ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 122 - (show annotations)
Mon Mar 12 15:10:25 2007 UTC (14 years, 7 months ago) by ph10
File size: 125213 byte(s)
Error occurred while calculating annotation data.
Documentation spring clean, and expansion of maint/README.
1 ChangeLog for PCRE
2 ------------------
4 Version 7.1 12-Mar-07
5 ---------------------
7 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
8 that is more "standard", making use of automake and other Autotools. There
9 is some re-arrangement of the files and adjustment of comments consequent
10 on this.
12 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
13 for recursive directory scanning broke on some systems because the files
14 are not scanned in any specific order and on different systems the order
15 was different. A call to "sort" has been inserted into RunGrepTest for the
16 approprate test as a short-term fix. In the longer term there may be an
17 alternative.
19 3. I had an email from Eric Raymond about problems translating some of PCRE's
20 man pages to HTML (despite the fact that I distribute HTML pages, some
21 people do their own conversions for various reasons). The problems
22 concerned the use of low-level troff macros .br and .in. I have therefore
23 removed all such uses from the man pages (some were redundant, some could
24 be replaced by .nf/.fi pairs). The 132html script that I use to generate
25 HTML has been updated to handle .nf/.fi and to complain if it encounters
26 .br or .in.
28 4. Updated comments in configure.ac that get placed in config.h.in and also
29 arranged for config.h to be included in the distribution, with the name
30 config.h.generic, for the benefit of those who have to compile without
31 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
33 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
34 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
35 makevp.bat for the current PCRE, using the additional files !compile.txt,
36 !linklib.txt, and pcregexp.pas.
38 6. A Windows user reported a minor discrepancy with test 2, which turned out
39 to be caused by a trailing space on an input line that had got lost in his
40 copy. The trailing space was an accident, so I've just removed it.
42 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
43 that is needed.
45 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
46 as "const" (a) because they are and (b) because it helps the PHP
47 maintainers who have recently made a script to detect big data structures
48 in the php code that should be moved to the .rodata section. I remembered
49 to update Builducptable as well, so it won't revert if ucptable.h is ever
50 re-created.
52 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
53 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
54 order to be able to cut out the UTF-8 tables in the latter when UTF-8
55 support is not required. This saves 1.5-2K of code, which is important in
56 some applications.
58 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
59 so as not to refer to the tables, even though these functions will never be
60 called when UTF-8 support is disabled. Otherwise there are problems with a
61 shared library.
63 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
65 (a) It was defining its arguments as char * instead of void *.
67 (b) It was assuming that all moves were upwards in memory; this was true
68 a long time ago when I wrote it, but is no longer the case.
70 The emulated memove() is provided for those environments that have neither
71 memmove() nor bcopy(). I didn't think anyone used it these days, but that
72 is clearly not the case, as these two bugs were recently reported.
74 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
75 and Detrail to create the HTML documentation, the .txt form of the man
76 pages, and it removes trailing spaces from listed files. It also creates
77 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
78 case, it wraps all the #defines with #ifndefs. This script should be run
79 before "make dist".
81 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
82 with Unicode property support.
84 (a) For a maximizing quantifier, if the two different cases of the
85 character were of different lengths in their UTF-8 codings (there are
86 some cases like this - I found 11), and the matching function had to
87 back up over a mixture of the two cases, it incorrectly assumed they
88 were both the same length.
90 (b) When PCRE was configured to use the heap rather than the stack for
91 recursion during matching, it was not correctly preserving the data for
92 the other case of a UTF-8 character when checking ahead for a match
93 while processing a minimizing repeat. If the check also involved
94 matching a wide character, but failed, corruption could cause an
95 erroneous result when trying to check for a repeat of the original
96 character.
98 13. Some tidying changes to the testing mechanism:
100 (a) The RunTest script now detects the internal link size and whether there
101 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
102 values substituted by "configure". (The RunGrepTest script already did
103 this for UTF-8.) The configure.ac script no longer substitutes the
104 relevant variables.
106 (b) The debugging options /B and /D in pcretest show the compiled bytecode
107 with length and offset values. This means that the output is different
108 for different internal link sizes. Test 2 is skipped for link sizes
109 other than 2 because of this, bypassing the problem. Unfortunately,
110 there was also a test in test 3 (the locale tests) that used /B and
111 failed for link sizes other than 2. Rather than cut the whole test out,
112 I have added a new /Z option to pcretest that replaces the length and
113 offset values with spaces. This is now used to make test 3 independent
114 of link size. (Test 2 will be tidied up later.)
116 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
117 segmentation fault instead of returning the appropriate error message.
120 Version 7.0 19-Dec-06
121 ---------------------
123 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
124 moving to gcc 4.1.1.
126 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
127 sys/time.h, which is documented as needed for this function. It doesn't
128 seem to matter on Linux, but it showed up on some releases of OS X.
130 3. It seems that there are systems where bytes whose values are greater than
131 127 match isprint() in the "C" locale. The "C" locale should be the
132 default when a C program starts up. In most systems, only ASCII printing
133 characters match isprint(). This difference caused the output from pcretest
134 to vary, making some of the tests fail. I have changed pcretest so that:
136 (a) When it is outputting text in the compiled version of a pattern, bytes
137 other than 32-126 are always shown as hex escapes.
139 (b) When it is outputting text that is a matched part of a subject string,
140 it does the same, unless a different locale has been set for the match
141 (using the /L modifier). In this case, it uses isprint() to decide.
143 4. Fixed a major bug that caused incorrect computation of the amount of memory
144 required for a compiled pattern when options that changed within the
145 pattern affected the logic of the preliminary scan that determines the
146 length. The relevant options are -x, and -i in UTF-8 mode. The result was
147 that the computed length was too small. The symptoms of this bug were
148 either the PCRE error "internal error: code overflow" from pcre_compile(),
149 or a glibc crash with a message such as "pcretest: free(): invalid next
150 size (fast)". Examples of patterns that provoked this bug (shown in
151 pcretest format) are:
153 /(?-x: )/x
154 /(?x)(?-x: \s*#\s*)/
155 /((?i)[\x{c0}])/8
156 /(?i:[\x{c0}])/8
158 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
159 is now done differently.
161 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
162 wrapper classes; (b) implement a new function in the C++ scanner that is
163 more efficient than the old way of doing things because it avoids levels of
164 recursion in the regex matching; (c) add a paragraph to the documentation
165 for the FullMatch() function.
167 6. The escape sequence \n was being treated as whatever was defined as
168 "newline". Not only was this contrary to the documentation, which states
169 that \n is character 10 (hex 0A), but it also went horribly wrong when
170 "newline" was defined as CRLF. This has been fixed.
172 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
173 was being set to -1 for the "end of line" case (supposedly a value that no
174 character can have). Though this value is never used (the check for end of
175 line is "zero bytes in current character"), it caused compiler complaints.
176 I've changed it to 0xffffffff.
178 8. In pcre_version.c, the version string was being built by a sequence of
179 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
180 string (as it is for production releases) called a macro with an empty
181 argument. The C standard says the result of this is undefined. The gcc
182 compiler treats it as an empty string (which was what was wanted) but it is
183 reported that Visual C gives an error. The source has been hacked around to
184 avoid this problem.
186 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
187 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
188 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
189 of them did).
191 10. Originally, pcretest opened its input and output without "b"; then I was
192 told that "b" was needed in some environments, so it was added for release
193 5.0 to both the input and output. (It makes no difference on Unix-like
194 systems.) Later I was told that it is wrong for the input on Windows. I've
195 now abstracted the modes into two macros, to make it easier to fiddle with
196 them, and removed "b" from the input mode under Windows.
198 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
200 12. Added -help and --help to pcretest as an official way of being reminded
201 of the options.
203 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
204 and pcrecpp.cc because they annoy compilers at high warning levels.
206 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
208 15. Fixed an occurrence of == in configure.ac that should have been = (shell
209 scripts are not C programs :-) and which was not noticed because it works
210 on Linux.
212 16. pcretest is supposed to handle any length of pattern and data line (as one
213 line or as a continued sequence of lines) by extending its input buffer if
214 necessary. This feature was broken for very long pattern lines, leading to
215 a string of junk being passed to pcre_compile() if the pattern was longer
216 than about 50K.
218 17. I have done a major re-factoring of the way pcre_compile() computes the
219 amount of memory needed for a compiled pattern. Previously, there was code
220 that made a preliminary scan of the pattern in order to do this. That was
221 OK when PCRE was new, but as the facilities have expanded, it has become
222 harder and harder to keep it in step with the real compile phase, and there
223 have been a number of bugs (see for example, 4 above). I have now found a
224 cunning way of running the real compile function in a "fake" mode that
225 enables it to compute how much memory it would need, while actually only
226 ever using a few hundred bytes of working memory and without too many
227 tests of the mode. This should make future maintenance and development
228 easier. A side effect of this work is that the limit of 200 on the nesting
229 depth of parentheses has been removed (though this was never a serious
230 limitation, I suspect). However, there is a downside: pcre_compile() now
231 runs more slowly than before (30% or more, depending on the pattern). I
232 hope this isn't a big issue. There is no effect on runtime performance.
234 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
235 newline (only possible for the last line of a file) and it was a
236 pattern that set a locale (followed by /Lsomething), pcretest crashed.
238 19. Added additional timing features to pcretest. (1) The -tm option now times
239 matching only, not compiling. (2) Both -t and -tm can be followed, as a
240 separate command line item, by a number that specifies the number of
241 repeats to use when timing. The default is 50000; this gives better
242 precision, but takes uncomfortably long for very large patterns.
244 20. Extended pcre_study() to be more clever in cases where a branch of a
245 subpattern has no definite first character. For example, (a*|b*)[cd] would
246 previously give no result from pcre_study(). Now it recognizes that the
247 first character must be a, b, c, or d.
249 21. There was an incorrect error "recursive call could loop indefinitely" if
250 a subpattern (or the entire pattern) that was being tested for matching an
251 empty string contained only one non-empty item after a nested subpattern.
252 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
253 incorrectly, because the \d was being skipped in the check.
255 22. The pcretest program now has a new pattern option /B and a command line
256 option -b, which is equivalent to adding /B to every pattern. This causes
257 it to show the compiled bytecode, without the additional information that
258 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
259 is the same as /B/I).
261 23. A new optimization is now able automatically to treat some sequences such
262 as a*b as a*+b. More specifically, if something simple (such as a character
263 or a simple class like \d) has an unlimited quantifier, and is followed by
264 something that cannot possibly match the quantified thing, the quantifier
265 is automatically "possessified".
267 24. A recursive reference to a subpattern whose number was greater than 39
268 went wrong under certain circumstances in UTF-8 mode. This bug could also
269 have affected the operation of pcre_study().
271 25. Realized that a little bit of performance could be had by replacing
272 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
274 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
276 27. Possessive quantifiers such as a++ were previously implemented by turning
277 them into atomic groups such as ($>a+). Now they have their own opcodes,
278 which improves performance. This includes the automatically created ones
279 from 23 above.
281 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
282 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
283 the first matched character to be a colon. This applied both to named and
284 numbered groups.
286 29. The ucpinternal.h header file was missing its idempotency #ifdef.
288 30. I was sent a "project" file called libpcre.a.dev which I understand makes
289 building PCRE on Windows easier, so I have included it in the distribution.
291 31. There is now a check in pcretest against a ridiculously large number being
292 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
293 loop, the loop is abandoned.
295 32. Forward references to subpatterns in conditions such as (?(2)...) where
296 subpattern 2 is defined later cause pcre_compile() to search forwards in
297 the pattern for the relevant set of parentheses. This search went wrong
298 when there were unescaped parentheses in a character class, parentheses
299 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
301 33. "Subroutine" calls and backreferences were previously restricted to
302 referencing subpatterns earlier in the regex. This restriction has now
303 been removed.
305 34. Added a number of extra features that are going to be in Perl 5.10. On the
306 whole, these are just syntactic alternatives for features that PCRE had
307 previously implemented using the Python syntax or my own invention. The
308 other formats are all retained for compatibility.
310 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
311 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
312 also .NET compatible.
314 (b) A recursion or subroutine call to a named group can now be defined as
315 (?&name) as well as (?P>name).
317 (c) A backreference to a named group can now be defined as \k<name> or
318 \k'name' as well as (?P=name). The new forms, as well as being in Perl
319 5.10, are also .NET compatible.
321 (d) A conditional reference to a named group can now use the syntax
322 (?(<name>) or (?('name') as well as (?(name).
324 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
325 groups (named and numbered) that are never evaluated inline, but can be
326 called as "subroutines" from elsewhere. In effect, the DEFINE condition
327 is always false. There may be only one alternative in such a group.
329 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
330 as the simple (?(R). The condition is true only if the most recent
331 recursion is that of the given number or name. It does not search out
332 through the entire recursion stack.
334 (g) The escape \gN or \g{N} has been added, where N is a positive or
335 negative number, specifying an absolute or relative reference.
337 35. Tidied to get rid of some further signed/unsigned compiler warnings and
338 some "unreachable code" warnings.
340 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
341 things, this adds five new scripts.
343 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
344 There were also incompatibilities regarding the handling of \Q..\E inside
345 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
346 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
348 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
349 matches an empty string, and forcibly breaks the loop. There were bugs in
350 this code in non-simple cases. For a pattern such as ^(a()*)* matched
351 against aaaa the result was just "a" rather than "aaaa", for example. Two
352 separate and independent bugs (that affected different cases) have been
353 fixed.
355 39. Refactored the code to abolish the use of different opcodes for small
356 capturing bracket numbers. This is a tidy that I avoided doing when I
357 removed the limit on the number of capturing brackets for 3.5 back in 2001.
358 The new approach is not only tidier, it makes it possible to reduce the
359 memory needed to fix the previous bug (38).
361 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
362 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
363 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
364 mode.
366 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
367 report.
369 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
370 copy construction and assignment in the C++ wrapper.
372 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
373 couple of bugs that could have given wrong results in the "--newline=crlf"
374 case.
376 44. Added a number of casts and did some reorganization of signed/unsigned int
377 variables following suggestions from Dair Grant. Also renamed the variable
378 "this" as "item" because it is a C++ keyword.
380 45. Arranged for dftables to add
382 #include "pcre_internal.h"
384 to pcre_chartables.c because without it, gcc 4.x may remove the array
385 definition from the final binary if PCRE is built into a static library and
386 dead code stripping is activated.
388 46. For an unanchored pattern, if a match attempt fails at the start of a
389 newline sequence, and the newline setting is CRLF or ANY, and the next two
390 characters are CRLF, advance by two characters instead of one.
393 Version 6.7 04-Jul-06
394 ---------------------
396 1. In order to handle tests when input lines are enormously long, pcretest has
397 been re-factored so that it automatically extends its buffers when
398 necessary. The code is crude, but this _is_ just a test program. The
399 default size has been increased from 32K to 50K.
401 2. The code in pcre_study() was using the value of the re argument before
402 testing it for NULL. (Of course, in any sensible call of the function, it
403 won't be NULL.)
405 3. The memmove() emulation function in pcre_internal.h, which is used on
406 systems that lack both memmove() and bcopy() - that is, hardly ever -
407 was missing a "static" storage class specifier.
409 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
410 containing an extended class (one that cannot be represented by a bitmap
411 because it contains high-valued characters or Unicode property items, e.g.
412 [\pZ]). Almost always one would set UTF-8 mode when processing such a
413 pattern, but PCRE should not loop if you do not (it no longer does).
414 [Detail: two cases were found: (a) a repeated subpattern containing an
415 extended class; (b) a recursive reference to a subpattern that followed a
416 previous extended class. It wasn't skipping over the extended class
417 correctly when UTF-8 mode was not set.]
419 5. A negated single-character class was not being recognized as fixed-length
420 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
421 compile error "lookbehind assertion is not fixed length".
423 6. The RunPerlTest auxiliary script was showing an unexpected difference
424 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
425 write a Perl script that can interpret lines of an input file either as
426 byte characters or as UTF-8, which is what "perltest" was being required to
427 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
428 can't do is switch easily at run time between having the "use utf8;" pragma
429 or not. In the end, I fudged it by using the RunPerlTest script to insert
430 "use utf8;" explicitly for the UTF-8 tests.
432 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
433 the end of the subject string, contrary to the documentation and to what
434 Perl does. This was true of both matching functions. Now it matches only at
435 the start of the subject and immediately after *internal* newlines.
437 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
438 a pointer to an int instead of a pointer to an unsigned long int. This
439 caused problems on 64-bit systems.
441 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
442 instance of the 'standard' template library not being so standard".
444 10. There was no check on the number of named subpatterns nor the maximum
445 length of a subpattern name. The product of these values is used to compute
446 the size of the memory block for a compiled pattern. By supplying a very
447 long subpattern name and a large number of named subpatterns, the size
448 computation could be caused to overflow. This is now prevented by limiting
449 the length of names to 32 characters, and the number of named subpatterns
450 to 10,000.
452 11. Subpatterns that are repeated with specific counts have to be replicated in
453 the compiled pattern. The size of memory for this was computed from the
454 length of the subpattern and the repeat count. The latter is limited to
455 65535, but there was no limit on the former, meaning that integer overflow
456 could in principle occur. The compiled length of a repeated subpattern is
457 now limited to 30,000 bytes in order to prevent this.
459 12. Added the optional facility to have named substrings with the same name.
461 13. Added the ability to use a named substring as a condition, using the
462 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
463 are numbers (not recommended). Forward references are permitted.
465 14. Added forward references in named backreferences (if you see what I mean).
467 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
468 pattern could run off the end of the subject. For example, the pattern
469 "(?s)(.{1,5})"8 did this with the subject "ab".
471 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
472 PCRE_CASELESS was set when matching characters that were quantified with ?
473 or *.
475 17. A character class other than a single negated character that had a minimum
476 but no maximum quantifier - for example [ab]{6,} - was not handled
477 correctly by pce_dfa_exec(). It would match only one character.
479 18. A valid (though odd) pattern that looked like a POSIX character
480 class but used an invalid character after [ (for example [[,abc,]]) caused
481 pcre_compile() to give the error "Failed: internal error: code overflow" or
482 in some cases to crash with a glibc free() error. This could even happen if
483 the pattern terminated after [[ but there just happened to be a sequence of
484 letters, a binary zero, and a closing ] in the memory that followed.
486 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
487 over the years. Originally (before any Unicode support), just the bottom 8
488 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
489 output from "man perlunicode" includes this:
491 The regular expression compiler produces polymorphic opcodes. That
492 is, the pattern adapts to the data and automatically switches to
493 the Unicode character scheme when presented with Unicode data--or
494 instead uses a traditional byte scheme when presented with byte
495 data.
497 Sadly, a wide octal escape does not cause a switch, and in a string with
498 no other multibyte characters, these octal escapes are treated as before.
499 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
500 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
501 Unicode string.
503 I have not perpetrated such confusion in PCRE. Up till now, it took just
504 the bottom 8 bits, as in old Perl. I have now made octal escapes with
505 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
506 translate to the appropriate multibyte character.
508 29. Applied some refactoring to reduce the number of warnings from Microsoft
509 and Borland compilers. This has included removing the fudge introduced
510 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
511 a warning about an unused variable.
513 21. PCRE has not included VT (character 0x0b) in the set of whitespace
514 characters since release 4.0, because Perl (from release 5.004) does not.
515 [Or at least, is documented not to: some releases seem to be in conflict
516 with the documentation.] However, when a pattern was studied with
517 pcre_study() and all its branches started with \s, PCRE still included VT
518 as a possible starting character. Of course, this did no harm; it just
519 caused an unnecessary match attempt.
521 22. Removed a now-redundant internal flag bit that recorded the fact that case
522 dependency changed within the pattern. This was once needed for "required
523 byte" processing, but is no longer used. This recovers a now-scarce options
524 bit. Also moved the least significant internal flag bit to the most-
525 significant bit of the word, which was not previously used (hangover from
526 the days when it was an int rather than a uint) to free up another bit for
527 the future.
529 23. Added support for CRLF line endings as well as CR and LF. As well as the
530 default being selectable at build time, it can now be changed at runtime
531 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
532 specify that it is scanning data with non-default line endings.
534 24. Changed the definition of CXXLINK to make it agree with the definition of
535 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
537 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
538 recursions. This makes a big different to stack usage for some patterns.
540 26. If a subpattern containing a named recursion or subroutine reference such
541 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
542 the space required for the compiled pattern went wrong and gave too small a
543 value. Depending on the environment, this could lead to "Failed: internal
544 error: code overflow at offset 49" or "glibc detected double free or
545 corruption" errors.
547 27. Applied patches from Google (a) to support the new newline modes and (b) to
548 advance over multibyte UTF-8 characters in GlobalReplace.
550 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
551 difference for some implementation of PCRE in some Windows version.
553 29. Added some extra testing facilities to pcretest:
555 \q<number> in a data line sets the "match limit" value
556 \Q<number> in a data line sets the "match recursion limt" value
557 -S <number> sets the stack size, where <number> is in megabytes
559 The -S option isn't available for Windows.
562 Version 6.6 06-Feb-06
563 ---------------------
565 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
566 in pcreposix.h. I have copied the definition from pcre.h.
568 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
569 because pcre.h is no longer a built file.
571 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
572 not normally included in the compiled code.
575 Version 6.5 01-Feb-06
576 ---------------------
578 1. When using the partial match feature with pcre_dfa_exec(), it was not
579 anchoring the second and subsequent partial matches at the new starting
580 point. This could lead to incorrect results. For example, with the pattern
581 /1234/, partially matching against "123" and then "a4" gave a match.
583 2. Changes to pcregrep:
585 (a) All non-match returns from pcre_exec() were being treated as failures
586 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
587 error message is output. Some extra information is given for the
589 probably the only errors that are likely to be caused by users (by
590 specifying a regex that has nested indefinite repeats, for instance).
591 If there are more than 20 of these errors, pcregrep is abandoned.
593 (b) A binary zero was treated as data while matching, but terminated the
594 output line if it was written out. This has been fixed: binary zeroes
595 are now no different to any other data bytes.
597 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
598 used to set a locale for matching. The --locale=xxxx long option has
599 been added (no short equivalent) to specify a locale explicitly on the
600 pcregrep command, overriding the environment variables.
602 (d) When -B was used with -n, some line numbers in the output were one less
603 than they should have been.
605 (e) Added the -o (--only-matching) option.
607 (f) If -A or -C was used with -c (count only), some lines of context were
608 accidentally printed for the final match.
610 (g) Added the -H (--with-filename) option.
612 (h) The combination of options -rh failed to suppress file names for files
613 that were found from directory arguments.
615 (i) Added the -D (--devices) and -d (--directories) options.
617 (j) Added the -F (--fixed-strings) option.
619 (k) Allow "-" to be used as a file name for -f as well as for a data file.
621 (l) Added the --colo(u)r option.
623 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
624 is not present by default.
626 3. A nasty bug was discovered in the handling of recursive patterns, that is,
627 items such as (?R) or (?1), when the recursion could match a number of
628 alternatives. If it matched one of the alternatives, but subsequently,
629 outside the recursion, there was a failure, the code tried to back up into
630 the recursion. However, because of the way PCRE is implemented, this is not
631 possible, and the result was an incorrect result from the match.
633 In order to prevent this happening, the specification of recursion has
634 been changed so that all such subpatterns are automatically treated as
635 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
637 4. I had overlooked the fact that, in some locales, there are characters for
638 which isalpha() is true but neither isupper() nor islower() are true. In
639 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
640 and ordfeminine) are like this. This affected the treatment of \w and \W
641 when they appeared in character classes, but not when they appeared outside
642 a character class. The bit map for "word" characters is now created
643 separately from the results of isalnum() instead of just taking it from the
644 upper, lower, and digit maps. (Plus the underscore character, of course.)
646 5. The above bug also affected the handling of POSIX character classes such as
647 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
648 permanent tables. Instead, the bit maps for such a class were previously
649 created as the appropriate unions of the upper, lower, and digit bitmaps.
650 Now they are created by subtraction from the [[:word:]] class, which has
651 its own bitmap.
653 6. The [[:blank:]] character class matches horizontal, but not vertical space.
654 It is created by subtracting the vertical space characters (\x09, \x0a,
655 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
656 subtraction was done in the overall bitmap for a character class, meaning
657 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
658 be recognized. This bug has been fixed.
660 7. Patches from the folks at Google:
662 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
663 real life, but is still worth protecting against".
665 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
666 regular expressions".
668 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
669 have it.
671 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
672 "configure" and the latter not, in order to fix a problem somebody had
673 with compiling the Arg class on HP-UX.
675 (e) Improve the error-handling of the C++ wrapper a little bit.
677 (f) New tests for checking recursion limiting.
679 8. The pcre_memmove() function, which is used only if the environment does not
680 have a standard memmove() function (and is therefore rarely compiled),
681 contained two bugs: (a) use of int instead of size_t, and (b) it was not
682 returning a result (though PCRE never actually uses the result).
684 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
685 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
686 returned instead of calling malloc() with an overflowing number that would
687 most likely cause subsequent chaos.
689 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
691 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
692 with this option is matched, the nmatch and pmatch options of regexec() are
693 ignored.
695 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
696 provided in case anyone wants to the the POSIX interface with UTF-8
697 strings.
699 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
700 C++ linking (needed for some HP-UX environments).
702 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
703 (unused parameter) and in the pcre_printint() function (omitted "default"
704 switch label when the default is to do nothing).
706 15. Added some code to make it possible, when PCRE is compiled as a C++
707 library, to replace subject pointers for pcre_exec() with a smart pointer
708 class, thus making it possible to process discontinuous strings.
710 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
711 much the same function. They were added by different people who were trying
712 to make PCRE easy to compile on non-Unix systems. It has been suggested
713 that PCRE_EXPORT be abolished now that there is more automatic apparatus
714 for compiling on Windows systems. I have therefore replaced it with
715 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
716 defaults to "extern" for C or "extern C" for C++, which works fine on
717 Unix-like systems. It is now possible to override the value of PCRE_DATA_
718 SCOPE with something explicit in config.h. In addition:
720 (a) pcreposix.h still had just "extern" instead of either of these macros;
721 I have replaced it with PCRE_DATA_SCOPE.
723 (b) Functions such as _pcre_xclass(), which are internal to the library,
724 but external in the C sense, all had PCRE_EXPORT in their definitions.
725 This is apparently wrong for the Windows case, so I have removed it.
726 (It makes no difference on Unix-like systems.)
728 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
729 of recursive calls to match(). This is different to MATCH_LIMIT because
730 that limits the total number of calls to match(), not all of which increase
731 the depth of recursion. Limiting the recursion depth limits the amount of
732 stack (or heap if NO_RECURSE is set) that is used. The default can be set
733 when PCRE is compiled, and changed at run time. A patch from Google adds
734 this functionality to the C++ interface.
736 18. Changes to the handling of Unicode character properties:
738 (a) Updated the table to Unicode 4.1.0.
740 (b) Recognize characters that are not in the table as "Cn" (undefined).
742 (c) I revised the way the table is implemented to a much improved format
743 which includes recognition of ranges. It now supports the ranges that
744 are defined in UnicodeData.txt, and it also amalgamates other
745 characters into ranges. This has reduced the number of entries in the
746 table from around 16,000 to around 3,000, thus reducing its size
747 considerably. I realized I did not need to use a tree structure after
748 all - a binary chop search is just as efficient. Having reduced the
749 number of entries, I extended their size from 6 bytes to 8 bytes to
750 allow for more data.
752 (d) Added support for Unicode script names via properties such as \p{Han}.
754 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
755 matching that character.
757 20. When matching a repeated Unicode property with a minimum greater than zero,
758 (for example \pL{2,}), PCRE could look past the end of the subject if it
759 reached it while seeking the minimum number of characters. This could
760 happen only if some of the characters were more than one byte long, because
761 there is a check for at least the minimum number of bytes.
763 21. Refactored the implementation of \p and \P so as to be more general, to
764 allow for more different types of property in future. This has changed the
765 compiled form incompatibly. Anybody with saved compiled patterns that use
766 \p or \P will have to recompile them.
768 22. Added "Any" and "L&" to the supported property types.
770 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
771 but give a compile time error if the value is greater than 0xff.
773 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
774 accidentally not being installed or uninstalled.
776 25. The pcre.h file was built from pcre.h.in, but the only changes that were
777 made were to insert the current release number. This seemed silly, because
778 it made things harder for people building PCRE on systems that don't run
779 "configure". I have turned pcre.h into a distributed file, no longer built
780 by "configure", with the version identification directly included. There is
781 no longer a pcre.h.in file.
783 However, this change necessitated a change to the pcre-config script as
784 well. It is built from pcre-config.in, and one of the substitutions was the
785 release number. I have updated configure.ac so that ./configure now finds
786 the release number by grepping pcre.h.
788 26. Added the ability to run the tests under valgrind.
791 Version 6.4 05-Sep-05
792 ---------------------
794 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
795 "--" to be printed when multiple files were scanned, even when none of the
796 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
797 consider it to be a bug, and have restored the previous behaviour.
799 2. A couple of code tidies to get rid of compiler warnings.
801 3. The pcretest program used to cheat by referring to symbols in the library
802 whose names begin with _pcre_. These are internal symbols that are not
803 really supposed to be visible externally, and in some environments it is
804 possible to suppress them. The cheating is now confined to including
805 certain files from the library's source, which is a bit cleaner.
807 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
808 file's purpose clearer.
810 5. Reorganized pcre_ucp_findchar().
813 Version 6.3 15-Aug-05
814 ---------------------
816 1. The file libpcre.pc.in did not have general read permission in the tarball.
818 2. There were some problems when building without C++ support:
820 (a) If C++ support was not built, "make install" and "make test" still
821 tried to test it.
823 (b) There were problems when the value of CXX was explicitly set. Some
824 changes have been made to try to fix these, and ...
826 (c) --disable-cpp can now be used to explicitly disable C++ support.
828 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
829 backslash in a target when C++ was disabled. This confuses some
830 versions of "make", apparently. Using an intermediate variable solves
831 this. (Same for CPP_LOBJ.)
833 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
834 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
835 necessary on certain architectures.
837 4. Added a setting of -export-symbols-regex to the link command to remove
838 those symbols that are exported in the C sense, but actually are local
839 within the library, and not documented. Their names all begin with
840 "_pcre_". This is not a perfect job, because (a) we have to except some
841 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
842 available (and never for static libraries). I have made a note to try to
843 find a way round (a) in the future.
846 Version 6.2 01-Aug-05
847 ---------------------
849 1. There was no test for integer overflow of quantifier values. A construction
850 such as {1111111111111111} would give undefined results. What is worse, if
851 a minimum quantifier for a parenthesized subpattern overflowed and became
852 negative, the calculation of the memory size went wrong. This could have
853 led to memory overwriting.
855 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
857 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
858 operating environments where this matters.
860 4. Applied Giuseppe Maxia's patch to add additional features for controlling
861 PCRE options from within the C++ wrapper.
863 5. Named capturing subpatterns were not being correctly counted when a pattern
864 was compiled. This caused two problems: (a) If there were more than 100
865 such subpatterns, the calculation of the memory needed for the whole
866 compiled pattern went wrong, leading to an overflow error. (b) Numerical
867 back references of the form \12, where the number was greater than 9, were
868 not recognized as back references, even though there were sufficient
869 previous subpatterns.
871 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
872 versions of gcc, e.g. 2.95.4.
875 Version 6.1 21-Jun-05
876 ---------------------
878 1. There was one reference to the variable "posix" in pcretest.c that was not
879 surrounded by "#if !defined NOPOSIX".
881 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
882 the cross-check on the old pcre_info() function, for the benefit of the
883 cut-down version of PCRE that is currently imported into Exim.
885 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
886 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
887 patterns that start (?i) but I don't think that matters. The use of (?i) is
888 just an example; this all applies to the other options as well.
890 4. Since libtool seems to echo the compile commands it is issuing, the output
891 from "make" can be reduced a bit by putting "@" in front of each libtool
892 compile command.
894 5. Patch from the folks at Google for configure.in to be a bit more thorough
895 in checking for a suitable C++ installation before trying to compile the
896 C++ stuff. This should fix a reported problem when a compiler was present,
897 but no suitable headers.
899 6. The man pages all had just "PCRE" as their title. I have changed them to
900 be the relevant file name. I have also arranged that these names are
901 retained in the file doc/pcre.txt, which is a concatenation in text format
902 of all the man pages except the little individual ones for each function.
904 7. The NON-UNIX-USE file had not been updated for the different set of source
905 files that come with release 6. I also added a few comments about the C++
906 wrapper.
909 Version 6.0 07-Jun-05
910 ---------------------
912 1. Some minor internal re-organization to help with my DFA experiments.
914 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
915 didn't matter for the library itself when fully configured, but did matter
916 when compiling without UCP support, or within Exim, where the ucp files are
917 not imported.
919 3. Refactoring of the library code to split up the various functions into
920 different source modules. The addition of the new DFA matching code (see
921 below) to a single monolithic source would have made it really too
922 unwieldy, quite apart from causing all the code to be include in a
923 statically linked application, when only some functions are used. This is
924 relevant even without the DFA addition now that patterns can be compiled in
925 one application and matched in another.
927 The downside of splitting up is that there have to be some external
928 functions and data tables that are used internally in different modules of
929 the library but which are not part of the API. These have all had their
930 names changed to start with "_pcre_" so that they are unlikely to clash
931 with other external names.
933 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
934 a different (DFA) algorithm. Although it is slower than the original
935 function, it does have some advantages for certain types of matching
936 problem.
938 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
939 including restarting after a partial match.
941 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
942 defined when compiling for Windows was sent to me. I have put it into the
943 code, though I have no means of testing or verifying it.
945 7. Added the pcre_refcount() auxiliary function.
947 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
948 match before or at the first newline in the subject string. In pcretest,
949 the /f option on a pattern can be used to set this.
951 9. A repeated \w when used in UTF-8 mode with characters greater than 256
952 would behave wrongly. This has been present in PCRE since release 4.0.
954 10. A number of changes to the pcregrep command:
956 (a) Refactored how -x works; insert ^(...)$ instead of setting
957 PCRE_ANCHORED and checking the length, in preparation for adding
958 something similar for -w.
960 (b) Added the -w (match as a word) option.
962 (c) Refactored the way lines are read and buffered so as to have more
963 than one at a time available.
965 (d) Implemented a pcregrep test script.
967 (e) Added the -M (multiline match) option. This allows patterns to match
968 over several lines of the subject. The buffering ensures that at least
969 8K, or the rest of the document (whichever is the shorter) is available
970 for matching (and similarly the previous 8K for lookbehind assertions).
972 (f) Changed the --help output so that it now says
974 -w, --word-regex(p)
976 instead of two lines, one with "regex" and the other with "regexp"
977 because that confused at least one person since the short forms are the
978 same. (This required a bit of code, as the output is generated
979 automatically from a table. It wasn't just a text change.)
981 (g) -- can be used to terminate pcregrep options if the next thing isn't an
982 option but starts with a hyphen. Could be a pattern or a path name
983 starting with a hyphen, for instance.
985 (h) "-" can be given as a file name to represent stdin.
987 (i) When file names are being printed, "(standard input)" is used for
988 the standard input, for compatibility with GNU grep. Previously
989 "<stdin>" was used.
991 (j) The option --label=xxx can be used to supply a name to be used for
992 stdin when file names are being printed. There is no short form.
994 (k) Re-factored the options decoding logic because we are going to add
995 two more options that take data. Such options can now be given in four
996 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
998 (l) Added the -A, -B, and -C options for requesting that lines of context
999 around matches be printed.
1001 (m) Added the -L option to print the names of files that do not contain
1002 any matching lines, that is, the complement of -l.
1004 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1005 continue to scan other files.
1007 (o) The -s option was incorrectly implemented. For compatibility with other
1008 greps, it now suppresses the error message for a non-existent or non-
1009 accessible file (but not the return code). There is a new option called
1010 -q that suppresses the output of matching lines, which was what -s was
1011 previously doing.
1013 (p) Added --include and --exclude options to specify files for inclusion
1014 and exclusion when recursing.
1016 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1017 Hopefully, it now does.
1019 12. Missing cast in pcre_study().
1021 13. Added an "uninstall" target to the makefile.
1023 14. Replaced "extern" in the function prototypes in Makefile.in with
1024 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1025 world, but is set differently for Windows.
1027 15. Added a second compiling function called pcre_compile2(). The only
1028 difference is that it has an extra argument, which is a pointer to an
1029 integer error code. When there is a compile-time failure, this is set
1030 non-zero, in addition to the error test pointer being set to point to an
1031 error message. The new argument may be NULL if no error number is required
1032 (but then you may as well call pcre_compile(), which is now just a
1033 wrapper). This facility is provided because some applications need a
1034 numeric error indication, but it has also enabled me to tidy up the way
1035 compile-time errors are handled in the POSIX wrapper.
1037 16. Added VPATH=.libs to the makefile; this should help when building with one
1038 prefix path and installing with another. (Or so I'm told by someone who
1039 knows more about this stuff than I do.)
1041 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1042 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1043 match everything, including newlines. This is not POSIX-compatible, but
1044 somebody wanted the feature. From pcretest it can be activated by using
1045 both the P and the s flags.
1047 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1049 19. libpcre.pc was being incorrectly installed as executable.
1051 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1052 it now also looks for '\r' so that it will work unmodified on Windows.
1054 21. Added Google's contributed C++ wrapper to the distribution.
1056 22. Added some untidy missing memory free() calls in pcretest, to keep
1057 Electric Fence happy when testing.
1061 Version 5.0 13-Sep-04
1062 ---------------------
1064 1. Internal change: literal characters are no longer packed up into items
1065 containing multiple characters in a single byte-string. Each character
1066 is now matched using a separate opcode. However, there may be more than one
1067 byte in the character in UTF-8 mode.
1069 2. The pcre_callout_block structure has two new fields: pattern_position and
1070 next_item_length. These contain the offset in the pattern to the next match
1071 item, and its length, respectively.
1073 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1074 insertion of callouts before each pattern item. Added the /C option to
1075 pcretest to make use of this.
1077 4. On the advice of a Windows user, the lines
1079 #if defined(_WIN32) || defined(WIN32)
1080 _setmode( _fileno( stdout ), 0x8000 );
1081 #endif /* defined(_WIN32) || defined(WIN32) */
1083 have been added to the source of pcretest. This apparently does useful
1084 magic in relation to line terminators.
1086 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1087 for the benefit of those environments where the "b" makes a difference.
1089 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1090 to know about it. I have put a hack into configure.in that adds in code
1091 to set GCC=yes if CC=icc. This seems to end up at a point in the
1092 generated configure script that is early enough to affect the setting of
1093 compiler options, which is what is needed, but I have no means of testing
1094 whether it really works. (The user who reported this had patched the
1095 generated configure script, which of course I cannot do.)
1097 LATER: After change 22 below (new libtool files), the configure script
1098 seems to know about icc (and also ecc). Therefore, I have commented out
1099 this hack in configure.in.
1101 7. Added support for pkg-config (2 patches were sent in).
1103 8. Negated POSIX character classes that used a combination of internal tables
1104 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1105 [[:^ascii]]. Typically, they would match almost any characters. The other
1106 POSIX classes were not broken in this way.
1108 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1109 to find the match, as PCRE was deluded into thinking that the match had to
1110 start at the start point or following a newline. The same bug applied to
1111 patterns with negative forward assertions or any backward assertions
1112 preceding ".*" at the start, unless the pattern required a fixed first
1113 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1115 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1116 starting at the last subject character, bytes beyond the end of the subject
1117 string were read.
1119 11. Renamed the variable "class" as "classbits" to make life easier for C++
1120 users. (Previously there was a macro definition, but it apparently wasn't
1121 enough.)
1123 12. Added the new field "tables" to the extra data so that tables can be passed
1124 in at exec time, or the internal tables can be re-selected. This allows
1125 a compiled regex to be saved and re-used at a later time by a different
1126 program that might have everything at different addresses.
1128 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1129 -R library as well as a -L library.
1131 14. The debugging options of pcretest (-d on the command line or D on a
1132 pattern) showed incorrect output for anything following an extended class
1133 that contained multibyte characters and which was followed by a quantifier.
1135 15. Added optional support for general category Unicode character properties
1136 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1137 support. It adds about 90K to the size of the library. The meanings of the
1138 inbuilt class escapes such as \d and \s have NOT been changed.
1140 16. Updated pcredemo.c to include calls to free() to release the memory for the
1141 compiled pattern.
1143 17. The generated file chartables.c was being created in the source directory
1144 instead of in the building directory. This caused the build to fail if the
1145 source directory was different from the building directory, and was
1146 read-only.
1148 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1149 file. No doubt somebody will tell me if they don't make sense... Also added
1150 Dan Mooney's comments about building on OpenVMS.
1152 19. Added support for partial matching via the PCRE_PARTIAL option for
1153 pcre_exec() and the \P data escape in pcretest.
1155 20. Extended pcretest with 3 new pattern features:
1157 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1158 write the compiled pattern to the file whose name is "rest-of-line".
1159 This is a straight binary dump of the data, with the saved pointer to
1160 the character tables forced to be NULL. The study data, if any, is
1161 written too. After writing, pcretest reads a new pattern.
1163 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1164 compiled pattern from the given file. There must not be any
1165 occurrences of "<" in the file name (pretty unlikely); if there are,
1166 pcretest will instead treat the initial "<" as a pattern delimiter.
1167 After reading in the pattern, pcretest goes on to read data lines as
1168 usual.
1170 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1171 and 16-bit fields in a compiled pattern, to simulate a pattern that
1172 was compiled on a host of opposite endianness.
1174 21. The pcre-exec() function can now cope with patterns that were compiled on
1175 hosts of opposite endianness, with this restriction:
1177 As for any compiled expression that is saved and used later, the tables
1178 pointer field cannot be preserved; the extra_data field in the arguments
1179 to pcre_exec() should be used to pass in a tables address if a value
1180 other than the default internal tables were used at compile time.
1182 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1183 now diagnosed as an error. Previously, most of the time, a negative number
1184 would have been treated as zero, but if in addition "ovector" was passed as
1185 NULL, a crash could occur.
1187 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1188 new versions from the libtool 1.5 distribution (the last one is a copy of
1189 a file called libtool.m4). This seems to have fixed the need to patch
1190 "configure" to support Darwin 1.3 (which I used to do). However, I still
1191 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1192 workstation).
1194 24. Changed the PCRE licence to be the more standard "BSD" licence.
1197 Version 4.5 01-Dec-03
1198 ---------------------
1200 1. There has been some re-arrangement of the code for the match() function so
1201 that it can be compiled in a version that does not call itself recursively.
1202 Instead, it keeps those local variables that need separate instances for
1203 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1204 needs to "recurse". Keeping track of where control must go is done by means
1205 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1206 hide most of the details from the main code, and operates only if
1207 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1208 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1209 operating.
1211 To make it easier for callers to provide specially tailored get/free
1212 functions for this usage, two new functions, pcre_stack_malloc, and
1213 pcre_stack_free, are used. They are always called in strict stacking order,
1214 and the size of block requested is always the same.
1216 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1217 PCRE has been compiled to use the stack or the heap for recursion. The
1218 -C option of pcretest uses this to show which version is compiled.
1220 A new data escape \S, is added to pcretest; it causes the amounts of store
1221 obtained and freed by both kinds of malloc/free at match time to be added
1222 to the output.
1224 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1225 what's available on my current Linux desktop machine.
1227 3. When matching a UTF-8 string, the test for a valid string at the start has
1228 been extended. If start_offset is not zero, PCRE now checks that it points
1229 to a byte that is the start of a UTF-8 character. If not, it returns
1230 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1231 this is necessary because there may be backward assertions in the pattern.
1232 When matching the same subject several times, it may save resources to use
1233 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1235 4. The code for checking the validity of UTF-8 strings has been tightened so
1236 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1237 containing "overlong sequences".
1239 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1240 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1241 should have been "&", but it just so happened that all the cases this let
1242 through by mistake were picked up later in the function.
1244 6. I had used a variable called "isblank" - this is a C99 function, causing
1245 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1247 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1248 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1249 so the version is shown; (c) stop "make test" echoing "./RunTest".
1251 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1253 9. The prototype for memmove() for systems that don't have it was using
1254 size_t, but the inclusion of the header that defines size_t was later. I've
1255 moved the #includes for the C headers earlier to avoid this.
1257 10. Added some adjustments to the code to make it easier to compiler on certain
1258 special systems:
1260 (a) Some "const" qualifiers were missing.
1261 (b) Added the macro EXPORT before all exported functions; by default this
1262 is defined to be empty.
1263 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1264 that it reads its output file name as an argument instead of writing
1265 to the standard output and assuming this can be redirected.
1267 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1268 class containing characters with values greater than 255, PCRE compilation
1269 went into a loop.
1271 12. A recursive reference to a subpattern that was within another subpattern
1272 that had a minimum quantifier of zero caused PCRE to crash. For example,
1273 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1274 recursion. If the recursively-called subpattern itself had a zero repeat,
1275 that was OK.
1277 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1278 buffer into which it was copied (for escape processing) was still set at
1279 1024, so long lines caused crashes.
1281 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1282 "internal error: code overflow...". This applied to any character class
1283 that was followed by a possessive quantifier.
1285 15. Modified the Makefile to add libpcre.la as a prerequisite for
1286 libpcreposix.la because I was told this is needed for a parallel build to
1287 work.
1289 16. If a pattern that contained .* following optional items at the start was
1290 studied, the wrong optimizing data was generated, leading to matching
1291 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1292 matching string must start with a or b or c. The correct conclusion for
1293 this pattern is that a match can start with any character.
1296 Version 4.4 13-Aug-03
1297 ---------------------
1299 1. In UTF-8 mode, a character class containing characters with values between
1300 127 and 255 was not handled correctly if the compiled pattern was studied.
1301 In fixing this, I have also improved the studying algorithm for such
1302 classes (slightly).
1304 2. Three internal functions had redundant arguments passed to them. Removal
1305 might give a very teeny performance improvement.
1307 3. Documentation bug: the value of the capture_top field in a callout is *one
1308 more than* the number of the hightest numbered captured substring.
1310 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1311 in incorrectly linking with a previously installed version. They now link
1312 explicitly with libpcre.la.
1314 5. configure.in no longer needs to recognize Cygwin specially.
1316 6. A problem in pcre.in for Windows platforms is fixed.
1318 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1319 pcretest, it used to include the size of the study block as part of its
1320 output. Unfortunately, the structure contains a field that has a different
1321 size on different hardware architectures. This meant that the tests that
1322 showed this size failed. As the block is currently always of a fixed size,
1323 this information isn't actually particularly useful in pcretest output, so
1324 I have just removed it.
1326 8. Three pre-processor statements accidentally did not start in column 1.
1327 Sadly, there are *still* compilers around that complain, even though
1328 standard C has not required this for well over a decade. Sigh.
1330 9. In pcretest, the code for checking callouts passed small integers in the
1331 callout_data field, which is a void * field. However, some picky compilers
1332 complained about the casts involved for this on 64-bit systems. Now
1333 pcretest passes the address of the small integer instead, which should get
1334 rid of the warnings.
1336 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1337 both compile and run time, and gives an error if an invalid UTF-8 sequence
1338 is found. There is a option for disabling this check in cases where the
1339 string is known to be correct and/or the maximum performance is wanted.
1341 11. In response to a bug report, I changed one line in Makefile.in from
1343 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1344 to
1345 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1347 to look similar to other lines, but I have no way of telling whether this
1348 is the right thing to do, as I do not use Windows. No doubt I'll get told
1349 if it's wrong...
1352 Version 4.3 21-May-03
1353 ---------------------
1355 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1356 Makefile.
1358 2. Some refactoring to improve the quality of the code:
1360 (i) The utf8_table... variables are now declared "const".
1362 (ii) The code for \cx, which used the "case flipping" table to upper case
1363 lower case letters, now just substracts 32. This is ASCII-specific,
1364 but the whole concept of \cx is ASCII-specific, so it seems
1365 reasonable.
1367 (iii) PCRE was using its character types table to recognize decimal and
1368 hexadecimal digits in the pattern. This is silly, because it handles
1369 only 0-9, a-f, and A-F, but the character types table is locale-
1370 specific, which means strange things might happen. A private
1371 table is now used for this - though it costs 256 bytes, a table is
1372 much faster than multiple explicit tests. Of course, the standard
1373 character types table is still used for matching digits in subject
1374 strings against \d.
1376 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1377 ending in _t are). So I've renamed it as ESC_tee.
1379 3. The first argument for regexec() in the POSIX wrapper should have been
1380 defined as "const".
1382 4. Changed pcretest to use malloc() for its buffers so that they can be
1383 Electric Fenced for debugging.
1385 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1386 to read one or more bytes before the start of the subject string. Often this
1387 had no effect on PCRE's behaviour, but in some circumstances it could
1388 provoke a segmentation fault.
1390 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1391 to try to read one or more bytes before the start of the subject string.
1393 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1394 UTF-8 support could misbehave in various ways if the subject string
1395 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1396 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1397 back over UTF-8 characters.)
1400 Version 4.2 14-Apr-03
1401 ---------------------
1403 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1405 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1406 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1407 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1408 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1409 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1411 Note: automatic setting of the BUILD variables is not yet working
1412 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1413 compile-time but not at link-time
1414 [LINK]: use for linking executables only
1415 make different versions for Windows and non-Windows
1416 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1417 libraries
1418 [LINK_FOR_BUILD]: new variable
1419 [OBJEXT]: use throughout
1420 [EXEEXT]: use throughout
1421 <winshared>: new target
1422 <wininstall>: new target
1423 <dftables.o>: use native compiler
1424 <dftables>: use native linker
1425 <install>: handle Windows platform correctly
1426 <clean>: ditto
1427 <check>: ditto
1428 copy DLL to top builddir before testing
1430 As part of these changes, -no-undefined was removed again. This was reported
1431 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1432 in any case.
1434 3. Some tidies to get rid of compiler warnings:
1436 . In the match_data structure, match_limit was an unsigned long int, whereas
1437 match_call_count was an int. I've made them both unsigned long ints.
1439 . In pcretest the fact that a const uschar * doesn't automatically cast to
1440 a void * provoked a warning.
1442 . Turning on some more compiler warnings threw up some "shadow" variables
1443 and a few more missing casts.
1445 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1446 option, a class that contained a single character with a value between 128
1447 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1449 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1450 option, a class that contained several characters, but with at least one
1451 whose value was between 128 and 255 caused PCRE to crash.
1454 Version 4.1 12-Mar-03
1455 ---------------------
1457 1. Compiling with gcc -pedantic found a couple of places where casts were
1458 needed, and a string in dftables.c that was longer than standard compilers are
1459 required to support.
1461 2. Compiling with Sun's compiler found a few more places where the code could
1462 be tidied up in order to avoid warnings.
1464 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1465 first of these names is deprecated in the latest Autoconf in favour of the name
1466 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1467 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1468 analogy I have changed it to CFLAGS_FOR_BUILD.
1470 4. Added -no-undefined to the linking command in the Makefile, because this is
1471 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1472 linking step for the pcreposix library.
1474 5. PCRE was failing to diagnose the case of two named groups with the same
1475 name.
1477 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1478 literal character that is needed in the subject for a match, and scans along to
1479 ensure that it is present before embarking on the full matching process. This
1480 saves time in cases of nested unlimited repeats that are never going to match.
1481 Problem: the scan can take a lot of time if the subject is very long (e.g.
1482 megabytes), thus penalizing straightforward matches. It is now done only if the
1483 amount of subject to be scanned is less than 1000 bytes.
1485 7. A lesser problem with the same optimization is that it was recording the
1486 first character of an anchored pattern as "needed", thus provoking a search
1487 right along the subject, even when the first match of the pattern was going to
1488 fail. The "needed" character is now not set for anchored patterns, unless it
1489 follows something in the pattern that is of non-fixed length. Thus, it still
1490 fulfils its original purpose of finding quick non-matches in cases of nested
1491 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1494 Version 4.0 17-Feb-03
1495 ---------------------
1497 1. If a comment in an extended regex that started immediately after a meta-item
1498 extended to the end of string, PCRE compiled incorrect data. This could lead to
1499 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
1501 2. Moved to autoconf 2.53 and libtool 1.4.2.
1503 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
1504 the special perltest8 script is no longer needed - all the tests can be run
1505 from a single perltest script.
1507 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
1508 by \s. It has now been removed in PCRE. This means it isn't recognized as
1509 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
1510 class [:space:] *does* include VT, thereby creating a mess.
1512 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
1513 space and tab.
1515 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
1516 its new features into the main test script, reducing the number of scripts.
1518 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
1519 were backward compatible, and made the (?i) apply to the whole pattern, as if
1520 /i were given. Now it behaves more logically, and applies the option setting
1521 only to what follows. PCRE has been changed to follow suit. However, if it
1522 finds options settings right at the start of the pattern, it extracts them into
1523 the global options, as before. Thus, they show up in the info data.
1525 8. Added support for the \Q...\E escape sequence. Characters in between are
1526 treated as literals. This is slightly different from Perl in that $ and @ are
1527 also handled as literals inside the quotes. In Perl, they will cause variable
1528 interpolation. Note the following examples:
1530 Pattern PCRE matches Perl matches
1532 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
1533 \Qabc\$xyz\E abc\$xyz abc\$xyz
1534 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
1536 For compatibility with Perl, \Q...\E sequences are recognized inside character
1537 classes as well as outside them.
1539 9. Re-organized 3 code statements in pcretest to avoid "overflow in
1540 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
1541 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
1542 signed/unsigned warnings.
1544 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
1545 option for pcretest, so I've replaced it by a simple function that does just
1546 that job.
1548 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
1549 "pcregrep -".
1551 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
1552 Java package. This provides some syntactic sugar for simple cases of what my
1553 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
1554 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
1555 item, you can use this simplified notation. Note that only makes sense with
1556 greedy quantifiers. Consequently, the use of the possessive quantifier forces
1557 greediness, whatever the setting of the PCRE_UNGREEDY option.
1559 13. A change of greediness default within a pattern was not taking effect at
1560 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
1561 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
1562 was abstracted outside.
1564 14. PCRE now supports the \G assertion. It is true when the current matching
1565 position is at the start point of the match. This differs from \A when the
1566 starting offset is non-zero. Used with the /g option of pcretest (or similar
1567 code), it works in the same way as it does for Perl's /g option. If all
1568 alternatives of a regex begin with \G, the expression is anchored to the start
1569 match position, and the "anchored" flag is set in the compiled expression.
1571 15. Some bugs concerning the handling of certain option changes within patterns
1572 have been fixed. These applied to options other than (?ims). For example,
1573 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
1574 been the other way round. Some of this was related to change 7 above.
1576 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
1577 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
1578 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
1579 POSIX classes only within a class (e.g. /[[:alpha:]]/).
1581 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
1582 mode. Unlike ".", it always matches newline, whatever the setting of
1583 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
1584 assertions. Perl allows it, but it doesn't (in general) work because it can't
1585 calculate the length of the lookbehind. At least, that's the case for Perl
1586 5.8.0 - I've been told they are going to document that it doesn't work in
1587 future.
1589 18. Added an error diagnosis for escapes that PCRE does not support: these are
1590 \L, \l, \N, \P, \p, \U, \u, and \X.
1592 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
1593 reading past the end of the pattern in cases such as /[abcd/.
1595 20. PCRE was getting more memory than necessary for patterns with classes that
1596 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
1598 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
1599 compiling PCRE for use with Virtual Pascal.
1601 22. Small fix to the Makefile to make it work properly if the build is done
1602 outside the source tree.
1604 23. Added a new extension: a condition to go with recursion. If a conditional
1605 subpattern starts with (?(R) the "true" branch is used if recursion has
1606 happened, whereas the "false" branch is used only at the top level.
1608 24. When there was a very long string of literal characters (over 255 bytes
1609 without UTF support, over 250 bytes with UTF support), the computation of how
1610 much memory was required could be incorrect, leading to segfaults or other
1611 strange effects.
1613 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
1614 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
1615 there was a subsequent back reference to those brackets. This meant that, for
1616 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
1617 possible to check for precisely this case. All we can do is abandon the
1618 optimization if .* occurs inside capturing brackets when there are any back
1619 references whatsoever. (See below for a better fix that came later.)
1621 26. The handling of the optimization for finding the first character of a
1622 non-anchored pattern, and for finding a character that is required later in the
1623 match were failing in some cases. This didn't break the matching; it just
1624 failed to optimize when it could. The way this is done has been re-implemented.
1626 27. Fixed typo in error message for invalid (?R item (it said "(?p").
1628 28. Added a new feature that provides some of the functionality that Perl
1629 provides with (?{...}). The facility is termed a "callout". The way it is done
1630 in PCRE is for the caller to provide an optional function, by setting
1631 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
1632 global variable. By default it is unset, which disables all calling out. To get
1633 the function called, the regex must include (?C) at appropriate points. This
1634 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
1635 This provides a means of identifying different callout points. When PCRE
1636 reaches such a point in the regex, if pcre_callout has been set, the external
1637 function is called. It is provided with data in a structure called
1638 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
1639 matching continues; if it returns a non-zero value, the match at the current
1640 point fails. However, backtracking will occur if possible. [This was changed
1641 later and other features added - see item 49 below.]
1643 29. pcretest is upgraded to test the callout functionality. It provides a
1644 callout function that displays information. By default, it shows the start of
1645 the match and the current position in the text. There are some new data escapes
1646 to vary what happens:
1648 \C+ in addition, show current contents of captured substrings
1649 \C- do not supply a callout function
1650 \C!n return 1 when callout number n is reached
1651 \C!n!m return 1 when callout number n is reached for the mth time
1653 30. If pcregrep was called with the -l option and just a single file name, it
1654 output "<stdin>" if a match was found, instead of the file name.
1656 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
1657 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
1658 pcre_exec(). This saves a malloc/free per call. The default value of
1659 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
1660 when configuring.
1662 32. The default maximum size of a compiled pattern is 64K. There have been a
1663 few cases of people hitting this limit. The code now uses macros to handle the
1664 storing of links as offsets within the compiled pattern. It defaults to 2-byte
1665 links, but this can be changed to 3 or 4 bytes by --with-link-size when
1666 configuring. Tests 2 and 5 work only with 2-byte links because they output
1667 debugging information about compiled patterns.
1669 33. Internal code re-arrangements:
1671 (a) Moved the debugging function for printing out a compiled regex into
1672 its own source file (printint.c) and used #include to pull it into
1673 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
1674 separate copies.
1676 (b) Defined the list of op-code names for debugging as a macro in
1677 internal.h so that it is next to the definition of the opcodes.
1679 (c) Defined a table of op-code lengths for simpler skipping along compiled
1680 code. This is again a macro in internal.h so that it is next to the
1681 definition of the opcodes.
1683 34. Added support for recursive calls to individual subpatterns, along the
1684 lines of Robin Houston's patch (but implemented somewhat differently).
1686 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
1687 allow it to read and process whole directories in Win32. This code was
1688 contributed by Lionel Fourquaux; it has not been tested by me.
1690 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
1691 used to name a group. Names consist of alphanumerics and underscores, and must
1692 be unique. Back references use the syntax (?P=name) and recursive calls use
1693 (?P>name) which is a PCRE extension to the Python extension. Groups still have
1694 numbers. The function pcre_fullinfo() can be used after compilation to extract
1695 a name/number map. There are three relevant calls:
1697 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
1698 PCRE_INFO_NAMECOUNT yields the number of entries
1699 PCRE_INFO_NAMETABLE yields a pointer to the map.
1701 The map is a vector of fixed-size entries. The size of each entry depends on
1702 the length of the longest name used. The first two bytes of each entry are the
1703 group number, most significant byte first. There follows the corresponding
1704 name, zero terminated. The names are in alphabetical order.
1706 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
1707 case instead of 255. Making it the same both with and without UTF-8 support
1708 means that the same test output works with both.
1710 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
1711 calling malloc() with a zero argument.
1713 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
1714 optimization. I've improved things by keeping a bitmap of backreferences with
1715 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
1716 fact referenced, the optimization can be applied. It is unlikely that a
1717 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
1718 the match to follow \n) will appear inside brackets with a number greater than
1719 31, but if it does, any back reference > 31 suppresses the optimization.
1721 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
1722 of disabling numbered capturing parentheses. Any opening parenthesis that is
1723 not followed by ? behaves as if it were followed by ?: but named parentheses
1724 can still be used for capturing (and they will acquire numbers in the usual
1725 way).
1727 41. Redesigned the return codes from the match() function into yes/no/error so
1728 that errors can be passed back from deep inside the nested calls. A malloc
1729 failure while inside a recursive subpattern call now causes the
1730 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
1732 42. It is now possible to set a limit on the number of times the match()
1733 function is called in a call to pcre_exec(). This facility makes it possible to
1734 limit the amount of recursion and backtracking, though not in a directly
1735 obvious way, because the match() function is used in a number of different
1736 circumstances. The count starts from zero for each position in the subject
1737 string (for non-anchored patterns). The default limit is, for compatibility, a
1738 large number, namely 10 000 000. You can change this in two ways:
1740 (a) When configuring PCRE before making, you can use --with-match-limit=n
1741 to set a default value for the compiled library.
1743 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
1744 a different value is set. See 45 below.
1746 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1748 43. Added a new function pcre_config(int, void *) to enable run-time extraction
1749 of things that can be changed at compile time. The first argument specifies
1750 what is wanted and the second points to where the information is to be placed.
1751 The current list of available information is:
1755 The output is an integer that is set to one if UTF-8 support is available;
1756 otherwise it is set to zero.
1760 The output is an integer that it set to the value of the code that is used for
1761 newline. It is either LF (10) or CR (13).
1765 The output is an integer that contains the number of bytes used for internal
1766 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
1770 The output is an integer that contains the threshold above which the POSIX
1771 interface uses malloc() for output vectors. See item 31 above.
1775 The output is an unsigned integer that contains the default limit of the number
1776 of match() calls in a pcre_exec() execution. See 42 above.
1778 44. pcretest has been upgraded by the addition of the -C option. This causes it
1779 to extract all the available output from the new pcre_config() function, and to
1780 output it. The program then exits immediately.
1782 45. A need has arisen to pass over additional data with calls to pcre_exec() in
1783 order to support additional features. One way would have been to define
1784 pcre_exec2() (for example) with extra arguments, but this would not have been
1785 extensible, and would also have required all calls to the original function to
1786 be mapped to the new one. Instead, I have chosen to extend the mechanism that
1787 is used for passing in "extra" data from pcre_study().
1789 The pcre_extra structure is now exposed and defined in pcre.h. It currently
1790 contains the following fields:
1792 flags a bitmap indicating which of the following fields are set
1793 study_data opaque data from pcre_study()
1794 match_limit a way of specifying a limit on match() calls for a specific
1795 call to pcre_exec()
1796 callout_data data for callouts (see 49 below)
1798 The flag bits are also defined in pcre.h, and are
1804 The pcre_study() function now returns one of these new pcre_extra blocks, with
1805 the actual study data pointed to by the study_data field, and the
1806 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
1807 before. That is, this change is entirely upwards-compatible and requires no
1808 change to existing code.
1810 If you want to pass in additional data to pcre_exec(), you can either place it
1811 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
1812 block.
1814 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
1815 data string contains the escape sequence \M, pcretest calls pcre_exec() several
1816 times with different match limits, until it finds the minimum value needed for
1817 pcre_exec() to complete. The value is then output. This can be instructive; for
1818 most simple matches the number is quite small, but for pathological cases it
1819 gets very large very quickly.
1821 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
1822 returns the size of the data block pointed to by the study_data field in a
1823 pcre_extra block, that is, the value that was passed as the argument to
1824 pcre_malloc() when PCRE was getting memory in which to place the information
1825 created by pcre_study(). The fourth argument should point to a size_t variable.
1826 pcretest has been extended so that this information is shown after a successful
1827 pcre_study() call when information about the compiled regex is being displayed.
1829 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
1830 because what follows is always an absolute path. (Later: it turns out that this
1831 is more than cosmetic for MinGW, because it doesn't like empty path
1832 components.)
1834 49. Some changes have been made to the callout feature (see 28 above):
1836 (i) A callout function now has three choices for what it returns:
1838 0 => success, carry on matching
1839 > 0 => failure at this point, but backtrack if possible
1840 < 0 => serious error, return this value from pcre_exec()
1842 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
1843 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
1844 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
1845 use by callout functions. It will never be used by PCRE itself.
1847 (ii) The pcre_extra structure (see 45 above) has a void * field called
1848 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
1849 pcre_callout_block structure has a field of the same name. The contents of
1850 the field passed in the pcre_extra structure are passed to the callout
1851 function in the corresponding field in the callout block. This makes it
1852 easier to use the same callout-containing regex from multiple threads. For
1853 testing, the pcretest program has a new data escape
1855 \C*n pass the number n (may be negative) as callout_data
1857 If the callout function in pcretest receives a non-zero value as
1858 callout_data, it returns that value.
1860 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
1861 there were some redundant $(CFLAGS) in commands that are now specified as
1862 $(LINK), which already includes $(CFLAGS).
1864 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
1865 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
1866 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
1867 one-byte characters throughout. Note that case-insensitive matching applies
1868 only to characters whose values are less than 256. PCRE doesn't support the
1869 notion of cases for higher-valued characters.
1871 (i) A character class whose characters are all within 0-255 is handled as
1872 a bit map, and the map is inverted for negative classes. Previously, a
1873 character > 255 always failed to match such a class; however it should
1874 match if the class was a negative one (e.g. [^ab]). This has been fixed.
1876 (ii) A negated character class with a single character < 255 is coded as
1877 "not this character" (OP_NOT). This wasn't working properly when the test
1878 character was multibyte, either singly or repeated.
1880 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
1881 mode, for example: \x{100}{2,3}.
1883 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
1884 singly or repeated) now correctly test multibyte characters. However,
1885 PCRE doesn't recognize any characters with values greater than 255 as
1886 digits, spaces, or word characters. Such characters always match \D, \S,
1887 and \W, and never match \d, \s, or \w.
1889 (v) Classes may now contain characters and character ranges with values
1890 greater than 255. For example: [ab\x{100}-\x{400}].
1892 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
1893 PCRE in UTF-8 mode.
1895 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
1896 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
1897 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
1898 value.)
1900 53. The single man page has become too large. I have therefore split it up into
1901 a number of separate man pages. These also give rise to individual HTML pages;
1902 these are now put in a separate directory, and there is an index.html page that
1903 lists them all. Some hyperlinking between the pages has been installed.
1905 54. Added convenience functions for handling named capturing parentheses.
1907 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
1908 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
1909 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
1910 are faulted.
1912 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
1913 calling configure. These values are used when compiling the dftables.c program
1914 which is run to generate the source of the default character tables. They
1915 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
1916 you will need to set these values.
1918 57. Updated the building process for Windows DLL, as provided by Fred Cox.
1921 Version 3.9 02-Jan-02
1922 ---------------------
1924 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
1926 2. If --disable-static was given, the building process failed when trying to
1927 build pcretest and pcregrep. (For some reason it was using libtool to compile
1928 them, which is not right, as they aren't part of the library.)
1931 Version 3.8 18-Dec-01
1932 ---------------------
1934 1. The experimental UTF-8 code was completely screwed up. It was packing the
1935 bytes in the wrong order. How dumb can you get?
1938 Version 3.7 29-Oct-01
1939 ---------------------
1941 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
1942 This caused pcretest, when used on the test data, to segfault. Unfortunately,
1943 this didn't happen under Solaris 8, where I normally test things.
1945 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
1946 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
1947 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
1948 this fix an hour or so after the initial 3.7 release.)
1951 Version 3.6 23-Oct-01
1952 ---------------------
1954 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
1955 offsets passed as NULL with zero offset count.
1957 2. The config.guess and config.sub files had not been updated when I moved to
1958 the latest autoconf.
1961 Version 3.5 15-Aug-01
1962 ---------------------
1964 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
1965 had been forgotten.
1967 2. By using declared but undefined structures, we can avoid using "void"
1968 definitions in pcre.h while keeping the internal definitions of the structures
1969 private.
1971 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
1972 user point of view, this means that both static and shared libraries are built
1973 by default, but this can be individually controlled. More of the work of
1974 handling this static/shared cases is now inside libtool instead of PCRE's make
1975 file.
1977 4. The pcretest utility is now installed along with pcregrep because it is
1978 useful for users (to test regexs) and by doing this, it automatically gets
1979 relinked by libtool. The documentation has been turned into a man page, so
1980 there are now .1, .txt, and .html versions in /doc.
1982 5. Upgrades to pcregrep:
1983 (i) Added long-form option names like gnu grep.
1984 (ii) Added --help to list all options with an explanatory phrase.
1985 (iii) Added -r, --recursive to recurse into sub-directories.
1986 (iv) Added -f, --file to read patterns from a file.
1988 6. pcre_exec() was referring to its "code" argument before testing that
1989 argument for NULL (and giving an error if it was NULL).
1991 7. Upgraded Makefile.in to allow for compiling in a different directory from
1992 the source directory.
1994 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
1995 options bits, the pointer it was passed was to an int instead of to an unsigned
1996 long int. This mattered only on 64-bit systems.
1998 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
1999 generated) instead of pcre.in, which it its source. Also made the same change
2000 in several of the .c files.
2002 10. A new release of gcc defines printf() as a macro, which broke pcretest
2003 because it had an ifdef in the middle of a string argument for printf(). Fixed
2004 by using separate calls to printf().
2006 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2007 script, to force use of CR or LF instead of \n in the source. On non-Unix
2008 systems, the value can be set in config.h.
2010 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2011 absolute limit. Changed the text of the error message to make this clear, and
2012 likewise updated the man page.
2014 13. The limit of 99 on the number of capturing subpatterns has been removed.
2015 The new limit is 65535, which I hope will not be a "real" limit.
2018 Version 3.4 22-Aug-00
2019 ---------------------
2021 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2023 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2026 Version 3.3 01-Aug-00
2027 ---------------------
2029 1. If an octal character was given, but the value was greater than \377, it
2030 was not getting masked to the least significant bits, as documented. This could
2031 lead to crashes in some systems.
2033 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2034 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2036 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2037 These just pass their arguments on to (pcre_free)(), but they are provided
2038 because some uses of PCRE bind it to non-C systems that can call its functions,
2039 but cannot call free() or pcre_free() directly.
2041 4. Add "make test" as a synonym for "make check". Corrected some comments in
2042 the Makefile.
2044 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2045 Makefile.
2047 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2048 command called pgrep for grepping around the active processes.
2050 7. Added the beginnings of support for UTF-8 character strings.
2052 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2053 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2054 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2055 out for the ar command.)
2058 Version 3.2 12-May-00
2059 ---------------------
2061 This is purely a bug fixing release.
2063 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2064 of ZA. This was just one example of several cases that could provoke this bug,
2065 which was introduced by change 9 of version 2.00. The code for breaking
2066 infinite loops after an iteration that matches an empty string was't working
2067 correctly.
2069 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2070 when matched against abbab (for example). After matching an empty string, it
2071 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2072 caused it to match further down the string than it should.
2074 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2075 was there because it doesn't seem to be needed, and it causes trouble on some
2076 systems, as it is not a Standard C header. It has been removed.
2078 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2079 were reported on the Macintosh. The changes were from
2081 while ((c = *(++ptr)) != 0 && c != '\n');
2082 to
2083 while ((c = *(++ptr)) != 0 && c != '\n') ;
2085 Totally extraordinary, but if that's what it takes...
2087 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2088 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2089 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2090 assumes the way PCRE uses memmove() (always moving upwards).
2092 6. PCRE is being used in one environment where strchr() is not available. There
2093 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2094 faster code anyway.
2097 Version 3.1 09-Feb-00
2098 ---------------------
2100 The only change in this release is the fixing of some bugs in Makefile.in for
2101 the "install" target:
2103 (1) It was failing to install pcreposix.h.
2105 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2108 Version 3.0 01-Feb-00
2109 ---------------------
2111 1. Add support for the /+ modifier to perltest (to output $` like it does in
2112 pcretest).
2114 2. Add support for the /g modifier to perltest.
2116 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2117 matches null strings.
2119 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2120 pattern. Perl treats empty patterns specially - it reuses the most recent
2121 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2122 effect.
2124 5. The POSIX interface was broken in that it was just handing over the POSIX
2125 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2126 required a bigger vector, with some working space on the end. This means that
2127 the POSIX wrapper now has to get and free some memory, and copy the results.
2129 6. Added some simple autoconf support, placing the test data and the
2130 documentation in separate directories, re-organizing some of the
2131 information files, and making it build pcre-config (a GNU standard). Also added
2132 libtool support for building PCRE as a shared library, which is now the
2133 default.
2135 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2136 09 are not valid octal constants. Single digits will be used for minor values
2137 less than 10.
2139 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2140 existing programs that set these in the POSIX interface can use PCRE without
2141 modification.
2143 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2144 return all that pcre_info() returns, plus additional data. The pcre_info()
2145 function is retained for compatibility, but is considered to be obsolete.
2147 10. Added experimental recursion feature (?R) to handle one common case that
2148 Perl 5.6 will be able to do with (?p{...}).
2150 11. Added support for POSIX character classes like [:alpha:], which Perl is
2151 adopting.
2154 Version 2.08 31-Aug-99
2155 ----------------------
2157 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2158 trying to match at the startoffset position, but instead was moving forward to
2159 the next newline as if a previous match had failed.
2161 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2162 and could get into a loop if a null string was matched other than at the start
2163 of the subject.
2165 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2166 be distinguished at compile time, and for completeness also added PCRE_DATE.
2168 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2169 in GnuWin32 environments.
2172 Version 2.07 29-Jul-99
2173 ----------------------
2175 1. The documentation is now supplied in plain text form and HTML as well as in
2176 the form of man page sources.
2178 2. C++ compilers don't like assigning (void *) values to other pointer types.
2179 In particular this affects malloc(). Although there is no problem in Standard
2180 C, I've put in casts to keep C++ compilers happy.
2182 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2183 should be (const char *).
2185 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2186 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2187 However, I haven't made this a standard facility. The documentation doesn't
2188 mention it, and the Makefile doesn't support it.
2190 5. The Makefile now contains an "install" target, with editable destinations at
2191 the top of the file. The pcretest program is not installed.
2193 6. pgrep -V now gives the PCRE version number and date.
2195 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2196 causing the entire string to be ignored, instead of just the last character.
2198 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2199 non-matching string, it can take a very, very long time, even for strings of
2200 quite modest length, because of the nested recursion. PCRE now does better in
2201 some of these cases. It does this by remembering the last required literal
2202 character in the pattern, and pre-searching the subject to ensure it is present
2203 before running the real match. In other words, it applies a heuristic to detect
2204 some types of certain failure quickly, and in the above example, if presented
2205 with a string that has no trailing " it gives "no match" very quickly.
2207 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2208 other alternatives are tried instead.
2211 Version 2.06 09-Jun-99
2212 ----------------------
2214 1. Change pcretest's output for amount of store used to show just the code
2215 space, because the remainder (the data block) varies in size between 32-bit and
2216 64-bit systems.
2218 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2219 start matching at. This allows lookbehinds to work when searching for multiple
2220 occurrences in a string.
2222 3. Added additional options to pcretest for testing multiple occurrences:
2224 /+ outputs the rest of the string that follows a match
2225 /g loops for multiple occurrences, using the new startoffset argument
2226 /G loops for multiple occurrences by passing an incremented pointer
2228 4. PCRE wasn't doing the "first character" optimization for patterns starting
2229 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2230 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2231 the letter 'x'. On long subject strings, this gives a significant speed-up.
2234 Version 2.05 21-Apr-99
2235 ----------------------
2237 1. Changed the type of magic_number from int to long int so that it works
2238 properly on 16-bit systems.
2240 2. Fixed a bug which caused patterns starting with .* not to work correctly
2241 when the subject string contained newline characters. PCRE was assuming
2242 anchoring for such patterns in all cases, which is not correct because .* will
2243 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2244 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2245 must be retried after every newline in the subject.
2248 Version 2.04 18-Feb-99
2249 ----------------------
2251 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2252 computation of the store needed to hold the pattern was incorrect (too large).
2253 If such patterns were nested a few deep, this could multiply and become a real
2254 problem.
2256 2. Added /M option to pcretest to show the memory requirement of a specific
2257 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2259 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2260 compiled in such a way that the backtracking after subsequent failure was
2261 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2262 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2265 Version 2.03 02-Feb-99
2266 ----------------------
2268 1. Fixed typo and small mistake in man page.
2270 2. Added 4th condition (GPL supersedes if conflict) and created separate
2271 LICENCE file containing the conditions.
2273 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2274 Perl, that is the internal \ allows the delimiter to be included in the
2275 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2276 the final delimiter, add \ to the end of the pattern (to test the error).
2278 4. Added the convenience functions for extracting substrings after a successful
2279 match. Updated pcretest to make it able to test these functions.
2282 Version 2.02 14-Jan-99
2283 ----------------------
2285 1. Initialized the working variables associated with each extraction so that
2286 their saving and restoring doesn't refer to uninitialized store.
2288 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2289 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2290 fix the problem.
2292 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2293 calls, and wasn't printing the correct value for compiling calls. Increased the
2294 default value of LOOPREPEAT, and the number of significant figures in the
2295 times.
2297 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2299 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2300 a building problem on Windows NT with a FAT file system.
2303 Version 2.01 21-Oct-98
2304 ----------------------
2306 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2307 to character tables built by pcre_maketables() in the current locale. If NULL
2308 is passed, the default tables are used.
2311 Version 2.00 24-Sep-98
2312 ----------------------
2314 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2315 it any more.
2317 2. Allow quantification of (?>) groups, and make it work correctly.
2319 3. The first character computation wasn't working for (?>) groups.
2321 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2322 end of the subject) and add 5.005's \z, which really does match only at the
2323 very end of the subject.
2325 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2327 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2328 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2329 localized options. All options to pcre_study() were also removed.
2331 7. Add other new features from 5.005:
2333 $(?<= positive lookbehind
2334 $(?<! negative lookbehind
2335 (?imsx-imsx) added the unsetting capability
2336 such a setting is global if at outer level; local otherwise
2337 (?imsx-imsx:) non-capturing groups with option setting
2338 (?(cond)re|re) conditional pattern matching
2340 A backreference to itself in a repeated group matches the previous
2341 captured string.
2343 8. General tidying up of studying (both automatic and via "study")
2344 consequential on the addition of new assertions.
2346 9. As in 5.005, unlimited repeated groups that could match an empty substring
2347 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2348 runtime if any iteration does actually match an empty substring.
2350 10. Include the RunTest script in the distribution.
2352 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2353 discrepancies, some of which were old and were also with respect to 5.004. They
2354 have now been fixed.
2357 Version 1.09 28-Apr-98
2358 ----------------------
2360 1. A negated single character class followed by a quantifier with a minimum
2361 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2362 program crashes, or just wrong answers. This did not apply to negated classes
2363 containing more than one character, or to minima other than one.
2366 Version 1.08 27-Mar-98
2367 ----------------------
2369 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2371 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2372 latter must appear before anything that relies on it in the pattern.
2375 Version 1.07 16-Feb-98
2376 ----------------------
2378 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2379 repeat of a potentially empty string).
2382 Version 1.06 23-Jan-98
2383 ----------------------
2385 1. Added Markus Oberhumer's little patches for C++.
2387 2. Literal strings longer than 255 characters were broken.
2390 Version 1.05 23-Dec-97
2391 ----------------------
2393 1. Negated character classes containing more than one character were failing if
2394 PCRE_CASELESS was set at run time.
2397 Version 1.04 19-Dec-97
2398 ----------------------
2400 1. Corrected the man page, where some "const" qualifiers had been omitted.
2402 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2403 input syntax.
2405 3. Fixed memory leak which occurred when a regex with back references was
2406 matched with an offsets vector that wasn't big enough. The temporary memory
2407 that is used in this case wasn't being freed if the match failed.
2409 4. Tidied pcretest to ensure it frees memory that it gets.
2411 5. Temporary memory was being obtained in the case where the passed offsets
2412 vector was exactly big enough.
2414 6. Corrected definition of offsetof() from change 5 below.
2416 7. I had screwed up change 6 below and broken the rules for the use of
2417 setjmp(). Now fixed.
2420 Version 1.03 18-Dec-97
2421 ----------------------
2423 1. A erroneous regex with a missing opening parenthesis was correctly
2424 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2425 on some systems.
2427 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2428 it was reported that one broken compiler failed on the former because "code" is
2429 also an independent variable.
2431 3. The erroneous regex a[]b caused an array overrun reference.
2433 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2434 fail on data ending with that character. (It was going on too far, and checking
2435 the next character, typically a binary zero.) This was specific to the
2436 optimized code for single-character negative classes.
2438 5. Added a contributed patch from the TIN world which does the following:
2440 + Add an undef for memmove, in case the the system defines a macro for it.
2442 + Add a definition of offsetof(), in case there isn't one. (I don't know
2443 the reason behind this - offsetof() is part of the ANSI standard - but
2444 it does no harm).
2446 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2447 most of the places where whitespace preceded '#'. I have given up and
2448 allowed the remaining 2 cases to be at the margin.
2450 + Rename some variables in pcre to eliminate shadowing. This seems very
2451 pedantic, but does no harm, of course.
2453 6. Moved the call to setjmp() into its own function, to get rid of warnings
2454 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2456 7. Constructs such as \d{8,} were compiling into the equivalent of
2457 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2458 outcome, but in this particular case used more store than had been allocated,
2459 which caused the bug to be discovered because it threw up an internal error.
2461 8. The debugging code in both pcre and pcretest for outputting the compiled
2462 form of a regex was going wrong in the case of back references followed by
2463 curly-bracketed repeats.
2466 Version 1.02 12-Dec-97
2467 ----------------------
2469 1. Typos in pcre.3 and comments in the source fixed.
2471 2. Applied a contributed patch to get rid of places where it used to remove
2472 'const' from variables, and fixed some signed/unsigned and uninitialized
2473 variable warnings.
2475 3. Added the "runtest" target to Makefile.
2477 4. Set default compiler flag to -O2 rather than just -O.
2480 Version 1.01 19-Nov-97
2481 ----------------------
2483 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2484 like /([ab]*)*/, that is, for classes with more than one character in them.
2486 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2487 as /((?>a*))*/ (a PCRE_EXTRA facility).
2490 Version 1.00 18-Nov-97
2491 ----------------------
2493 1. Added compile-time macros to support systems such as SunOS4 which don't have
2494 memmove() or strerror() but have other things that can be used instead.
2496 2. Arranged that "make clean" removes the executables.
2499 Version 0.99 27-Oct-97
2500 ----------------------
2502 1. Fixed bug in code for optimizing classes with only one character. It was
2503 initializing a 32-byte map regardless, which could cause it to run off the end
2504 of the memory it had got.
2506 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
2509 Version 0.98 22-Oct-97
2510 ----------------------
2512 1. Fixed bug in code for handling temporary memory usage when there are more
2513 back references than supplied space in the ovector. This could cause segfaults.
2516 Version 0.97 21-Oct-97
2517 ----------------------
2519 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
2521 2. Optimized negated single characters not to use a bit map.
2523 3. Brought error texts together as macro definitions; clarified some of them;
2524 fixed one that was wrong - it said "range out of order" when it meant "invalid
2525 escape sequence".
2527 4. Changed some char * arguments to const char *.
2529 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
2531 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
2532 pcretest.
2535 Version 0.96 16-Oct-97
2536 ----------------------
2538 1. Added a simple "pgrep" utility to the distribution.
2540 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
2541 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
2542 where "ddd" means "one or more decimal digits".
2544 3. Fixed serious bug. If a pattern had a back reference, but the call to
2545 pcre_exec() didn't supply a large enough ovector to record the related
2546 identifying subpattern, the match always failed. PCRE now remembers the number
2547 of the largest back reference, and gets some temporary memory in which to save
2548 the offsets during matching if necessary, in order to ensure that
2549 backreferences always work.
2551 4. Increased the compatibility with Perl in a number of ways:
2553 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
2554 to request this handling. The option can be set at compile or exec time.
2556 (b) $ matches before a terminating newline by default; an option
2557 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
2558 mode). The option can be set at compile or exec time.
2560 (c) The handling of \ followed by a digit other than 0 is now supposed to be
2561 the same as Perl's. If the decimal number it represents is less than 10
2562 or there aren't that many previous left capturing parentheses, an octal
2563 escape is read. Inside a character class, it's always an octal escape,
2564 even if it is a single digit.
2566 (d) An escaped but undefined alphabetic character is taken as a literal,
2567 unless PCRE_EXTRA is set. Currently this just reserves the remaining
2568 escapes.
2570 (e) {0} is now permitted. (The previous item is removed from the compiled
2571 pattern).
2573 5. Changed all the names of code files so that the basic parts are no longer
2574 than 10 characters, and abolished the teeny "globals.c" file.
2576 6. Changed the handling of character classes; they are now done with a 32-byte
2577 bit map always.
2579 7. Added the -d and /D options to pcretest to make it possible to look at the
2580 internals of compilation without having to recompile pcre.
2583 Version 0.95 23-Sep-97
2584 ----------------------
2586 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
2587 \x20 at the start of a run of normal characters. These were being treated as
2588 real characters, instead of the source characters being re-checked.
2591 Version 0.94 18-Sep-97
2592 ----------------------
2594 1. The functions are now thread-safe, with the caveat that the global variables
2595 containing pointers to malloc() and free() or alternative functions are the
2596 same for all threads.
2598 2. Get pcre_study() to generate a bitmap of initial characters for non-
2599 anchored patterns when this is possible, and use it if passed to pcre_exec().
2602 Version 0.93 15-Sep-97
2603 ----------------------
2605 1. /(b)|(:+)/ was computing an incorrect first character.
2607 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
2608 but not actually doing anything yet.
2610 3. Treat "-" characters in classes that cannot be part of ranges as literals,
2611 as Perl does (e.g. [-az] or [az-]).
2613 4. Set the anchored flag if a branch starts with .* or .*? because that tests
2614 all possible positions.
2616 5. Split up into different modules to avoid including unneeded functions in a
2617 compiled binary. However, compile and exec are still in one module. The "study"
2618 function is split off.
2620 6. The character tables are now in a separate module whose source is generated
2621 by an auxiliary program - but can then be edited by hand if required. There are
2622 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
2623 toupper() in the code.
2625 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
2626 make them global. Abolish the function for setting them, as the caller can now
2627 set them directly.
2630 Version 0.92 11-Sep-97
2631 ----------------------
2633 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
2634 (e.g. /a{1,3}/) was broken (I mis-optimized it).
2636 2. Caseless matching was not working in character classes if the characters in
2637 the pattern were in upper case.
2639 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
2641 4. Make PCRE_ANCHORED public and accept as a compile option.
2643 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
2644 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
2645 pass them.
2647 6. Give an error if bad option bits passed at compile or run time.
2649 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
2650 pcretest to cause it to pass that flag.
2652 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
2653 options, and the first character, if set.
2655 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
2658 Version 0.91 10-Sep-97
2659 ----------------------
2661 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
2662 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
2664 2. PCRE was looping on encountering an indefinitely repeated back reference to
2665 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
2666 Perl does - treats the match as successful.
2668 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5