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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 166 - (show annotations) (download)
Wed May 9 14:48:28 2007 UTC (7 years, 6 months ago) by ph10
File size: 128097 byte(s)
Add (?-n) and (?+n) relative references.

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

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