large patterns.
351 20. Extended pcre_study() to be more clever in cases where a branch of a
352 subpattern has no definite first character. For example, (a*|b*)[cd] would
353 previously give no result from pcre_study(). Now it recognizes that the
354 first character must be a, b, c, or d.
356 21. There was an incorrect error "recursive call could loop indefinitely" if
357 a subpattern (or the entire pattern) that was being tested for matching an
358 empty string contained only one non-empty item after a nested subpattern.
359 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
360 incorrectly, because the \d was being skipped in the check.
362 22. The pcretest program now has a new pattern option /B and a command line
363 option -b, which is equivalent to adding /B to every pattern. This causes
364 it to show the compiled bytecode, without the additional information that
365 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
366 is the same as /B/I).
368 23. A new optimization is now able automatically to treat some sequences such
369 as a*b as a*+b. More specifically, if something simple (such as a character
370 or a simple class like \d) has an unlimited quantifier, and is followed by
371 something that cannot possibly match the quantified thing, the quantifier
372 is automatically "possessified".
374 24. A recursive reference to a subpattern whose number was greater than 39
375 went wrong under certain circumstances in UTF-8 mode. This bug could also
376 have affected the operation of pcre_study().
378 25. Realized that a little bit of performance could be had by replacing
379 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
381 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
383 27. Possessive quantifiers such as a++ were previously implemented by turning
384 them into atomic groups such as ($>a+). Now they have their own opcodes,
385 which improves performance. This includes the automatically created ones
388 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
389 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
390 the first matched character to be a colon. This applied both to named and
393 29. The ucpinternal.h header file was missing its idempotency #ifdef.
395 30. I was sent a "project" file called libpcre.a.dev which I understand makes
396 building PCRE on Windows easier, so I have included it in the distribution.
398 31. There is now a check in pcretest against a ridiculously large number being
399 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
400 loop, the loop is abandoned.
402 32. Forward references to subpatterns in conditions such as (?(2)...) where
403 subpattern 2 is defined later cause pcre_compile() to search forwards in
404 the pattern for the relevant set of parentheses. This search went wrong
405 when there were unescaped parentheses in a character class, parentheses
406 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
408 33. "Subroutine" calls and backreferences were previously restricted to
409 referencing subpatterns earlier in the regex. This restriction has now
412 34. Added a number of extra features that are going to be in Perl 5.10. On the
413 whole, these are just syntactic alternatives for features that PCRE had
414 previously implemented using the Python syntax or my own invention. The
415 other formats are all retained for compatibility.
417 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
418 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
419 also .NET compatible.
421 (b) A recursion or subroutine call to a named group can now be defined as
422 (?&name) as well as (?P>name).
424 (c) A backreference to a named group can now be defined as \k<name> or
425 \k'name' as well as (?P=name). The new forms, as well as being in Perl
426 5.10, are also .NET compatible.
428 (d) A conditional reference to a named group can now use the syntax
429 (?(<name>) or (?('name') as well as (?(name).
431 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
432 groups (named and numbered) that are never evaluated inline, but can be
433 called as "subroutines" from elsewhere. In effect, the DEFINE condition
434 is always false. There may be only one alternative in such a group.
436 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
437 as the simple (?(R). The condition is true only if the most recent
438 recursion is that of the given number or name. It does not search out
439 through the entire recursion stack.
441 (g) The escape \gN or \g{N} has been added, where N is a positive or
442 negative number, specifying an absolute or relative reference.
444 35. Tidied to get rid of some further signed/unsigned compiler warnings and
445 some "unreachable code" warnings.
447 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
448 things, this adds five new scripts.
450 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
451 There were also incompatibilities regarding the handling of \Q..\E inside
452 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
453 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
455 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
456 matches an empty string, and forcibly breaks the loop. There were bugs in
457 this code in non-simple cases. For a pattern such as ^(a()*)* matched
458 against aaaa the result was just "a" rather than "aaaa", for example. Two
459 separate and independent bugs (that affected different cases) have been
462 39. Refactored the code to abolish the use of different opcodes for small
463 capturing bracket numbers. This is a tidy that I avoided doing when I
464 removed the limit on the number of capturing brackets for 3.5 back in 2001.
465 The new approach is not only tidier, it makes it possible to reduce the
466 memory needed to fix the previous bug (38).
468 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
469 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
470 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
473 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
476 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
477 copy construction and assignment in the C++ wrapper.
479 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
480 couple of bugs that could have given wrong results in the "--newline=crlf"
483 44. Added a number of casts and did some reorganization of signed/unsigned int
484 variables following suggestions from Dair Grant. Also renamed the variable
485 "this" as "item" because it is a C++ keyword.
487 45. Arranged for dftables to add
489 #include "pcre_internal.h"
491 to pcre_chartables.c because without it, gcc 4.x may remove the array
492 definition from the final binary if PCRE is built into a static library and
493 dead code stripping is activated.
495 46. For an unanchored pattern, if a match attempt fails at the start of a
496 newline sequence, and the newline setting is CRLF or ANY, and the next two
497 characters are CRLF, advance by two characters instead of one.
500 Version 6.7 04-Jul-06
501 ---------------------
503 1. In order to handle tests when input lines are enormously long, pcretest has
504 been re-factored so that it automatically extends its buffers when
505 necessary. The code is crude, but this _is_ just a test program. The
506 default size has been increased from 32K to 50K.
508 2. The code in pcre_study() was using the value of the re argument before
509 testing it for NULL. (Of course, in any sensible call of the function, it
512 3. The memmove() emulation function in pcre_internal.h, which is used on
513 systems that lack both memmove() and bcopy() - that is, hardly ever -
514 was missing a "static" storage class specifier.
516 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
517 containing an extended class (one that cannot be represented by a bitmap
518 because it contains high-valued characters or Unicode property items, e.g.
519 [\pZ]). Almost always one would set UTF-8 mode when processing such a
520 pattern, but PCRE should not loop if you do not (it no longer does).
521 [Detail: two cases were found: (a) a repeated subpattern containing an
522 extended class; (b) a recursive reference to a subpattern that followed a
523 previous extended class. It wasn't skipping over the extended class
524 correctly when UTF-8 mode was not set.]
526 5. A negated single-character class was not being recognized as fixed-length
527 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
528 compile error "lookbehind assertion is not fixed length".
530 6. The RunPerlTest auxiliary script was showing an unexpected difference
531 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
532 write a Perl script that can interpret lines of an input file either as
533 byte characters or as UTF-8, which is what "perltest" was being required to
534 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
535 can't do is switch easily at run time between having the "use utf8;" pragma
536 or not. In the end, I fudged it by using the RunPerlTest script to insert
537 "use utf8;" explicitly for the UTF-8 tests.
539 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
540 the end of the subject string, contrary to the documentation and to what
541 Perl does. This was true of both matching functions. Now it matches only at
542 the start of the subject and immediately after *internal* newlines.
544 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
545 a pointer to an int instead of a pointer to an unsigned long int. This
546 caused problems on 64-bit systems.
548 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
549 instance of the 'standard' template library not being so standard".
551 10. There was no check on the number of named subpatterns nor the maximum
552 length of a subpattern name. The product of these values is used to compute
553 the size of the memory block for a compiled pattern. By supplying a very
554 long subpattern name and a large number of named subpatterns, the size
555 computation could be caused to overflow. This is now prevented by limiting
556 the length of names to 32 characters, and the number of named subpatterns
559 11. Subpatterns that are repeated with specific counts have to be replicated in
560 the compiled pattern. The size of memory for this was computed from the
561 length of the subpattern and the repeat count. The latter is limited to
562 65535, but there was no limit on the former, meaning that integer overflow
563 could in principle occur. The compiled length of a repeated subpattern is
564 now limited to 30,000 bytes in order to prevent this.
566 12. Added the optional facility to have named substrings with the same name.
568 13. Added the ability to use a named substring as a condition, using the
569 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
570 are numbers (not recommended). Forward references are permitted.
572 14. Added forward references in named backreferences (if you see what I mean).
574 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
575 pattern could run off the end of the subject. For example, the pattern
576 "(?s)(.{1,5})"8 did this with the subject "ab".
578 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
579 PCRE_CASELESS was set when matching characters that were quantified with ?
582 17. A character class other than a single negated character that had a minimum
583 but no maximum quantifier - for example [ab]{6,} - was not handled
584 correctly by pce_dfa_exec(). It would match only one character.
586 18. A valid (though odd) pattern that looked like a POSIX character
587 class but used an invalid character after [ (for example [[,abc,]]) caused
588 pcre_compile() to give the error "Failed: internal error: code overflow" or
589 in some cases to crash with a glibc free() error. This could even happen if
590 the pattern terminated after [[ but there just happened to be a sequence of
591 letters, a binary zero, and a closing ] in the memory that followed.
593 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
594 over the years. Originally (before any Unicode support), just the bottom 8
595 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
596 output from "man perlunicode" includes this:
598 The regular expression compiler produces polymorphic opcodes. That
599 is, the pattern adapts to the data and automatically switches to
600 the Unicode character scheme when presented with Unicode data--or
601 instead uses a traditional byte scheme when presented with byte
604 Sadly, a wide octal escape does not cause a switch, and in a string with
605 no other multibyte characters, these octal escapes are treated as before.
606 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
607 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
610 I have not perpetrated such confusion in PCRE. Up till now, it took just
611 the bottom 8 bits, as in old Perl. I have now made octal escapes with
612 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
613 translate to the appropriate multibyte character.
615 29. Applied some refactoring to reduce the number of warnings from Microsoft
616 and Borland compilers. This has included removing the fudge introduced
617 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
618 a warning about an unused variable.
620 21. PCRE has not included VT (character 0x0b) in the set of whitespace
621 characters since release 4.0, because Perl (from release 5.004) does not.
622 [Or at least, is documented not to: some releases seem to be in conflict
623 with the documentation.] However, when a pattern was studied with
624 pcre_study() and all its branches started with \s, PCRE still included VT
625 as a possible starting character. Of course, this did no harm; it just
626 caused an unnecessary match attempt.
628 22. Removed a now-redundant internal flag bit that recorded the fact that case
629 dependency changed within the pattern. This was once needed for "required
630 byte" processing, but is no longer used. This recovers a now-scarce options
631 bit. Also moved the least significant internal flag bit to the most-
632 significant bit of the word, which was not previously used (hangover from
633 the days when it was an int rather than a uint) to free up another bit for
636 23. Added support for CRLF line endings as well as CR and LF. As well as the
637 default being selectable at build time, it can now be changed at runtime
638 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
639 specify that it is scanning data with non-default line endings.
641 24. Changed the definition of CXXLINK to make it agree with the definition of
642 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
644 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
645 recursions. This makes a big different to stack usage for some patterns.
647 26. If a subpattern containing a named recursion or subroutine reference such
648 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
649 the space required for the compiled pattern went wrong and gave too small a
650 value. Depending on the environment, this could lead to "Failed: internal
651 error: code overflow at offset 49" or "glibc detected double free or
654 27. Applied patches from Google (a) to support the new newline modes and (b) to
655 advance over multibyte UTF-8 characters in GlobalReplace.
657 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
658 difference for some implementation of PCRE in some Windows version.
660 29. Added some extra testing facilities to pcretest:
662 \q<number> in a data line sets the "match limit" value
663 \Q<number> in a data line sets the "match recursion limt" value
664 -S <number> sets the stack size, where <number> is in megabytes
666 The -S option isn't available for Windows.
669 Version 6.6 06-Feb-06
670 ---------------------
672 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
673 in pcreposix.h. I have copied the definition from pcre.h.
675 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
676 because pcre.h is no longer a built file.
678 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
679 not normally included in the compiled code.
682 Version 6.5 01-Feb-06
683 ---------------------
685 1. When using the partial match feature with pcre_dfa_exec(), it was not
686 anchoring the second and subsequent partial matches at the new starting
687 point. This could lead to incorrect results. For example, with the pattern
688 /1234/, partially matching against "123" and then "a4" gave a match.
690 2. Changes to pcregrep:
692 (a) All non-match returns from pcre_exec() were being treated as failures
693 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
694 error message is output. Some extra information is given for the
695 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
696 probably the only errors that are likely to be caused by users (by
697 specifying a regex that has nested indefinite repeats, for instance).
698 If there are more than 20 of these errors, pcregrep is abandoned.
700 (b) A binary zero was treated as data while matching, but terminated the
701 output line if it was written out. This has been fixed: binary zeroes
702 are now no different to any other data bytes.
704 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
705 used to set a locale for matching. The --locale=xxxx long option has
706 been added (no short equivalent) to specify a locale explicitly on the
707 pcregrep command, overriding the environment variables.
709 (d) When -B was used with -n, some line numbers in the output were one less
710 than they should have been.
712 (e) Added the -o (--only-matching) option.
714 (f) If -A or -C was used with -c (count only), some lines of context were
715 accidentally printed for the final match.
717 (g) Added the -H (--with-filename) option.
719 (h) The combination of options -rh failed to suppress file names for files
720 that were found from directory arguments.
722 (i) Added the -D (--devices) and -d (--directories) options.
724 (j) Added the -F (--fixed-strings) option.
726 (k) Allow "-" to be used as a file name for -f as well as for a data file.
728 (l) Added the --colo(u)r option.
730 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
731 is not present by default.
733 3. A nasty bug was discovered in the handling of recursive patterns, that is,
734 items such as (?R) or (?1), when the recursion could match a number of
735 alternatives. If it matched one of the alternatives, but subsequently,
736 outside the recursion, there was a failure, the code tried to back up into
737 the recursion. However, because of the way PCRE is implemented, this is not
738 possible, and the result was an incorrect result from the match.
740 In order to prevent this happening, the specification of recursion has
741 been changed so that all such subpatterns are automatically treated as
742 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
744 4. I had overlooked the fact that, in some locales, there are characters for
745 which isalpha() is true but neither isupper() nor islower() are true. In
746 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
747 and ordfeminine) are like this. This affected the treatment of \w and \W
748 when they appeared in character classes, but not when they appeared outside
749 a character class. The bit map for "word" characters is now created
750 separately from the results of isalnum() instead of just taking it from the
751 upper, lower, and digit maps. (Plus the underscore character, of course.)
753 5. The above bug also affected the handling of POSIX character classes such as
754 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
755 permanent tables. Instead, the bit maps for such a class were previously
756 created as the appropriate unions of the upper, lower, and digit bitmaps.
757 Now they are created by subtraction from the [[:word:]] class, which has
760 6. The [[:blank:]] character class matches horizontal, but not vertical space.
761 It is created by subtracting the vertical space characters (\x09, \x0a,
762 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
763 subtraction was done in the overall bitmap for a character class, meaning
764 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
765 be recognized. This bug has been fixed.
767 7. Patches from the folks at Google:
769 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
770 real life, but is still worth protecting against".
772 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
773 regular expressions".
775 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
778 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
779 "configure" and the latter not, in order to fix a problem somebody had
780 with compiling the Arg class on HP-UX.
782 (e) Improve the error-handling of the C++ wrapper a little bit.
784 (f) New tests for checking recursion limiting.
786 8. The pcre_memmove() function, which is used only if the environment does not
787 have a standard memmove() function (and is therefore rarely compiled),
788 contained two bugs: (a) use of int instead of size_t, and (b) it was not
789 returning a result (though PCRE never actually uses the result).
791 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
792 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
793 returned instead of calling malloc() with an overflowing number that would
794 most likely cause subsequent chaos.
796 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
798 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
799 with this option is matched, the nmatch and pmatch options of regexec() are
802 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
803 provided in case anyone wants to the the POSIX interface with UTF-8
806 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
807 C++ linking (needed for some HP-UX environments).
809 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
810 (unused parameter) and in the pcre_printint() function (omitted "default"
811 switch label when the default is to do nothing).
813 15. Added some code to make it possible, when PCRE is compiled as a C++
814 library, to replace subject pointers for pcre_exec() with a smart pointer
815 class, thus making it possible to process discontinuous strings.
817 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
818 much the same function. They were added by different people who were trying
819 to make PCRE easy to compile on non-Unix systems. It has been suggested
820 that PCRE_EXPORT be abolished now that there is more automatic apparatus
821 for compiling on Windows systems. I have therefore replaced it with
822 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
823 defaults to "extern" for C or "extern C" for C++, which works fine on
824 Unix-like systems. It is now possible to override the value of PCRE_DATA_
825 SCOPE with something explicit in config.h. In addition:
827 (a) pcreposix.h still had just "extern" instead of either of these macros;
828 I have replaced it with PCRE_DATA_SCOPE.
830 (b) Functions such as _pcre_xclass(), which are internal to the library,
831 but external in the C sense, all had PCRE_EXPORT in their definitions.
832 This is apparently wrong for the Windows case, so I have removed it.
833 (It makes no difference on Unix-like systems.)
835 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
836 of recursive calls to match(). This is different to MATCH_LIMIT because
837 that limits the total number of calls to match(), not all of which increase
838 the depth of recursion. Limiting the recursion depth limits the amount of
839 stack (or heap if NO_RECURSE is set) that is used. The default can be set
840 when PCRE is compiled, and changed at run time. A patch from Google adds
841 this functionality to the C++ interface.
843 18. Changes to the handling of Unicode character properties:
845 (a) Updated the table to Unicode 4.1.0.
847 (b) Recognize characters that are not in the table as "Cn" (undefined).
849 (c) I revised the way the table is implemented to a much improved format
850 which includes recognition of ranges. It now supports the ranges that
851 are defined in UnicodeData.txt, and it also amalgamates other
852 characters into ranges. This has reduced the number of entries in the
853 table from around 16,000 to around 3,000, thus reducing its size
854 considerably. I realized I did not need to use a tree structure after
855 all - a binary chop search is just as efficient. Having reduced the
856 number of entries, I extended their size from 6 bytes to 8 bytes to
859 (d) Added support for Unicode script names via properties such as \p{Han}.
861 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
862 matching that character.
864 20. When matching a repeated Unicode property with a minimum greater than zero,
865 (for example \pL{2,}), PCRE could look past the end of the subject if it
866 reached it while seeking the minimum number of characters. This could
867 happen only if some of the characters were more than one byte long, because
868 there is a check for at least the minimum number of bytes.
870 21. Refactored the implementation of \p and \P so as to be more general, to
871 allow for more different types of property in future. This has changed the
872 compiled form incompatibly. Anybody with saved compiled patterns that use
873 \p or \P will have to recompile them.
875 22. Added "Any" and "L&" to the supported property types.
877 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
878 but give a compile time error if the value is greater than 0xff.
880 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
881 accidentally not being installed or uninstalled.
883 25. The pcre.h file was built from pcre.h.in, but the only changes that were
884 made were to insert the current release number. This seemed silly, because
885 it made things harder for people building PCRE on systems that don't run
886 "configure". I have turned pcre.h into a distributed file, no longer built
887 by "configure", with the version identification directly included. There is
888 no longer a pcre.h.in file.
890 However, this change necessitated a change to the pcre-config script as
891 well. It is built from pcre-config.in, and one of the substitutions was the
892 release number. I have updated configure.ac so that ./configure now finds
893 the release number by grepping pcre.h.
895 26. Added the ability to run the tests under valgrind.
898 Version 6.4 05-Sep-05
899 ---------------------
901 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
902 "--" to be printed when multiple files were scanned, even when none of the
903 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
904 consider it to be a bug, and have restored the previous behaviour.
906 2. A couple of code tidies to get rid of compiler warnings.
908 3. The pcretest program used to cheat by referring to symbols in the library
909 whose names begin with _pcre_. These are internal symbols that are not
910 really supposed to be visible externally, and in some environments it is
911 possible to suppress them. The cheating is now confined to including
912 certain files from the library's source, which is a bit cleaner.
914 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
915 file's purpose clearer.
917 5. Reorganized pcre_ucp_findchar().
920 Version 6.3 15-Aug-05
921 ---------------------
923 1. The file libpcre.pc.in did not have general read permission in the tarball.
925 2. There were some problems when building without C++ support:
927 (a) If C++ support was not built, "make install" and "make test" still
930 (b) There were problems when the value of CXX was explicitly set. Some
931 changes have been made to try to fix these, and ...
933 (c) --disable-cpp can now be used to explicitly disable C++ support.
935 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
936 backslash in a target when C++ was disabled. This confuses some
937 versions of "make", apparently. Using an intermediate variable solves
938 this. (Same for CPP_LOBJ.)
940 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
941 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
942 necessary on certain architectures.
944 4. Added a setting of -export-symbols-regex to the link command to remove
945 those symbols that are exported in the C sense, but actually are local
946 within the library, and not documented. Their names all begin with
947 "_pcre_". This is not a perfect job, because (a) we have to except some
948 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
949 available (and never for static libraries). I have made a note to try to
950 find a way round (a) in the future.
953 Version 6.2 01-Aug-05
954 ---------------------
956 1. There was no test for integer overflow of quantifier values. A construction
957 such as {1111111111111111} would give undefined results. What is worse, if
958 a minimum quantifier for a parenthesized subpattern overflowed and became
959 negative, the calculation of the memory size went wrong. This could have
960 led to memory overwriting.
962 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
964 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
965 operating environments where this matters.
967 4. Applied Giuseppe Maxia's patch to add additional features for controlling
968 PCRE options from within the C++ wrapper.
970 5. Named capturing subpatterns were not being correctly counted when a pattern
971 was compiled. This caused two problems: (a) If there were more than 100
972 such subpatterns, the calculation of the memory needed for the whole
973 compiled pattern went wrong, leading to an overflow error. (b) Numerical
974 back references of the form \12, where the number was greater than 9, were
975 not recognized as back references, even though there were sufficient
976 previous subpatterns.
978 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
979 versions of gcc, e.g. 2.95.4.
982 Version 6.1 21-Jun-05
983 ---------------------
985 1. There was one reference to the variable "posix" in pcretest.c that was not
986 surrounded by "#if !defined NOPOSIX".
988 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
989 the cross-check on the old pcre_info() function, for the benefit of the
990 cut-down version of PCRE that is currently imported into Exim.
992 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
993 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
994 patterns that start (?i) but I don't think that matters. The use of (?i) is
995 just an example; this all applies to the other options as well.
997 4. Since libtool seems to echo the compile commands it is issuing, the output
998 from "make" can be reduced a bit by putting "@" in front of each libtool
1001 5. Patch from the folks at Google for configure.in to be a bit more thorough
1002 in checking for a suitable C++ installation before trying to compile the
1003 C++ stuff. This should fix a reported problem when a compiler was present,
1004 but no suitable headers.
1006 6. The man pages all had just "PCRE" as their title. I have changed them to
1007 be the relevant file name. I have also arranged that these names are
1008 retained in the file doc/pcre.txt, which is a concatenation in text format
1009 of all the man pages except the little individual ones for each function.
1011 7. The NON-UNIX-USE file had not been updated for the different set of source
1012 files that come with release 6. I also added a few comments about the C++
1016 Version 6.0 07-Jun-05
1017 ---------------------
1019 1. Some minor internal re-organization to help with my DFA experiments.
1021 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1022 didn't matter for the library itself when fully configured, but did matter
1023 when compiling without UCP support, or within Exim, where the ucp files are
1026 3. Refactoring of the library code to split up the various functions into
1027 different source modules. The addition of the new DFA matching code (see
1028 below) to a single monolithic source would have made it really too
1029 unwieldy, quite apart from causing all the code to be include in a
1030 statically linked application, when only some functions are used. This is
1031 relevant even without the DFA addition now that patterns can be compiled in
1032 one application and matched in another.
1034 The downside of splitting up is that there have to be some external
1035 functions and data tables that are used internally in different modules of
1036 the library but which are not part of the API. These have all had their
1037 names changed to start with "_pcre_" so that they are unlikely to clash
1038 with other external names.
1040 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1041 a different (DFA) algorithm. Although it is slower than the original
1042 function, it does have some advantages for certain types of matching
1045 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1046 including restarting after a partial match.
1048 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1049 defined when compiling for Windows was sent to me. I have put it into the
1050 code, though I have no means of testing or verifying it.
1052 7. Added the pcre_refcount() auxiliary function.
1054 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1055 match before or at the first newline in the subject string. In pcretest,
1056 the /f option on a pattern can be used to set this.
1058 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1059 would behave wrongly. This has been present in PCRE since release 4.0.
1061 10. A number of changes to the pcregrep command:
1063 (a) Refactored how -x works; insert ^(...)$ instead of setting
1064 PCRE_ANCHORED and checking the length, in preparation for adding
1065 something similar for -w.
1067 (b) Added the -w (match as a word) option.
1069 (c) Refactored the way lines are read and buffered so as to have more
1070 than one at a time available.
1072 (d) Implemented a pcregrep test script.
1074 (e) Added the -M (multiline match) option. This allows patterns to match
1075 over several lines of the subject. The buffering ensures that at least
1076 8K, or the rest of the document (whichever is the shorter) is available
1077 for matching (and similarly the previous 8K for lookbehind assertions).
1079 (f) Changed the --help output so that it now says
1083 instead of two lines, one with "regex" and the other with "regexp"
1084 because that confused at least one person since the short forms are the
1085 same. (This required a bit of code, as the output is generated
1086 automatically from a table. It wasn't just a text change.)
1088 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1089 option but starts with a hyphen. Could be a pattern or a path name
1090 starting with a hyphen, for instance.
1092 (h) "-" can be given as a file name to represent stdin.
1094 (i) When file names are being printed, "(standard input)" is used for
1095 the standard input, for compatibility with GNU grep. Previously
1098 (j) The option --label=xxx can be used to supply a name to be used for
1099 stdin when file names are being printed. There is no short form.
1101 (k) Re-factored the options decoding logic because we are going to add
1102 two more options that take data. Such options can now be given in four
1103 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1105 (l) Added the -A, -B, and -C options for requesting that lines of context
1106 around matches be printed.
1108 (m) Added the -L option to print the names of files that do not contain
1109 any matching lines, that is, the complement of -l.
1111 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1112 continue to scan other files.
1114 (o) The -s option was incorrectly implemented. For compatibility with other
1115 greps, it now suppresses the error message for a non-existent or non-
1116 accessible file (but not the return code). There is a new option called
1117 -q that suppresses the output of matching lines, which was what -s was
1120 (p) Added --include and --exclude options to specify files for inclusion
1121 and exclusion when recursing.
1123 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1124 Hopefully, it now does.
1126 12. Missing cast in pcre_study().
1128 13. Added an "uninstall" target to the makefile.
1130 14. Replaced "extern" in the function prototypes in Makefile.in with
1131 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1132 world, but is set differently for Windows.
1134 15. Added a second compiling function called pcre_compile2(). The only
1135 difference is that it has an extra argument, which is a pointer to an
1136 integer error code. When there is a compile-time failure, this is set
1137 non-zero, in addition to the error test pointer being set to point to an
1138 error message. The new argument may be NULL if no error number is required
1139 (but then you may as well call pcre_compile(), which is now just a
1140 wrapper). This facility is provided because some applications need a
1141 numeric error indication, but it has also enabled me to tidy up the way
1142 compile-time errors are handled in the POSIX wrapper.
1144 16. Added VPATH=.libs to the makefile; this should help when building with one
1145 prefix path and installing with another. (Or so I'm told by someone who
1146 knows more about this stuff than I do.)
1148 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1149 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1150 match everything, including newlines. This is not POSIX-compatible, but
1151 somebody wanted the feature. From pcretest it can be activated by using
1152 both the P and the s flags.
1154 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1156 19. libpcre.pc was being incorrectly installed as executable.
1158 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1159 it now also looks for '\r' so that it will work unmodified on Windows.
1161 21. Added Google's contributed C++ wrapper to the distribution.
1163 22. Added some untidy missing memory free() calls in pcretest, to keep
1164 Electric Fence happy when testing.
1168 Version 5.0 13-Sep-04
1169 ---------------------
1171 1. Internal change: literal characters are no longer packed up into items
1172 containing multiple characters in a single byte-string. Each character
1173 is now matched using a separate opcode. However, there may be more than one
1174 byte in the character in UTF-8 mode.
1176 2. The pcre_callout_block structure has two new fields: pattern_position and
1177 next_item_length. These contain the offset in the pattern to the next match
1178 item, and its length, respectively.
1180 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1181 insertion of callouts before each pattern item. Added the /C option to
1182 pcretest to make use of this.
1184 4. On the advice of a Windows user, the lines
1186 #if defined(_WIN32) || defined(WIN32)
1187 _setmode( _fileno( stdout ), 0x8000 );
1188 #endif /* defined(_WIN32) || defined(WIN32) */
1190 have been added to the source of pcretest. This apparently does useful
1191 magic in relation to line terminators.
1193 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1194 for the benefit of those environments where the "b" makes a difference.
1196 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1197 to know about it. I have put a hack into configure.in that adds in code
1198 to set GCC=yes if CC=icc. This seems to end up at a point in the
1199 generated configure script that is early enough to affect the setting of
1200 compiler options, which is what is needed, but I have no means of testing
1201 whether it really works. (The user who reported this had patched the
1202 generated configure script, which of course I cannot do.)
1204 LATER: After change 22 below (new libtool files), the configure script
1205 seems to know about icc (and also ecc). Therefore, I have commented out
1206 this hack in configure.in.
1208 7. Added support for pkg-config (2 patches were sent in).
1210 8. Negated POSIX character classes that used a combination of internal tables
1211 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1212 [[:^ascii]]. Typically, they would match almost any characters. The other
1213 POSIX classes were not broken in this way.
1215 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1216 to find the match, as PCRE was deluded into thinking that the match had to
1217 start at the start point or following a newline. The same bug applied to
1218 patterns with negative forward assertions or any backward assertions
1219 preceding ".*" at the start, unless the pattern required a fixed first
1220 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1222 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1223 starting at the last subject character, bytes beyond the end of the subject
1226 11. Renamed the variable "class" as "classbits" to make life easier for C++
1227 users. (Previously there was a macro definition, but it apparently wasn't
1230 12. Added the new field "tables" to the extra data so that tables can be passed
1231 in at exec time, or the internal tables can be re-selected. This allows
1232 a compiled regex to be saved and re-used at a later time by a different
1233 program that might have everything at different addresses.
1235 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1236 -R library as well as a -L library.
1238 14. The debugging options of pcretest (-d on the command line or D on a
1239 pattern) showed incorrect output for anything following an extended class
1240 that contained multibyte characters and which was followed by a quantifier.
1242 15. Added optional support for general category Unicode character properties
1243 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1244 support. It adds about 90K to the size of the library. The meanings of the
1245 inbuilt class escapes such as \d and \s have NOT been changed.
1247 16. Updated pcredemo.c to include calls to free() to release the memory for the
1250 17. The generated file chartables.c was being created in the source directory
1251 instead of in the building directory. This caused the build to fail if the
1252 source directory was different from the building directory, and was
1255 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1256 file. No doubt somebody will tell me if they don't make sense... Also added
1257 Dan Mooney's comments about building on OpenVMS.
1259 19. Added support for partial matching via the PCRE_PARTIAL option for
1260 pcre_exec() and the \P data escape in pcretest.
1262 20. Extended pcretest with 3 new pattern features:
1264 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1265 write the compiled pattern to the file whose name is "rest-of-line".
1266 This is a straight binary dump of the data, with the saved pointer to
1267 the character tables forced to be NULL. The study data, if any, is
1268 written too. After writing, pcretest reads a new pattern.
1270 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1271 compiled pattern from the given file. There must not be any
1272 occurrences of "<" in the file name (pretty unlikely); if there are,
1273 pcretest will instead treat the initial "<" as a pattern delimiter.
1274 After reading in the pattern, pcretest goes on to read data lines as
1277 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1278 and 16-bit fields in a compiled pattern, to simulate a pattern that
1279 was compiled on a host of opposite endianness.
1281 21. The pcre-exec() function can now cope with patterns that were compiled on
1282 hosts of opposite endianness, with this restriction:
1284 As for any compiled expression that is saved and used later, the tables
1285 pointer field cannot be preserved; the extra_data field in the arguments
1286 to pcre_exec() should be used to pass in a tables address if a value
1287 other than the default internal tables were used at compile time.
1289 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1290 now diagnosed as an error. Previously, most of the time, a negative number
1291 would have been treated as zero, but if in addition "ovector" was passed as
1292 NULL, a crash could occur.
1294 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1295 new versions from the libtool 1.5 distribution (the last one is a copy of
1296 a file called libtool.m4). This seems to have fixed the need to patch
1297 "configure" to support Darwin 1.3 (which I used to do). However, I still
1298 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1301 24. Changed the PCRE licence to be the more standard "BSD" licence.
1304 Version 4.5 01-Dec-03
1305 ---------------------
1307 1. There has been some re-arrangement of the code for the match() function so
1308 that it can be compiled in a version that does not call itself recursively.
1309 Instead, it keeps those local variables that need separate instances for
1310 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1311 needs to "recurse". Keeping track of where control must go is done by means
1312 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1313 hide most of the details from the main code, and operates only if
1314 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1315 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1318 To make it easier for callers to provide specially tailored get/free
1319 functions for this usage, two new functions, pcre_stack_malloc, and
1320 pcre_stack_free, are used. They are always called in strict stacking order,
1321 and the size of block requested is always the same.
1323 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1324 PCRE has been compiled to use the stack or the heap for recursion. The
1325 -C option of pcretest uses this to show which version is compiled.
1327 A new data escape \S, is added to pcretest; it causes the amounts of store
1328 obtained and freed by both kinds of malloc/free at match time to be added
1331 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1332 what's available on my current Linux desktop machine.
1334 3. When matching a UTF-8 string, the test for a valid string at the start has
1335 been extended. If start_offset is not zero, PCRE now checks that it points
1336 to a byte that is the start of a UTF-8 character. If not, it returns
1337 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1338 this is necessary because there may be backward assertions in the pattern.
1339 When matching the same subject several times, it may save resources to use
1340 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1342 4. The code for checking the validity of UTF-8 strings has been tightened so
1343 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1344 containing "overlong sequences".
1346 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1347 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1348 should have been "&", but it just so happened that all the cases this let
1349 through by mistake were picked up later in the function.
1351 6. I had used a variable called "isblank" - this is a C99 function, causing
1352 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1354 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1355 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1356 so the version is shown; (c) stop "make test" echoing "./RunTest".
1358 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1360 9. The prototype for memmove() for systems that don't have it was using
1361 size_t, but the inclusion of the header that defines size_t was later. I've
1362 moved the #includes for the C headers earlier to avoid this.
1364 10. Added some adjustments to the code to make it easier to compiler on certain
1367 (a) Some "const" qualifiers were missing.
1368 (b) Added the macro EXPORT before all exported functions; by default this
1369 is defined to be empty.
1370 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1371 that it reads its output file name as an argument instead of writing
1372 to the standard output and assuming this can be redirected.
1374 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1375 class containing characters with values greater than 255, PCRE compilation
1378 12. A recursive reference to a subpattern that was within another subpattern
1379 that had a minimum quantifier of zero caused PCRE to crash. For example,
1380 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1381 recursion. If the recursively-called subpattern itself had a zero repeat,
1384 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1385 buffer into which it was copied (for escape processing) was still set at
1386 1024, so long lines caused crashes.
1388 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1389 "internal error: code overflow...". This applied to any character class
1390 that was followed by a possessive quantifier.
1392 15. Modified the Makefile to add libpcre.la as a prerequisite for
1393 libpcreposix.la because I was told this is needed for a parallel build to
1396 16. If a pattern that contained .* following optional items at the start was
1397 studied, the wrong optimizing data was generated, leading to matching
1398 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1399 matching string must start with a or b or c. The correct conclusion for
1400 this pattern is that a match can start with any character.
1403 Version 4.4 13-Aug-03
1404 ---------------------
1406 1. In UTF-8 mode, a character class containing characters with values between
1407 127 and 255 was not handled correctly if the compiled pattern was studied.
1408 In fixing this, I have also improved the studying algorithm for such
1411 2. Three internal functions had redundant arguments passed to them. Removal
1412 might give a very teeny performance improvement.
1414 3. Documentation bug: the value of the capture_top field in a callout is *one
1415 more than* the number of the hightest numbered captured substring.
1417 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1418 in incorrectly linking with a previously installed version. They now link
1419 explicitly with libpcre.la.
1421 5. configure.in no longer needs to recognize Cygwin specially.
1423 6. A problem in pcre.in for Windows platforms is fixed.
1425 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1426 pcretest, it used to include the size of the study block as part of its
1427 output. Unfortunately, the structure contains a field that has a different
1428 size on different hardware architectures. This meant that the tests that
1429 showed this size failed. As the block is currently always of a fixed size,
1430 this information isn't actually particularly useful in pcretest output, so
1431 I have just removed it.
1433 8. Three pre-processor statements accidentally did not start in column 1.
1434 Sadly, there are *still* compilers around that complain, even though
1435 standard C has not required this for well over a decade. Sigh.
1437 9. In pcretest, the code for checking callouts passed small integers in the
1438 callout_data field, which is a void * field. However, some picky compilers
1439 complained about the casts involved for this on 64-bit systems. Now
1440 pcretest passes the address of the small integer instead, which should get
1441 rid of the warnings.
1443 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1444 both compile and run time, and gives an error if an invalid UTF-8 sequence
1445 is found. There is a option for disabling this check in cases where the
1446 string is known to be correct and/or the maximum performance is wanted.
1448 11. In response to a bug report, I changed one line in Makefile.in from
1450 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1452 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1454 to look similar to other lines, but I have no way of telling whether this
1455 is the right thing to do, as I do not use Windows. No doubt I'll get told
1459 Version 4.3 21-May-03
1460 ---------------------
1462 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1465 2. Some refactoring to improve the quality of the code:
1467 (i) The utf8_table... variables are now declared "const".
1469 (ii) The code for \cx, which used the "case flipping" table to upper case
1470 lower case letters, now just substracts 32. This is ASCII-specific,
1471 but the whole concept of \cx is ASCII-specific, so it seems
1474 (iii) PCRE was using its character types table to recognize decimal and
1475 hexadecimal digits in the pattern. This is silly, because it handles
1476 only 0-9, a-f, and A-F, but the character types table is locale-
1477 specific, which means strange things might happen. A private
1478 table is now used for this - though it costs 256 bytes, a table is
1479 much faster than multiple explicit tests. Of course, the standard
1480 character types table is still used for matching digits in subject
1483 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1484 ending in _t are). So I've renamed it as ESC_tee.
1486 3. The first argument for regexec() in the POSIX wrapper should have been
1489 4. Changed pcretest to use malloc() for its buffers so that they can be
1490 Electric Fenced for debugging.
1492 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1493 to read one or more bytes before the start of the subject string. Often this
1494 had no effect on PCRE's behaviour, but in some circumstances it could
1495 provoke a segmentation fault.
1497 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1498 to try to read one or more bytes before the start of the subject string.
1500 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1501 UTF-8 support could misbehave in various ways if the subject string
1502 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1503 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1504 back over UTF-8 characters.)
1507 Version 4.2 14-Apr-03
1508 ---------------------
1510 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1512 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1513 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1514 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1515 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1516 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1518 Note: automatic setting of the BUILD variables is not yet working
1519 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1520 compile-time but not at link-time
1521 [LINK]: use for linking executables only
1522 make different versions for Windows and non-Windows
1523 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1525 [LINK_FOR_BUILD]: new variable
1526 [OBJEXT]: use throughout
1527 [EXEEXT]: use throughout
1528 <winshared>: new target
1529 <wininstall>: new target
1530 <dftables.o>: use native compiler
1531 <dftables>: use native linker
1532 <install>: handle Windows platform correctly
1535 copy DLL to top builddir before testing
1537 As part of these changes, -no-undefined was removed again. This was reported
1538 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1541 3. Some tidies to get rid of compiler warnings:
1543 . In the match_data structure, match_limit was an unsigned long int, whereas
1544 match_call_count was an int. I've made them both unsigned long ints.
1546 . In pcretest the fact that a const uschar * doesn't automatically cast to
1547 a void * provoked a warning.
1549 . Turning on some more compiler warnings threw up some "shadow" variables
1550 and a few more missing casts.
1552 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1553 option, a class that contained a single character with a value between 128
1554 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1556 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1557 option, a class that contained several characters, but with at least one
1558 whose value was between 128 and 255 caused PCRE to crash.
1561 Version 4.1 12-Mar-03
1562 ---------------------
1564 1. Compiling with gcc -pedantic found a couple of places where casts were
1565 needed, and a string in dftables.c that was longer than standard compilers are
1566 required to support.
1568 2. Compiling with Sun's compiler found a few more places where the code could
1569 be tidied up in order to avoid warnings.
1571 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1572 first of these names is deprecated in the latest Autoconf in favour of the name
1573 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1574 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1575 analogy I have changed it to CFLAGS_FOR_BUILD.
1577 4. Added -no-undefined to the linking command in the Makefile, because this is
1578 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1579 linking step for the pcreposix library.
1581 5. PCRE was failing to diagnose the case of two named groups with the same
1584 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1585 literal character that is needed in the subject for a match, and scans along to
1586 ensure that it is present before embarking on the full matching process. This
1587 saves time in cases of nested unlimited repeats that are never going to match.
1588 Problem: the scan can take a lot of time if the subject is very long (e.g.
1589 megabytes), thus penalizing straightforward matches. It is now done only if the
1590 amount of subject to be scanned is less than 1000 bytes.
1592 7. A lesser problem with the same optimization is that it was recording the
1593 first character of an anchored pattern as "needed", thus provoking a search
1594 right along the subject, even when the first match of the pattern was going to
1595 fail. The "needed" character is now not set for anchored patterns, unless it
1596 follows something in the pattern that is of non-fixed length. Thus, it still
1597 fulfils its original purpose of finding quick non-matches in cases of nested
1598 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1601 Version 4.0 17-Feb-03
1602 ---------------------
1604 1. If a comment in an extended regex that started immediately after a meta-item
1605 extended to the end of string, PCRE compiled incorrect data. This could lead to
1606 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
1608 2. Moved to autoconf 2.53 and libtool 1.4.2.
1610 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
1611 the special perltest8 script is no longer needed - all the tests can be run
1612 from a single perltest script.
1614 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
1615 by \s. It has now been removed in PCRE. This means it isn't recognized as
1616 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
1617 class [:space:] *does* include VT, thereby creating a mess.
1619 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
1622 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
1623 its new features into the main test script, reducing the number of scripts.
1625 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
1626 were backward compatible, and made the (?i) apply to the whole pattern, as if
1627 /i were given. Now it behaves more logically, and applies the option setting
1628 only to what follows. PCRE has been changed to follow suit. However, if it
1629 finds options settings right at the start of the pattern, it extracts them into
1630 the global options, as before. Thus, they show up in the info data.
1632 8. Added support for the \Q...\E escape sequence. Characters in between are
1633 treated as literals. This is slightly different from Perl in that $ and @ are
1634 also handled as literals inside the quotes. In Perl, they will cause variable
1635 interpolation. Note the following examples:
1637 Pattern PCRE matches Perl matches
1639 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
1640 \Qabc\$xyz\E abc\$xyz abc\$xyz
1641 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
1643 For compatibility with Perl, \Q...\E sequences are recognized inside character
1644 classes as well as outside them.
1646 9. Re-organized 3 code statements in pcretest to avoid "overflow in
1647 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
1648 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
1649 signed/unsigned warnings.
1651 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
1652 option for pcretest, so I've replaced it by a simple function that does just
1655 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
1658 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
1659 Java package. This provides some syntactic sugar for simple cases of what my
1660 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
1661 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
1662 item, you can use this simplified notation. Note that only makes sense with
1663 greedy quantifiers. Consequently, the use of the possessive quantifier forces
1664 greediness, whatever the setting of the PCRE_UNGREEDY option.
1666 13. A change of greediness default within a pattern was not taking effect at
1667 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
1668 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
1669 was abstracted outside.
1671 14. PCRE now supports the \G assertion. It is true when the current matching
1672 position is at the start point of the match. This differs from \A when the
1673 starting offset is non-zero. Used with the /g option of pcretest (or similar
1674 code), it works in the same way as it does for Perl's /g option. If all
1675 alternatives of a regex begin with \G, the expression is anchored to the start
1676 match position, and the "anchored" flag is set in the compiled expression.
1678 15. Some bugs concerning the handling of certain option changes within patterns
1679 have been fixed. These applied to options other than (?ims). For example,
1680 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
1681 been the other way round. Some of this was related to change 7 above.
1683 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
1684 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
1685 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
1686 POSIX classes only within a class (e.g. /[[:alpha:]]/).
1688 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
1689 mode. Unlike ".", it always matches newline, whatever the setting of
1690 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
1691 assertions. Perl allows it, but it doesn't (in general) work because it can't
1692 calculate the length of the lookbehind. At least, that's the case for Perl
1693 5.8.0 - I've been told they are going to document that it doesn't work in
1696 18. Added an error diagnosis for escapes that PCRE does not support: these are
1697 \L, \l, \N, \P, \p, \U, \u, and \X.
1699 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
1700 reading past the end of the pattern in cases such as /[abcd/.
1702 20. PCRE was getting more memory than necessary for patterns with classes that
1703 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
1705 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
1706 compiling PCRE for use with Virtual Pascal.
1708 22. Small fix to the Makefile to make it work properly if the build is done
1709 outside the source tree.
1711 23. Added a new extension: a condition to go with recursion. If a conditional
1712 subpattern starts with (?(R) the "true" branch is used if recursion has
1713 happened, whereas the "false" branch is used only at the top level.
1715 24. When there was a very long string of literal characters (over 255 bytes
1716 without UTF support, over 250 bytes with UTF support), the computation of how
1717 much memory was required could be incorrect, leading to segfaults or other
1720 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
1721 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
1722 there was a subsequent back reference to those brackets. This meant that, for
1723 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
1724 possible to check for precisely this case. All we can do is abandon the
1725 optimization if .* occurs inside capturing brackets when there are any back
1726 references whatsoever. (See below for a better fix that came later.)
1728 26. The handling of the optimization for finding the first character of a
1729 non-anchored pattern, and for finding a character that is required later in the
1730 match were failing in some cases. This didn't break the matching; it just
1731 failed to optimize when it could. The way this is done has been re-implemented.
1733 27. Fixed typo in error message for invalid (?R item (it said "(?p").
1735 28. Added a new feature that provides some of the functionality that Perl
1736 provides with (?{...}). The facility is termed a "callout". The way it is done
1737 in PCRE is for the caller to provide an optional function, by setting
1738 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
1739 global variable. By default it is unset, which disables all calling out. To get
1740 the function called, the regex must include (?C) at appropriate points. This
1741 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
1742 This provides a means of identifying different callout points. When PCRE
1743 reaches such a point in the regex, if pcre_callout has been set, the external
1744 function is called. It is provided with data in a structure called
1745 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
1746 matching continues; if it returns a non-zero value, the match at the current
1747 point fails. However, backtracking will occur if possible. [This was changed
1748 later and other features added - see item 49 below.]
1750 29. pcretest is upgraded to test the callout functionality. It provides a
1751 callout function that displays information. By default, it shows the start of
1752 the match and the current position in the text. There are some new data escapes
1753 to vary what happens:
1755 \C+ in addition, show current contents of captured substrings
1756 \C- do not supply a callout function
1757 \C!n return 1 when callout number n is reached
1758 \C!n!m return 1 when callout number n is reached for the mth time
1760 30. If pcregrep was called with the -l option and just a single file name, it
1761 output "<stdin>" if a match was found, instead of the file name.
1763 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
1764 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
1765 pcre_exec(). This saves a malloc/free per call. The default value of
1766 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
1769 32. The default maximum size of a compiled pattern is 64K. There have been a
1770 few cases of people hitting this limit. The code now uses macros to handle the
1771 storing of links as offsets within the compiled pattern. It defaults to 2-byte
1772 links, but this can be changed to 3 or 4 bytes by --with-link-size when
1773 configuring. Tests 2 and 5 work only with 2-byte links because they output
1774 debugging information about compiled patterns.
1776 33. Internal code re-arrangements:
1778 (a) Moved the debugging function for printing out a compiled regex into
1779 its own source file (printint.c) and used #include to pull it into
1780 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
1783 (b) Defined the list of op-code names for debugging as a macro in
1784 internal.h so that it is next to the definition of the opcodes.
1786 (c) Defined a table of op-code lengths for simpler skipping along compiled
1787 code. This is again a macro in internal.h so that it is next to the
1788 definition of the opcodes.
1790 34. Added support for recursive calls to individual subpatterns, along the
1791 lines of Robin Houston's patch (but implemented somewhat differently).
1793 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
1794 allow it to read and process whole directories in Win32. This code was
1795 contributed by Lionel Fourquaux; it has not been tested by me.
1797 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
1798 used to name a group. Names consist of alphanumerics and underscores, and must
1799 be unique. Back references use the syntax (?P=name) and recursive calls use
1800 (?P>name) which is a PCRE extension to the Python extension. Groups still have
1801 numbers. The function pcre_fullinfo() can be used after compilation to extract
1802 a name/number map. There are three relevant calls:
1804 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
1805 PCRE_INFO_NAMECOUNT yields the number of entries
1806 PCRE_INFO_NAMETABLE yields a pointer to the map.
1808 The map is a vector of fixed-size entries. The size of each entry depends on
1809 the length of the longest name used. The first two bytes of each entry are the
1810 group number, most significant byte first. There follows the corresponding
1811 name, zero terminated. The names are in alphabetical order.
1813 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
1814 case instead of 255. Making it the same both with and without UTF-8 support
1815 means that the same test output works with both.
1817 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
1818 calling malloc() with a zero argument.
1820 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
1821 optimization. I've improved things by keeping a bitmap of backreferences with
1822 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
1823 fact referenced, the optimization can be applied. It is unlikely that a
1824 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
1825 the match to follow \n) will appear inside brackets with a number greater than
1826 31, but if it does, any back reference > 31 suppresses the optimization.
1828 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
1829 of disabling numbered capturing parentheses. Any opening parenthesis that is
1830 not followed by ? behaves as if it were followed by ?: but named parentheses
1831 can still be used for capturing (and they will acquire numbers in the usual
1834 41. Redesigned the return codes from the match() function into yes/no/error so
1835 that errors can be passed back from deep inside the nested calls. A malloc
1836 failure while inside a recursive subpattern call now causes the
1837 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
1839 42. It is now possible to set a limit on the number of times the match()
1840 function is called in a call to pcre_exec(). This facility makes it possible to
1841 limit the amount of recursion and backtracking, though not in a directly
1842 obvious way, because the match() function is used in a number of different
1843 circumstances. The count starts from zero for each position in the subject
1844 string (for non-anchored patterns). The default limit is, for compatibility, a
1845 large number, namely 10 000 000. You can change this in two ways:
1847 (a) When configuring PCRE before making, you can use --with-match-limit=n
1848 to set a default value for the compiled library.
1850 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
1851 a different value is set. See 45 below.
1853 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1855 43. Added a new function pcre_config(int, void *) to enable run-time extraction
1856 of things that can be changed at compile time. The first argument specifies
1857 what is wanted and the second points to where the information is to be placed.
1858 The current list of available information is:
1862 The output is an integer that is set to one if UTF-8 support is available;
1863 otherwise it is set to zero.
1867 The output is an integer that it set to the value of the code that is used for
1868 newline. It is either LF (10) or CR (13).
1870 PCRE_CONFIG_LINK_SIZE
1872 The output is an integer that contains the number of bytes used for internal
1873 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
1875 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
1877 The output is an integer that contains the threshold above which the POSIX
1878 interface uses malloc() for output vectors. See item 31 above.
1880 PCRE_CONFIG_MATCH_LIMIT
1882 The output is an unsigned integer that contains the default limit of the number
1883 of match() calls in a pcre_exec() execution. See 42 above.
1885 44. pcretest has been upgraded by the addition of the -C option. This causes it
1886 to extract all the available output from the new pcre_config() function, and to
1887 output it. The program then exits immediately.
1889 45. A need has arisen to pass over additional data with calls to pcre_exec() in
1890 order to support additional features. One way would have been to define
1891 pcre_exec2() (for example) with extra arguments, but this would not have been
1892 extensible, and would also have required all calls to the original function to
1893 be mapped to the new one. Instead, I have chosen to extend the mechanism that
1894 is used for passing in "extra" data from pcre_study().
1896 The pcre_extra structure is now exposed and defined in pcre.h. It currently
1897 contains the following fields:
1899 flags a bitmap indicating which of the following fields are set
1900 study_data opaque data from pcre_study()
1901 match_limit a way of specifying a limit on match() calls for a specific
1903 callout_data data for callouts (see 49 below)
1905 The flag bits are also defined in pcre.h, and are
1907 PCRE_EXTRA_STUDY_DATA
1908 PCRE_EXTRA_MATCH_LIMIT
1909 PCRE_EXTRA_CALLOUT_DATA
1911 The pcre_study() function now returns one of these new pcre_extra blocks, with
1912 the actual study data pointed to by the study_data field, and the
1913 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
1914 before. That is, this change is entirely upwards-compatible and requires no
1915 change to existing code.
1917 If you want to pass in additional data to pcre_exec(), you can either place it
1918 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
1921 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
1922 data string contains the escape sequence \M, pcretest calls pcre_exec() several
1923 times with different match limits, until it finds the minimum value needed for
1924 pcre_exec() to complete. The value is then output. This can be instructive; for
1925 most simple matches the number is quite small, but for pathological cases it
1926 gets very large very quickly.
1928 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
1929 returns the size of the data block pointed to by the study_data field in a
1930 pcre_extra block, that is, the value that was passed as the argument to
1931 pcre_malloc() when PCRE was getting memory in which to place the information
1932 created by pcre_study(). The fourth argument should point to a size_t variable.
1933 pcretest has been extended so that this information is shown after a successful
1934 pcre_study() call when information about the compiled regex is being displayed.
1936 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
1937 because what follows is always an absolute path. (Later: it turns out that this
1938 is more than cosmetic for MinGW, because it doesn't like empty path
1941 49. Some changes have been made to the callout feature (see 28 above):
1943 (i) A callout function now has three choices for what it returns:
1945 0 => success, carry on matching
1946 > 0 => failure at this point, but backtrack if possible
1947 < 0 => serious error, return this value from pcre_exec()
1949 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
1950 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
1951 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
1952 use by callout functions. It will never be used by PCRE itself.
1954 (ii) The pcre_extra structure (see 45 above) has a void * field called
1955 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
1956 pcre_callout_block structure has a field of the same name. The contents of
1957 the field passed in the pcre_extra structure are passed to the callout
1958 function in the corresponding field in the callout block. This makes it
1959 easier to use the same callout-containing regex from multiple threads. For
1960 testing, the pcretest program has a new data escape
1962 \C*n pass the number n (may be negative) as callout_data
1964 If the callout function in pcretest receives a non-zero value as
1965 callout_data, it returns that value.
1967 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
1968 there were some redundant $(CFLAGS) in commands that are now specified as
1969 $(LINK), which already includes $(CFLAGS).
1971 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
1972 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
1973 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
1974 one-byte characters throughout. Note that case-insensitive matching applies
1975 only to characters whose values are less than 256. PCRE doesn't support the
1976 notion of cases for higher-valued characters.
1978 (i) A character class whose characters are all within 0-255 is handled as
1979 a bit map, and the map is inverted for negative classes. Previously, a
1980 character > 255 always failed to match such a class; however it should
1981 match if the class was a negative one (e.g. [^ab]). This has been fixed.
1983 (ii) A negated character class with a single character < 255 is coded as
1984 "not this character" (OP_NOT). This wasn't working properly when the test
1985 character was multibyte, either singly or repeated.
1987 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
1988 mode, for example: \x{100}{2,3}.
1990 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
1991 singly or repeated) now correctly test multibyte characters. However,
1992 PCRE doesn't recognize any characters with values greater than 255 as
1993 digits, spaces, or word characters. Such characters always match \D, \S,
1994 and \W, and never match \d, \s, or \w.
1996 (v) Classes may now contain characters and character ranges with values
1997 greater than 255. For example: [ab\x{100}-\x{400}].
1999 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2002 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2003 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2004 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2007 53. The single man page has become too large. I have therefore split it up into
2008 a number of separate man pages. These also give rise to individual HTML pages;
2009 these are now put in a separate directory, and there is an index.html page that
2010 lists them all. Some hyperlinking between the pages has been installed.
2012 54. Added convenience functions for handling named capturing parentheses.
2014 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2015 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2016 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2019 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2020 calling configure. These values are used when compiling the dftables.c program
2021 which is run to generate the source of the default character tables. They
2022 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2023 you will need to set these values.
2025 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2028 Version 3.9 02-Jan-02
2029 ---------------------
2031 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2033 2. If --disable-static was given, the building process failed when trying to
2034 build pcretest and pcregrep. (For some reason it was using libtool to compile
2035 them, which is not right, as they aren't part of the library.)
2038 Version 3.8 18-Dec-01
2039 ---------------------
2041 1. The experimental UTF-8 code was completely screwed up. It was packing the
2042 bytes in the wrong order. How dumb can you get?
2045 Version 3.7 29-Oct-01
2046 ---------------------
2048 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2049 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2050 this didn't happen under Solaris 8, where I normally test things.
2052 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2053 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2054 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2055 this fix an hour or so after the initial 3.7 release.)
2058 Version 3.6 23-Oct-01
2059 ---------------------
2061 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2062 offsets passed as NULL with zero offset count.
2064 2. The config.guess and config.sub files had not been updated when I moved to
2065 the latest autoconf.
2068 Version 3.5 15-Aug-01
2069 ---------------------
2071 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2074 2. By using declared but undefined structures, we can avoid using "void"
2075 definitions in pcre.h while keeping the internal definitions of the structures
2078 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2079 user point of view, this means that both static and shared libraries are built
2080 by default, but this can be individually controlled. More of the work of
2081 handling this static/shared cases is now inside libtool instead of PCRE's make
2084 4. The pcretest utility is now installed along with pcregrep because it is
2085 useful for users (to test regexs) and by doing this, it automatically gets
2086 relinked by libtool. The documentation has been turned into a man page, so
2087 there are now .1, .txt, and .html versions in /doc.
2089 5. Upgrades to pcregrep:
2090 (i) Added long-form option names like gnu grep.
2091 (ii) Added --help to list all options with an explanatory phrase.
2092 (iii) Added -r, --recursive to recurse into sub-directories.
2093 (iv) Added -f, --file to read patterns from a file.
2095 6. pcre_exec() was referring to its "code" argument before testing that
2096 argument for NULL (and giving an error if it was NULL).
2098 7. Upgraded Makefile.in to allow for compiling in a different directory from
2099 the source directory.
2101 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2102 options bits, the pointer it was passed was to an int instead of to an unsigned
2103 long int. This mattered only on 64-bit systems.
2105 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2106 generated) instead of pcre.in, which it its source. Also made the same change
2107 in several of the .c files.
2109 10. A new release of gcc defines printf() as a macro, which broke pcretest
2110 because it had an ifdef in the middle of a string argument for printf(). Fixed
2111 by using separate calls to printf().
2113 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2114 script, to force use of CR or LF instead of \n in the source. On non-Unix
2115 systems, the value can be set in config.h.
2117 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2118 absolute limit. Changed the text of the error message to make this clear, and
2119 likewise updated the man page.
2121 13. The limit of 99 on the number of capturing subpatterns has been removed.
2122 The new limit is 65535, which I hope will not be a "real" limit.
2125 Version 3.4 22-Aug-00
2126 ---------------------
2128 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2130 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2133 Version 3.3 01-Aug-00
2134 ---------------------
2136 1. If an octal character was given, but the value was greater than \377, it
2137 was not getting masked to the least significant bits, as documented. This could
2138 lead to crashes in some systems.
2140 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2141 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2143 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2144 These just pass their arguments on to (pcre_free)(), but they are provided
2145 because some uses of PCRE bind it to non-C systems that can call its functions,
2146 but cannot call free() or pcre_free() directly.
2148 4. Add "make test" as a synonym for "make check". Corrected some comments in
2151 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2154 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2155 command called pgrep for grepping around the active processes.
2157 7. Added the beginnings of support for UTF-8 character strings.
2159 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2160 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2161 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2162 out for the ar command.)
2165 Version 3.2 12-May-00
2166 ---------------------
2168 This is purely a bug fixing release.
2170 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2171 of ZA. This was just one example of several cases that could provoke this bug,
2172 which was introduced by change 9 of version 2.00. The code for breaking
2173 infinite loops after an iteration that matches an empty string was't working
2176 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2177 when matched against abbab (for example). After matching an empty string, it
2178 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2179 caused it to match further down the string than it should.
2181 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2182 was there because it doesn't seem to be needed, and it causes trouble on some
2183 systems, as it is not a Standard C header. It has been removed.
2185 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2186 were reported on the Macintosh. The changes were from
2188 while ((c = *(++ptr)) != 0 && c != '\n');
2190 while ((c = *(++ptr)) != 0 && c != '\n') ;
2192 Totally extraordinary, but if that's what it takes...
2194 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2195 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2196 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2197 assumes the way PCRE uses memmove() (always moving upwards).
2199 6. PCRE is being used in one environment where strchr() is not available. There
2200 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2204 Version 3.1 09-Feb-00
2205 ---------------------
2207 The only change in this release is the fixing of some bugs in Makefile.in for
2208 the "install" target:
2210 (1) It was failing to install pcreposix.h.
2212 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2215 Version 3.0 01-Feb-00
2216 ---------------------
2218 1. Add support for the /+ modifier to perltest (to output $` like it does in
2221 2. Add support for the /g modifier to perltest.
2223 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2224 matches null strings.
2226 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2227 pattern. Perl treats empty patterns specially - it reuses the most recent
2228 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2231 5. The POSIX interface was broken in that it was just handing over the POSIX
2232 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2233 required a bigger vector, with some working space on the end. This means that
2234 the POSIX wrapper now has to get and free some memory, and copy the results.
2236 6. Added some simple autoconf support, placing the test data and the
2237 documentation in separate directories, re-organizing some of the
2238 information files, and making it build pcre-config (a GNU standard). Also added
2239 libtool support for building PCRE as a shared library, which is now the
2242 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2243 09 are not valid octal constants. Single digits will be used for minor values
2246 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2247 existing programs that set these in the POSIX interface can use PCRE without
2250 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2251 return all that pcre_info() returns, plus additional data. The pcre_info()
2252 function is retained for compatibility, but is considered to be obsolete.
2254 10. Added experimental recursion feature (?R) to handle one common case that
2255 Perl 5.6 will be able to do with (?p{...}).
2257 11. Added support for POSIX character classes like [:alpha:], which Perl is
2261 Version 2.08 31-Aug-99
2262 ----------------------
2264 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2265 trying to match at the startoffset position, but instead was moving forward to
2266 the next newline as if a previous match had failed.
2268 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2269 and could get into a loop if a null string was matched other than at the start
2272 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2273 be distinguished at compile time, and for completeness also added PCRE_DATE.
2275 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2276 in GnuWin32 environments.
2279 Version 2.07 29-Jul-99
2280 ----------------------
2282 1. The documentation is now supplied in plain text form and HTML as well as in
2283 the form of man page sources.
2285 2. C++ compilers don't like assigning (void *) values to other pointer types.
2286 In particular this affects malloc(). Although there is no problem in Standard
2287 C, I've put in casts to keep C++ compilers happy.
2289 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2290 should be (const char *).
2292 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2293 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2294 However, I haven't made this a standard facility. The documentation doesn't
2295 mention it, and the Makefile doesn't support it.
2297 5. The Makefile now contains an "install" target, with editable destinations at
2298 the top of the file. The pcretest program is not installed.
2300 6. pgrep -V now gives the PCRE version number and date.
2302 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2303 causing the entire string to be ignored, instead of just the last character.
2305 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2306 non-matching string, it can take a very, very long time, even for strings of
2307 quite modest length, because of the nested recursion. PCRE now does better in
2308 some of these cases. It does this by remembering the last required literal
2309 character in the pattern, and pre-searching the subject to ensure it is present
2310 before running the real match. In other words, it applies a heuristic to detect
2311 some types of certain failure quickly, and in the above example, if presented
2312 with a string that has no trailing " it gives "no match" very quickly.
2314 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2315 other alternatives are tried instead.
2318 Version 2.06 09-Jun-99
2319 ----------------------
2321 1. Change pcretest's output for amount of store used to show just the code
2322 space, because the remainder (the data block) varies in size between 32-bit and
2325 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2326 start matching at. This allows lookbehinds to work when searching for multiple
2327 occurrences in a string.
2329 3. Added additional options to pcretest for testing multiple occurrences:
2331 /+ outputs the rest of the string that follows a match
2332 /g loops for multiple occurrences, using the new startoffset argument
2333 /G loops for multiple occurrences by passing an incremented pointer
2335 4. PCRE wasn't doing the "first character" optimization for patterns starting
2336 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2337 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2338 the letter 'x'. On long subject strings, this gives a significant speed-up.
2341 Version 2.05 21-Apr-99
2342 ----------------------
2344 1. Changed the type of magic_number from int to long int so that it works
2345 properly on 16-bit systems.
2347 2. Fixed a bug which caused patterns starting with .* not to work correctly
2348 when the subject string contained newline characters. PCRE was assuming
2349 anchoring for such patterns in all cases, which is not correct because .* will
2350 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2351 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2352 must be retried after every newline in the subject.
2355 Version 2.04 18-Feb-99
2356 ----------------------
2358 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2359 computation of the store needed to hold the pattern was incorrect (too large).
2360 If such patterns were nested a few deep, this could multiply and become a real
2363 2. Added /M option to pcretest to show the memory requirement of a specific
2364 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2366 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2367 compiled in such a way that the backtracking after subsequent failure was
2368 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2369 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2372 Version 2.03 02-Feb-99
2373 ----------------------
2375 1. Fixed typo and small mistake in man page.
2377 2. Added 4th condition (GPL supersedes if conflict) and created separate
2378 LICENCE file containing the conditions.
2380 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2381 Perl, that is the internal \ allows the delimiter to be included in the
2382 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2383 the final delimiter, add \ to the end of the pattern (to test the error).
2385 4. Added the convenience functions for extracting substrings after a successful
2386 match. Updated pcretest to make it able to test these functions.
2389 Version 2.02 14-Jan-99
2390 ----------------------
2392 1. Initialized the working variables associated with each extraction so that
2393 their saving and restoring doesn't refer to uninitialized store.
2395 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2396 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2399 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2400 calls, and wasn't printing the correct value for compiling calls. Increased the
2401 default value of LOOPREPEAT, and the number of significant figures in the
2404 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2406 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2407 a building problem on Windows NT with a FAT file system.
2410 Version 2.01 21-Oct-98
2411 ----------------------
2413 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2414 to character tables built by pcre_maketables() in the current locale. If NULL
2415 is passed, the default tables are used.
2418 Version 2.00 24-Sep-98
2419 ----------------------
2421 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2424 2. Allow quantification of (?>) groups, and make it work correctly.
2426 3. The first character computation wasn't working for (?>) groups.
2428 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2429 end of the subject) and add 5.005's \z, which really does match only at the
2430 very end of the subject.
2432 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2434 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2435 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2436 localized options. All options to pcre_study() were also removed.
2438 7. Add other new features from 5.005:
2440 $(?<= positive lookbehind
2441 $(?<! negative lookbehind
2442 (?imsx-imsx) added the unsetting capability
2443 such a setting is global if at outer level; local otherwise
2444 (?imsx-imsx:) non-capturing groups with option setting
2445 (?(cond)re|re) conditional pattern matching
2447 A backreference to itself in a repeated group matches the previous
2450 8. General tidying up of studying (both automatic and via "study")
2451 consequential on the addition of new assertions.
2453 9. As in 5.005, unlimited repeated groups that could match an empty substring
2454 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2455 runtime if any iteration does actually match an empty substring.
2457 10. Include the RunTest script in the distribution.
2459 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2460 discrepancies, some of which were old and were also with respect to 5.004. They
2461 have now been fixed.
2464 Version 1.09 28-Apr-98
2465 ----------------------
2467 1. A negated single character class followed by a quantifier with a minimum
2468 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2469 program crashes, or just wrong answers. This did not apply to negated classes
2470 containing more than one character, or to minima other than one.
2473 Version 1.08 27-Mar-98
2474 ----------------------
2476 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2478 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2479 latter must appear before anything that relies on it in the pattern.
2482 Version 1.07 16-Feb-98
2483 ----------------------
2485 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2486 repeat of a potentially empty string).
2489 Version 1.06 23-Jan-98
2490 ----------------------
2492 1. Added Markus Oberhumer's little patches for C++.
2494 2. Literal strings longer than 255 characters were broken.
2497 Version 1.05 23-Dec-97
2498 ----------------------
2500 1. Negated character classes containing more than one character were failing if
2501 PCRE_CASELESS was set at run time.
2504 Version 1.04 19-Dec-97
2505 ----------------------
2507 1. Corrected the man page, where some "const" qualifiers had been omitted.
2509 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2512 3. Fixed memory leak which occurred when a regex with back references was
2513 matched with an offsets vector that wasn't big enough. The temporary memory
2514 that is used in this case wasn't being freed if the match failed.
2516 4. Tidied pcretest to ensure it frees memory that it gets.
2518 5. Temporary memory was being obtained in the case where the passed offsets
2519 vector was exactly big enough.
2521 6. Corrected definition of offsetof() from change 5 below.
2523 7. I had screwed up change 6 below and broken the rules for the use of
2524 setjmp(). Now fixed.
2527 Version 1.03 18-Dec-97
2528 ----------------------
2530 1. A erroneous regex with a missing opening parenthesis was correctly
2531 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2534 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2535 it was reported that one broken compiler failed on the former because "code" is
2536 also an independent variable.
2538 3. The erroneous regex a[]b caused an array overrun reference.
2540 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2541 fail on data ending with that character. (It was going on too far, and checking
2542 the next character, typically a binary zero.) This was specific to the
2543 optimized code for single-character negative classes.
2545 5. Added a contributed patch from the TIN world which does the following:
2547 + Add an undef for memmove, in case the the system defines a macro for it.
2549 + Add a definition of offsetof(), in case there isn't one. (I don't know
2550 the reason behind this - offsetof() is part of the ANSI standard - but
2553 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2554 most of the places where whitespace preceded '#'. I have given up and
2555 allowed the remaining 2 cases to be at the margin.
2557 + Rename some variables in pcre to eliminate shadowing. This seems very
2558 pedantic, but does no harm, of course.
2560 6. Moved the call to setjmp() into its own function, to get rid of warnings
2561 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2563 7. Constructs such as \d{8,} were compiling into the equivalent of
2564 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2565 outcome, but in this particular case used more store than had been allocated,
2566 which caused the bug to be discovered because it threw up an internal error.
2568 8. The debugging code in both pcre and pcretest for outputting the compiled
2569 form of a regex was going wrong in the case of back references followed by
2570 curly-bracketed repeats.
2573 Version 1.02 12-Dec-97
2574 ----------------------
2576 1. Typos in pcre.3 and comments in the source fixed.
2578 2. Applied a contributed patch to get rid of places where it used to remove
2579 'const' from variables, and fixed some signed/unsigned and uninitialized
2582 3. Added the "runtest" target to Makefile.
2584 4. Set default compiler flag to -O2 rather than just -O.
2587 Version 1.01 19-Nov-97
2588 ----------------------
2590 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2591 like /([ab]*)*/, that is, for classes with more than one character in them.
2593 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2594 as /((?>a*))*/ (a PCRE_EXTRA facility).
2597 Version 1.00 18-Nov-97
2598 ----------------------
2600 1. Added compile-time macros to support systems such as SunOS4 which don't have
2601 memmove() or strerror() but have other things that can be used instead.
2603 2. Arranged that "make clean" removes the executables.
2606 Version 0.99 27-Oct-97
2607 ----------------------
2609 1. Fixed bug in code for optimizing classes with only one character. It was
2610 initializing a 32-byte map regardless, which could cause it to run off the end
2611 of the memory it had got.
2613 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
2616 Version 0.98 22-Oct-97
2617 ----------------------
2619 1. Fixed bug in code for handling temporary memory usage when there are more
2620 back references than supplied space in the ovector. This could cause segfaults.
2623 Version 0.97 21-Oct-97
2624 ----------------------
2626 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
2628 2. Optimized negated single characters not to use a bit map.
2630 3. Brought error texts together as macro definitions; clarified some of them;
2631 fixed one that was wrong - it said "range out of order" when it meant "invalid
2634 4. Changed some char * arguments to const char *.
2636 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
2638 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
2642 Version 0.96 16-Oct-97
2643 ----------------------
2645 1. Added a simple "pgrep" utility to the distribution.
2647 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
2648 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
2649 where "ddd" means "one or more decimal digits".
2651 3. Fixed serious bug. If a pattern had a back reference, but the call to
2652 pcre_exec() didn't supply a large enough ovector to record the related
2653 identifying subpattern, the match always failed. PCRE now remembers the number
2654 of the largest back reference, and gets some temporary memory in which to save
2655 the offsets during matching if necessary, in order to ensure that
2656 backreferences always work.
2658 4. Increased the compatibility with Perl in a number of ways:
2660 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
2661 to request this handling. The option can be set at compile or exec time.
2663 (b) $ matches before a terminating newline by default; an option
2664 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
2665 mode). The option can be set at compile or exec time.
2667 (c) The handling of \ followed by a digit other than 0 is now supposed to be
2668 the same as Perl's. If the decimal number it represents is less than 10
2669 or there aren't that many previous left capturing parentheses, an octal
2670 escape is read. Inside a character class, it's always an octal escape,
2671 even if it is a single digit.
2673 (d) An escaped but undefined alphabetic character is taken as a literal,
2674 unless PCRE_EXTRA is set. Currently this just reserves the remaining
2677 (e) {0} is now permitted. (The previous item is removed from the compiled
2680 5. Changed all the names of code files so that the basic parts are no longer
2681 than 10 characters, and abolished the teeny "globals.c" file.
2683 6. Changed the handling of character classes; they are now done with a 32-byte
2686 7. Added the -d and /D options to pcretest to make it possible to look at the
2687 internals of compilation without having to recompile pcre.
2690 Version 0.95 23-Sep-97
2691 ----------------------
2693 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
2694 \x20 at the start of a run of normal characters. These were being treated as
2695 real characters, instead of the source characters being re-checked.
2698 Version 0.94 18-Sep-97
2699 ----------------------
2701 1. The functions are now thread-safe, with the caveat that the global variables
2702 containing pointers to malloc() and free() or alternative functions are the
2703 same for all threads.
2705 2. Get pcre_study() to generate a bitmap of initial characters for non-
2706 anchored patterns when this is possible, and use it if passed to pcre_exec().
2709 Version 0.93 15-Sep-97
2710 ----------------------
2712 1. /(b)|(:+)/ was computing an incorrect first character.
2714 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
2715 but not actually doing anything yet.
2717 3. Treat "-" characters in classes that cannot be part of ranges as literals,
2718 as Perl does (e.g. [-az] or [az-]).
2720 4. Set the anchored flag if a branch starts with .* or .*? because that tests
2721 all possible positions.
2723 5. Split up into different modules to avoid including unneeded functions in a
2724 compiled binary. However, compile and exec are still in one module. The "study"
2725 function is split off.
2727 6. The character tables are now in a separate module whose source is generated
2728 by an auxiliary program - but can then be edited by hand if required. There are
2729 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
2730 toupper() in the code.
2732 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
2733 make them global. Abolish the function for setting them, as the caller can now
2737 Version 0.92 11-Sep-97
2738 ----------------------
2740 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
2741 (e.g. /a{1,3}/) was broken (I mis-optimized it).
2743 2. Caseless matching was not working in character classes if the characters in
2744 the pattern were in upper case.
2746 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
2748 4. Make PCRE_ANCHORED public and accept as a compile option.
2750 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
2751 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
2754 6. Give an error if bad option bits passed at compile or run time.
2756 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
2757 pcretest to cause it to pass that flag.
2759 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
2760 options, and the first character, if set.
2762 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
2765 Version 0.91 10-Sep-97
2766 ----------------------
2768 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
2769 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
2771 2. PCRE was looping on encountering an indefinitely repeated back reference to
2772 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
2773 Perl does - treats the match as successful.