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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 168 - (show annotations) (download)
Tue May 29 15:18:18 2007 UTC (7 years, 6 months ago) by ph10
File size: 128348 byte(s)
Add support for the Perl 5.10 \K facility.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.2 01-May-07
5 ---------------------
6
7 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
8 which is apparently normally available under Windows.
9
10 2. Re-jig the pcregrep tests with different newline settings in an attempt
11 to make them independent of the local environment's newline setting.
12
13 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
14
15 4. Some of the "internals" tests were previously cut out when the link size
16 was not 2, because the output contained actual offsets. The recent new
17 "Z" feature of pcretest means that these can be cut out, making the tests
18 usable with all link sizes.
19
20 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
21 stack recursion. This gives a massive performance boost under BSD, but just
22 a small improvement under Linux. However, it saves one field in the frame
23 in all cases.
24
25 6. Added more features from the forthcoming Perl 5.10:
26
27 (a) (?-n) (where n is a string of digits) is a relative subroutine or
28 recursion call. It refers to the nth most recently opened parentheses.
29
30 (b) (?+n) is also a relative subroutine call; it refers to the nth next
31 to be opened parentheses.
32
33 (c) Conditions that refer to capturing parentheses can be specified
34 relatively, for example, (?(-2)... or (?(+3)...
35
36 (d) \K resets the start of the current match so that everything before
37 is not part of it.
38
39
40 Version 7.1 24-Apr-07
41 ---------------------
42
43 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
44 that is more "standard", making use of automake and other Autotools. There
45 is some re-arrangement of the files and adjustment of comments consequent
46 on this.
47
48 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
49 for recursive directory scanning broke on some systems because the files
50 are not scanned in any specific order and on different systems the order
51 was different. A call to "sort" has been inserted into RunGrepTest for the
52 approprate test as a short-term fix. In the longer term there may be an
53 alternative.
54
55 3. I had an email from Eric Raymond about problems translating some of PCRE's
56 man pages to HTML (despite the fact that I distribute HTML pages, some
57 people do their own conversions for various reasons). The problems
58 concerned the use of low-level troff macros .br and .in. I have therefore
59 removed all such uses from the man pages (some were redundant, some could
60 be replaced by .nf/.fi pairs). The 132html script that I use to generate
61 HTML has been updated to handle .nf/.fi and to complain if it encounters
62 .br or .in.
63
64 4. Updated comments in configure.ac that get placed in config.h.in and also
65 arranged for config.h to be included in the distribution, with the name
66 config.h.generic, for the benefit of those who have to compile without
67 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
68
69 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
70 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
71 makevp.bat for the current PCRE, using the additional files
72 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
73
74 6. A Windows user reported a minor discrepancy with test 2, which turned out
75 to be caused by a trailing space on an input line that had got lost in his
76 copy. The trailing space was an accident, so I've just removed it.
77
78 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
79 that is needed.
80
81 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
82 as "const" (a) because they are and (b) because it helps the PHP
83 maintainers who have recently made a script to detect big data structures
84 in the php code that should be moved to the .rodata section. I remembered
85 to update Builducptable as well, so it won't revert if ucptable.h is ever
86 re-created.
87
88 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
89 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
90 order to be able to cut out the UTF-8 tables in the latter when UTF-8
91 support is not required. This saves 1.5-2K of code, which is important in
92 some applications.
93
94 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
95 so as not to refer to the tables, even though these functions will never be
96 called when UTF-8 support is disabled. Otherwise there are problems with a
97 shared library.
98
99 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
100
101 (a) It was defining its arguments as char * instead of void *.
102
103 (b) It was assuming that all moves were upwards in memory; this was true
104 a long time ago when I wrote it, but is no longer the case.
105
106 The emulated memove() is provided for those environments that have neither
107 memmove() nor bcopy(). I didn't think anyone used it these days, but that
108 is clearly not the case, as these two bugs were recently reported.
109
110 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
111 and Detrail to create the HTML documentation, the .txt form of the man
112 pages, and it removes trailing spaces from listed files. It also creates
113 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
114 case, it wraps all the #defines with #ifndefs. This script should be run
115 before "make dist".
116
117 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
118 with Unicode property support.
119
120 (a) For a maximizing quantifier, if the two different cases of the
121 character were of different lengths in their UTF-8 codings (there are
122 some cases like this - I found 11), and the matching function had to
123 back up over a mixture of the two cases, it incorrectly assumed they
124 were both the same length.
125
126 (b) When PCRE was configured to use the heap rather than the stack for
127 recursion during matching, it was not correctly preserving the data for
128 the other case of a UTF-8 character when checking ahead for a match
129 while processing a minimizing repeat. If the check also involved
130 matching a wide character, but failed, corruption could cause an
131 erroneous result when trying to check for a repeat of the original
132 character.
133
134 13. Some tidying changes to the testing mechanism:
135
136 (a) The RunTest script now detects the internal link size and whether there
137 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
138 values substituted by "configure". (The RunGrepTest script already did
139 this for UTF-8.) The configure.ac script no longer substitutes the
140 relevant variables.
141
142 (b) The debugging options /B and /D in pcretest show the compiled bytecode
143 with length and offset values. This means that the output is different
144 for different internal link sizes. Test 2 is skipped for link sizes
145 other than 2 because of this, bypassing the problem. Unfortunately,
146 there was also a test in test 3 (the locale tests) that used /B and
147 failed for link sizes other than 2. Rather than cut the whole test out,
148 I have added a new /Z option to pcretest that replaces the length and
149 offset values with spaces. This is now used to make test 3 independent
150 of link size. (Test 2 will be tidied up later.)
151
152 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
153 segmentation fault instead of returning the appropriate error message.
154
155 15. In multiline mode when the newline sequence was set to "any", the pattern
156 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
157 This doesn't seem right; it now treats the CRLF combination as the line
158 ending, and so does not match in that case. It's only a pattern such as ^$
159 that would hit this one: something like ^ABC$ would have failed after \r
160 and then tried again after \r\n.
161
162 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
163 in an attempt to make files that differ only in their line terminators
164 compare equal. This works on Linux.
165
166 17. Under certain error circumstances pcregrep might try to free random memory
167 as it exited. This is now fixed, thanks to valgrind.
168
169 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
170 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
171 was because its rules for how to advance for /g after matching an empty
172 string at the end of a line did not allow for this case. They now check for
173 it specially.
174
175 20. pcretest is supposed to handle patterns and data of any length, by
176 extending its buffers when necessary. It was getting this wrong when the
177 buffer for a data line had to be extended.
178
179 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
180 CRLF as a newline sequence.
181
182 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
183 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
184 I have nevertheless tidied it up.
185
186 23. Added some casts to kill warnings from HP-UX ia64 compiler.
187
188 24. Added a man page for pcre-config.
189
190
191 Version 7.0 19-Dec-06
192 ---------------------
193
194 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
195 moving to gcc 4.1.1.
196
197 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
198 sys/time.h, which is documented as needed for this function. It doesn't
199 seem to matter on Linux, but it showed up on some releases of OS X.
200
201 3. It seems that there are systems where bytes whose values are greater than
202 127 match isprint() in the "C" locale. The "C" locale should be the
203 default when a C program starts up. In most systems, only ASCII printing
204 characters match isprint(). This difference caused the output from pcretest
205 to vary, making some of the tests fail. I have changed pcretest so that:
206
207 (a) When it is outputting text in the compiled version of a pattern, bytes
208 other than 32-126 are always shown as hex escapes.
209
210 (b) When it is outputting text that is a matched part of a subject string,
211 it does the same, unless a different locale has been set for the match
212 (using the /L modifier). In this case, it uses isprint() to decide.
213
214 4. Fixed a major bug that caused incorrect computation of the amount of memory
215 required for a compiled pattern when options that changed within the
216 pattern affected the logic of the preliminary scan that determines the
217 length. The relevant options are -x, and -i in UTF-8 mode. The result was
218 that the computed length was too small. The symptoms of this bug were
219 either the PCRE error "internal error: code overflow" from pcre_compile(),
220 or a glibc crash with a message such as "pcretest: free(): invalid next
221 size (fast)". Examples of patterns that provoked this bug (shown in
222 pcretest format) are:
223
224 /(?-x: )/x
225 /(?x)(?-x: \s*#\s*)/
226 /((?i)[\x{c0}])/8
227 /(?i:[\x{c0}])/8
228
229 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
230 is now done differently.
231
232 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
233 wrapper classes; (b) implement a new function in the C++ scanner that is
234 more efficient than the old way of doing things because it avoids levels of
235 recursion in the regex matching; (c) add a paragraph to the documentation
236 for the FullMatch() function.
237
238 6. The escape sequence \n was being treated as whatever was defined as
239 "newline". Not only was this contrary to the documentation, which states
240 that \n is character 10 (hex 0A), but it also went horribly wrong when
241 "newline" was defined as CRLF. This has been fixed.
242
243 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
244 was being set to -1 for the "end of line" case (supposedly a value that no
245 character can have). Though this value is never used (the check for end of
246 line is "zero bytes in current character"), it caused compiler complaints.
247 I've changed it to 0xffffffff.
248
249 8. In pcre_version.c, the version string was being built by a sequence of
250 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
251 string (as it is for production releases) called a macro with an empty
252 argument. The C standard says the result of this is undefined. The gcc
253 compiler treats it as an empty string (which was what was wanted) but it is
254 reported that Visual C gives an error. The source has been hacked around to
255 avoid this problem.
256
257 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
258 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
259 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
260 of them did).
261
262 10. Originally, pcretest opened its input and output without "b"; then I was
263 told that "b" was needed in some environments, so it was added for release
264 5.0 to both the input and output. (It makes no difference on Unix-like
265 systems.) Later I was told that it is wrong for the input on Windows. I've
266 now abstracted the modes into two macros, to make it easier to fiddle with
267 them, and removed "b" from the input mode under Windows.
268
269 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
270
271 12. Added -help and --help to pcretest as an official way of being reminded
272 of the options.
273
274 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
275 and pcrecpp.cc because they annoy compilers at high warning levels.
276
277 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
278
279 15. Fixed an occurrence of == in configure.ac that should have been = (shell
280 scripts are not C programs :-) and which was not noticed because it works
281 on Linux.
282
283 16. pcretest is supposed to handle any length of pattern and data line (as one
284 line or as a continued sequence of lines) by extending its input buffer if
285 necessary. This feature was broken for very long pattern lines, leading to
286 a string of junk being passed to pcre_compile() if the pattern was longer
287 than about 50K.
288
289 17. I have done a major re-factoring of the way pcre_compile() computes the
290 amount of memory needed for a compiled pattern. Previously, there was code
291 that made a preliminary scan of the pattern in order to do this. That was
292 OK when PCRE was new, but as the facilities have expanded, it has become
293 harder and harder to keep it in step with the real compile phase, and there
294 have been a number of bugs (see for example, 4 above). I have now found a
295 cunning way of running the real compile function in a "fake" mode that
296 enables it to compute how much memory it would need, while actually only
297 ever using a few hundred bytes of working memory and without too many
298 tests of the mode. This should make future maintenance and development
299 easier. A side effect of this work is that the limit of 200 on the nesting
300 depth of parentheses has been removed (though this was never a serious
301 limitation, I suspect). However, there is a downside: pcre_compile() now
302 runs more slowly than before (30% or more, depending on the pattern). I
303 hope this isn't a big issue. There is no effect on runtime performance.
304
305 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
306 newline (only possible for the last line of a file) and it was a
307 pattern that set a locale (followed by /Lsomething), pcretest crashed.
308
309 19. Added additional timing features to pcretest. (1) The -tm option now times
310 matching only, not compiling. (2) Both -t and -tm can be followed, as a
311 separate command line item, by a number that specifies the number of
312 repeats to use when timing. The default is 50000; this gives better
313 precision, but takes uncomfortably long for very large patterns.
314
315 20. Extended pcre_study() to be more clever in cases where a branch of a
316 subpattern has no definite first character. For example, (a*|b*)[cd] would
317 previously give no result from pcre_study(). Now it recognizes that the
318 first character must be a, b, c, or d.
319
320 21. There was an incorrect error "recursive call could loop indefinitely" if
321 a subpattern (or the entire pattern) that was being tested for matching an
322 empty string contained only one non-empty item after a nested subpattern.
323 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
324 incorrectly, because the \d was being skipped in the check.
325
326 22. The pcretest program now has a new pattern option /B and a command line
327 option -b, which is equivalent to adding /B to every pattern. This causes
328 it to show the compiled bytecode, without the additional information that
329 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
330 is the same as /B/I).
331
332 23. A new optimization is now able automatically to treat some sequences such
333 as a*b as a*+b. More specifically, if something simple (such as a character
334 or a simple class like \d) has an unlimited quantifier, and is followed by
335 something that cannot possibly match the quantified thing, the quantifier
336 is automatically "possessified".
337
338 24. A recursive reference to a subpattern whose number was greater than 39
339 went wrong under certain circumstances in UTF-8 mode. This bug could also
340 have affected the operation of pcre_study().
341
342 25. Realized that a little bit of performance could be had by replacing
343 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
344
345 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
346
347 27. Possessive quantifiers such as a++ were previously implemented by turning
348 them into atomic groups such as ($>a+). Now they have their own opcodes,
349 which improves performance. This includes the automatically created ones
350 from 23 above.
351
352 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
353 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
354 the first matched character to be a colon. This applied both to named and
355 numbered groups.
356
357 29. The ucpinternal.h header file was missing its idempotency #ifdef.
358
359 30. I was sent a "project" file called libpcre.a.dev which I understand makes
360 building PCRE on Windows easier, so I have included it in the distribution.
361
362 31. There is now a check in pcretest against a ridiculously large number being
363 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
364 loop, the loop is abandoned.
365
366 32. Forward references to subpatterns in conditions such as (?(2)...) where
367 subpattern 2 is defined later cause pcre_compile() to search forwards in
368 the pattern for the relevant set of parentheses. This search went wrong
369 when there were unescaped parentheses in a character class, parentheses
370 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
371
372 33. "Subroutine" calls and backreferences were previously restricted to
373 referencing subpatterns earlier in the regex. This restriction has now
374 been removed.
375
376 34. Added a number of extra features that are going to be in Perl 5.10. On the
377 whole, these are just syntactic alternatives for features that PCRE had
378 previously implemented using the Python syntax or my own invention. The
379 other formats are all retained for compatibility.
380
381 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
382 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
383 also .NET compatible.
384
385 (b) A recursion or subroutine call to a named group can now be defined as
386 (?&name) as well as (?P>name).
387
388 (c) A backreference to a named group can now be defined as \k<name> or
389 \k'name' as well as (?P=name). The new forms, as well as being in Perl
390 5.10, are also .NET compatible.
391
392 (d) A conditional reference to a named group can now use the syntax
393 (?(<name>) or (?('name') as well as (?(name).
394
395 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
396 groups (named and numbered) that are never evaluated inline, but can be
397 called as "subroutines" from elsewhere. In effect, the DEFINE condition
398 is always false. There may be only one alternative in such a group.
399
400 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
401 as the simple (?(R). The condition is true only if the most recent
402 recursion is that of the given number or name. It does not search out
403 through the entire recursion stack.
404
405 (g) The escape \gN or \g{N} has been added, where N is a positive or
406 negative number, specifying an absolute or relative reference.
407
408 35. Tidied to get rid of some further signed/unsigned compiler warnings and
409 some "unreachable code" warnings.
410
411 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
412 things, this adds five new scripts.
413
414 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
415 There were also incompatibilities regarding the handling of \Q..\E inside
416 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
417 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
418
419 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
420 matches an empty string, and forcibly breaks the loop. There were bugs in
421 this code in non-simple cases. For a pattern such as ^(a()*)* matched
422 against aaaa the result was just "a" rather than "aaaa", for example. Two
423 separate and independent bugs (that affected different cases) have been
424 fixed.
425
426 39. Refactored the code to abolish the use of different opcodes for small
427 capturing bracket numbers. This is a tidy that I avoided doing when I
428 removed the limit on the number of capturing brackets for 3.5 back in 2001.
429 The new approach is not only tidier, it makes it possible to reduce the
430 memory needed to fix the previous bug (38).
431
432 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
433 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
434 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
435 mode.
436
437 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
438 report.
439
440 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
441 copy construction and assignment in the C++ wrapper.
442
443 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
444 couple of bugs that could have given wrong results in the "--newline=crlf"
445 case.
446
447 44. Added a number of casts and did some reorganization of signed/unsigned int
448 variables following suggestions from Dair Grant. Also renamed the variable
449 "this" as "item" because it is a C++ keyword.
450
451 45. Arranged for dftables to add
452
453 #include "pcre_internal.h"
454
455 to pcre_chartables.c because without it, gcc 4.x may remove the array
456 definition from the final binary if PCRE is built into a static library and
457 dead code stripping is activated.
458
459 46. For an unanchored pattern, if a match attempt fails at the start of a
460 newline sequence, and the newline setting is CRLF or ANY, and the next two
461 characters are CRLF, advance by two characters instead of one.
462
463
464 Version 6.7 04-Jul-06
465 ---------------------
466
467 1. In order to handle tests when input lines are enormously long, pcretest has
468 been re-factored so that it automatically extends its buffers when
469 necessary. The code is crude, but this _is_ just a test program. The
470 default size has been increased from 32K to 50K.
471
472 2. The code in pcre_study() was using the value of the re argument before
473 testing it for NULL. (Of course, in any sensible call of the function, it
474 won't be NULL.)
475
476 3. The memmove() emulation function in pcre_internal.h, which is used on
477 systems that lack both memmove() and bcopy() - that is, hardly ever -
478 was missing a "static" storage class specifier.
479
480 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
481 containing an extended class (one that cannot be represented by a bitmap
482 because it contains high-valued characters or Unicode property items, e.g.
483 [\pZ]). Almost always one would set UTF-8 mode when processing such a
484 pattern, but PCRE should not loop if you do not (it no longer does).
485 [Detail: two cases were found: (a) a repeated subpattern containing an
486 extended class; (b) a recursive reference to a subpattern that followed a
487 previous extended class. It wasn't skipping over the extended class
488 correctly when UTF-8 mode was not set.]
489
490 5. A negated single-character class was not being recognized as fixed-length
491 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
492 compile error "lookbehind assertion is not fixed length".
493
494 6. The RunPerlTest auxiliary script was showing an unexpected difference
495 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
496 write a Perl script that can interpret lines of an input file either as
497 byte characters or as UTF-8, which is what "perltest" was being required to
498 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
499 can't do is switch easily at run time between having the "use utf8;" pragma
500 or not. In the end, I fudged it by using the RunPerlTest script to insert
501 "use utf8;" explicitly for the UTF-8 tests.
502
503 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
504 the end of the subject string, contrary to the documentation and to what
505 Perl does. This was true of both matching functions. Now it matches only at
506 the start of the subject and immediately after *internal* newlines.
507
508 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
509 a pointer to an int instead of a pointer to an unsigned long int. This
510 caused problems on 64-bit systems.
511
512 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
513 instance of the 'standard' template library not being so standard".
514
515 10. There was no check on the number of named subpatterns nor the maximum
516 length of a subpattern name. The product of these values is used to compute
517 the size of the memory block for a compiled pattern. By supplying a very
518 long subpattern name and a large number of named subpatterns, the size
519 computation could be caused to overflow. This is now prevented by limiting
520 the length of names to 32 characters, and the number of named subpatterns
521 to 10,000.
522
523 11. Subpatterns that are repeated with specific counts have to be replicated in
524 the compiled pattern. The size of memory for this was computed from the
525 length of the subpattern and the repeat count. The latter is limited to
526 65535, but there was no limit on the former, meaning that integer overflow
527 could in principle occur. The compiled length of a repeated subpattern is
528 now limited to 30,000 bytes in order to prevent this.
529
530 12. Added the optional facility to have named substrings with the same name.
531
532 13. Added the ability to use a named substring as a condition, using the
533 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
534 are numbers (not recommended). Forward references are permitted.
535
536 14. Added forward references in named backreferences (if you see what I mean).
537
538 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
539 pattern could run off the end of the subject. For example, the pattern
540 "(?s)(.{1,5})"8 did this with the subject "ab".
541
542 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
543 PCRE_CASELESS was set when matching characters that were quantified with ?
544 or *.
545
546 17. A character class other than a single negated character that had a minimum
547 but no maximum quantifier - for example [ab]{6,} - was not handled
548 correctly by pce_dfa_exec(). It would match only one character.
549
550 18. A valid (though odd) pattern that looked like a POSIX character
551 class but used an invalid character after [ (for example [[,abc,]]) caused
552 pcre_compile() to give the error "Failed: internal error: code overflow" or
553 in some cases to crash with a glibc free() error. This could even happen if
554 the pattern terminated after [[ but there just happened to be a sequence of
555 letters, a binary zero, and a closing ] in the memory that followed.
556
557 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
558 over the years. Originally (before any Unicode support), just the bottom 8
559 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
560 output from "man perlunicode" includes this:
561
562 The regular expression compiler produces polymorphic opcodes. That
563 is, the pattern adapts to the data and automatically switches to
564 the Unicode character scheme when presented with Unicode data--or
565 instead uses a traditional byte scheme when presented with byte
566 data.
567
568 Sadly, a wide octal escape does not cause a switch, and in a string with
569 no other multibyte characters, these octal escapes are treated as before.
570 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
571 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
572 Unicode string.
573
574 I have not perpetrated such confusion in PCRE. Up till now, it took just
575 the bottom 8 bits, as in old Perl. I have now made octal escapes with
576 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
577 translate to the appropriate multibyte character.
578
579 29. Applied some refactoring to reduce the number of warnings from Microsoft
580 and Borland compilers. This has included removing the fudge introduced
581 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
582 a warning about an unused variable.
583
584 21. PCRE has not included VT (character 0x0b) in the set of whitespace
585 characters since release 4.0, because Perl (from release 5.004) does not.
586 [Or at least, is documented not to: some releases seem to be in conflict
587 with the documentation.] However, when a pattern was studied with
588 pcre_study() and all its branches started with \s, PCRE still included VT
589 as a possible starting character. Of course, this did no harm; it just
590 caused an unnecessary match attempt.
591
592 22. Removed a now-redundant internal flag bit that recorded the fact that case
593 dependency changed within the pattern. This was once needed for "required
594 byte" processing, but is no longer used. This recovers a now-scarce options
595 bit. Also moved the least significant internal flag bit to the most-
596 significant bit of the word, which was not previously used (hangover from
597 the days when it was an int rather than a uint) to free up another bit for
598 the future.
599
600 23. Added support for CRLF line endings as well as CR and LF. As well as the
601 default being selectable at build time, it can now be changed at runtime
602 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
603 specify that it is scanning data with non-default line endings.
604
605 24. Changed the definition of CXXLINK to make it agree with the definition of
606 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
607
608 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
609 recursions. This makes a big different to stack usage for some patterns.
610
611 26. If a subpattern containing a named recursion or subroutine reference such
612 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
613 the space required for the compiled pattern went wrong and gave too small a
614 value. Depending on the environment, this could lead to "Failed: internal
615 error: code overflow at offset 49" or "glibc detected double free or
616 corruption" errors.
617
618 27. Applied patches from Google (a) to support the new newline modes and (b) to
619 advance over multibyte UTF-8 characters in GlobalReplace.
620
621 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
622 difference for some implementation of PCRE in some Windows version.
623
624 29. Added some extra testing facilities to pcretest:
625
626 \q<number> in a data line sets the "match limit" value
627 \Q<number> in a data line sets the "match recursion limt" value
628 -S <number> sets the stack size, where <number> is in megabytes
629
630 The -S option isn't available for Windows.
631
632
633 Version 6.6 06-Feb-06
634 ---------------------
635
636 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
637 in pcreposix.h. I have copied the definition from pcre.h.
638
639 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
640 because pcre.h is no longer a built file.
641
642 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
643 not normally included in the compiled code.
644
645
646 Version 6.5 01-Feb-06
647 ---------------------
648
649 1. When using the partial match feature with pcre_dfa_exec(), it was not
650 anchoring the second and subsequent partial matches at the new starting
651 point. This could lead to incorrect results. For example, with the pattern
652 /1234/, partially matching against "123" and then "a4" gave a match.
653
654 2. Changes to pcregrep:
655
656 (a) All non-match returns from pcre_exec() were being treated as failures
657 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
658 error message is output. Some extra information is given for the
659 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
660 probably the only errors that are likely to be caused by users (by
661 specifying a regex that has nested indefinite repeats, for instance).
662 If there are more than 20 of these errors, pcregrep is abandoned.
663
664 (b) A binary zero was treated as data while matching, but terminated the
665 output line if it was written out. This has been fixed: binary zeroes
666 are now no different to any other data bytes.
667
668 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
669 used to set a locale for matching. The --locale=xxxx long option has
670 been added (no short equivalent) to specify a locale explicitly on the
671 pcregrep command, overriding the environment variables.
672
673 (d) When -B was used with -n, some line numbers in the output were one less
674 than they should have been.
675
676 (e) Added the -o (--only-matching) option.
677
678 (f) If -A or -C was used with -c (count only), some lines of context were
679 accidentally printed for the final match.
680
681 (g) Added the -H (--with-filename) option.
682
683 (h) The combination of options -rh failed to suppress file names for files
684 that were found from directory arguments.
685
686 (i) Added the -D (--devices) and -d (--directories) options.
687
688 (j) Added the -F (--fixed-strings) option.
689
690 (k) Allow "-" to be used as a file name for -f as well as for a data file.
691
692 (l) Added the --colo(u)r option.
693
694 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
695 is not present by default.
696
697 3. A nasty bug was discovered in the handling of recursive patterns, that is,
698 items such as (?R) or (?1), when the recursion could match a number of
699 alternatives. If it matched one of the alternatives, but subsequently,
700 outside the recursion, there was a failure, the code tried to back up into
701 the recursion. However, because of the way PCRE is implemented, this is not
702 possible, and the result was an incorrect result from the match.
703
704 In order to prevent this happening, the specification of recursion has
705 been changed so that all such subpatterns are automatically treated as
706 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
707
708 4. I had overlooked the fact that, in some locales, there are characters for
709 which isalpha() is true but neither isupper() nor islower() are true. In
710 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
711 and ordfeminine) are like this. This affected the treatment of \w and \W
712 when they appeared in character classes, but not when they appeared outside
713 a character class. The bit map for "word" characters is now created
714 separately from the results of isalnum() instead of just taking it from the
715 upper, lower, and digit maps. (Plus the underscore character, of course.)
716
717 5. The above bug also affected the handling of POSIX character classes such as
718 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
719 permanent tables. Instead, the bit maps for such a class were previously
720 created as the appropriate unions of the upper, lower, and digit bitmaps.
721 Now they are created by subtraction from the [[:word:]] class, which has
722 its own bitmap.
723
724 6. The [[:blank:]] character class matches horizontal, but not vertical space.
725 It is created by subtracting the vertical space characters (\x09, \x0a,
726 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
727 subtraction was done in the overall bitmap for a character class, meaning
728 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
729 be recognized. This bug has been fixed.
730
731 7. Patches from the folks at Google:
732
733 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
734 real life, but is still worth protecting against".
735
736 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
737 regular expressions".
738
739 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
740 have it.
741
742 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
743 "configure" and the latter not, in order to fix a problem somebody had
744 with compiling the Arg class on HP-UX.
745
746 (e) Improve the error-handling of the C++ wrapper a little bit.
747
748 (f) New tests for checking recursion limiting.
749
750 8. The pcre_memmove() function, which is used only if the environment does not
751 have a standard memmove() function (and is therefore rarely compiled),
752 contained two bugs: (a) use of int instead of size_t, and (b) it was not
753 returning a result (though PCRE never actually uses the result).
754
755 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
756 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
757 returned instead of calling malloc() with an overflowing number that would
758 most likely cause subsequent chaos.
759
760 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
761
762 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
763 with this option is matched, the nmatch and pmatch options of regexec() are
764 ignored.
765
766 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
767 provided in case anyone wants to the the POSIX interface with UTF-8
768 strings.
769
770 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
771 C++ linking (needed for some HP-UX environments).
772
773 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
774 (unused parameter) and in the pcre_printint() function (omitted "default"
775 switch label when the default is to do nothing).
776
777 15. Added some code to make it possible, when PCRE is compiled as a C++
778 library, to replace subject pointers for pcre_exec() with a smart pointer
779 class, thus making it possible to process discontinuous strings.
780
781 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
782 much the same function. They were added by different people who were trying
783 to make PCRE easy to compile on non-Unix systems. It has been suggested
784 that PCRE_EXPORT be abolished now that there is more automatic apparatus
785 for compiling on Windows systems. I have therefore replaced it with
786 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
787 defaults to "extern" for C or "extern C" for C++, which works fine on
788 Unix-like systems. It is now possible to override the value of PCRE_DATA_
789 SCOPE with something explicit in config.h. In addition:
790
791 (a) pcreposix.h still had just "extern" instead of either of these macros;
792 I have replaced it with PCRE_DATA_SCOPE.
793
794 (b) Functions such as _pcre_xclass(), which are internal to the library,
795 but external in the C sense, all had PCRE_EXPORT in their definitions.
796 This is apparently wrong for the Windows case, so I have removed it.
797 (It makes no difference on Unix-like systems.)
798
799 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
800 of recursive calls to match(). This is different to MATCH_LIMIT because
801 that limits the total number of calls to match(), not all of which increase
802 the depth of recursion. Limiting the recursion depth limits the amount of
803 stack (or heap if NO_RECURSE is set) that is used. The default can be set
804 when PCRE is compiled, and changed at run time. A patch from Google adds
805 this functionality to the C++ interface.
806
807 18. Changes to the handling of Unicode character properties:
808
809 (a) Updated the table to Unicode 4.1.0.
810
811 (b) Recognize characters that are not in the table as "Cn" (undefined).
812
813 (c) I revised the way the table is implemented to a much improved format
814 which includes recognition of ranges. It now supports the ranges that
815 are defined in UnicodeData.txt, and it also amalgamates other
816 characters into ranges. This has reduced the number of entries in the
817 table from around 16,000 to around 3,000, thus reducing its size
818 considerably. I realized I did not need to use a tree structure after
819 all - a binary chop search is just as efficient. Having reduced the
820 number of entries, I extended their size from 6 bytes to 8 bytes to
821 allow for more data.
822
823 (d) Added support for Unicode script names via properties such as \p{Han}.
824
825 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
826 matching that character.
827
828 20. When matching a repeated Unicode property with a minimum greater than zero,
829 (for example \pL{2,}), PCRE could look past the end of the subject if it
830 reached it while seeking the minimum number of characters. This could
831 happen only if some of the characters were more than one byte long, because
832 there is a check for at least the minimum number of bytes.
833
834 21. Refactored the implementation of \p and \P so as to be more general, to
835 allow for more different types of property in future. This has changed the
836 compiled form incompatibly. Anybody with saved compiled patterns that use
837 \p or \P will have to recompile them.
838
839 22. Added "Any" and "L&" to the supported property types.
840
841 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
842 but give a compile time error if the value is greater than 0xff.
843
844 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
845 accidentally not being installed or uninstalled.
846
847 25. The pcre.h file was built from pcre.h.in, but the only changes that were
848 made were to insert the current release number. This seemed silly, because
849 it made things harder for people building PCRE on systems that don't run
850 "configure". I have turned pcre.h into a distributed file, no longer built
851 by "configure", with the version identification directly included. There is
852 no longer a pcre.h.in file.
853
854 However, this change necessitated a change to the pcre-config script as
855 well. It is built from pcre-config.in, and one of the substitutions was the
856 release number. I have updated configure.ac so that ./configure now finds
857 the release number by grepping pcre.h.
858
859 26. Added the ability to run the tests under valgrind.
860
861
862 Version 6.4 05-Sep-05
863 ---------------------
864
865 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
866 "--" to be printed when multiple files were scanned, even when none of the
867 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
868 consider it to be a bug, and have restored the previous behaviour.
869
870 2. A couple of code tidies to get rid of compiler warnings.
871
872 3. The pcretest program used to cheat by referring to symbols in the library
873 whose names begin with _pcre_. These are internal symbols that are not
874 really supposed to be visible externally, and in some environments it is
875 possible to suppress them. The cheating is now confined to including
876 certain files from the library's source, which is a bit cleaner.
877
878 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
879 file's purpose clearer.
880
881 5. Reorganized pcre_ucp_findchar().
882
883
884 Version 6.3 15-Aug-05
885 ---------------------
886
887 1. The file libpcre.pc.in did not have general read permission in the tarball.
888
889 2. There were some problems when building without C++ support:
890
891 (a) If C++ support was not built, "make install" and "make test" still
892 tried to test it.
893
894 (b) There were problems when the value of CXX was explicitly set. Some
895 changes have been made to try to fix these, and ...
896
897 (c) --disable-cpp can now be used to explicitly disable C++ support.
898
899 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
900 backslash in a target when C++ was disabled. This confuses some
901 versions of "make", apparently. Using an intermediate variable solves
902 this. (Same for CPP_LOBJ.)
903
904 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
905 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
906 necessary on certain architectures.
907
908 4. Added a setting of -export-symbols-regex to the link command to remove
909 those symbols that are exported in the C sense, but actually are local
910 within the library, and not documented. Their names all begin with
911 "_pcre_". This is not a perfect job, because (a) we have to except some
912 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
913 available (and never for static libraries). I have made a note to try to
914 find a way round (a) in the future.
915
916
917 Version 6.2 01-Aug-05
918 ---------------------
919
920 1. There was no test for integer overflow of quantifier values. A construction
921 such as {1111111111111111} would give undefined results. What is worse, if
922 a minimum quantifier for a parenthesized subpattern overflowed and became
923 negative, the calculation of the memory size went wrong. This could have
924 led to memory overwriting.
925
926 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
927
928 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
929 operating environments where this matters.
930
931 4. Applied Giuseppe Maxia's patch to add additional features for controlling
932 PCRE options from within the C++ wrapper.
933
934 5. Named capturing subpatterns were not being correctly counted when a pattern
935 was compiled. This caused two problems: (a) If there were more than 100
936 such subpatterns, the calculation of the memory needed for the whole
937 compiled pattern went wrong, leading to an overflow error. (b) Numerical
938 back references of the form \12, where the number was greater than 9, were
939 not recognized as back references, even though there were sufficient
940 previous subpatterns.
941
942 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
943 versions of gcc, e.g. 2.95.4.
944
945
946 Version 6.1 21-Jun-05
947 ---------------------
948
949 1. There was one reference to the variable "posix" in pcretest.c that was not
950 surrounded by "#if !defined NOPOSIX".
951
952 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
953 the cross-check on the old pcre_info() function, for the benefit of the
954 cut-down version of PCRE that is currently imported into Exim.
955
956 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
957 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
958 patterns that start (?i) but I don't think that matters. The use of (?i) is
959 just an example; this all applies to the other options as well.
960
961 4. Since libtool seems to echo the compile commands it is issuing, the output
962 from "make" can be reduced a bit by putting "@" in front of each libtool
963 compile command.
964
965 5. Patch from the folks at Google for configure.in to be a bit more thorough
966 in checking for a suitable C++ installation before trying to compile the
967 C++ stuff. This should fix a reported problem when a compiler was present,
968 but no suitable headers.
969
970 6. The man pages all had just "PCRE" as their title. I have changed them to
971 be the relevant file name. I have also arranged that these names are
972 retained in the file doc/pcre.txt, which is a concatenation in text format
973 of all the man pages except the little individual ones for each function.
974
975 7. The NON-UNIX-USE file had not been updated for the different set of source
976 files that come with release 6. I also added a few comments about the C++
977 wrapper.
978
979
980 Version 6.0 07-Jun-05
981 ---------------------
982
983 1. Some minor internal re-organization to help with my DFA experiments.
984
985 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
986 didn't matter for the library itself when fully configured, but did matter
987 when compiling without UCP support, or within Exim, where the ucp files are
988 not imported.
989
990 3. Refactoring of the library code to split up the various functions into
991 different source modules. The addition of the new DFA matching code (see
992 below) to a single monolithic source would have made it really too
993 unwieldy, quite apart from causing all the code to be include in a
994 statically linked application, when only some functions are used. This is
995 relevant even without the DFA addition now that patterns can be compiled in
996 one application and matched in another.
997
998 The downside of splitting up is that there have to be some external
999 functions and data tables that are used internally in different modules of
1000 the library but which are not part of the API. These have all had their
1001 names changed to start with "_pcre_" so that they are unlikely to clash
1002 with other external names.
1003
1004 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1005 a different (DFA) algorithm. Although it is slower than the original
1006 function, it does have some advantages for certain types of matching
1007 problem.
1008
1009 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1010 including restarting after a partial match.
1011
1012 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1013 defined when compiling for Windows was sent to me. I have put it into the
1014 code, though I have no means of testing or verifying it.
1015
1016 7. Added the pcre_refcount() auxiliary function.
1017
1018 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1019 match before or at the first newline in the subject string. In pcretest,
1020 the /f option on a pattern can be used to set this.
1021
1022 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1023 would behave wrongly. This has been present in PCRE since release 4.0.
1024
1025 10. A number of changes to the pcregrep command:
1026
1027 (a) Refactored how -x works; insert ^(...)$ instead of setting
1028 PCRE_ANCHORED and checking the length, in preparation for adding
1029 something similar for -w.
1030
1031 (b) Added the -w (match as a word) option.
1032
1033 (c) Refactored the way lines are read and buffered so as to have more
1034 than one at a time available.
1035
1036 (d) Implemented a pcregrep test script.
1037
1038 (e) Added the -M (multiline match) option. This allows patterns to match
1039 over several lines of the subject. The buffering ensures that at least
1040 8K, or the rest of the document (whichever is the shorter) is available
1041 for matching (and similarly the previous 8K for lookbehind assertions).
1042
1043 (f) Changed the --help output so that it now says
1044
1045 -w, --word-regex(p)
1046
1047 instead of two lines, one with "regex" and the other with "regexp"
1048 because that confused at least one person since the short forms are the
1049 same. (This required a bit of code, as the output is generated
1050 automatically from a table. It wasn't just a text change.)
1051
1052 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1053 option but starts with a hyphen. Could be a pattern or a path name
1054 starting with a hyphen, for instance.
1055
1056 (h) "-" can be given as a file name to represent stdin.
1057
1058 (i) When file names are being printed, "(standard input)" is used for
1059 the standard input, for compatibility with GNU grep. Previously
1060 "<stdin>" was used.
1061
1062 (j) The option --label=xxx can be used to supply a name to be used for
1063 stdin when file names are being printed. There is no short form.
1064
1065 (k) Re-factored the options decoding logic because we are going to add
1066 two more options that take data. Such options can now be given in four
1067 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1068
1069 (l) Added the -A, -B, and -C options for requesting that lines of context
1070 around matches be printed.
1071
1072 (m) Added the -L option to print the names of files that do not contain
1073 any matching lines, that is, the complement of -l.
1074
1075 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1076 continue to scan other files.
1077
1078 (o) The -s option was incorrectly implemented. For compatibility with other
1079 greps, it now suppresses the error message for a non-existent or non-
1080 accessible file (but not the return code). There is a new option called
1081 -q that suppresses the output of matching lines, which was what -s was
1082 previously doing.
1083
1084 (p) Added --include and --exclude options to specify files for inclusion
1085 and exclusion when recursing.
1086
1087 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1088 Hopefully, it now does.
1089
1090 12. Missing cast in pcre_study().
1091
1092 13. Added an "uninstall" target to the makefile.
1093
1094 14. Replaced "extern" in the function prototypes in Makefile.in with
1095 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1096 world, but is set differently for Windows.
1097
1098 15. Added a second compiling function called pcre_compile2(). The only
1099 difference is that it has an extra argument, which is a pointer to an
1100 integer error code. When there is a compile-time failure, this is set
1101 non-zero, in addition to the error test pointer being set to point to an
1102 error message. The new argument may be NULL if no error number is required
1103 (but then you may as well call pcre_compile(), which is now just a
1104 wrapper). This facility is provided because some applications need a
1105 numeric error indication, but it has also enabled me to tidy up the way
1106 compile-time errors are handled in the POSIX wrapper.
1107
1108 16. Added VPATH=.libs to the makefile; this should help when building with one
1109 prefix path and installing with another. (Or so I'm told by someone who
1110 knows more about this stuff than I do.)
1111
1112 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1113 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1114 match everything, including newlines. This is not POSIX-compatible, but
1115 somebody wanted the feature. From pcretest it can be activated by using
1116 both the P and the s flags.
1117
1118 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1119
1120 19. libpcre.pc was being incorrectly installed as executable.
1121
1122 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1123 it now also looks for '\r' so that it will work unmodified on Windows.
1124
1125 21. Added Google's contributed C++ wrapper to the distribution.
1126
1127 22. Added some untidy missing memory free() calls in pcretest, to keep
1128 Electric Fence happy when testing.
1129
1130
1131
1132 Version 5.0 13-Sep-04
1133 ---------------------
1134
1135 1. Internal change: literal characters are no longer packed up into items
1136 containing multiple characters in a single byte-string. Each character
1137 is now matched using a separate opcode. However, there may be more than one
1138 byte in the character in UTF-8 mode.
1139
1140 2. The pcre_callout_block structure has two new fields: pattern_position and
1141 next_item_length. These contain the offset in the pattern to the next match
1142 item, and its length, respectively.
1143
1144 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1145 insertion of callouts before each pattern item. Added the /C option to
1146 pcretest to make use of this.
1147
1148 4. On the advice of a Windows user, the lines
1149
1150 #if defined(_WIN32) || defined(WIN32)
1151 _setmode( _fileno( stdout ), 0x8000 );
1152 #endif /* defined(_WIN32) || defined(WIN32) */
1153
1154 have been added to the source of pcretest. This apparently does useful
1155 magic in relation to line terminators.
1156
1157 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1158 for the benefit of those environments where the "b" makes a difference.
1159
1160 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1161 to know about it. I have put a hack into configure.in that adds in code
1162 to set GCC=yes if CC=icc. This seems to end up at a point in the
1163 generated configure script that is early enough to affect the setting of
1164 compiler options, which is what is needed, but I have no means of testing
1165 whether it really works. (The user who reported this had patched the
1166 generated configure script, which of course I cannot do.)
1167
1168 LATER: After change 22 below (new libtool files), the configure script
1169 seems to know about icc (and also ecc). Therefore, I have commented out
1170 this hack in configure.in.
1171
1172 7. Added support for pkg-config (2 patches were sent in).
1173
1174 8. Negated POSIX character classes that used a combination of internal tables
1175 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1176 [[:^ascii]]. Typically, they would match almost any characters. The other
1177 POSIX classes were not broken in this way.
1178
1179 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1180 to find the match, as PCRE was deluded into thinking that the match had to
1181 start at the start point or following a newline. The same bug applied to
1182 patterns with negative forward assertions or any backward assertions
1183 preceding ".*" at the start, unless the pattern required a fixed first
1184 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1185
1186 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1187 starting at the last subject character, bytes beyond the end of the subject
1188 string were read.
1189
1190 11. Renamed the variable "class" as "classbits" to make life easier for C++
1191 users. (Previously there was a macro definition, but it apparently wasn't
1192 enough.)
1193
1194 12. Added the new field "tables" to the extra data so that tables can be passed
1195 in at exec time, or the internal tables can be re-selected. This allows
1196 a compiled regex to be saved and re-used at a later time by a different
1197 program that might have everything at different addresses.
1198
1199 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1200 -R library as well as a -L library.
1201
1202 14. The debugging options of pcretest (-d on the command line or D on a
1203 pattern) showed incorrect output for anything following an extended class
1204 that contained multibyte characters and which was followed by a quantifier.
1205
1206 15. Added optional support for general category Unicode character properties
1207 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1208 support. It adds about 90K to the size of the library. The meanings of the
1209 inbuilt class escapes such as \d and \s have NOT been changed.
1210
1211 16. Updated pcredemo.c to include calls to free() to release the memory for the
1212 compiled pattern.
1213
1214 17. The generated file chartables.c was being created in the source directory
1215 instead of in the building directory. This caused the build to fail if the
1216 source directory was different from the building directory, and was
1217 read-only.
1218
1219 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1220 file. No doubt somebody will tell me if they don't make sense... Also added
1221 Dan Mooney's comments about building on OpenVMS.
1222
1223 19. Added support for partial matching via the PCRE_PARTIAL option for
1224 pcre_exec() and the \P data escape in pcretest.
1225
1226 20. Extended pcretest with 3 new pattern features:
1227
1228 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1229 write the compiled pattern to the file whose name is "rest-of-line".
1230 This is a straight binary dump of the data, with the saved pointer to
1231 the character tables forced to be NULL. The study data, if any, is
1232 written too. After writing, pcretest reads a new pattern.
1233
1234 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1235 compiled pattern from the given file. There must not be any
1236 occurrences of "<" in the file name (pretty unlikely); if there are,
1237 pcretest will instead treat the initial "<" as a pattern delimiter.
1238 After reading in the pattern, pcretest goes on to read data lines as
1239 usual.
1240
1241 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1242 and 16-bit fields in a compiled pattern, to simulate a pattern that
1243 was compiled on a host of opposite endianness.
1244
1245 21. The pcre-exec() function can now cope with patterns that were compiled on
1246 hosts of opposite endianness, with this restriction:
1247
1248 As for any compiled expression that is saved and used later, the tables
1249 pointer field cannot be preserved; the extra_data field in the arguments
1250 to pcre_exec() should be used to pass in a tables address if a value
1251 other than the default internal tables were used at compile time.
1252
1253 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1254 now diagnosed as an error. Previously, most of the time, a negative number
1255 would have been treated as zero, but if in addition "ovector" was passed as
1256 NULL, a crash could occur.
1257
1258 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1259 new versions from the libtool 1.5 distribution (the last one is a copy of
1260 a file called libtool.m4). This seems to have fixed the need to patch
1261 "configure" to support Darwin 1.3 (which I used to do). However, I still
1262 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1263 workstation).
1264
1265 24. Changed the PCRE licence to be the more standard "BSD" licence.
1266
1267
1268 Version 4.5 01-Dec-03
1269 ---------------------
1270
1271 1. There has been some re-arrangement of the code for the match() function so
1272 that it can be compiled in a version that does not call itself recursively.
1273 Instead, it keeps those local variables that need separate instances for
1274 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1275 needs to "recurse". Keeping track of where control must go is done by means
1276 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1277 hide most of the details from the main code, and operates only if
1278 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1279 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1280 operating.
1281
1282 To make it easier for callers to provide specially tailored get/free
1283 functions for this usage, two new functions, pcre_stack_malloc, and
1284 pcre_stack_free, are used. They are always called in strict stacking order,
1285 and the size of block requested is always the same.
1286
1287 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1288 PCRE has been compiled to use the stack or the heap for recursion. The
1289 -C option of pcretest uses this to show which version is compiled.
1290
1291 A new data escape \S, is added to pcretest; it causes the amounts of store
1292 obtained and freed by both kinds of malloc/free at match time to be added
1293 to the output.
1294
1295 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1296 what's available on my current Linux desktop machine.
1297
1298 3. When matching a UTF-8 string, the test for a valid string at the start has
1299 been extended. If start_offset is not zero, PCRE now checks that it points
1300 to a byte that is the start of a UTF-8 character. If not, it returns
1301 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1302 this is necessary because there may be backward assertions in the pattern.
1303 When matching the same subject several times, it may save resources to use
1304 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1305
1306 4. The code for checking the validity of UTF-8 strings has been tightened so
1307 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1308 containing "overlong sequences".
1309
1310 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1311 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1312 should have been "&", but it just so happened that all the cases this let
1313 through by mistake were picked up later in the function.
1314
1315 6. I had used a variable called "isblank" - this is a C99 function, causing
1316 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1317
1318 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1319 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1320 so the version is shown; (c) stop "make test" echoing "./RunTest".
1321
1322 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1323
1324 9. The prototype for memmove() for systems that don't have it was using
1325 size_t, but the inclusion of the header that defines size_t was later. I've
1326 moved the #includes for the C headers earlier to avoid this.
1327
1328 10. Added some adjustments to the code to make it easier to compiler on certain
1329 special systems:
1330
1331 (a) Some "const" qualifiers were missing.
1332 (b) Added the macro EXPORT before all exported functions; by default this
1333 is defined to be empty.
1334 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1335 that it reads its output file name as an argument instead of writing
1336 to the standard output and assuming this can be redirected.
1337
1338 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1339 class containing characters with values greater than 255, PCRE compilation
1340 went into a loop.
1341
1342 12. A recursive reference to a subpattern that was within another subpattern
1343 that had a minimum quantifier of zero caused PCRE to crash. For example,
1344 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1345 recursion. If the recursively-called subpattern itself had a zero repeat,
1346 that was OK.
1347
1348 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1349 buffer into which it was copied (for escape processing) was still set at
1350 1024, so long lines caused crashes.
1351
1352 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1353 "internal error: code overflow...". This applied to any character class
1354 that was followed by a possessive quantifier.
1355
1356 15. Modified the Makefile to add libpcre.la as a prerequisite for
1357 libpcreposix.la because I was told this is needed for a parallel build to
1358 work.
1359
1360 16. If a pattern that contained .* following optional items at the start was
1361 studied, the wrong optimizing data was generated, leading to matching
1362 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1363 matching string must start with a or b or c. The correct conclusion for
1364 this pattern is that a match can start with any character.
1365
1366
1367 Version 4.4 13-Aug-03
1368 ---------------------
1369
1370 1. In UTF-8 mode, a character class containing characters with values between
1371 127 and 255 was not handled correctly if the compiled pattern was studied.
1372 In fixing this, I have also improved the studying algorithm for such
1373 classes (slightly).
1374
1375 2. Three internal functions had redundant arguments passed to them. Removal
1376 might give a very teeny performance improvement.
1377
1378 3. Documentation bug: the value of the capture_top field in a callout is *one
1379 more than* the number of the hightest numbered captured substring.
1380
1381 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1382 in incorrectly linking with a previously installed version. They now link
1383 explicitly with libpcre.la.
1384
1385 5. configure.in no longer needs to recognize Cygwin specially.
1386
1387 6. A problem in pcre.in for Windows platforms is fixed.
1388
1389 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1390 pcretest, it used to include the size of the study block as part of its
1391 output. Unfortunately, the structure contains a field that has a different
1392 size on different hardware architectures. This meant that the tests that
1393 showed this size failed. As the block is currently always of a fixed size,
1394 this information isn't actually particularly useful in pcretest output, so
1395 I have just removed it.
1396
1397 8. Three pre-processor statements accidentally did not start in column 1.
1398 Sadly, there are *still* compilers around that complain, even though
1399 standard C has not required this for well over a decade. Sigh.
1400
1401 9. In pcretest, the code for checking callouts passed small integers in the
1402 callout_data field, which is a void * field. However, some picky compilers
1403 complained about the casts involved for this on 64-bit systems. Now
1404 pcretest passes the address of the small integer instead, which should get
1405 rid of the warnings.
1406
1407 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1408 both compile and run time, and gives an error if an invalid UTF-8 sequence
1409 is found. There is a option for disabling this check in cases where the
1410 string is known to be correct and/or the maximum performance is wanted.
1411
1412 11. In response to a bug report, I changed one line in Makefile.in from
1413
1414 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1415 to
1416 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1417
1418 to look similar to other lines, but I have no way of telling whether this
1419 is the right thing to do, as I do not use Windows. No doubt I'll get told
1420 if it's wrong...
1421
1422
1423 Version 4.3 21-May-03
1424 ---------------------
1425
1426 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1427 Makefile.
1428
1429 2. Some refactoring to improve the quality of the code:
1430
1431 (i) The utf8_table... variables are now declared "const".
1432
1433 (ii) The code for \cx, which used the "case flipping" table to upper case
1434 lower case letters, now just substracts 32. This is ASCII-specific,
1435 but the whole concept of \cx is ASCII-specific, so it seems
1436 reasonable.
1437
1438 (iii) PCRE was using its character types table to recognize decimal and
1439 hexadecimal digits in the pattern. This is silly, because it handles
1440 only 0-9, a-f, and A-F, but the character types table is locale-
1441 specific, which means strange things might happen. A private
1442 table is now used for this - though it costs 256 bytes, a table is
1443 much faster than multiple explicit tests. Of course, the standard
1444 character types table is still used for matching digits in subject
1445 strings against \d.
1446
1447 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1448 ending in _t are). So I've renamed it as ESC_tee.
1449
1450 3. The first argument for regexec() in the POSIX wrapper should have been
1451 defined as "const".
1452
1453 4. Changed pcretest to use malloc() for its buffers so that they can be
1454 Electric Fenced for debugging.
1455
1456 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1457 to read one or more bytes before the start of the subject string. Often this
1458 had no effect on PCRE's behaviour, but in some circumstances it could
1459 provoke a segmentation fault.
1460
1461 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1462 to try to read one or more bytes before the start of the subject string.
1463
1464 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1465 UTF-8 support could misbehave in various ways if the subject string
1466 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1467 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1468 back over UTF-8 characters.)
1469
1470
1471 Version 4.2 14-Apr-03
1472 ---------------------
1473
1474 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1475
1476 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1477 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1478 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1479 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1480 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1481 and BUILD_EXEEXT
1482 Note: automatic setting of the BUILD variables is not yet working
1483 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1484 compile-time but not at link-time
1485 [LINK]: use for linking executables only
1486 make different versions for Windows and non-Windows
1487 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1488 libraries
1489 [LINK_FOR_BUILD]: new variable
1490 [OBJEXT]: use throughout
1491 [EXEEXT]: use throughout
1492 <winshared>: new target
1493 <wininstall>: new target
1494 <dftables.o>: use native compiler
1495 <dftables>: use native linker
1496 <install>: handle Windows platform correctly
1497 <clean>: ditto
1498 <check>: ditto
1499 copy DLL to top builddir before testing
1500
1501 As part of these changes, -no-undefined was removed again. This was reported
1502 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1503 in any case.
1504
1505 3. Some tidies to get rid of compiler warnings:
1506
1507 . In the match_data structure, match_limit was an unsigned long int, whereas
1508 match_call_count was an int. I've made them both unsigned long ints.
1509
1510 . In pcretest the fact that a const uschar * doesn't automatically cast to
1511 a void * provoked a warning.
1512
1513 . Turning on some more compiler warnings threw up some "shadow" variables
1514 and a few more missing casts.
1515
1516 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1517 option, a class that contained a single character with a value between 128
1518 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1519
1520 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1521 option, a class that contained several characters, but with at least one
1522 whose value was between 128 and 255 caused PCRE to crash.
1523
1524
1525 Version 4.1 12-Mar-03
1526 ---------------------
1527
1528 1. Compiling with gcc -pedantic found a couple of places where casts were
1529 needed, and a string in dftables.c that was longer than standard compilers are
1530 required to support.
1531
1532 2. Compiling with Sun's compiler found a few more places where the code could
1533 be tidied up in order to avoid warnings.
1534
1535 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1536 first of these names is deprecated in the latest Autoconf in favour of the name
1537 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1538 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1539 analogy I have changed it to CFLAGS_FOR_BUILD.
1540
1541 4. Added -no-undefined to the linking command in the Makefile, because this is
1542 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1543 linking step for the pcreposix library.
1544
1545 5. PCRE was failing to diagnose the case of two named groups with the same
1546 name.
1547
1548 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1549 literal character that is needed in the subject for a match, and scans along to
1550 ensure that it is present before embarking on the full matching process. This
1551 saves time in cases of nested unlimited repeats that are never going to match.
1552 Problem: the scan can take a lot of time if the subject is very long (e.g.
1553 megabytes), thus penalizing straightforward matches. It is now done only if the
1554 amount of subject to be scanned is less than 1000 bytes.
1555
1556 7. A lesser problem with the same optimization is that it was recording the
1557 first character of an anchored pattern as "needed", thus provoking a search
1558 right along the subject, even when the first match of the pattern was going to
1559 fail. The "needed" character is now not set for anchored patterns, unless it
1560 follows something in the pattern that is of non-fixed length. Thus, it still
1561 fulfils its original purpose of finding quick non-matches in cases of nested
1562 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1563
1564
1565 Version 4.0 17-Feb-03
1566 ---------------------
1567
1568 1. If a comment in an extended regex that started immediately after a meta-item
1569 extended to the end of string, PCRE compiled incorrect data. This could lead to
1570 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
1571
1572 2. Moved to autoconf 2.53 and libtool 1.4.2.
1573
1574 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
1575 the special perltest8 script is no longer needed - all the tests can be run
1576 from a single perltest script.
1577
1578 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
1579 by \s. It has now been removed in PCRE. This means it isn't recognized as
1580 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
1581 class [:space:] *does* include VT, thereby creating a mess.
1582
1583 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
1584 space and tab.
1585
1586 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
1587 its new features into the main test script, reducing the number of scripts.
1588
1589 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
1590 were backward compatible, and made the (?i) apply to the whole pattern, as if
1591 /i were given. Now it behaves more logically, and applies the option setting
1592 only to what follows. PCRE has been changed to follow suit. However, if it
1593 finds options settings right at the start of the pattern, it extracts them into
1594 the global options, as before. Thus, they show up in the info data.
1595
1596 8. Added support for the \Q...\E escape sequence. Characters in between are
1597 treated as literals. This is slightly different from Perl in that $ and @ are
1598 also handled as literals inside the quotes. In Perl, they will cause variable
1599 interpolation. Note the following examples:
1600
1601 Pattern PCRE matches Perl matches
1602
1603 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
1604 \Qabc\$xyz\E abc\$xyz abc\$xyz
1605 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
1606
1607 For compatibility with Perl, \Q...\E sequences are recognized inside character
1608 classes as well as outside them.
1609
1610 9. Re-organized 3 code statements in pcretest to avoid "overflow in
1611 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
1612 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
1613 signed/unsigned warnings.
1614
1615 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
1616 option for pcretest, so I've replaced it by a simple function that does just
1617 that job.
1618
1619 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
1620 "pcregrep -".
1621
1622 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
1623 Java package. This provides some syntactic sugar for simple cases of what my
1624 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
1625 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
1626 item, you can use this simplified notation. Note that only makes sense with
1627 greedy quantifiers. Consequently, the use of the possessive quantifier forces
1628 greediness, whatever the setting of the PCRE_UNGREEDY option.
1629
1630 13. A change of greediness default within a pattern was not taking effect at
1631 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
1632 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
1633 was abstracted outside.
1634
1635 14. PCRE now supports the \G assertion. It is true when the current matching
1636 position is at the start point of the match. This differs from \A when the
1637 starting offset is non-zero. Used with the /g option of pcretest (or similar
1638 code), it works in the same way as it does for Perl's /g option. If all
1639 alternatives of a regex begin with \G, the expression is anchored to the start
1640 match position, and the "anchored" flag is set in the compiled expression.
1641
1642 15. Some bugs concerning the handling of certain option changes within patterns
1643 have been fixed. These applied to options other than (?ims). For example,
1644 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
1645 been the other way round. Some of this was related to change 7 above.
1646
1647 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
1648 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
1649 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
1650 POSIX classes only within a class (e.g. /[[:alpha:]]/).
1651
1652 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
1653 mode. Unlike ".", it always matches newline, whatever the setting of
1654 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
1655 assertions. Perl allows it, but it doesn't (in general) work because it can't
1656 calculate the length of the lookbehind. At least, that's the case for Perl
1657 5.8.0 - I've been told they are going to document that it doesn't work in
1658 future.
1659
1660 18. Added an error diagnosis for escapes that PCRE does not support: these are
1661 \L, \l, \N, \P, \p, \U, \u, and \X.
1662
1663 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
1664 reading past the end of the pattern in cases such as /[abcd/.
1665
1666 20. PCRE was getting more memory than necessary for patterns with classes that
1667 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
1668
1669 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
1670 compiling PCRE for use with Virtual Pascal.
1671
1672 22. Small fix to the Makefile to make it work properly if the build is done
1673 outside the source tree.
1674
1675 23. Added a new extension: a condition to go with recursion. If a conditional
1676 subpattern starts with (?(R) the "true" branch is used if recursion has
1677 happened, whereas the "false" branch is used only at the top level.
1678
1679 24. When there was a very long string of literal characters (over 255 bytes
1680 without UTF support, over 250 bytes with UTF support), the computation of how
1681 much memory was required could be incorrect, leading to segfaults or other
1682 strange effects.
1683
1684 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
1685 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
1686 there was a subsequent back reference to those brackets. This meant that, for
1687 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
1688 possible to check for precisely this case. All we can do is abandon the
1689 optimization if .* occurs inside capturing brackets when there are any back
1690 references whatsoever. (See below for a better fix that came later.)
1691
1692 26. The handling of the optimization for finding the first character of a
1693 non-anchored pattern, and for finding a character that is required later in the
1694 match were failing in some cases. This didn't break the matching; it just
1695 failed to optimize when it could. The way this is done has been re-implemented.
1696
1697 27. Fixed typo in error message for invalid (?R item (it said "(?p").
1698
1699 28. Added a new feature that provides some of the functionality that Perl
1700 provides with (?{...}). The facility is termed a "callout". The way it is done
1701 in PCRE is for the caller to provide an optional function, by setting
1702 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
1703 global variable. By default it is unset, which disables all calling out. To get
1704 the function called, the regex must include (?C) at appropriate points. This
1705 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
1706 This provides a means of identifying different callout points. When PCRE
1707 reaches such a point in the regex, if pcre_callout has been set, the external
1708 function is called. It is provided with data in a structure called
1709 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
1710 matching continues; if it returns a non-zero value, the match at the current
1711 point fails. However, backtracking will occur if possible. [This was changed
1712 later and other features added - see item 49 below.]
1713
1714 29. pcretest is upgraded to test the callout functionality. It provides a
1715 callout function that displays information. By default, it shows the start of
1716 the match and the current position in the text. There are some new data escapes
1717 to vary what happens:
1718
1719 \C+ in addition, show current contents of captured substrings
1720 \C- do not supply a callout function
1721 \C!n return 1 when callout number n is reached
1722 \C!n!m return 1 when callout number n is reached for the mth time
1723
1724 30. If pcregrep was called with the -l option and just a single file name, it
1725 output "<stdin>" if a match was found, instead of the file name.
1726
1727 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
1728 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
1729 pcre_exec(). This saves a malloc/free per call. The default value of
1730 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
1731 when configuring.
1732
1733 32. The default maximum size of a compiled pattern is 64K. There have been a
1734 few cases of people hitting this limit. The code now uses macros to handle the
1735 storing of links as offsets within the compiled pattern. It defaults to 2-byte
1736 links, but this can be changed to 3 or 4 bytes by --with-link-size when
1737 configuring. Tests 2 and 5 work only with 2-byte links because they output
1738 debugging information about compiled patterns.
1739
1740 33. Internal code re-arrangements:
1741
1742 (a) Moved the debugging function for printing out a compiled regex into
1743 its own source file (printint.c) and used #include to pull it into
1744 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
1745 separate copies.
1746
1747 (b) Defined the list of op-code names for debugging as a macro in
1748 internal.h so that it is next to the definition of the opcodes.
1749
1750 (c) Defined a table of op-code lengths for simpler skipping along compiled
1751 code. This is again a macro in internal.h so that it is next to the
1752 definition of the opcodes.
1753
1754 34. Added support for recursive calls to individual subpatterns, along the
1755 lines of Robin Houston's patch (but implemented somewhat differently).
1756
1757 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
1758 allow it to read and process whole directories in Win32. This code was
1759 contributed by Lionel Fourquaux; it has not been tested by me.
1760
1761 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
1762 used to name a group. Names consist of alphanumerics and underscores, and must
1763 be unique. Back references use the syntax (?P=name) and recursive calls use
1764 (?P>name) which is a PCRE extension to the Python extension. Groups still have
1765 numbers. The function pcre_fullinfo() can be used after compilation to extract
1766 a name/number map. There are three relevant calls:
1767
1768 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
1769 PCRE_INFO_NAMECOUNT yields the number of entries
1770 PCRE_INFO_NAMETABLE yields a pointer to the map.
1771
1772 The map is a vector of fixed-size entries. The size of each entry depends on
1773 the length of the longest name used. The first two bytes of each entry are the
1774 group number, most significant byte first. There follows the corresponding
1775 name, zero terminated. The names are in alphabetical order.
1776
1777 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
1778 case instead of 255. Making it the same both with and without UTF-8 support
1779 means that the same test output works with both.
1780
1781 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
1782 calling malloc() with a zero argument.
1783
1784 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
1785 optimization. I've improved things by keeping a bitmap of backreferences with
1786 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
1787 fact referenced, the optimization can be applied. It is unlikely that a
1788 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
1789 the match to follow \n) will appear inside brackets with a number greater than
1790 31, but if it does, any back reference > 31 suppresses the optimization.
1791
1792 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
1793 of disabling numbered capturing parentheses. Any opening parenthesis that is
1794 not followed by ? behaves as if it were followed by ?: but named parentheses
1795 can still be used for capturing (and they will acquire numbers in the usual
1796 way).
1797
1798 41. Redesigned the return codes from the match() function into yes/no/error so
1799 that errors can be passed back from deep inside the nested calls. A malloc
1800 failure while inside a recursive subpattern call now causes the
1801 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
1802
1803 42. It is now possible to set a limit on the number of times the match()
1804 function is called in a call to pcre_exec(). This facility makes it possible to
1805 limit the amount of recursion and backtracking, though not in a directly
1806 obvious way, because the match() function is used in a number of different
1807 circumstances. The count starts from zero for each position in the subject
1808 string (for non-anchored patterns). The default limit is, for compatibility, a
1809 large number, namely 10 000 000. You can change this in two ways:
1810
1811 (a) When configuring PCRE before making, you can use --with-match-limit=n
1812 to set a default value for the compiled library.
1813
1814 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
1815 a different value is set. See 45 below.
1816
1817 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1818
1819 43. Added a new function pcre_config(int, void *) to enable run-time extraction
1820 of things that can be changed at compile time. The first argument specifies
1821 what is wanted and the second points to where the information is to be placed.
1822 The current list of available information is:
1823
1824 PCRE_CONFIG_UTF8
1825
1826 The output is an integer that is set to one if UTF-8 support is available;
1827 otherwise it is set to zero.
1828
1829 PCRE_CONFIG_NEWLINE
1830
1831 The output is an integer that it set to the value of the code that is used for
1832 newline. It is either LF (10) or CR (13).
1833
1834 PCRE_CONFIG_LINK_SIZE
1835
1836 The output is an integer that contains the number of bytes used for internal
1837 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
1838
1839 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
1840
1841 The output is an integer that contains the threshold above which the POSIX
1842 interface uses malloc() for output vectors. See item 31 above.
1843
1844 PCRE_CONFIG_MATCH_LIMIT
1845
1846 The output is an unsigned integer that contains the default limit of the number
1847 of match() calls in a pcre_exec() execution. See 42 above.
1848
1849 44. pcretest has been upgraded by the addition of the -C option. This causes it
1850 to extract all the available output from the new pcre_config() function, and to
1851 output it. The program then exits immediately.
1852
1853 45. A need has arisen to pass over additional data with calls to pcre_exec() in
1854 order to support additional features. One way would have been to define
1855 pcre_exec2() (for example) with extra arguments, but this would not have been
1856 extensible, and would also have required all calls to the original function to
1857 be mapped to the new one. Instead, I have chosen to extend the mechanism that
1858 is used for passing in "extra" data from pcre_study().
1859
1860 The pcre_extra structure is now exposed and defined in pcre.h. It currently
1861 contains the following fields:
1862
1863 flags a bitmap indicating which of the following fields are set
1864 study_data opaque data from pcre_study()
1865 match_limit a way of specifying a limit on match() calls for a specific
1866 call to pcre_exec()
1867 callout_data data for callouts (see 49 below)
1868
1869 The flag bits are also defined in pcre.h, and are
1870
1871 PCRE_EXTRA_STUDY_DATA
1872 PCRE_EXTRA_MATCH_LIMIT
1873 PCRE_EXTRA_CALLOUT_DATA
1874
1875 The pcre_study() function now returns one of these new pcre_extra blocks, with
1876 the actual study data pointed to by the study_data field, and the
1877 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
1878 before. That is, this change is entirely upwards-compatible and requires no
1879 change to existing code.
1880
1881 If you want to pass in additional data to pcre_exec(), you can either place it
1882 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
1883 block.
1884
1885 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
1886 data string contains the escape sequence \M, pcretest calls pcre_exec() several
1887 times with different match limits, until it finds the minimum value needed for
1888 pcre_exec() to complete. The value is then output. This can be instructive; for
1889 most simple matches the number is quite small, but for pathological cases it
1890 gets very large very quickly.
1891
1892 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
1893 returns the size of the data block pointed to by the study_data field in a
1894 pcre_extra block, that is, the value that was passed as the argument to
1895 pcre_malloc() when PCRE was getting memory in which to place the information
1896 created by pcre_study(). The fourth argument should point to a size_t variable.
1897 pcretest has been extended so that this information is shown after a successful
1898 pcre_study() call when information about the compiled regex is being displayed.
1899
1900 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
1901 because what follows is always an absolute path. (Later: it turns out that this
1902 is more than cosmetic for MinGW, because it doesn't like empty path
1903 components.)
1904
1905 49. Some changes have been made to the callout feature (see 28 above):
1906
1907 (i) A callout function now has three choices for what it returns:
1908
1909 0 => success, carry on matching
1910 > 0 => failure at this point, but backtrack if possible
1911 < 0 => serious error, return this value from pcre_exec()
1912
1913 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
1914 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
1915 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
1916 use by callout functions. It will never be used by PCRE itself.
1917
1918 (ii) The pcre_extra structure (see 45 above) has a void * field called
1919 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
1920 pcre_callout_block structure has a field of the same name. The contents of
1921 the field passed in the pcre_extra structure are passed to the callout
1922 function in the corresponding field in the callout block. This makes it
1923 easier to use the same callout-containing regex from multiple threads. For
1924 testing, the pcretest program has a new data escape
1925
1926 \C*n pass the number n (may be negative) as callout_data
1927
1928 If the callout function in pcretest receives a non-zero value as
1929 callout_data, it returns that value.
1930
1931 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
1932 there were some redundant $(CFLAGS) in commands that are now specified as
1933 $(LINK), which already includes $(CFLAGS).
1934
1935 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
1936 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
1937 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
1938 one-byte characters throughout. Note that case-insensitive matching applies
1939 only to characters whose values are less than 256. PCRE doesn't support the
1940 notion of cases for higher-valued characters.
1941
1942 (i) A character class whose characters are all within 0-255 is handled as
1943 a bit map, and the map is inverted for negative classes. Previously, a
1944 character > 255 always failed to match such a class; however it should
1945 match if the class was a negative one (e.g. [^ab]). This has been fixed.
1946
1947 (ii) A negated character class with a single character < 255 is coded as
1948 "not this character" (OP_NOT). This wasn't working properly when the test
1949 character was multibyte, either singly or repeated.
1950
1951 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
1952 mode, for example: \x{100}{2,3}.
1953
1954 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
1955 singly or repeated) now correctly test multibyte characters. However,
1956 PCRE doesn't recognize any characters with values greater than 255 as
1957 digits, spaces, or word characters. Such characters always match \D, \S,
1958 and \W, and never match \d, \s, or \w.
1959
1960 (v) Classes may now contain characters and character ranges with values
1961 greater than 255. For example: [ab\x{100}-\x{400}].
1962
1963 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
1964 PCRE in UTF-8 mode.
1965
1966 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
1967 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
1968 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
1969 value.)
1970
1971 53. The single man page has become too large. I have therefore split it up into
1972 a number of separate man pages. These also give rise to individual HTML pages;
1973 these are now put in a separate directory, and there is an index.html page that
1974 lists them all. Some hyperlinking between the pages has been installed.
1975
1976 54. Added convenience functions for handling named capturing parentheses.
1977
1978 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
1979 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
1980 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
1981 are faulted.
1982
1983 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
1984 calling configure. These values are used when compiling the dftables.c program
1985 which is run to generate the source of the default character tables. They
1986 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
1987 you will need to set these values.
1988
1989 57. Updated the building process for Windows DLL, as provided by Fred Cox.
1990
1991
1992 Version 3.9 02-Jan-02
1993 ---------------------
1994
1995 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
1996
1997 2. If --disable-static was given, the building process failed when trying to
1998 build pcretest and pcregrep. (For some reason it was using libtool to compile
1999 them, which is not right, as they aren't part of the library.)
2000
2001
2002 Version 3.8 18-Dec-01
2003 ---------------------
2004
2005 1. The experimental UTF-8 code was completely screwed up. It was packing the
2006 bytes in the wrong order. How dumb can you get?
2007
2008
2009 Version 3.7 29-Oct-01
2010 ---------------------
2011
2012 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2013 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2014 this didn't happen under Solaris 8, where I normally test things.
2015
2016 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2017 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2018 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2019 this fix an hour or so after the initial 3.7 release.)
2020
2021
2022 Version 3.6 23-Oct-01
2023 ---------------------
2024
2025 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2026 offsets passed as NULL with zero offset count.
2027
2028 2. The config.guess and config.sub files had not been updated when I moved to
2029 the latest autoconf.
2030
2031
2032 Version 3.5 15-Aug-01
2033 ---------------------
2034
2035 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2036 had been forgotten.
2037
2038 2. By using declared but undefined structures, we can avoid using "void"
2039 definitions in pcre.h while keeping the internal definitions of the structures
2040 private.
2041
2042 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2043 user point of view, this means that both static and shared libraries are built
2044 by default, but this can be individually controlled. More of the work of
2045 handling this static/shared cases is now inside libtool instead of PCRE's make
2046 file.
2047
2048 4. The pcretest utility is now installed along with pcregrep because it is
2049 useful for users (to test regexs) and by doing this, it automatically gets
2050 relinked by libtool. The documentation has been turned into a man page, so
2051 there are now .1, .txt, and .html versions in /doc.
2052
2053 5. Upgrades to pcregrep:
2054 (i) Added long-form option names like gnu grep.
2055 (ii) Added --help to list all options with an explanatory phrase.
2056 (iii) Added -r, --recursive to recurse into sub-directories.
2057 (iv) Added -f, --file to read patterns from a file.
2058
2059 6. pcre_exec() was referring to its "code" argument before testing that
2060 argument for NULL (and giving an error if it was NULL).
2061
2062 7. Upgraded Makefile.in to allow for compiling in a different directory from
2063 the source directory.
2064
2065 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2066 options bits, the pointer it was passed was to an int instead of to an unsigned
2067 long int. This mattered only on 64-bit systems.
2068
2069 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2070 generated) instead of pcre.in, which it its source. Also made the same change
2071 in several of the .c files.
2072
2073 10. A new release of gcc defines printf() as a macro, which broke pcretest
2074 because it had an ifdef in the middle of a string argument for printf(). Fixed
2075 by using separate calls to printf().
2076
2077 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2078 script, to force use of CR or LF instead of \n in the source. On non-Unix
2079 systems, the value can be set in config.h.
2080
2081 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2082 absolute limit. Changed the text of the error message to make this clear, and
2083 likewise updated the man page.
2084
2085 13. The limit of 99 on the number of capturing subpatterns has been removed.
2086 The new limit is 65535, which I hope will not be a "real" limit.
2087
2088
2089 Version 3.4 22-Aug-00
2090 ---------------------
2091
2092 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2093
2094 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2095
2096
2097 Version 3.3 01-Aug-00
2098 ---------------------
2099
2100 1. If an octal character was given, but the value was greater than \377, it
2101 was not getting masked to the least significant bits, as documented. This could
2102 lead to crashes in some systems.
2103
2104 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2105 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2106
2107 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2108 These just pass their arguments on to (pcre_free)(), but they are provided
2109 because some uses of PCRE bind it to non-C systems that can call its functions,
2110 but cannot call free() or pcre_free() directly.
2111
2112 4. Add "make test" as a synonym for "make check". Corrected some comments in
2113 the Makefile.
2114
2115 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2116 Makefile.
2117
2118 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2119 command called pgrep for grepping around the active processes.
2120
2121 7. Added the beginnings of support for UTF-8 character strings.
2122
2123 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2124 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2125 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2126 out for the ar command.)
2127
2128
2129 Version 3.2 12-May-00
2130 ---------------------
2131
2132 This is purely a bug fixing release.
2133
2134 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2135 of ZA. This was just one example of several cases that could provoke this bug,
2136 which was introduced by change 9 of version 2.00. The code for breaking
2137 infinite loops after an iteration that matches an empty string was't working
2138 correctly.
2139
2140 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2141 when matched against abbab (for example). After matching an empty string, it
2142 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2143 caused it to match further down the string than it should.
2144
2145 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2146 was there because it doesn't seem to be needed, and it causes trouble on some
2147 systems, as it is not a Standard C header. It has been removed.
2148
2149 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2150 were reported on the Macintosh. The changes were from
2151
2152 while ((c = *(++ptr)) != 0 && c != '\n');
2153 to
2154 while ((c = *(++ptr)) != 0 && c != '\n') ;
2155
2156 Totally extraordinary, but if that's what it takes...
2157
2158 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2159 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2160 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2161 assumes the way PCRE uses memmove() (always moving upwards).
2162
2163 6. PCRE is being used in one environment where strchr() is not available. There
2164 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2165 faster code anyway.
2166
2167
2168 Version 3.1 09-Feb-00
2169 ---------------------
2170
2171 The only change in this release is the fixing of some bugs in Makefile.in for
2172 the "install" target:
2173
2174 (1) It was failing to install pcreposix.h.
2175
2176 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2177
2178
2179 Version 3.0 01-Feb-00
2180 ---------------------
2181
2182 1. Add support for the /+ modifier to perltest (to output $` like it does in
2183 pcretest).
2184
2185 2. Add support for the /g modifier to perltest.
2186
2187 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2188 matches null strings.
2189
2190 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2191 pattern. Perl treats empty patterns specially - it reuses the most recent
2192 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2193 effect.
2194
2195 5. The POSIX interface was broken in that it was just handing over the POSIX
2196 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2197 required a bigger vector, with some working space on the end. This means that
2198 the POSIX wrapper now has to get and free some memory, and copy the results.
2199
2200 6. Added some simple autoconf support, placing the test data and the
2201 documentation in separate directories, re-organizing some of the
2202 information files, and making it build pcre-config (a GNU standard). Also added
2203 libtool support for building PCRE as a shared library, which is now the
2204 default.
2205
2206 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2207 09 are not valid octal constants. Single digits will be used for minor values
2208 less than 10.
2209
2210 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2211 existing programs that set these in the POSIX interface can use PCRE without
2212 modification.
2213
2214 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2215 return all that pcre_info() returns, plus additional data. The pcre_info()
2216 function is retained for compatibility, but is considered to be obsolete.
2217
2218 10. Added experimental recursion feature (?R) to handle one common case that
2219 Perl 5.6 will be able to do with (?p{...}).
2220
2221 11. Added support for POSIX character classes like [:alpha:], which Perl is
2222 adopting.
2223
2224
2225 Version 2.08 31-Aug-99
2226 ----------------------
2227
2228 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2229 trying to match at the startoffset position, but instead was moving forward to
2230 the next newline as if a previous match had failed.
2231
2232 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2233 and could get into a loop if a null string was matched other than at the start
2234 of the subject.
2235
2236 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2237 be distinguished at compile time, and for completeness also added PCRE_DATE.
2238
2239 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2240 in GnuWin32 environments.
2241
2242
2243 Version 2.07 29-Jul-99
2244 ----------------------
2245
2246 1. The documentation is now supplied in plain text form and HTML as well as in
2247 the form of man page sources.
2248
2249 2. C++ compilers don't like assigning (void *) values to other pointer types.
2250 In particular this affects malloc(). Although there is no problem in Standard
2251 C, I've put in casts to keep C++ compilers happy.
2252
2253 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2254 should be (const char *).
2255
2256 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2257 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2258 However, I haven't made this a standard facility. The documentation doesn't
2259 mention it, and the Makefile doesn't support it.
2260
2261 5. The Makefile now contains an "install" target, with editable destinations at
2262 the top of the file. The pcretest program is not installed.
2263
2264 6. pgrep -V now gives the PCRE version number and date.
2265
2266 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2267 causing the entire string to be ignored, instead of just the last character.
2268
2269 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2270 non-matching string, it can take a very, very long time, even for strings of
2271 quite modest length, because of the nested recursion. PCRE now does better in
2272 some of these cases. It does this by remembering the last required literal
2273 character in the pattern, and pre-searching the subject to ensure it is present
2274 before running the real match. In other words, it applies a heuristic to detect
2275 some types of certain failure quickly, and in the above example, if presented
2276 with a string that has no trailing " it gives "no match" very quickly.
2277
2278 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2279 other alternatives are tried instead.
2280
2281
2282 Version 2.06 09-Jun-99
2283 ----------------------
2284
2285 1. Change pcretest's output for amount of store used to show just the code
2286 space, because the remainder (the data block) varies in size between 32-bit and
2287 64-bit systems.
2288
2289 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2290 start matching at. This allows lookbehinds to work when searching for multiple
2291 occurrences in a string.
2292
2293 3. Added additional options to pcretest for testing multiple occurrences:
2294
2295 /+ outputs the rest of the string that follows a match
2296 /g loops for multiple occurrences, using the new startoffset argument
2297 /G loops for multiple occurrences by passing an incremented pointer
2298
2299 4. PCRE wasn't doing the "first character" optimization for patterns starting
2300 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2301 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2302 the letter 'x'. On long subject strings, this gives a significant speed-up.
2303
2304
2305 Version 2.05 21-Apr-99
2306 ----------------------
2307
2308 1. Changed the type of magic_number from int to long int so that it works
2309 properly on 16-bit systems.
2310
2311 2. Fixed a bug which caused patterns starting with .* not to work correctly
2312 when the subject string contained newline characters. PCRE was assuming
2313 anchoring for such patterns in all cases, which is not correct because .* will
2314 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2315 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2316 must be retried after every newline in the subject.
2317
2318
2319 Version 2.04 18-Feb-99
2320 ----------------------
2321
2322 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2323 computation of the store needed to hold the pattern was incorrect (too large).
2324 If such patterns were nested a few deep, this could multiply and become a real
2325 problem.
2326
2327 2. Added /M option to pcretest to show the memory requirement of a specific
2328 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2329
2330 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2331 compiled in such a way that the backtracking after subsequent failure was
2332 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2333 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2334
2335
2336 Version 2.03 02-Feb-99
2337 ----------------------
2338
2339 1. Fixed typo and small mistake in man page.
2340
2341 2. Added 4th condition (GPL supersedes if conflict) and created separate
2342 LICENCE file containing the conditions.
2343
2344 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2345 Perl, that is the internal \ allows the delimiter to be included in the
2346 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2347 the final delimiter, add \ to the end of the pattern (to test the error).
2348
2349 4. Added the convenience functions for extracting substrings after a successful
2350 match. Updated pcretest to make it able to test these functions.
2351
2352
2353 Version 2.02 14-Jan-99
2354 ----------------------
2355
2356 1. Initialized the working variables associated with each extraction so that
2357 their saving and restoring doesn't refer to uninitialized store.
2358
2359 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2360 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2361 fix the problem.
2362
2363 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2364 calls, and wasn't printing the correct value for compiling calls. Increased the
2365 default value of LOOPREPEAT, and the number of significant figures in the
2366 times.
2367
2368 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2369
2370 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2371 a building problem on Windows NT with a FAT file system.
2372
2373
2374 Version 2.01 21-Oct-98
2375 ----------------------
2376
2377 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2378 to character tables built by pcre_maketables() in the current locale. If NULL
2379 is passed, the default tables are used.
2380
2381
2382 Version 2.00 24-Sep-98
2383 ----------------------
2384
2385 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2386 it any more.
2387
2388 2. Allow quantification of (?>) groups, and make it work correctly.
2389
2390 3. The first character computation wasn't working for (?>) groups.
2391
2392 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2393 end of the subject) and add 5.005's \z, which really does match only at the
2394 very end of the subject.
2395
2396 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2397
2398 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2399 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2400 localized options. All options to pcre_study() were also removed.
2401
2402 7. Add other new features from 5.005:
2403
2404 $(?<= positive lookbehind
2405 $(?<! negative lookbehind
2406 (?imsx-imsx) added the unsetting capability
2407 such a setting is global if at outer level; local otherwise
2408 (?imsx-imsx:) non-capturing groups with option setting
2409 (?(cond)re|re) conditional pattern matching
2410
2411 A backreference to itself in a repeated group matches the previous
2412 captured string.
2413
2414 8. General tidying up of studying (both automatic and via "study")
2415 consequential on the addition of new assertions.
2416
2417 9. As in 5.005, unlimited repeated groups that could match an empty substring
2418 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2419 runtime if any iteration does actually match an empty substring.
2420
2421 10. Include the RunTest script in the distribution.
2422
2423 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2424 discrepancies, some of which were old and were also with respect to 5.004. They
2425 have now been fixed.
2426
2427
2428 Version 1.09 28-Apr-98
2429 ----------------------
2430
2431 1. A negated single character class followed by a quantifier with a minimum
2432 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2433 program crashes, or just wrong answers. This did not apply to negated classes
2434 containing more than one character, or to minima other than one.
2435
2436
2437 Version 1.08 27-Mar-98
2438 ----------------------
2439
2440 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2441
2442 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2443 latter must appear before anything that relies on it in the pattern.
2444
2445
2446 Version 1.07 16-Feb-98
2447 ----------------------
2448
2449 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2450 repeat of a potentially empty string).
2451
2452
2453 Version 1.06 23-Jan-98
2454 ----------------------
2455
2456 1. Added Markus Oberhumer's little patches for C++.
2457
2458 2. Literal strings longer than 255 characters were broken.
2459
2460
2461 Version 1.05 23-Dec-97
2462 ----------------------
2463
2464 1. Negated character classes containing more than one character were failing if
2465 PCRE_CASELESS was set at run time.
2466
2467
2468 Version 1.04 19-Dec-97
2469 ----------------------
2470
2471 1. Corrected the man page, where some "const" qualifiers had been omitted.
2472
2473 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2474 input syntax.
2475
2476 3. Fixed memory leak which occurred when a regex with back references was
2477 matched with an offsets vector that wasn't big enough. The temporary memory
2478 that is used in this case wasn't being freed if the match failed.
2479
2480 4. Tidied pcretest to ensure it frees memory that it gets.
2481
2482 5. Temporary memory was being obtained in the case where the passed offsets
2483 vector was exactly big enough.
2484
2485 6. Corrected definition of offsetof() from change 5 below.
2486
2487 7. I had screwed up change 6 below and broken the rules for the use of
2488 setjmp(). Now fixed.
2489
2490
2491 Version 1.03 18-Dec-97
2492 ----------------------
2493
2494 1. A erroneous regex with a missing opening parenthesis was correctly
2495 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2496 on some systems.
2497
2498 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2499 it was reported that one broken compiler failed on the former because "code" is
2500 also an independent variable.
2501
2502 3. The erroneous regex a[]b caused an array overrun reference.
2503
2504 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2505 fail on data ending with that character. (It was going on too far, and checking
2506 the next character, typically a binary zero.) This was specific to the
2507 optimized code for single-character negative classes.
2508
2509 5. Added a contributed patch from the TIN world which does the following:
2510
2511 + Add an undef for memmove, in case the the system defines a macro for it.
2512
2513 + Add a definition of offsetof(), in case there isn't one. (I don't know
2514 the reason behind this - offsetof() is part of the ANSI standard - but
2515 it does no harm).
2516
2517 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2518 most of the places where whitespace preceded '#'. I have given up and
2519 allowed the remaining 2 cases to be at the margin.
2520
2521 + Rename some variables in pcre to eliminate shadowing. This seems very
2522 pedantic, but does no harm, of course.
2523
2524 6. Moved the call to setjmp() into its own function, to get rid of warnings
2525 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2526
2527 7. Constructs such as \d{8,} were compiling into the equivalent of
2528 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2529 outcome, but in this particular case used more store than had been allocated,
2530 which caused the bug to be discovered because it threw up an internal error.
2531
2532 8. The debugging code in both pcre and pcretest for outputting the compiled
2533 form of a regex was going wrong in the case of back references followed by
2534 curly-bracketed repeats.
2535
2536
2537 Version 1.02 12-Dec-97
2538 ----------------------
2539
2540 1. Typos in pcre.3 and comments in the source fixed.
2541
2542 2. Applied a contributed patch to get rid of places where it used to remove
2543 'const' from variables, and fixed some signed/unsigned and uninitialized
2544 variable warnings.
2545
2546 3. Added the "runtest" target to Makefile.
2547
2548 4. Set default compiler flag to -O2 rather than just -O.
2549
2550
2551 Version 1.01 19-Nov-97
2552 ----------------------
2553
2554 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2555 like /([ab]*)*/, that is, for classes with more than one character in them.
2556
2557 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2558 as /((?>a*))*/ (a PCRE_EXTRA facility).
2559
2560
2561 Version 1.00 18-Nov-97
2562 ----------------------
2563
2564 1. Added compile-time macros to support systems such as SunOS4 which don't have
2565 memmove() or strerror() but have other things that can be used instead.
2566
2567 2. Arranged that "make clean" removes the executables.
2568
2569
2570 Version 0.99 27-Oct-97
2571 ----------------------
2572
2573 1. Fixed bug in code for optimizing classes with only one character. It was
2574 initializing a 32-byte map regardless, which could cause it to run off the end
2575 of the memory it had got.
2576
2577 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
2578
2579
2580 Version 0.98 22-Oct-97
2581 ----------------------
2582
2583 1. Fixed bug in code for handling temporary memory usage when there are more
2584 back references than supplied space in the ovector. This could cause segfaults.
2585
2586
2587 Version 0.97 21-Oct-97
2588 ----------------------
2589
2590 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
2591
2592 2. Optimized negated single characters not to use a bit map.
2593
2594 3. Brought error texts together as macro definitions; clarified some of them;
2595 fixed one that was wrong - it said "range out of order" when it meant "invalid
2596 escape sequence".
2597
2598 4. Changed some char * arguments to const char *.
2599
2600 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
2601
2602 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
2603 pcretest.
2604
2605
2606 Version 0.96 16-Oct-97
2607 ----------------------
2608
2609 1. Added a simple "pgrep" utility to the distribution.
2610
2611 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
2612 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
2613 where "ddd" means "one or more decimal digits".
2614
2615 3. Fixed serious bug. If a pattern had a back reference, but the call to
2616 pcre_exec() didn't supply a large enough ovector to record the related
2617 identifying subpattern, the match always failed. PCRE now remembers the number
2618 of the largest back reference, and gets some temporary memory in which to save
2619 the offsets during matching if necessary, in order to ensure that
2620 backreferences always work.
2621
2622 4. Increased the compatibility with Perl in a number of ways:
2623
2624 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
2625 to request this handling. The option can be set at compile or exec time.
2626
2627 (b) $ matches before a terminating newline by default; an option
2628 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
2629 mode). The option can be set at compile or exec time.
2630
2631 (c) The handling of \ followed by a digit other than 0 is now supposed to be
2632 the same as Perl's. If the decimal number it represents is less than 10
2633 or there aren't that many previous left capturing parentheses, an octal
2634 escape is read. Inside a character class, it's always an octal escape,
2635 even if it is a single digit.
2636
2637 (d) An escaped but undefined alphabetic character is taken as a literal,
2638 unless PCRE_EXTRA is set. Currently this just reserves the remaining
2639 escapes.
2640
2641 (e) {0} is now permitted. (The previous item is removed from the compiled
2642 pattern).
2643
2644 5. Changed all the names of code files so that the basic parts are no longer
2645 than 10 characters, and abolished the teeny "globals.c" file.
2646
2647 6. Changed the handling of character classes; they are now done with a 32-byte
2648 bit map always.
2649
2650 7. Added the -d and /D options to pcretest to make it possible to look at the
2651 internals of compilation without having to recompile pcre.
2652
2653
2654 Version 0.95 23-Sep-97
2655 ----------------------
2656
2657 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
2658 \x20 at the start of a run of normal characters. These were being treated as
2659 real characters, instead of the source characters being re-checked.
2660
2661
2662 Version 0.94 18-Sep-97
2663 ----------------------
2664
2665 1. The functions are now thread-safe, with the caveat that the global variables
2666 containing pointers to malloc() and free() or alternative functions are the
2667 same for all threads.
2668
2669 2. Get pcre_study() to generate a bitmap of initial characters for non-
2670 anchored patterns when this is possible, and use it if passed to pcre_exec().
2671
2672
2673 Version 0.93 15-Sep-97
2674 ----------------------
2675
2676 1. /(b)|(:+)/ was computing an incorrect first character.
2677
2678 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
2679 but not actually doing anything yet.
2680
2681 3. Treat "-" characters in classes that cannot be part of ranges as literals,
2682 as Perl does (e.g. [-az] or [az-]).
2683
2684 4. Set the anchored flag if a branch starts with .* or .*? because that tests
2685 all possible positions.
2686
2687 5. Split up into different modules to avoid including unneeded functions in a
2688 compiled binary. However, compile and exec are still in one module. The "study"
2689 function is split off.
2690
2691 6. The character tables are now in a separate module whose source is generated
2692 by an auxiliary program - but can then be edited by hand if required. There are
2693 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
2694 toupper() in the code.
2695
2696 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
2697 make them global. Abolish the function for setting them, as the caller can now
2698 set them directly.
2699
2700
2701 Version 0.92 11-Sep-97
2702 ----------------------
2703
2704 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
2705 (e.g. /a{1,3}/) was broken (I mis-optimized it).
2706
2707 2. Caseless matching was not working in character classes if the characters in
2708 the pattern were in upper case.
2709
2710 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
2711
2712 4. Make PCRE_ANCHORED public and accept as a compile option.
2713
2714 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
2715 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
2716 pass them.
2717
2718 6. Give an error if bad option bits passed at compile or run time.
2719
2720 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
2721 pcretest to cause it to pass that flag.
2722
2723 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
2724 options, and the first character, if set.
2725
2726 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
2727
2728
2729 Version 0.91 10-Sep-97
2730 ----------------------
2731
2732 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
2733 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
2734
2735 2. PCRE was looping on encountering an indefinitely repeated back reference to
2736 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
2737 Perl does - treats the match as successful.
2738
2739 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12