/[pcre]/code/tags/pcre-7.3/ChangeLog
ViewVC logotype

Contents of /code/tags/pcre-7.3/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 214 - (show annotations) (download)
Wed Aug 15 14:08:10 2007 UTC (7 years, 1 month ago) by ph10
Original Path: code/trunk/ChangeLog
File size: 135692 byte(s)
Fixed another looking-too-far-back-in-non-UTF-8-mode bug.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 7.3 09-Aug-07
5 ---------------------
6
7 1. In the rejigging of the build system that eventually resulted in 7.1, the
8 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
9 brackets there is not right, since it causes compilers to look for an
10 installed pcre.h, not the version that is in the source that is being
11 compiled (which of course may be different). I have changed it back to:
12
13 #include "pcre.h"
14
15 I have a vague recollection that the change was concerned with compiling in
16 different directories, but in the new build system, that is taken care of
17 by the VPATH setting the Makefile.
18
19 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
20 when the subject happened to end in the byte 0x85 (e.g. if the last
21 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
22 characters but of course it shouldn't be taken as a newline when it is part
23 of another character. The bug was that, for an unlimited repeat of . in
24 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
25 characters when looking for a newline.
26
27 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
28
29 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
30 in debug output.
31
32 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
33 long printing in the pcrecpp unittest when running under MinGW.
34
35 6. ESC_K was left out of the EBCDIC table.
36
37 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
38 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
39 limit also applies to "virtual nesting" when a pattern is recursive, and in
40 this case 1000 isn't so big. I have been able to remove this limit at the
41 expense of backing off one optimization in certain circumstances. Normally,
42 when pcre_exec() would call its internal match() function recursively and
43 immediately return the result unconditionally, it uses a "tail recursion"
44 feature to save stack. However, when a subpattern that can match an empty
45 string has an unlimited repetition quantifier, it no longer makes this
46 optimization. That gives it a stack frame in which to save the data for
47 checking that an empty string has been matched. Previously this was taken
48 from the 1000-entry workspace that had been reserved. So now there is no
49 explicit limit, but more stack is used.
50
51 8. Applied Daniel's patches to solve problems with the import/export magic
52 syntax that is required for Windows, and which was going wrong for the
53 pcreposix and pcrecpp parts of the library. These were overlooked when this
54 problem was solved for the main library.
55
56 9. There were some crude static tests to avoid integer overflow when computing
57 the size of patterns that contain repeated groups with explicit upper
58 limits. As the maximum quantifier is 65535, the maximum group length was
59 set at 30,000 so that the product of these two numbers did not overflow a
60 32-bit integer. However, it turns out that people want to use groups that
61 are longer than 30,000 bytes (though not repeat them that many times).
62 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
63 made it possible to implement the integer overflow checks in a much more
64 dynamic way, which I have now done. The artificial limitation on group
65 length has been removed - we now have only the limit on the total length of
66 the compiled pattern, which depends on the LINK_SIZE setting.
67
68 10. Fixed a bug in the documentation for get/copy named substring when
69 duplicate names are permitted. If none of the named substrings are set, the
70 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
71 empty string.
72
73 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
74 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
75 because the ] is interpreted as the first data character and the
76 terminating ] is not found. PCRE has been made compatible with Perl in this
77 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
78 cause memory overwriting.
79
80 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
81 string has been matched (to stop an infinite loop). It was not recognizing
82 a conditional subpattern that could match an empty string if that
83 subpattern was within another subpattern. For example, it looped when
84 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
85 condition was not nested. This bug has been fixed.
86
87 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
88 past the start of the subject in the presence of bytes with the top bit
89 set, for example "\x8aBCD".
90
91 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
92 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
93
94 14. Optimized (?!) to (*FAIL).
95
96 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
97 This restricts code points to be within the range 0 to 0x10FFFF, excluding
98 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
99 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
100 does: it's just the validity check that is more restrictive.
101
102 16. Inserted checks for integer overflows during escape sequence (backslash)
103 processing, and also fixed erroneous offset values for syntax errors during
104 backslash processing.
105
106 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
107 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
108
109
110 Version 7.2 19-Jun-07
111 ---------------------
112
113 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
114 which is apparently normally available under Windows.
115
116 2. Re-jig the pcregrep tests with different newline settings in an attempt
117 to make them independent of the local environment's newline setting.
118
119 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
120
121 4. Some of the "internals" tests were previously cut out when the link size
122 was not 2, because the output contained actual offsets. The recent new
123 "Z" feature of pcretest means that these can be cut out, making the tests
124 usable with all link sizes.
125
126 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
127 stack recursion. This gives a massive performance boost under BSD, but just
128 a small improvement under Linux. However, it saves one field in the frame
129 in all cases.
130
131 6. Added more features from the forthcoming Perl 5.10:
132
133 (a) (?-n) (where n is a string of digits) is a relative subroutine or
134 recursion call. It refers to the nth most recently opened parentheses.
135
136 (b) (?+n) is also a relative subroutine call; it refers to the nth next
137 to be opened parentheses.
138
139 (c) Conditions that refer to capturing parentheses can be specified
140 relatively, for example, (?(-2)... or (?(+3)...
141
142 (d) \K resets the start of the current match so that everything before
143 is not part of it.
144
145 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
146
147 (f) \g{name} is another synonym - part of Perl 5.10's unification of
148 reference syntax.
149
150 (g) (?| introduces a group in which the numbering of parentheses in each
151 alternative starts with the same number.
152
153 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
154
155 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
156 PCRE_INFO_JCHANGED.
157
158 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
159 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
160 for detecting groups that can match an empty string.
161
162 9. A pattern with a very large number of alternatives (more than several
163 hundred) was running out of internal workspace during the pre-compile
164 phase, where pcre_compile() figures out how much memory will be needed. A
165 bit of new cunning has reduced the workspace needed for groups with
166 alternatives. The 1000-alternative test pattern now uses 12 bytes of
167 workspace instead of running out of the 4096 that are available.
168
169 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
170
171 11. Applied patch from Google to remove an optimization that didn't quite work.
172 The report of the bug said:
173
174 pcrecpp::RE("a*").FullMatch("aaa") matches, while
175 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
176 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
177
178 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
179 it matched the wrong number of bytes.
180
181
182 Version 7.1 24-Apr-07
183 ---------------------
184
185 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
186 that is more "standard", making use of automake and other Autotools. There
187 is some re-arrangement of the files and adjustment of comments consequent
188 on this.
189
190 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
191 for recursive directory scanning broke on some systems because the files
192 are not scanned in any specific order and on different systems the order
193 was different. A call to "sort" has been inserted into RunGrepTest for the
194 approprate test as a short-term fix. In the longer term there may be an
195 alternative.
196
197 3. I had an email from Eric Raymond about problems translating some of PCRE's
198 man pages to HTML (despite the fact that I distribute HTML pages, some
199 people do their own conversions for various reasons). The problems
200 concerned the use of low-level troff macros .br and .in. I have therefore
201 removed all such uses from the man pages (some were redundant, some could
202 be replaced by .nf/.fi pairs). The 132html script that I use to generate
203 HTML has been updated to handle .nf/.fi and to complain if it encounters
204 .br or .in.
205
206 4. Updated comments in configure.ac that get placed in config.h.in and also
207 arranged for config.h to be included in the distribution, with the name
208 config.h.generic, for the benefit of those who have to compile without
209 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
210
211 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
212 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
213 makevp.bat for the current PCRE, using the additional files
214 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
215
216 6. A Windows user reported a minor discrepancy with test 2, which turned out
217 to be caused by a trailing space on an input line that had got lost in his
218 copy. The trailing space was an accident, so I've just removed it.
219
220 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
221 that is needed.
222
223 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
224 as "const" (a) because they are and (b) because it helps the PHP
225 maintainers who have recently made a script to detect big data structures
226 in the php code that should be moved to the .rodata section. I remembered
227 to update Builducptable as well, so it won't revert if ucptable.h is ever
228 re-created.
229
230 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
231 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
232 order to be able to cut out the UTF-8 tables in the latter when UTF-8
233 support is not required. This saves 1.5-2K of code, which is important in
234 some applications.
235
236 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
237 so as not to refer to the tables, even though these functions will never be
238 called when UTF-8 support is disabled. Otherwise there are problems with a
239 shared library.
240
241 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
242
243 (a) It was defining its arguments as char * instead of void *.
244
245 (b) It was assuming that all moves were upwards in memory; this was true
246 a long time ago when I wrote it, but is no longer the case.
247
248 The emulated memove() is provided for those environments that have neither
249 memmove() nor bcopy(). I didn't think anyone used it these days, but that
250 is clearly not the case, as these two bugs were recently reported.
251
252 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
253 and Detrail to create the HTML documentation, the .txt form of the man
254 pages, and it removes trailing spaces from listed files. It also creates
255 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
256 case, it wraps all the #defines with #ifndefs. This script should be run
257 before "make dist".
258
259 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
260 with Unicode property support.
261
262 (a) For a maximizing quantifier, if the two different cases of the
263 character were of different lengths in their UTF-8 codings (there are
264 some cases like this - I found 11), and the matching function had to
265 back up over a mixture of the two cases, it incorrectly assumed they
266 were both the same length.
267
268 (b) When PCRE was configured to use the heap rather than the stack for
269 recursion during matching, it was not correctly preserving the data for
270 the other case of a UTF-8 character when checking ahead for a match
271 while processing a minimizing repeat. If the check also involved
272 matching a wide character, but failed, corruption could cause an
273 erroneous result when trying to check for a repeat of the original
274 character.
275
276 13. Some tidying changes to the testing mechanism:
277
278 (a) The RunTest script now detects the internal link size and whether there
279 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
280 values substituted by "configure". (The RunGrepTest script already did
281 this for UTF-8.) The configure.ac script no longer substitutes the
282 relevant variables.
283
284 (b) The debugging options /B and /D in pcretest show the compiled bytecode
285 with length and offset values. This means that the output is different
286 for different internal link sizes. Test 2 is skipped for link sizes
287 other than 2 because of this, bypassing the problem. Unfortunately,
288 there was also a test in test 3 (the locale tests) that used /B and
289 failed for link sizes other than 2. Rather than cut the whole test out,
290 I have added a new /Z option to pcretest that replaces the length and
291 offset values with spaces. This is now used to make test 3 independent
292 of link size. (Test 2 will be tidied up later.)
293
294 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
295 segmentation fault instead of returning the appropriate error message.
296
297 15. In multiline mode when the newline sequence was set to "any", the pattern
298 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
299 This doesn't seem right; it now treats the CRLF combination as the line
300 ending, and so does not match in that case. It's only a pattern such as ^$
301 that would hit this one: something like ^ABC$ would have failed after \r
302 and then tried again after \r\n.
303
304 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
305 in an attempt to make files that differ only in their line terminators
306 compare equal. This works on Linux.
307
308 17. Under certain error circumstances pcregrep might try to free random memory
309 as it exited. This is now fixed, thanks to valgrind.
310
311 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
312 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
313 was because its rules for how to advance for /g after matching an empty
314 string at the end of a line did not allow for this case. They now check for
315 it specially.
316
317 20. pcretest is supposed to handle patterns and data of any length, by
318 extending its buffers when necessary. It was getting this wrong when the
319 buffer for a data line had to be extended.
320
321 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
322 CRLF as a newline sequence.
323
324 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
325 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
326 I have nevertheless tidied it up.
327
328 23. Added some casts to kill warnings from HP-UX ia64 compiler.
329
330 24. Added a man page for pcre-config.
331
332
333 Version 7.0 19-Dec-06
334 ---------------------
335
336 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
337 moving to gcc 4.1.1.
338
339 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
340 sys/time.h, which is documented as needed for this function. It doesn't
341 seem to matter on Linux, but it showed up on some releases of OS X.
342
343 3. It seems that there are systems where bytes whose values are greater than
344 127 match isprint() in the "C" locale. The "C" locale should be the
345 default when a C program starts up. In most systems, only ASCII printing
346 characters match isprint(). This difference caused the output from pcretest
347 to vary, making some of the tests fail. I have changed pcretest so that:
348
349 (a) When it is outputting text in the compiled version of a pattern, bytes
350 other than 32-126 are always shown as hex escapes.
351
352 (b) When it is outputting text that is a matched part of a subject string,
353 it does the same, unless a different locale has been set for the match
354 (using the /L modifier). In this case, it uses isprint() to decide.
355
356 4. Fixed a major bug that caused incorrect computation of the amount of memory
357 required for a compiled pattern when options that changed within the
358 pattern affected the logic of the preliminary scan that determines the
359 length. The relevant options are -x, and -i in UTF-8 mode. The result was
360 that the computed length was too small. The symptoms of this bug were
361 either the PCRE error "internal error: code overflow" from pcre_compile(),
362 or a glibc crash with a message such as "pcretest: free(): invalid next
363 size (fast)". Examples of patterns that provoked this bug (shown in
364 pcretest format) are:
365
366 /(?-x: )/x
367 /(?x)(?-x: \s*#\s*)/
368 /((?i)[\x{c0}])/8
369 /(?i:[\x{c0}])/8
370
371 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
372 is now done differently.
373
374 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
375 wrapper classes; (b) implement a new function in the C++ scanner that is
376 more efficient than the old way of doing things because it avoids levels of
377 recursion in the regex matching; (c) add a paragraph to the documentation
378 for the FullMatch() function.
379
380 6. The escape sequence \n was being treated as whatever was defined as
381 "newline". Not only was this contrary to the documentation, which states
382 that \n is character 10 (hex 0A), but it also went horribly wrong when
383 "newline" was defined as CRLF. This has been fixed.
384
385 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
386 was being set to -1 for the "end of line" case (supposedly a value that no
387 character can have). Though this value is never used (the check for end of
388 line is "zero bytes in current character"), it caused compiler complaints.
389 I've changed it to 0xffffffff.
390
391 8. In pcre_version.c, the version string was being built by a sequence of
392 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
393 string (as it is for production releases) called a macro with an empty
394 argument. The C standard says the result of this is undefined. The gcc
395 compiler treats it as an empty string (which was what was wanted) but it is
396 reported that Visual C gives an error. The source has been hacked around to
397 avoid this problem.
398
399 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
400 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
401 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
402 of them did).
403
404 10. Originally, pcretest opened its input and output without "b"; then I was
405 told that "b" was needed in some environments, so it was added for release
406 5.0 to both the input and output. (It makes no difference on Unix-like
407 systems.) Later I was told that it is wrong for the input on Windows. I've
408 now abstracted the modes into two macros, to make it easier to fiddle with
409 them, and removed "b" from the input mode under Windows.
410
411 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
412
413 12. Added -help and --help to pcretest as an official way of being reminded
414 of the options.
415
416 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
417 and pcrecpp.cc because they annoy compilers at high warning levels.
418
419 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
420
421 15. Fixed an occurrence of == in configure.ac that should have been = (shell
422 scripts are not C programs :-) and which was not noticed because it works
423 on Linux.
424
425 16. pcretest is supposed to handle any length of pattern and data line (as one
426 line or as a continued sequence of lines) by extending its input buffer if
427 necessary. This feature was broken for very long pattern lines, leading to
428 a string of junk being passed to pcre_compile() if the pattern was longer
429 than about 50K.
430
431 17. I have done a major re-factoring of the way pcre_compile() computes the
432 amount of memory needed for a compiled pattern. Previously, there was code
433 that made a preliminary scan of the pattern in order to do this. That was
434 OK when PCRE was new, but as the facilities have expanded, it has become
435 harder and harder to keep it in step with the real compile phase, and there
436 have been a number of bugs (see for example, 4 above). I have now found a
437 cunning way of running the real compile function in a "fake" mode that
438 enables it to compute how much memory it would need, while actually only
439 ever using a few hundred bytes of working memory and without too many
440 tests of the mode. This should make future maintenance and development
441 easier. A side effect of this work is that the limit of 200 on the nesting
442 depth of parentheses has been removed (though this was never a serious
443 limitation, I suspect). However, there is a downside: pcre_compile() now
444 runs more slowly than before (30% or more, depending on the pattern). I
445 hope this isn't a big issue. There is no effect on runtime performance.
446
447 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
448 newline (only possible for the last line of a file) and it was a
449 pattern that set a locale (followed by /Lsomething), pcretest crashed.
450
451 19. Added additional timing features to pcretest. (1) The -tm option now times
452 matching only, not compiling. (2) Both -t and -tm can be followed, as a
453 separate command line item, by a number that specifies the number of
454 repeats to use when timing. The default is 50000; this gives better
455 precision, but takes uncomfortably long for very large patterns.
456
457 20. Extended pcre_study() to be more clever in cases where a branch of a
458 subpattern has no definite first character. For example, (a*|b*)[cd] would
459 previously give no result from pcre_study(). Now it recognizes that the
460 first character must be a, b, c, or d.
461
462 21. There was an incorrect error "recursive call could loop indefinitely" if
463 a subpattern (or the entire pattern) that was being tested for matching an
464 empty string contained only one non-empty item after a nested subpattern.
465 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
466 incorrectly, because the \d was being skipped in the check.
467
468 22. The pcretest program now has a new pattern option /B and a command line
469 option -b, which is equivalent to adding /B to every pattern. This causes
470 it to show the compiled bytecode, without the additional information that
471 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
472 is the same as /B/I).
473
474 23. A new optimization is now able automatically to treat some sequences such
475 as a*b as a*+b. More specifically, if something simple (such as a character
476 or a simple class like \d) has an unlimited quantifier, and is followed by
477 something that cannot possibly match the quantified thing, the quantifier
478 is automatically "possessified".
479
480 24. A recursive reference to a subpattern whose number was greater than 39
481 went wrong under certain circumstances in UTF-8 mode. This bug could also
482 have affected the operation of pcre_study().
483
484 25. Realized that a little bit of performance could be had by replacing
485 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
486
487 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
488
489 27. Possessive quantifiers such as a++ were previously implemented by turning
490 them into atomic groups such as ($>a+). Now they have their own opcodes,
491 which improves performance. This includes the automatically created ones
492 from 23 above.
493
494 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
495 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
496 the first matched character to be a colon. This applied both to named and
497 numbered groups.
498
499 29. The ucpinternal.h header file was missing its idempotency #ifdef.
500
501 30. I was sent a "project" file called libpcre.a.dev which I understand makes
502 building PCRE on Windows easier, so I have included it in the distribution.
503
504 31. There is now a check in pcretest against a ridiculously large number being
505 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
506 loop, the loop is abandoned.
507
508 32. Forward references to subpatterns in conditions such as (?(2)...) where
509 subpattern 2 is defined later cause pcre_compile() to search forwards in
510 the pattern for the relevant set of parentheses. This search went wrong
511 when there were unescaped parentheses in a character class, parentheses
512 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
513
514 33. "Subroutine" calls and backreferences were previously restricted to
515 referencing subpatterns earlier in the regex. This restriction has now
516 been removed.
517
518 34. Added a number of extra features that are going to be in Perl 5.10. On the
519 whole, these are just syntactic alternatives for features that PCRE had
520 previously implemented using the Python syntax or my own invention. The
521 other formats are all retained for compatibility.
522
523 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
524 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
525 also .NET compatible.
526
527 (b) A recursion or subroutine call to a named group can now be defined as
528 (?&name) as well as (?P>name).
529
530 (c) A backreference to a named group can now be defined as \k<name> or
531 \k'name' as well as (?P=name). The new forms, as well as being in Perl
532 5.10, are also .NET compatible.
533
534 (d) A conditional reference to a named group can now use the syntax
535 (?(<name>) or (?('name') as well as (?(name).
536
537 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
538 groups (named and numbered) that are never evaluated inline, but can be
539 called as "subroutines" from elsewhere. In effect, the DEFINE condition
540 is always false. There may be only one alternative in such a group.
541
542 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
543 as the simple (?(R). The condition is true only if the most recent
544 recursion is that of the given number or name. It does not search out
545 through the entire recursion stack.
546
547 (g) The escape \gN or \g{N} has been added, where N is a positive or
548 negative number, specifying an absolute or relative reference.
549
550 35. Tidied to get rid of some further signed/unsigned compiler warnings and
551 some "unreachable code" warnings.
552
553 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
554 things, this adds five new scripts.
555
556 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
557 There were also incompatibilities regarding the handling of \Q..\E inside
558 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
559 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
560
561 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
562 matches an empty string, and forcibly breaks the loop. There were bugs in
563 this code in non-simple cases. For a pattern such as ^(a()*)* matched
564 against aaaa the result was just "a" rather than "aaaa", for example. Two
565 separate and independent bugs (that affected different cases) have been
566 fixed.
567
568 39. Refactored the code to abolish the use of different opcodes for small
569 capturing bracket numbers. This is a tidy that I avoided doing when I
570 removed the limit on the number of capturing brackets for 3.5 back in 2001.
571 The new approach is not only tidier, it makes it possible to reduce the
572 memory needed to fix the previous bug (38).
573
574 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
575 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
576 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
577 mode.
578
579 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
580 report.
581
582 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
583 copy construction and assignment in the C++ wrapper.
584
585 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
586 couple of bugs that could have given wrong results in the "--newline=crlf"
587 case.
588
589 44. Added a number of casts and did some reorganization of signed/unsigned int
590 variables following suggestions from Dair Grant. Also renamed the variable
591 "this" as "item" because it is a C++ keyword.
592
593 45. Arranged for dftables to add
594
595 #include "pcre_internal.h"
596
597 to pcre_chartables.c because without it, gcc 4.x may remove the array
598 definition from the final binary if PCRE is built into a static library and
599 dead code stripping is activated.
600
601 46. For an unanchored pattern, if a match attempt fails at the start of a
602 newline sequence, and the newline setting is CRLF or ANY, and the next two
603 characters are CRLF, advance by two characters instead of one.
604
605
606 Version 6.7 04-Jul-06
607 ---------------------
608
609 1. In order to handle tests when input lines are enormously long, pcretest has
610 been re-factored so that it automatically extends its buffers when
611 necessary. The code is crude, but this _is_ just a test program. The
612 default size has been increased from 32K to 50K.
613
614 2. The code in pcre_study() was using the value of the re argument before
615 testing it for NULL. (Of course, in any sensible call of the function, it
616 won't be NULL.)
617
618 3. The memmove() emulation function in pcre_internal.h, which is used on
619 systems that lack both memmove() and bcopy() - that is, hardly ever -
620 was missing a "static" storage class specifier.
621
622 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
623 containing an extended class (one that cannot be represented by a bitmap
624 because it contains high-valued characters or Unicode property items, e.g.
625 [\pZ]). Almost always one would set UTF-8 mode when processing such a
626 pattern, but PCRE should not loop if you do not (it no longer does).
627 [Detail: two cases were found: (a) a repeated subpattern containing an
628 extended class; (b) a recursive reference to a subpattern that followed a
629 previous extended class. It wasn't skipping over the extended class
630 correctly when UTF-8 mode was not set.]
631
632 5. A negated single-character class was not being recognized as fixed-length
633 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
634 compile error "lookbehind assertion is not fixed length".
635
636 6. The RunPerlTest auxiliary script was showing an unexpected difference
637 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
638 write a Perl script that can interpret lines of an input file either as
639 byte characters or as UTF-8, which is what "perltest" was being required to
640 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
641 can't do is switch easily at run time between having the "use utf8;" pragma
642 or not. In the end, I fudged it by using the RunPerlTest script to insert
643 "use utf8;" explicitly for the UTF-8 tests.
644
645 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
646 the end of the subject string, contrary to the documentation and to what
647 Perl does. This was true of both matching functions. Now it matches only at
648 the start of the subject and immediately after *internal* newlines.
649
650 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
651 a pointer to an int instead of a pointer to an unsigned long int. This
652 caused problems on 64-bit systems.
653
654 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
655 instance of the 'standard' template library not being so standard".
656
657 10. There was no check on the number of named subpatterns nor the maximum
658 length of a subpattern name. The product of these values is used to compute
659 the size of the memory block for a compiled pattern. By supplying a very
660 long subpattern name and a large number of named subpatterns, the size
661 computation could be caused to overflow. This is now prevented by limiting
662 the length of names to 32 characters, and the number of named subpatterns
663 to 10,000.
664
665 11. Subpatterns that are repeated with specific counts have to be replicated in
666 the compiled pattern. The size of memory for this was computed from the
667 length of the subpattern and the repeat count. The latter is limited to
668 65535, but there was no limit on the former, meaning that integer overflow
669 could in principle occur. The compiled length of a repeated subpattern is
670 now limited to 30,000 bytes in order to prevent this.
671
672 12. Added the optional facility to have named substrings with the same name.
673
674 13. Added the ability to use a named substring as a condition, using the
675 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
676 are numbers (not recommended). Forward references are permitted.
677
678 14. Added forward references in named backreferences (if you see what I mean).
679
680 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
681 pattern could run off the end of the subject. For example, the pattern
682 "(?s)(.{1,5})"8 did this with the subject "ab".
683
684 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
685 PCRE_CASELESS was set when matching characters that were quantified with ?
686 or *.
687
688 17. A character class other than a single negated character that had a minimum
689 but no maximum quantifier - for example [ab]{6,} - was not handled
690 correctly by pce_dfa_exec(). It would match only one character.
691
692 18. A valid (though odd) pattern that looked like a POSIX character
693 class but used an invalid character after [ (for example [[,abc,]]) caused
694 pcre_compile() to give the error "Failed: internal error: code overflow" or
695 in some cases to crash with a glibc free() error. This could even happen if
696 the pattern terminated after [[ but there just happened to be a sequence of
697 letters, a binary zero, and a closing ] in the memory that followed.
698
699 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
700 over the years. Originally (before any Unicode support), just the bottom 8
701 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
702 output from "man perlunicode" includes this:
703
704 The regular expression compiler produces polymorphic opcodes. That
705 is, the pattern adapts to the data and automatically switches to
706 the Unicode character scheme when presented with Unicode data--or
707 instead uses a traditional byte scheme when presented with byte
708 data.
709
710 Sadly, a wide octal escape does not cause a switch, and in a string with
711 no other multibyte characters, these octal escapes are treated as before.
712 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
713 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
714 Unicode string.
715
716 I have not perpetrated such confusion in PCRE. Up till now, it took just
717 the bottom 8 bits, as in old Perl. I have now made octal escapes with
718 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
719 translate to the appropriate multibyte character.
720
721 29. Applied some refactoring to reduce the number of warnings from Microsoft
722 and Borland compilers. This has included removing the fudge introduced
723 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
724 a warning about an unused variable.
725
726 21. PCRE has not included VT (character 0x0b) in the set of whitespace
727 characters since release 4.0, because Perl (from release 5.004) does not.
728 [Or at least, is documented not to: some releases seem to be in conflict
729 with the documentation.] However, when a pattern was studied with
730 pcre_study() and all its branches started with \s, PCRE still included VT
731 as a possible starting character. Of course, this did no harm; it just
732 caused an unnecessary match attempt.
733
734 22. Removed a now-redundant internal flag bit that recorded the fact that case
735 dependency changed within the pattern. This was once needed for "required
736 byte" processing, but is no longer used. This recovers a now-scarce options
737 bit. Also moved the least significant internal flag bit to the most-
738 significant bit of the word, which was not previously used (hangover from
739 the days when it was an int rather than a uint) to free up another bit for
740 the future.
741
742 23. Added support for CRLF line endings as well as CR and LF. As well as the
743 default being selectable at build time, it can now be changed at runtime
744 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
745 specify that it is scanning data with non-default line endings.
746
747 24. Changed the definition of CXXLINK to make it agree with the definition of
748 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
749
750 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
751 recursions. This makes a big different to stack usage for some patterns.
752
753 26. If a subpattern containing a named recursion or subroutine reference such
754 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
755 the space required for the compiled pattern went wrong and gave too small a
756 value. Depending on the environment, this could lead to "Failed: internal
757 error: code overflow at offset 49" or "glibc detected double free or
758 corruption" errors.
759
760 27. Applied patches from Google (a) to support the new newline modes and (b) to
761 advance over multibyte UTF-8 characters in GlobalReplace.
762
763 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
764 difference for some implementation of PCRE in some Windows version.
765
766 29. Added some extra testing facilities to pcretest:
767
768 \q<number> in a data line sets the "match limit" value
769 \Q<number> in a data line sets the "match recursion limt" value
770 -S <number> sets the stack size, where <number> is in megabytes
771
772 The -S option isn't available for Windows.
773
774
775 Version 6.6 06-Feb-06
776 ---------------------
777
778 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
779 in pcreposix.h. I have copied the definition from pcre.h.
780
781 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
782 because pcre.h is no longer a built file.
783
784 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
785 not normally included in the compiled code.
786
787
788 Version 6.5 01-Feb-06
789 ---------------------
790
791 1. When using the partial match feature with pcre_dfa_exec(), it was not
792 anchoring the second and subsequent partial matches at the new starting
793 point. This could lead to incorrect results. For example, with the pattern
794 /1234/, partially matching against "123" and then "a4" gave a match.
795
796 2. Changes to pcregrep:
797
798 (a) All non-match returns from pcre_exec() were being treated as failures
799 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
800 error message is output. Some extra information is given for the
801 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
802 probably the only errors that are likely to be caused by users (by
803 specifying a regex that has nested indefinite repeats, for instance).
804 If there are more than 20 of these errors, pcregrep is abandoned.
805
806 (b) A binary zero was treated as data while matching, but terminated the
807 output line if it was written out. This has been fixed: binary zeroes
808 are now no different to any other data bytes.
809
810 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
811 used to set a locale for matching. The --locale=xxxx long option has
812 been added (no short equivalent) to specify a locale explicitly on the
813 pcregrep command, overriding the environment variables.
814
815 (d) When -B was used with -n, some line numbers in the output were one less
816 than they should have been.
817
818 (e) Added the -o (--only-matching) option.
819
820 (f) If -A or -C was used with -c (count only), some lines of context were
821 accidentally printed for the final match.
822
823 (g) Added the -H (--with-filename) option.
824
825 (h) The combination of options -rh failed to suppress file names for files
826 that were found from directory arguments.
827
828 (i) Added the -D (--devices) and -d (--directories) options.
829
830 (j) Added the -F (--fixed-strings) option.
831
832 (k) Allow "-" to be used as a file name for -f as well as for a data file.
833
834 (l) Added the --colo(u)r option.
835
836 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
837 is not present by default.
838
839 3. A nasty bug was discovered in the handling of recursive patterns, that is,
840 items such as (?R) or (?1), when the recursion could match a number of
841 alternatives. If it matched one of the alternatives, but subsequently,
842 outside the recursion, there was a failure, the code tried to back up into
843 the recursion. However, because of the way PCRE is implemented, this is not
844 possible, and the result was an incorrect result from the match.
845
846 In order to prevent this happening, the specification of recursion has
847 been changed so that all such subpatterns are automatically treated as
848 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
849
850 4. I had overlooked the fact that, in some locales, there are characters for
851 which isalpha() is true but neither isupper() nor islower() are true. In
852 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
853 and ordfeminine) are like this. This affected the treatment of \w and \W
854 when they appeared in character classes, but not when they appeared outside
855 a character class. The bit map for "word" characters is now created
856 separately from the results of isalnum() instead of just taking it from the
857 upper, lower, and digit maps. (Plus the underscore character, of course.)
858
859 5. The above bug also affected the handling of POSIX character classes such as
860 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
861 permanent tables. Instead, the bit maps for such a class were previously
862 created as the appropriate unions of the upper, lower, and digit bitmaps.
863 Now they are created by subtraction from the [[:word:]] class, which has
864 its own bitmap.
865
866 6. The [[:blank:]] character class matches horizontal, but not vertical space.
867 It is created by subtracting the vertical space characters (\x09, \x0a,
868 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
869 subtraction was done in the overall bitmap for a character class, meaning
870 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
871 be recognized. This bug has been fixed.
872
873 7. Patches from the folks at Google:
874
875 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
876 real life, but is still worth protecting against".
877
878 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
879 regular expressions".
880
881 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
882 have it.
883
884 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
885 "configure" and the latter not, in order to fix a problem somebody had
886 with compiling the Arg class on HP-UX.
887
888 (e) Improve the error-handling of the C++ wrapper a little bit.
889
890 (f) New tests for checking recursion limiting.
891
892 8. The pcre_memmove() function, which is used only if the environment does not
893 have a standard memmove() function (and is therefore rarely compiled),
894 contained two bugs: (a) use of int instead of size_t, and (b) it was not
895 returning a result (though PCRE never actually uses the result).
896
897 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
898 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
899 returned instead of calling malloc() with an overflowing number that would
900 most likely cause subsequent chaos.
901
902 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
903
904 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
905 with this option is matched, the nmatch and pmatch options of regexec() are
906 ignored.
907
908 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
909 provided in case anyone wants to the the POSIX interface with UTF-8
910 strings.
911
912 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
913 C++ linking (needed for some HP-UX environments).
914
915 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
916 (unused parameter) and in the pcre_printint() function (omitted "default"
917 switch label when the default is to do nothing).
918
919 15. Added some code to make it possible, when PCRE is compiled as a C++
920 library, to replace subject pointers for pcre_exec() with a smart pointer
921 class, thus making it possible to process discontinuous strings.
922
923 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
924 much the same function. They were added by different people who were trying
925 to make PCRE easy to compile on non-Unix systems. It has been suggested
926 that PCRE_EXPORT be abolished now that there is more automatic apparatus
927 for compiling on Windows systems. I have therefore replaced it with
928 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
929 defaults to "extern" for C or "extern C" for C++, which works fine on
930 Unix-like systems. It is now possible to override the value of PCRE_DATA_
931 SCOPE with something explicit in config.h. In addition:
932
933 (a) pcreposix.h still had just "extern" instead of either of these macros;
934 I have replaced it with PCRE_DATA_SCOPE.
935
936 (b) Functions such as _pcre_xclass(), which are internal to the library,
937 but external in the C sense, all had PCRE_EXPORT in their definitions.
938 This is apparently wrong for the Windows case, so I have removed it.
939 (It makes no difference on Unix-like systems.)
940
941 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
942 of recursive calls to match(). This is different to MATCH_LIMIT because
943 that limits the total number of calls to match(), not all of which increase
944 the depth of recursion. Limiting the recursion depth limits the amount of
945 stack (or heap if NO_RECURSE is set) that is used. The default can be set
946 when PCRE is compiled, and changed at run time. A patch from Google adds
947 this functionality to the C++ interface.
948
949 18. Changes to the handling of Unicode character properties:
950
951 (a) Updated the table to Unicode 4.1.0.
952
953 (b) Recognize characters that are not in the table as "Cn" (undefined).
954
955 (c) I revised the way the table is implemented to a much improved format
956 which includes recognition of ranges. It now supports the ranges that
957 are defined in UnicodeData.txt, and it also amalgamates other
958 characters into ranges. This has reduced the number of entries in the
959 table from around 16,000 to around 3,000, thus reducing its size
960 considerably. I realized I did not need to use a tree structure after
961 all - a binary chop search is just as efficient. Having reduced the
962 number of entries, I extended their size from 6 bytes to 8 bytes to
963 allow for more data.
964
965 (d) Added support for Unicode script names via properties such as \p{Han}.
966
967 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
968 matching that character.
969
970 20. When matching a repeated Unicode property with a minimum greater than zero,
971 (for example \pL{2,}), PCRE could look past the end of the subject if it
972 reached it while seeking the minimum number of characters. This could
973 happen only if some of the characters were more than one byte long, because
974 there is a check for at least the minimum number of bytes.
975
976 21. Refactored the implementation of \p and \P so as to be more general, to
977 allow for more different types of property in future. This has changed the
978 compiled form incompatibly. Anybody with saved compiled patterns that use
979 \p or \P will have to recompile them.
980
981 22. Added "Any" and "L&" to the supported property types.
982
983 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
984 but give a compile time error if the value is greater than 0xff.
985
986 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
987 accidentally not being installed or uninstalled.
988
989 25. The pcre.h file was built from pcre.h.in, but the only changes that were
990 made were to insert the current release number. This seemed silly, because
991 it made things harder for people building PCRE on systems that don't run
992 "configure". I have turned pcre.h into a distributed file, no longer built
993 by "configure", with the version identification directly included. There is
994 no longer a pcre.h.in file.
995
996 However, this change necessitated a change to the pcre-config script as
997 well. It is built from pcre-config.in, and one of the substitutions was the
998 release number. I have updated configure.ac so that ./configure now finds
999 the release number by grepping pcre.h.
1000
1001 26. Added the ability to run the tests under valgrind.
1002
1003
1004 Version 6.4 05-Sep-05
1005 ---------------------
1006
1007 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1008 "--" to be printed when multiple files were scanned, even when none of the
1009 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1010 consider it to be a bug, and have restored the previous behaviour.
1011
1012 2. A couple of code tidies to get rid of compiler warnings.
1013
1014 3. The pcretest program used to cheat by referring to symbols in the library
1015 whose names begin with _pcre_. These are internal symbols that are not
1016 really supposed to be visible externally, and in some environments it is
1017 possible to suppress them. The cheating is now confined to including
1018 certain files from the library's source, which is a bit cleaner.
1019
1020 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1021 file's purpose clearer.
1022
1023 5. Reorganized pcre_ucp_findchar().
1024
1025
1026 Version 6.3 15-Aug-05
1027 ---------------------
1028
1029 1. The file libpcre.pc.in did not have general read permission in the tarball.
1030
1031 2. There were some problems when building without C++ support:
1032
1033 (a) If C++ support was not built, "make install" and "make test" still
1034 tried to test it.
1035
1036 (b) There were problems when the value of CXX was explicitly set. Some
1037 changes have been made to try to fix these, and ...
1038
1039 (c) --disable-cpp can now be used to explicitly disable C++ support.
1040
1041 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1042 backslash in a target when C++ was disabled. This confuses some
1043 versions of "make", apparently. Using an intermediate variable solves
1044 this. (Same for CPP_LOBJ.)
1045
1046 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1047 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1048 necessary on certain architectures.
1049
1050 4. Added a setting of -export-symbols-regex to the link command to remove
1051 those symbols that are exported in the C sense, but actually are local
1052 within the library, and not documented. Their names all begin with
1053 "_pcre_". This is not a perfect job, because (a) we have to except some
1054 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1055 available (and never for static libraries). I have made a note to try to
1056 find a way round (a) in the future.
1057
1058
1059 Version 6.2 01-Aug-05
1060 ---------------------
1061
1062 1. There was no test for integer overflow of quantifier values. A construction
1063 such as {1111111111111111} would give undefined results. What is worse, if
1064 a minimum quantifier for a parenthesized subpattern overflowed and became
1065 negative, the calculation of the memory size went wrong. This could have
1066 led to memory overwriting.
1067
1068 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1069
1070 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1071 operating environments where this matters.
1072
1073 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1074 PCRE options from within the C++ wrapper.
1075
1076 5. Named capturing subpatterns were not being correctly counted when a pattern
1077 was compiled. This caused two problems: (a) If there were more than 100
1078 such subpatterns, the calculation of the memory needed for the whole
1079 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1080 back references of the form \12, where the number was greater than 9, were
1081 not recognized as back references, even though there were sufficient
1082 previous subpatterns.
1083
1084 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1085 versions of gcc, e.g. 2.95.4.
1086
1087
1088 Version 6.1 21-Jun-05
1089 ---------------------
1090
1091 1. There was one reference to the variable "posix" in pcretest.c that was not
1092 surrounded by "#if !defined NOPOSIX".
1093
1094 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1095 the cross-check on the old pcre_info() function, for the benefit of the
1096 cut-down version of PCRE that is currently imported into Exim.
1097
1098 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1099 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1100 patterns that start (?i) but I don't think that matters. The use of (?i) is
1101 just an example; this all applies to the other options as well.
1102
1103 4. Since libtool seems to echo the compile commands it is issuing, the output
1104 from "make" can be reduced a bit by putting "@" in front of each libtool
1105 compile command.
1106
1107 5. Patch from the folks at Google for configure.in to be a bit more thorough
1108 in checking for a suitable C++ installation before trying to compile the
1109 C++ stuff. This should fix a reported problem when a compiler was present,
1110 but no suitable headers.
1111
1112 6. The man pages all had just "PCRE" as their title. I have changed them to
1113 be the relevant file name. I have also arranged that these names are
1114 retained in the file doc/pcre.txt, which is a concatenation in text format
1115 of all the man pages except the little individual ones for each function.
1116
1117 7. The NON-UNIX-USE file had not been updated for the different set of source
1118 files that come with release 6. I also added a few comments about the C++
1119 wrapper.
1120
1121
1122 Version 6.0 07-Jun-05
1123 ---------------------
1124
1125 1. Some minor internal re-organization to help with my DFA experiments.
1126
1127 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1128 didn't matter for the library itself when fully configured, but did matter
1129 when compiling without UCP support, or within Exim, where the ucp files are
1130 not imported.
1131
1132 3. Refactoring of the library code to split up the various functions into
1133 different source modules. The addition of the new DFA matching code (see
1134 below) to a single monolithic source would have made it really too
1135 unwieldy, quite apart from causing all the code to be include in a
1136 statically linked application, when only some functions are used. This is
1137 relevant even without the DFA addition now that patterns can be compiled in
1138 one application and matched in another.
1139
1140 The downside of splitting up is that there have to be some external
1141 functions and data tables that are used internally in different modules of
1142 the library but which are not part of the API. These have all had their
1143 names changed to start with "_pcre_" so that they are unlikely to clash
1144 with other external names.
1145
1146 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1147 a different (DFA) algorithm. Although it is slower than the original
1148 function, it does have some advantages for certain types of matching
1149 problem.
1150
1151 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1152 including restarting after a partial match.
1153
1154 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1155 defined when compiling for Windows was sent to me. I have put it into the
1156 code, though I have no means of testing or verifying it.
1157
1158 7. Added the pcre_refcount() auxiliary function.
1159
1160 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1161 match before or at the first newline in the subject string. In pcretest,
1162 the /f option on a pattern can be used to set this.
1163
1164 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1165 would behave wrongly. This has been present in PCRE since release 4.0.
1166
1167 10. A number of changes to the pcregrep command:
1168
1169 (a) Refactored how -x works; insert ^(...)$ instead of setting
1170 PCRE_ANCHORED and checking the length, in preparation for adding
1171 something similar for -w.
1172
1173 (b) Added the -w (match as a word) option.
1174
1175 (c) Refactored the way lines are read and buffered so as to have more
1176 than one at a time available.
1177
1178 (d) Implemented a pcregrep test script.
1179
1180 (e) Added the -M (multiline match) option. This allows patterns to match
1181 over several lines of the subject. The buffering ensures that at least
1182 8K, or the rest of the document (whichever is the shorter) is available
1183 for matching (and similarly the previous 8K for lookbehind assertions).
1184
1185 (f) Changed the --help output so that it now says
1186
1187 -w, --word-regex(p)
1188
1189 instead of two lines, one with "regex" and the other with "regexp"
1190 because that confused at least one person since the short forms are the
1191 same. (This required a bit of code, as the output is generated
1192 automatically from a table. It wasn't just a text change.)
1193
1194 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1195 option but starts with a hyphen. Could be a pattern or a path name
1196 starting with a hyphen, for instance.
1197
1198 (h) "-" can be given as a file name to represent stdin.
1199
1200 (i) When file names are being printed, "(standard input)" is used for
1201 the standard input, for compatibility with GNU grep. Previously
1202 "<stdin>" was used.
1203
1204 (j) The option --label=xxx can be used to supply a name to be used for
1205 stdin when file names are being printed. There is no short form.
1206
1207 (k) Re-factored the options decoding logic because we are going to add
1208 two more options that take data. Such options can now be given in four
1209 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1210
1211 (l) Added the -A, -B, and -C options for requesting that lines of context
1212 around matches be printed.
1213
1214 (m) Added the -L option to print the names of files that do not contain
1215 any matching lines, that is, the complement of -l.
1216
1217 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1218 continue to scan other files.
1219
1220 (o) The -s option was incorrectly implemented. For compatibility with other
1221 greps, it now suppresses the error message for a non-existent or non-
1222 accessible file (but not the return code). There is a new option called
1223 -q that suppresses the output of matching lines, which was what -s was
1224 previously doing.
1225
1226 (p) Added --include and --exclude options to specify files for inclusion
1227 and exclusion when recursing.
1228
1229 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1230 Hopefully, it now does.
1231
1232 12. Missing cast in pcre_study().
1233
1234 13. Added an "uninstall" target to the makefile.
1235
1236 14. Replaced "extern" in the function prototypes in Makefile.in with
1237 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1238 world, but is set differently for Windows.
1239
1240 15. Added a second compiling function called pcre_compile2(). The only
1241 difference is that it has an extra argument, which is a pointer to an
1242 integer error code. When there is a compile-time failure, this is set
1243 non-zero, in addition to the error test pointer being set to point to an
1244 error message. The new argument may be NULL if no error number is required
1245 (but then you may as well call pcre_compile(), which is now just a
1246 wrapper). This facility is provided because some applications need a
1247 numeric error indication, but it has also enabled me to tidy up the way
1248 compile-time errors are handled in the POSIX wrapper.
1249
1250 16. Added VPATH=.libs to the makefile; this should help when building with one
1251 prefix path and installing with another. (Or so I'm told by someone who
1252 knows more about this stuff than I do.)
1253
1254 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1255 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1256 match everything, including newlines. This is not POSIX-compatible, but
1257 somebody wanted the feature. From pcretest it can be activated by using
1258 both the P and the s flags.
1259
1260 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1261
1262 19. libpcre.pc was being incorrectly installed as executable.
1263
1264 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1265 it now also looks for '\r' so that it will work unmodified on Windows.
1266
1267 21. Added Google's contributed C++ wrapper to the distribution.
1268
1269 22. Added some untidy missing memory free() calls in pcretest, to keep
1270 Electric Fence happy when testing.
1271
1272
1273
1274 Version 5.0 13-Sep-04
1275 ---------------------
1276
1277 1. Internal change: literal characters are no longer packed up into items
1278 containing multiple characters in a single byte-string. Each character
1279 is now matched using a separate opcode. However, there may be more than one
1280 byte in the character in UTF-8 mode.
1281
1282 2. The pcre_callout_block structure has two new fields: pattern_position and
1283 next_item_length. These contain the offset in the pattern to the next match
1284 item, and its length, respectively.
1285
1286 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
1287 insertion of callouts before each pattern item. Added the /C option to
1288 pcretest to make use of this.
1289
1290 4. On the advice of a Windows user, the lines
1291
1292 #if defined(_WIN32) || defined(WIN32)
1293 _setmode( _fileno( stdout ), 0x8000 );
1294 #endif /* defined(_WIN32) || defined(WIN32) */
1295
1296 have been added to the source of pcretest. This apparently does useful
1297 magic in relation to line terminators.
1298
1299 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
1300 for the benefit of those environments where the "b" makes a difference.
1301
1302 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
1303 to know about it. I have put a hack into configure.in that adds in code
1304 to set GCC=yes if CC=icc. This seems to end up at a point in the
1305 generated configure script that is early enough to affect the setting of
1306 compiler options, which is what is needed, but I have no means of testing
1307 whether it really works. (The user who reported this had patched the
1308 generated configure script, which of course I cannot do.)
1309
1310 LATER: After change 22 below (new libtool files), the configure script
1311 seems to know about icc (and also ecc). Therefore, I have commented out
1312 this hack in configure.in.
1313
1314 7. Added support for pkg-config (2 patches were sent in).
1315
1316 8. Negated POSIX character classes that used a combination of internal tables
1317 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
1318 [[:^ascii]]. Typically, they would match almost any characters. The other
1319 POSIX classes were not broken in this way.
1320
1321 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
1322 to find the match, as PCRE was deluded into thinking that the match had to
1323 start at the start point or following a newline. The same bug applied to
1324 patterns with negative forward assertions or any backward assertions
1325 preceding ".*" at the start, unless the pattern required a fixed first
1326 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
1327
1328 10. In UTF-8 mode, when moving forwards in the subject after a failed match
1329 starting at the last subject character, bytes beyond the end of the subject
1330 string were read.
1331
1332 11. Renamed the variable "class" as "classbits" to make life easier for C++
1333 users. (Previously there was a macro definition, but it apparently wasn't
1334 enough.)
1335
1336 12. Added the new field "tables" to the extra data so that tables can be passed
1337 in at exec time, or the internal tables can be re-selected. This allows
1338 a compiled regex to be saved and re-used at a later time by a different
1339 program that might have everything at different addresses.
1340
1341 13. Modified the pcre-config script so that, when run on Solaris, it shows a
1342 -R library as well as a -L library.
1343
1344 14. The debugging options of pcretest (-d on the command line or D on a
1345 pattern) showed incorrect output for anything following an extended class
1346 that contained multibyte characters and which was followed by a quantifier.
1347
1348 15. Added optional support for general category Unicode character properties
1349 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
1350 support. It adds about 90K to the size of the library. The meanings of the
1351 inbuilt class escapes such as \d and \s have NOT been changed.
1352
1353 16. Updated pcredemo.c to include calls to free() to release the memory for the
1354 compiled pattern.
1355
1356 17. The generated file chartables.c was being created in the source directory
1357 instead of in the building directory. This caused the build to fail if the
1358 source directory was different from the building directory, and was
1359 read-only.
1360
1361 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
1362 file. No doubt somebody will tell me if they don't make sense... Also added
1363 Dan Mooney's comments about building on OpenVMS.
1364
1365 19. Added support for partial matching via the PCRE_PARTIAL option for
1366 pcre_exec() and the \P data escape in pcretest.
1367
1368 20. Extended pcretest with 3 new pattern features:
1369
1370 (i) A pattern option of the form ">rest-of-line" causes pcretest to
1371 write the compiled pattern to the file whose name is "rest-of-line".
1372 This is a straight binary dump of the data, with the saved pointer to
1373 the character tables forced to be NULL. The study data, if any, is
1374 written too. After writing, pcretest reads a new pattern.
1375
1376 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
1377 compiled pattern from the given file. There must not be any
1378 occurrences of "<" in the file name (pretty unlikely); if there are,
1379 pcretest will instead treat the initial "<" as a pattern delimiter.
1380 After reading in the pattern, pcretest goes on to read data lines as
1381 usual.
1382
1383 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
1384 and 16-bit fields in a compiled pattern, to simulate a pattern that
1385 was compiled on a host of opposite endianness.
1386
1387 21. The pcre-exec() function can now cope with patterns that were compiled on
1388 hosts of opposite endianness, with this restriction:
1389
1390 As for any compiled expression that is saved and used later, the tables
1391 pointer field cannot be preserved; the extra_data field in the arguments
1392 to pcre_exec() should be used to pass in a tables address if a value
1393 other than the default internal tables were used at compile time.
1394
1395 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
1396 now diagnosed as an error. Previously, most of the time, a negative number
1397 would have been treated as zero, but if in addition "ovector" was passed as
1398 NULL, a crash could occur.
1399
1400 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
1401 new versions from the libtool 1.5 distribution (the last one is a copy of
1402 a file called libtool.m4). This seems to have fixed the need to patch
1403 "configure" to support Darwin 1.3 (which I used to do). However, I still
1404 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
1405 workstation).
1406
1407 24. Changed the PCRE licence to be the more standard "BSD" licence.
1408
1409
1410 Version 4.5 01-Dec-03
1411 ---------------------
1412
1413 1. There has been some re-arrangement of the code for the match() function so
1414 that it can be compiled in a version that does not call itself recursively.
1415 Instead, it keeps those local variables that need separate instances for
1416 each "recursion" in a frame on the heap, and gets/frees frames whenever it
1417 needs to "recurse". Keeping track of where control must go is done by means
1418 of setjmp/longjmp. The whole thing is implemented by a set of macros that
1419 hide most of the details from the main code, and operates only if
1420 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
1421 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
1422 operating.
1423
1424 To make it easier for callers to provide specially tailored get/free
1425 functions for this usage, two new functions, pcre_stack_malloc, and
1426 pcre_stack_free, are used. They are always called in strict stacking order,
1427 and the size of block requested is always the same.
1428
1429 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
1430 PCRE has been compiled to use the stack or the heap for recursion. The
1431 -C option of pcretest uses this to show which version is compiled.
1432
1433 A new data escape \S, is added to pcretest; it causes the amounts of store
1434 obtained and freed by both kinds of malloc/free at match time to be added
1435 to the output.
1436
1437 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
1438 what's available on my current Linux desktop machine.
1439
1440 3. When matching a UTF-8 string, the test for a valid string at the start has
1441 been extended. If start_offset is not zero, PCRE now checks that it points
1442 to a byte that is the start of a UTF-8 character. If not, it returns
1443 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
1444 this is necessary because there may be backward assertions in the pattern.
1445 When matching the same subject several times, it may save resources to use
1446 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
1447
1448 4. The code for checking the validity of UTF-8 strings has been tightened so
1449 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
1450 containing "overlong sequences".
1451
1452 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
1453 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
1454 should have been "&", but it just so happened that all the cases this let
1455 through by mistake were picked up later in the function.
1456
1457 6. I had used a variable called "isblank" - this is a C99 function, causing
1458 some compilers to warn. To avoid this, I renamed it (as "blankclass").
1459
1460 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
1461 prompting; (b) run "./pcretest /dev/null" at the start of the test script
1462 so the version is shown; (c) stop "make test" echoing "./RunTest".
1463
1464 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
1465
1466 9. The prototype for memmove() for systems that don't have it was using
1467 size_t, but the inclusion of the header that defines size_t was later. I've
1468 moved the #includes for the C headers earlier to avoid this.
1469
1470 10. Added some adjustments to the code to make it easier to compiler on certain
1471 special systems:
1472
1473 (a) Some "const" qualifiers were missing.
1474 (b) Added the macro EXPORT before all exported functions; by default this
1475 is defined to be empty.
1476 (c) Changed the dftables auxiliary program (that builds chartables.c) so
1477 that it reads its output file name as an argument instead of writing
1478 to the standard output and assuming this can be redirected.
1479
1480 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
1481 class containing characters with values greater than 255, PCRE compilation
1482 went into a loop.
1483
1484 12. A recursive reference to a subpattern that was within another subpattern
1485 that had a minimum quantifier of zero caused PCRE to crash. For example,
1486 (x(y(?2))z)? provoked this bug with a subject that got as far as the
1487 recursion. If the recursively-called subpattern itself had a zero repeat,
1488 that was OK.
1489
1490 13. In pcretest, the buffer for reading a data line was set at 30K, but the
1491 buffer into which it was copied (for escape processing) was still set at
1492 1024, so long lines caused crashes.
1493
1494 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
1495 "internal error: code overflow...". This applied to any character class
1496 that was followed by a possessive quantifier.
1497
1498 15. Modified the Makefile to add libpcre.la as a prerequisite for
1499 libpcreposix.la because I was told this is needed for a parallel build to
1500 work.
1501
1502 16. If a pattern that contained .* following optional items at the start was
1503 studied, the wrong optimizing data was generated, leading to matching
1504 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
1505 matching string must start with a or b or c. The correct conclusion for
1506 this pattern is that a match can start with any character.
1507
1508
1509 Version 4.4 13-Aug-03
1510 ---------------------
1511
1512 1. In UTF-8 mode, a character class containing characters with values between
1513 127 and 255 was not handled correctly if the compiled pattern was studied.
1514 In fixing this, I have also improved the studying algorithm for such
1515 classes (slightly).
1516
1517 2. Three internal functions had redundant arguments passed to them. Removal
1518 might give a very teeny performance improvement.
1519
1520 3. Documentation bug: the value of the capture_top field in a callout is *one
1521 more than* the number of the hightest numbered captured substring.
1522
1523 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
1524 in incorrectly linking with a previously installed version. They now link
1525 explicitly with libpcre.la.
1526
1527 5. configure.in no longer needs to recognize Cygwin specially.
1528
1529 6. A problem in pcre.in for Windows platforms is fixed.
1530
1531 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
1532 pcretest, it used to include the size of the study block as part of its
1533 output. Unfortunately, the structure contains a field that has a different
1534 size on different hardware architectures. This meant that the tests that
1535 showed this size failed. As the block is currently always of a fixed size,
1536 this information isn't actually particularly useful in pcretest output, so
1537 I have just removed it.
1538
1539 8. Three pre-processor statements accidentally did not start in column 1.
1540 Sadly, there are *still* compilers around that complain, even though
1541 standard C has not required this for well over a decade. Sigh.
1542
1543 9. In pcretest, the code for checking callouts passed small integers in the
1544 callout_data field, which is a void * field. However, some picky compilers
1545 complained about the casts involved for this on 64-bit systems. Now
1546 pcretest passes the address of the small integer instead, which should get
1547 rid of the warnings.
1548
1549 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
1550 both compile and run time, and gives an error if an invalid UTF-8 sequence
1551 is found. There is a option for disabling this check in cases where the
1552 string is known to be correct and/or the maximum performance is wanted.
1553
1554 11. In response to a bug report, I changed one line in Makefile.in from
1555
1556 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
1557 to
1558 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
1559
1560 to look similar to other lines, but I have no way of telling whether this
1561 is the right thing to do, as I do not use Windows. No doubt I'll get told
1562 if it's wrong...
1563
1564
1565 Version 4.3 21-May-03
1566 ---------------------
1567
1568 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
1569 Makefile.
1570
1571 2. Some refactoring to improve the quality of the code:
1572
1573 (i) The utf8_table... variables are now declared "const".
1574
1575 (ii) The code for \cx, which used the "case flipping" table to upper case
1576 lower case letters, now just substracts 32. This is ASCII-specific,
1577 but the whole concept of \cx is ASCII-specific, so it seems
1578 reasonable.
1579
1580 (iii) PCRE was using its character types table to recognize decimal and
1581 hexadecimal digits in the pattern. This is silly, because it handles
1582 only 0-9, a-f, and A-F, but the character types table is locale-
1583 specific, which means strange things might happen. A private
1584 table is now used for this - though it costs 256 bytes, a table is
1585 much faster than multiple explicit tests. Of course, the standard
1586 character types table is still used for matching digits in subject
1587 strings against \d.
1588
1589 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
1590 ending in _t are). So I've renamed it as ESC_tee.
1591
1592 3. The first argument for regexec() in the POSIX wrapper should have been
1593 defined as "const".
1594
1595 4. Changed pcretest to use malloc() for its buffers so that they can be
1596 Electric Fenced for debugging.
1597
1598 5. There were several places in the code where, in UTF-8 mode, PCRE would try
1599 to read one or more bytes before the start of the subject string. Often this
1600 had no effect on PCRE's behaviour, but in some circumstances it could
1601 provoke a segmentation fault.
1602
1603 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
1604 to try to read one or more bytes before the start of the subject string.
1605
1606 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
1607 UTF-8 support could misbehave in various ways if the subject string
1608 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
1609 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
1610 back over UTF-8 characters.)
1611
1612
1613 Version 4.2 14-Apr-03
1614 ---------------------
1615
1616 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
1617
1618 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
1619 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
1620 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
1621 [WIN_PREFIX]: new variable, "cyg" for Cygwin
1622 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
1623 and BUILD_EXEEXT
1624 Note: automatic setting of the BUILD variables is not yet working
1625 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
1626 compile-time but not at link-time
1627 [LINK]: use for linking executables only
1628 make different versions for Windows and non-Windows
1629 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
1630 libraries
1631 [LINK_FOR_BUILD]: new variable
1632 [OBJEXT]: use throughout
1633 [EXEEXT]: use throughout
1634 <winshared>: new target
1635 <wininstall>: new target
1636 <dftables.o>: use native compiler
1637 <dftables>: use native linker
1638 <install>: handle Windows platform correctly
1639 <clean>: ditto
1640 <check>: ditto
1641 copy DLL to top builddir before testing
1642
1643 As part of these changes, -no-undefined was removed again. This was reported
1644 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
1645 in any case.
1646
1647 3. Some tidies to get rid of compiler warnings:
1648
1649 . In the match_data structure, match_limit was an unsigned long int, whereas
1650 match_call_count was an int. I've made them both unsigned long ints.
1651
1652 . In pcretest the fact that a const uschar * doesn't automatically cast to
1653 a void * provoked a warning.
1654
1655 . Turning on some more compiler warnings threw up some "shadow" variables
1656 and a few more missing casts.
1657
1658 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
1659 option, a class that contained a single character with a value between 128
1660 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
1661
1662 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
1663 option, a class that contained several characters, but with at least one
1664 whose value was between 128 and 255 caused PCRE to crash.
1665
1666
1667 Version 4.1 12-Mar-03
1668 ---------------------
1669
1670 1. Compiling with gcc -pedantic found a couple of places where casts were
1671 needed, and a string in dftables.c that was longer than standard compilers are
1672 required to support.
1673
1674 2. Compiling with Sun's compiler found a few more places where the code could
1675 be tidied up in order to avoid warnings.
1676
1677 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
1678 first of these names is deprecated in the latest Autoconf in favour of the name
1679 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
1680 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
1681 analogy I have changed it to CFLAGS_FOR_BUILD.
1682
1683 4. Added -no-undefined to the linking command in the Makefile, because this is
1684 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
1685 linking step for the pcreposix library.
1686
1687 5. PCRE was failing to diagnose the case of two named groups with the same
1688 name.
1689
1690 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
1691 literal character that is needed in the subject for a match, and scans along to
1692 ensure that it is present before embarking on the full matching process. This
1693 saves time in cases of nested unlimited repeats that are never going to match.
1694 Problem: the scan can take a lot of time if the subject is very long (e.g.
1695 megabytes), thus penalizing straightforward matches. It is now done only if the
1696 amount of subject to be scanned is less than 1000 bytes.
1697
1698 7. A lesser problem with the same optimization is that it was recording the
1699 first character of an anchored pattern as "needed", thus provoking a search
1700 right along the subject, even when the first match of the pattern was going to
1701 fail. The "needed" character is now not set for anchored patterns, unless it
1702 follows something in the pattern that is of non-fixed length. Thus, it still
1703 fulfils its original purpose of finding quick non-matches in cases of nested
1704 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
1705
1706
1707 Version 4.0 17-Feb-03
1708 ---------------------
1709
1710 1. If a comment in an extended regex that started immediately after a meta-item
1711 extended to the end of string, PCRE compiled incorrect data. This could lead to
1712 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
1713
1714 2. Moved to autoconf 2.53 and libtool 1.4.2.
1715
1716 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
1717 the special perltest8 script is no longer needed - all the tests can be run
1718 from a single perltest script.
1719
1720 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
1721 by \s. It has now been removed in PCRE. This means it isn't recognized as
1722 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
1723 class [:space:] *does* include VT, thereby creating a mess.
1724
1725 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
1726 space and tab.
1727
1728 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
1729 its new features into the main test script, reducing the number of scripts.
1730
1731 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
1732 were backward compatible, and made the (?i) apply to the whole pattern, as if
1733 /i were given. Now it behaves more logically, and applies the option setting
1734 only to what follows. PCRE has been changed to follow suit. However, if it
1735 finds options settings right at the start of the pattern, it extracts them into
1736 the global options, as before. Thus, they show up in the info data.
1737
1738 8. Added support for the \Q...\E escape sequence. Characters in between are
1739 treated as literals. This is slightly different from Perl in that $ and @ are
1740 also handled as literals inside the quotes. In Perl, they will cause variable
1741 interpolation. Note the following examples:
1742
1743 Pattern PCRE matches Perl matches
1744
1745 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
1746 \Qabc\$xyz\E abc\$xyz abc\$xyz
1747 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
1748
1749 For compatibility with Perl, \Q...\E sequences are recognized inside character
1750 classes as well as outside them.
1751
1752 9. Re-organized 3 code statements in pcretest to avoid "overflow in
1753 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
1754 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
1755 signed/unsigned warnings.
1756
1757 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
1758 option for pcretest, so I've replaced it by a simple function that does just
1759 that job.
1760
1761 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
1762 "pcregrep -".
1763
1764 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
1765 Java package. This provides some syntactic sugar for simple cases of what my
1766 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
1767 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
1768 item, you can use this simplified notation. Note that only makes sense with
1769 greedy quantifiers. Consequently, the use of the possessive quantifier forces
1770 greediness, whatever the setting of the PCRE_UNGREEDY option.
1771
1772 13. A change of greediness default within a pattern was not taking effect at
1773 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
1774 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
1775 was abstracted outside.
1776
1777 14. PCRE now supports the \G assertion. It is true when the current matching
1778 position is at the start point of the match. This differs from \A when the
1779 starting offset is non-zero. Used with the /g option of pcretest (or similar
1780 code), it works in the same way as it does for Perl's /g option. If all
1781 alternatives of a regex begin with \G, the expression is anchored to the start
1782 match position, and the "anchored" flag is set in the compiled expression.
1783
1784 15. Some bugs concerning the handling of certain option changes within patterns
1785 have been fixed. These applied to options other than (?ims). For example,
1786 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
1787 been the other way round. Some of this was related to change 7 above.
1788
1789 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
1790 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
1791 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
1792 POSIX classes only within a class (e.g. /[[:alpha:]]/).
1793
1794 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
1795 mode. Unlike ".", it always matches newline, whatever the setting of
1796 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
1797 assertions. Perl allows it, but it doesn't (in general) work because it can't
1798 calculate the length of the lookbehind. At least, that's the case for Perl
1799 5.8.0 - I've been told they are going to document that it doesn't work in
1800 future.
1801
1802 18. Added an error diagnosis for escapes that PCRE does not support: these are
1803 \L, \l, \N, \P, \p, \U, \u, and \X.
1804
1805 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
1806 reading past the end of the pattern in cases such as /[abcd/.
1807
1808 20. PCRE was getting more memory than necessary for patterns with classes that
1809 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
1810
1811 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
1812 compiling PCRE for use with Virtual Pascal.
1813
1814 22. Small fix to the Makefile to make it work properly if the build is done
1815 outside the source tree.
1816
1817 23. Added a new extension: a condition to go with recursion. If a conditional
1818 subpattern starts with (?(R) the "true" branch is used if recursion has
1819 happened, whereas the "false" branch is used only at the top level.
1820
1821 24. When there was a very long string of literal characters (over 255 bytes
1822 without UTF support, over 250 bytes with UTF support), the computation of how
1823 much memory was required could be incorrect, leading to segfaults or other
1824 strange effects.
1825
1826 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
1827 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
1828 there was a subsequent back reference to those brackets. This meant that, for
1829 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
1830 possible to check for precisely this case. All we can do is abandon the
1831 optimization if .* occurs inside capturing brackets when there are any back
1832 references whatsoever. (See below for a better fix that came later.)
1833
1834 26. The handling of the optimization for finding the first character of a
1835 non-anchored pattern, and for finding a character that is required later in the
1836 match were failing in some cases. This didn't break the matching; it just
1837 failed to optimize when it could. The way this is done has been re-implemented.
1838
1839 27. Fixed typo in error message for invalid (?R item (it said "(?p").
1840
1841 28. Added a new feature that provides some of the functionality that Perl
1842 provides with (?{...}). The facility is termed a "callout". The way it is done
1843 in PCRE is for the caller to provide an optional function, by setting
1844 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
1845 global variable. By default it is unset, which disables all calling out. To get
1846 the function called, the regex must include (?C) at appropriate points. This
1847 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
1848 This provides a means of identifying different callout points. When PCRE
1849 reaches such a point in the regex, if pcre_callout has been set, the external
1850 function is called. It is provided with data in a structure called
1851 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
1852 matching continues; if it returns a non-zero value, the match at the current
1853 point fails. However, backtracking will occur if possible. [This was changed
1854 later and other features added - see item 49 below.]
1855
1856 29. pcretest is upgraded to test the callout functionality. It provides a
1857 callout function that displays information. By default, it shows the start of
1858 the match and the current position in the text. There are some new data escapes
1859 to vary what happens:
1860
1861 \C+ in addition, show current contents of captured substrings
1862 \C- do not supply a callout function
1863 \C!n return 1 when callout number n is reached
1864 \C!n!m return 1 when callout number n is reached for the mth time
1865
1866 30. If pcregrep was called with the -l option and just a single file name, it
1867 output "<stdin>" if a match was found, instead of the file name.
1868
1869 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
1870 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
1871 pcre_exec(). This saves a malloc/free per call. The default value of
1872 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
1873 when configuring.
1874
1875 32. The default maximum size of a compiled pattern is 64K. There have been a
1876 few cases of people hitting this limit. The code now uses macros to handle the
1877 storing of links as offsets within the compiled pattern. It defaults to 2-byte
1878 links, but this can be changed to 3 or 4 bytes by --with-link-size when
1879 configuring. Tests 2 and 5 work only with 2-byte links because they output
1880 debugging information about compiled patterns.
1881
1882 33. Internal code re-arrangements:
1883
1884 (a) Moved the debugging function for printing out a compiled regex into
1885 its own source file (printint.c) and used #include to pull it into
1886 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
1887 separate copies.
1888
1889 (b) Defined the list of op-code names for debugging as a macro in
1890 internal.h so that it is next to the definition of the opcodes.
1891
1892 (c) Defined a table of op-code lengths for simpler skipping along compiled
1893 code. This is again a macro in internal.h so that it is next to the
1894 definition of the opcodes.
1895
1896 34. Added support for recursive calls to individual subpatterns, along the
1897 lines of Robin Houston's patch (but implemented somewhat differently).
1898
1899 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
1900 allow it to read and process whole directories in Win32. This code was
1901 contributed by Lionel Fourquaux; it has not been tested by me.
1902
1903 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
1904 used to name a group. Names consist of alphanumerics and underscores, and must
1905 be unique. Back references use the syntax (?P=name) and recursive calls use
1906 (?P>name) which is a PCRE extension to the Python extension. Groups still have
1907 numbers. The function pcre_fullinfo() can be used after compilation to extract
1908 a name/number map. There are three relevant calls:
1909
1910 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
1911 PCRE_INFO_NAMECOUNT yields the number of entries
1912 PCRE_INFO_NAMETABLE yields a pointer to the map.
1913
1914 The map is a vector of fixed-size entries. The size of each entry depends on
1915 the length of the longest name used. The first two bytes of each entry are the
1916 group number, most significant byte first. There follows the corresponding
1917 name, zero terminated. The names are in alphabetical order.
1918
1919 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
1920 case instead of 255. Making it the same both with and without UTF-8 support
1921 means that the same test output works with both.
1922
1923 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
1924 calling malloc() with a zero argument.
1925
1926 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
1927 optimization. I've improved things by keeping a bitmap of backreferences with
1928 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
1929 fact referenced, the optimization can be applied. It is unlikely that a
1930 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
1931 the match to follow \n) will appear inside brackets with a number greater than
1932 31, but if it does, any back reference > 31 suppresses the optimization.
1933
1934 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
1935 of disabling numbered capturing parentheses. Any opening parenthesis that is
1936 not followed by ? behaves as if it were followed by ?: but named parentheses
1937 can still be used for capturing (and they will acquire numbers in the usual
1938 way).
1939
1940 41. Redesigned the return codes from the match() function into yes/no/error so
1941 that errors can be passed back from deep inside the nested calls. A malloc
1942 failure while inside a recursive subpattern call now causes the
1943 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
1944
1945 42. It is now possible to set a limit on the number of times the match()
1946 function is called in a call to pcre_exec(). This facility makes it possible to
1947 limit the amount of recursion and backtracking, though not in a directly
1948 obvious way, because the match() function is used in a number of different
1949 circumstances. The count starts from zero for each position in the subject
1950 string (for non-anchored patterns). The default limit is, for compatibility, a
1951 large number, namely 10 000 000. You can change this in two ways:
1952
1953 (a) When configuring PCRE before making, you can use --with-match-limit=n
1954 to set a default value for the compiled library.
1955
1956 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
1957 a different value is set. See 45 below.
1958
1959 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
1960
1961 43. Added a new function pcre_config(int, void *) to enable run-time extraction
1962 of things that can be changed at compile time. The first argument specifies
1963 what is wanted and the second points to where the information is to be placed.
1964 The current list of available information is:
1965
1966 PCRE_CONFIG_UTF8
1967
1968 The output is an integer that is set to one if UTF-8 support is available;
1969 otherwise it is set to zero.
1970
1971 PCRE_CONFIG_NEWLINE
1972
1973 The output is an integer that it set to the value of the code that is used for
1974 newline. It is either LF (10) or CR (13).
1975
1976 PCRE_CONFIG_LINK_SIZE
1977
1978 The output is an integer that contains the number of bytes used for internal
1979 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
1980
1981 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
1982
1983 The output is an integer that contains the threshold above which the POSIX
1984 interface uses malloc() for output vectors. See item 31 above.
1985
1986 PCRE_CONFIG_MATCH_LIMIT
1987
1988 The output is an unsigned integer that contains the default limit of the number
1989 of match() calls in a pcre_exec() execution. See 42 above.
1990
1991 44. pcretest has been upgraded by the addition of the -C option. This causes it
1992 to extract all the available output from the new pcre_config() function, and to
1993 output it. The program then exits immediately.
1994
1995 45. A need has arisen to pass over additional data with calls to pcre_exec() in
1996 order to support additional features. One way would have been to define
1997 pcre_exec2() (for example) with extra arguments, but this would not have been
1998 extensible, and would also have required all calls to the original function to
1999 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2000 is used for passing in "extra" data from pcre_study().
2001
2002 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2003 contains the following fields:
2004
2005 flags a bitmap indicating which of the following fields are set
2006 study_data opaque data from pcre_study()
2007 match_limit a way of specifying a limit on match() calls for a specific
2008 call to pcre_exec()
2009 callout_data data for callouts (see 49 below)
2010
2011 The flag bits are also defined in pcre.h, and are
2012
2013 PCRE_EXTRA_STUDY_DATA
2014 PCRE_EXTRA_MATCH_LIMIT
2015 PCRE_EXTRA_CALLOUT_DATA
2016
2017 The pcre_study() function now returns one of these new pcre_extra blocks, with
2018 the actual study data pointed to by the study_data field, and the
2019 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2020 before. That is, this change is entirely upwards-compatible and requires no
2021 change to existing code.
2022
2023 If you want to pass in additional data to pcre_exec(), you can either place it
2024 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2025 block.
2026
2027 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2028 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2029 times with different match limits, until it finds the minimum value needed for
2030 pcre_exec() to complete. The value is then output. This can be instructive; for
2031 most simple matches the number is quite small, but for pathological cases it
2032 gets very large very quickly.
2033
2034 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2035 returns the size of the data block pointed to by the study_data field in a
2036 pcre_extra block, that is, the value that was passed as the argument to
2037 pcre_malloc() when PCRE was getting memory in which to place the information
2038 created by pcre_study(). The fourth argument should point to a size_t variable.
2039 pcretest has been extended so that this information is shown after a successful
2040 pcre_study() call when information about the compiled regex is being displayed.
2041
2042 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2043 because what follows is always an absolute path. (Later: it turns out that this
2044 is more than cosmetic for MinGW, because it doesn't like empty path
2045 components.)
2046
2047 49. Some changes have been made to the callout feature (see 28 above):
2048
2049 (i) A callout function now has three choices for what it returns:
2050
2051 0 => success, carry on matching
2052 > 0 => failure at this point, but backtrack if possible
2053 < 0 => serious error, return this value from pcre_exec()
2054
2055 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2056 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2057 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2058 use by callout functions. It will never be used by PCRE itself.
2059
2060 (ii) The pcre_extra structure (see 45 above) has a void * field called
2061 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2062 pcre_callout_block structure has a field of the same name. The contents of
2063 the field passed in the pcre_extra structure are passed to the callout
2064 function in the corresponding field in the callout block. This makes it
2065 easier to use the same callout-containing regex from multiple threads. For
2066 testing, the pcretest program has a new data escape
2067
2068 \C*n pass the number n (may be negative) as callout_data
2069
2070 If the callout function in pcretest receives a non-zero value as
2071 callout_data, it returns that value.
2072
2073 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2074 there were some redundant $(CFLAGS) in commands that are now specified as
2075 $(LINK), which already includes $(CFLAGS).
2076
2077 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2078 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2079 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2080 one-byte characters throughout. Note that case-insensitive matching applies
2081 only to characters whose values are less than 256. PCRE doesn't support the
2082 notion of cases for higher-valued characters.
2083
2084 (i) A character class whose characters are all within 0-255 is handled as
2085 a bit map, and the map is inverted for negative classes. Previously, a
2086 character > 255 always failed to match such a class; however it should
2087 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2088
2089 (ii) A negated character class with a single character < 255 is coded as
2090 "not this character" (OP_NOT). This wasn't working properly when the test
2091 character was multibyte, either singly or repeated.
2092
2093 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2094 mode, for example: \x{100}{2,3}.
2095
2096 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2097 singly or repeated) now correctly test multibyte characters. However,
2098 PCRE doesn't recognize any characters with values greater than 255 as
2099 digits, spaces, or word characters. Such characters always match \D, \S,
2100 and \W, and never match \d, \s, or \w.
2101
2102 (v) Classes may now contain characters and character ranges with values
2103 greater than 255. For example: [ab\x{100}-\x{400}].
2104
2105 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2106 PCRE in UTF-8 mode.
2107
2108 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2109 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2110 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2111 value.)
2112
2113 53. The single man page has become too large. I have therefore split it up into
2114 a number of separate man pages. These also give rise to individual HTML pages;
2115 these are now put in a separate directory, and there is an index.html page that
2116 lists them all. Some hyperlinking between the pages has been installed.
2117
2118 54. Added convenience functions for handling named capturing parentheses.
2119
2120 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2121 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2122 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2123 are faulted.
2124
2125 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2126 calling configure. These values are used when compiling the dftables.c program
2127 which is run to generate the source of the default character tables. They
2128 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2129 you will need to set these values.
2130
2131 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2132
2133
2134 Version 3.9 02-Jan-02
2135 ---------------------
2136
2137 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2138
2139 2. If --disable-static was given, the building process failed when trying to
2140 build pcretest and pcregrep. (For some reason it was using libtool to compile
2141 them, which is not right, as they aren't part of the library.)
2142
2143
2144 Version 3.8 18-Dec-01
2145 ---------------------
2146
2147 1. The experimental UTF-8 code was completely screwed up. It was packing the
2148 bytes in the wrong order. How dumb can you get?
2149
2150
2151 Version 3.7 29-Oct-01
2152 ---------------------
2153
2154 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2155 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2156 this didn't happen under Solaris 8, where I normally test things.
2157
2158 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2159 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2160 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2161 this fix an hour or so after the initial 3.7 release.)
2162
2163
2164 Version 3.6 23-Oct-01
2165 ---------------------
2166
2167 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2168 offsets passed as NULL with zero offset count.
2169
2170 2. The config.guess and config.sub files had not been updated when I moved to
2171 the latest autoconf.
2172
2173
2174 Version 3.5 15-Aug-01
2175 ---------------------
2176
2177 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2178 had been forgotten.
2179
2180 2. By using declared but undefined structures, we can avoid using "void"
2181 definitions in pcre.h while keeping the internal definitions of the structures
2182 private.
2183
2184 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2185 user point of view, this means that both static and shared libraries are built
2186 by default, but this can be individually controlled. More of the work of
2187 handling this static/shared cases is now inside libtool instead of PCRE's make
2188 file.
2189
2190 4. The pcretest utility is now installed along with pcregrep because it is
2191 useful for users (to test regexs) and by doing this, it automatically gets
2192 relinked by libtool. The documentation has been turned into a man page, so
2193 there are now .1, .txt, and .html versions in /doc.
2194
2195 5. Upgrades to pcregrep:
2196 (i) Added long-form option names like gnu grep.
2197 (ii) Added --help to list all options with an explanatory phrase.
2198 (iii) Added -r, --recursive to recurse into sub-directories.
2199 (iv) Added -f, --file to read patterns from a file.
2200
2201 6. pcre_exec() was referring to its "code" argument before testing that
2202 argument for NULL (and giving an error if it was NULL).
2203
2204 7. Upgraded Makefile.in to allow for compiling in a different directory from
2205 the source directory.
2206
2207 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2208 options bits, the pointer it was passed was to an int instead of to an unsigned
2209 long int. This mattered only on 64-bit systems.
2210
2211 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2212 generated) instead of pcre.in, which it its source. Also made the same change
2213 in several of the .c files.
2214
2215 10. A new release of gcc defines printf() as a macro, which broke pcretest
2216 because it had an ifdef in the middle of a string argument for printf(). Fixed
2217 by using separate calls to printf().
2218
2219 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2220 script, to force use of CR or LF instead of \n in the source. On non-Unix
2221 systems, the value can be set in config.h.
2222
2223 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2224 absolute limit. Changed the text of the error message to make this clear, and
2225 likewise updated the man page.
2226
2227 13. The limit of 99 on the number of capturing subpatterns has been removed.
2228 The new limit is 65535, which I hope will not be a "real" limit.
2229
2230
2231 Version 3.4 22-Aug-00
2232 ---------------------
2233
2234 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2235
2236 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2237
2238
2239 Version 3.3 01-Aug-00
2240 ---------------------
2241
2242 1. If an octal character was given, but the value was greater than \377, it
2243 was not getting masked to the least significant bits, as documented. This could
2244 lead to crashes in some systems.
2245
2246 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2247 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2248
2249 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2250 These just pass their arguments on to (pcre_free)(), but they are provided
2251 because some uses of PCRE bind it to non-C systems that can call its functions,
2252 but cannot call free() or pcre_free() directly.
2253
2254 4. Add "make test" as a synonym for "make check". Corrected some comments in
2255 the Makefile.
2256
2257 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2258 Makefile.
2259
2260 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2261 command called pgrep for grepping around the active processes.
2262
2263 7. Added the beginnings of support for UTF-8 character strings.
2264
2265 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2266 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2267 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2268 out for the ar command.)
2269
2270
2271 Version 3.2 12-May-00
2272 ---------------------
2273
2274 This is purely a bug fixing release.
2275
2276 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
2277 of ZA. This was just one example of several cases that could provoke this bug,
2278 which was introduced by change 9 of version 2.00. The code for breaking
2279 infinite loops after an iteration that matches an empty string was't working
2280 correctly.
2281
2282 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
2283 when matched against abbab (for example). After matching an empty string, it
2284 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
2285 caused it to match further down the string than it should.
2286
2287 3. The code contained an inclusion of sys/types.h. It isn't clear why this
2288 was there because it doesn't seem to be needed, and it causes trouble on some
2289 systems, as it is not a Standard C header. It has been removed.
2290
2291 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
2292 were reported on the Macintosh. The changes were from
2293
2294 while ((c = *(++ptr)) != 0 && c != '\n');
2295 to
2296 while ((c = *(++ptr)) != 0 && c != '\n') ;
2297
2298 Totally extraordinary, but if that's what it takes...
2299
2300 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
2301 available. Added HAVE_BCOPY and an autoconf test for it; if neither
2302 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
2303 assumes the way PCRE uses memmove() (always moving upwards).
2304
2305 6. PCRE is being used in one environment where strchr() is not available. There
2306 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
2307 faster code anyway.
2308
2309
2310 Version 3.1 09-Feb-00
2311 ---------------------
2312
2313 The only change in this release is the fixing of some bugs in Makefile.in for
2314 the "install" target:
2315
2316 (1) It was failing to install pcreposix.h.
2317
2318 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
2319
2320
2321 Version 3.0 01-Feb-00
2322 ---------------------
2323
2324 1. Add support for the /+ modifier to perltest (to output $` like it does in
2325 pcretest).
2326
2327 2. Add support for the /g modifier to perltest.
2328
2329 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
2330 matches null strings.
2331
2332 4. Fix perltest so that it doesn't do unwanted things when fed an empty
2333 pattern. Perl treats empty patterns specially - it reuses the most recent
2334 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
2335 effect.
2336
2337 5. The POSIX interface was broken in that it was just handing over the POSIX
2338 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
2339 required a bigger vector, with some working space on the end. This means that
2340 the POSIX wrapper now has to get and free some memory, and copy the results.
2341
2342 6. Added some simple autoconf support, placing the test data and the
2343 documentation in separate directories, re-organizing some of the
2344 information files, and making it build pcre-config (a GNU standard). Also added
2345 libtool support for building PCRE as a shared library, which is now the
2346 default.
2347
2348 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
2349 09 are not valid octal constants. Single digits will be used for minor values
2350 less than 10.
2351
2352 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
2353 existing programs that set these in the POSIX interface can use PCRE without
2354 modification.
2355
2356 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
2357 return all that pcre_info() returns, plus additional data. The pcre_info()
2358 function is retained for compatibility, but is considered to be obsolete.
2359
2360 10. Added experimental recursion feature (?R) to handle one common case that
2361 Perl 5.6 will be able to do with (?p{...}).
2362
2363 11. Added support for POSIX character classes like [:alpha:], which Perl is
2364 adopting.
2365
2366
2367 Version 2.08 31-Aug-99
2368 ----------------------
2369
2370 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
2371 trying to match at the startoffset position, but instead was moving forward to
2372 the next newline as if a previous match had failed.
2373
2374 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
2375 and could get into a loop if a null string was matched other than at the start
2376 of the subject.
2377
2378 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
2379 be distinguished at compile time, and for completeness also added PCRE_DATE.
2380
2381 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
2382 in GnuWin32 environments.
2383
2384
2385 Version 2.07 29-Jul-99
2386 ----------------------
2387
2388 1. The documentation is now supplied in plain text form and HTML as well as in
2389 the form of man page sources.
2390
2391 2. C++ compilers don't like assigning (void *) values to other pointer types.
2392 In particular this affects malloc(). Although there is no problem in Standard
2393 C, I've put in casts to keep C++ compilers happy.
2394
2395 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
2396 should be (const char *).
2397
2398 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
2399 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
2400 However, I haven't made this a standard facility. The documentation doesn't
2401 mention it, and the Makefile doesn't support it.
2402
2403 5. The Makefile now contains an "install" target, with editable destinations at
2404 the top of the file. The pcretest program is not installed.
2405
2406 6. pgrep -V now gives the PCRE version number and date.
2407
2408 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
2409 causing the entire string to be ignored, instead of just the last character.
2410
2411 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
2412 non-matching string, it can take a very, very long time, even for strings of
2413 quite modest length, because of the nested recursion. PCRE now does better in
2414 some of these cases. It does this by remembering the last required literal
2415 character in the pattern, and pre-searching the subject to ensure it is present
2416 before running the real match. In other words, it applies a heuristic to detect
2417 some types of certain failure quickly, and in the above example, if presented
2418 with a string that has no trailing " it gives "no match" very quickly.
2419
2420 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
2421 other alternatives are tried instead.
2422
2423
2424 Version 2.06 09-Jun-99
2425 ----------------------
2426
2427 1. Change pcretest's output for amount of store used to show just the code
2428 space, because the remainder (the data block) varies in size between 32-bit and
2429 64-bit systems.
2430
2431 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
2432 start matching at. This allows lookbehinds to work when searching for multiple
2433 occurrences in a string.
2434
2435 3. Added additional options to pcretest for testing multiple occurrences:
2436
2437 /+ outputs the rest of the string that follows a match
2438 /g loops for multiple occurrences, using the new startoffset argument
2439 /G loops for multiple occurrences by passing an incremented pointer
2440
2441 4. PCRE wasn't doing the "first character" optimization for patterns starting
2442 with \b or \B, though it was doing it for other lookbehind assertions. That is,
2443 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
2444 the letter 'x'. On long subject strings, this gives a significant speed-up.
2445
2446
2447 Version 2.05 21-Apr-99
2448 ----------------------
2449
2450 1. Changed the type of magic_number from int to long int so that it works
2451 properly on 16-bit systems.
2452
2453 2. Fixed a bug which caused patterns starting with .* not to work correctly
2454 when the subject string contained newline characters. PCRE was assuming
2455 anchoring for such patterns in all cases, which is not correct because .* will
2456 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
2457 DOTALL is set at top level; otherwise it knows that patterns starting with .*
2458 must be retried after every newline in the subject.
2459
2460
2461 Version 2.04 18-Feb-99
2462 ----------------------
2463
2464 1. For parenthesized subpatterns with repeats whose minimum was zero, the
2465 computation of the store needed to hold the pattern was incorrect (too large).
2466 If such patterns were nested a few deep, this could multiply and become a real
2467 problem.
2468
2469 2. Added /M option to pcretest to show the memory requirement of a specific
2470 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
2471
2472 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
2473 compiled in such a way that the backtracking after subsequent failure was
2474 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
2475 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
2476
2477
2478 Version 2.03 02-Feb-99
2479 ----------------------
2480
2481 1. Fixed typo and small mistake in man page.
2482
2483 2. Added 4th condition (GPL supersedes if conflict) and created separate
2484 LICENCE file containing the conditions.
2485
2486 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
2487 Perl, that is the internal \ allows the delimiter to be included in the
2488 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
2489 the final delimiter, add \ to the end of the pattern (to test the error).
2490
2491 4. Added the convenience functions for extracting substrings after a successful
2492 match. Updated pcretest to make it able to test these functions.
2493
2494
2495 Version 2.02 14-Jan-99
2496 ----------------------
2497
2498 1. Initialized the working variables associated with each extraction so that
2499 their saving and restoring doesn't refer to uninitialized store.
2500
2501 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
2502 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
2503 fix the problem.
2504
2505 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
2506 calls, and wasn't printing the correct value for compiling calls. Increased the
2507 default value of LOOPREPEAT, and the number of significant figures in the
2508 times.
2509
2510 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
2511
2512 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
2513 a building problem on Windows NT with a FAT file system.
2514
2515
2516 Version 2.01 21-Oct-98
2517 ----------------------
2518
2519 1. Changed the API for pcre_compile() to allow for the provision of a pointer
2520 to character tables built by pcre_maketables() in the current locale. If NULL
2521 is passed, the default tables are used.
2522
2523
2524 Version 2.00 24-Sep-98
2525 ----------------------
2526
2527 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
2528 it any more.
2529
2530 2. Allow quantification of (?>) groups, and make it work correctly.
2531
2532 3. The first character computation wasn't working for (?>) groups.
2533
2534 4. Correct the implementation of \Z (it is permitted to match on the \n at the
2535 end of the subject) and add 5.005's \z, which really does match only at the
2536 very end of the subject.
2537
2538 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
2539
2540 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
2541 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
2542 localized options. All options to pcre_study() were also removed.
2543
2544 7. Add other new features from 5.005:
2545
2546 $(?<= positive lookbehind
2547 $(?<! negative lookbehind
2548 (?imsx-imsx) added the unsetting capability
2549 such a setting is global if at outer level; local otherwise
2550 (?imsx-imsx:) non-capturing groups with option setting
2551 (?(cond)re|re) conditional pattern matching
2552
2553 A backreference to itself in a repeated group matches the previous
2554 captured string.
2555
2556 8. General tidying up of studying (both automatic and via "study")
2557 consequential on the addition of new assertions.
2558
2559 9. As in 5.005, unlimited repeated groups that could match an empty substring
2560 are no longer faulted at compile time. Instead, the loop is forcibly broken at
2561 runtime if any iteration does actually match an empty substring.
2562
2563 10. Include the RunTest script in the distribution.
2564
2565 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
2566 discrepancies, some of which were old and were also with respect to 5.004. They
2567 have now been fixed.
2568
2569
2570 Version 1.09 28-Apr-98
2571 ----------------------
2572
2573 1. A negated single character class followed by a quantifier with a minimum
2574 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
2575 program crashes, or just wrong answers. This did not apply to negated classes
2576 containing more than one character, or to minima other than one.
2577
2578
2579 Version 1.08 27-Mar-98
2580 ----------------------
2581
2582 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
2583
2584 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
2585 latter must appear before anything that relies on it in the pattern.
2586
2587
2588 Version 1.07 16-Feb-98
2589 ----------------------
2590
2591 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
2592 repeat of a potentially empty string).
2593
2594
2595 Version 1.06 23-Jan-98
2596 ----------------------
2597
2598 1. Added Markus Oberhumer's little patches for C++.
2599
2600 2. Literal strings longer than 255 characters were broken.
2601
2602
2603 Version 1.05 23-Dec-97
2604 ----------------------
2605
2606 1. Negated character classes containing more than one character were failing if
2607 PCRE_CASELESS was set at run time.
2608
2609
2610 Version 1.04 19-Dec-97
2611 ----------------------
2612
2613 1. Corrected the man page, where some "const" qualifiers had been omitted.
2614
2615 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
2616 input syntax.
2617
2618 3. Fixed memory leak which occurred when a regex with back references was
2619 matched with an offsets vector that wasn't big enough. The temporary memory
2620 that is used in this case wasn't being freed if the match failed.
2621
2622 4. Tidied pcretest to ensure it frees memory that it gets.
2623
2624 5. Temporary memory was being obtained in the case where the passed offsets
2625 vector was exactly big enough.
2626
2627 6. Corrected definition of offsetof() from change 5 below.
2628
2629 7. I had screwed up change 6 below and broken the rules for the use of
2630 setjmp(). Now fixed.
2631
2632
2633 Version 1.03 18-Dec-97
2634 ----------------------
2635
2636 1. A erroneous regex with a missing opening parenthesis was correctly
2637 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
2638 on some systems.
2639
2640 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
2641 it was reported that one broken compiler failed on the former because "code" is
2642 also an independent variable.
2643
2644 3. The erroneous regex a[]b caused an array overrun reference.
2645
2646 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
2647 fail on data ending with that character. (It was going on too far, and checking
2648 the next character, typically a binary zero.) This was specific to the
2649 optimized code for single-character negative classes.
2650
2651 5. Added a contributed patch from the TIN world which does the following:
2652
2653 + Add an undef for memmove, in case the the system defines a macro for it.
2654
2655 + Add a definition of offsetof(), in case there isn't one. (I don't know
2656 the reason behind this - offsetof() is part of the ANSI standard - but
2657 it does no harm).
2658
2659 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
2660 most of the places where whitespace preceded '#'. I have given up and
2661 allowed the remaining 2 cases to be at the margin.
2662
2663 + Rename some variables in pcre to eliminate shadowing. This seems very
2664 pedantic, but does no harm, of course.
2665
2666 6. Moved the call to setjmp() into its own function, to get rid of warnings
2667 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
2668
2669 7. Constructs such as \d{8,} were compiling into the equivalent of
2670 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
2671 outcome, but in this particular case used more store than had been allocated,
2672 which caused the bug to be discovered because it threw up an internal error.
2673
2674 8. The debugging code in both pcre and pcretest for outputting the compiled
2675 form of a regex was going wrong in the case of back references followed by
2676 curly-bracketed repeats.
2677
2678
2679 Version 1.02 12-Dec-97
2680 ----------------------
2681
2682 1. Typos in pcre.3 and comments in the source fixed.
2683
2684 2. Applied a contributed patch to get rid of places where it used to remove
2685 'const' from variables, and fixed some signed/unsigned and uninitialized
2686 variable warnings.
2687
2688 3. Added the "runtest" target to Makefile.
2689
2690 4. Set default compiler flag to -O2 rather than just -O.
2691
2692
2693 Version 1.01 19-Nov-97
2694 ----------------------
2695
2696 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
2697 like /([ab]*)*/, that is, for classes with more than one character in them.
2698
2699 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
2700 as /((?>a*))*/ (a PCRE_EXTRA facility).
2701
2702
2703 Version 1.00 18-Nov-97
2704 ----------------------
2705
2706 1. Added compile-time macros to support systems such as SunOS4 which don't have
2707 memmove() or strerror() but have other things that can be used instead.
2708
2709 2. Arranged that "make clean" removes the executables.
2710
2711
2712 Version 0.99 27-Oct-97
2713 ----------------------
2714
2715 1. Fixed bug in code for optimizing classes with only one character. It was
2716 initializing a 32-byte map regardless, which could cause it to run off the end
2717 of the memory it had got.
2718
2719 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
2720
2721
2722 Version 0.98 22-Oct-97
2723 ----------------------
2724
2725 1. Fixed bug in code for handling temporary memory usage when there are more
2726 back references than supplied space in the ovector. This could cause segfaults.
2727
2728
2729 Version 0.97 21-Oct-97
2730 ----------------------
2731
2732 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
2733
2734 2. Optimized negated single characters not to use a bit map.
2735
2736 3. Brought error texts together as macro definitions; clarified some of them;
2737 fixed one that was wrong - it said "range out of order" when it meant "invalid
2738 escape sequence".
2739
2740 4. Changed some char * arguments to const char *.
2741
2742 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
2743
2744 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
2745 pcretest.
2746
2747
2748 Version 0.96 16-Oct-97
2749 ----------------------
2750
2751 1. Added a simple "pgrep" utility to the distribution.
2752
2753 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
2754 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
2755 where "ddd" means "one or more decimal digits".
2756
2757 3. Fixed serious bug. If a pattern had a back reference, but the call to
2758 pcre_exec() didn't supply a large enough ovector to record the related
2759 identifying subpattern, the match always failed. PCRE now remembers the number
2760 of the largest back reference, and gets some temporary memory in which to save
2761 the offsets during matching if necessary, in order to ensure that
2762 backreferences always work.
2763
2764 4. Increased the compatibility with Perl in a number of ways:
2765
2766 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
2767 to request this handling. The option can be set at compile or exec time.
2768
2769 (b) $ matches before a terminating newline by default; an option
2770 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
2771 mode). The option can be set at compile or exec time.
2772
2773 (c) The handling of \ followed by a digit other than 0 is now supposed to be
2774 the same as Perl's. If the decimal number it represents is less than 10
2775 or there aren't that many previous left capturing parentheses, an octal
2776 escape is read. Inside a character class, it's always an octal escape,
2777 even if it is a single digit.
2778
2779 (d) An escaped but undefined alphabetic character is taken as a literal,
2780 unless PCRE_EXTRA is set. Currently this just reserves the remaining
2781 escapes.
2782
2783 (e) {0} is now permitted. (The previous item is removed from the compiled
2784 pattern).
2785
2786 5. Changed all the names of code files so that the basic parts are no longer
2787 than 10 characters, and abolished the teeny "globals.c" file.
2788
2789 6. Changed the handling of character classes; they are now done with a 32-byte
2790 bit map always.
2791
2792 7. Added the -d and /D options to pcretest to make it possible to look at the
2793 internals of compilation without having to recompile pcre.
2794
2795
2796 Version 0.95 23-Sep-97
2797 ----------------------
2798
2799 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
2800 \x20 at the start of a run of normal characters. These were being treated as
2801 real characters, instead of the source characters being re-checked.
2802
2803
2804 Version 0.94 18-Sep-97
2805 ----------------------
2806
2807 1. The functions are now thread-safe, with the caveat that the global variables
2808 containing pointers to malloc() and free() or alternative functions are the
2809 same for all threads.
2810
2811 2. Get pcre_study() to generate a bitmap of initial characters for non-
2812 anchored patterns when this is possible, and use it if passed to pcre_exec().
2813
2814
2815 Version 0.93 15-Sep-97
2816 ----------------------
2817
2818 1. /(b)|(:+)/ was computing an incorrect first character.
2819
2820 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
2821 but not actually doing anything yet.
2822
2823 3. Treat "-" characters in classes that cannot be part of ranges as literals,
2824 as Perl does (e.g. [-az] or [az-]).
2825
2826 4. Set the anchored flag if a branch starts with .* or .*? because that tests
2827 all possible positions.
2828
2829 5. Split up into different modules to avoid including unneeded functions in a
2830 compiled binary. However, compile and exec are still in one module. The "study"
2831 function is split off.
2832
2833 6. The character tables are now in a separate module whose source is generated
2834 by an auxiliary program - but can then be edited by hand if required. There are
2835 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
2836 toupper() in the code.
2837
2838 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
2839 make them global. Abolish the function for setting them, as the caller can now
2840 set them directly.
2841
2842
2843 Version 0.92 11-Sep-97
2844 ----------------------
2845
2846 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
2847 (e.g. /a{1,3}/) was broken (I mis-optimized it).
2848
2849 2. Caseless matching was not working in character classes if the characters in
2850 the pattern were in upper case.
2851
2852 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
2853
2854 4. Make PCRE_ANCHORED public and accept as a compile option.
2855
2856 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
2857 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
2858 pass them.
2859
2860 6. Give an error if bad option bits passed at compile or run time.
2861
2862 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
2863 pcretest to cause it to pass that flag.
2864
2865 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
2866 options, and the first character, if set.
2867
2868 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
2869
2870
2871 Version 0.91 10-Sep-97
2872 ----------------------
2873
2874 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
2875 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
2876
2877 2. PCRE was looping on encountering an indefinitely repeated back reference to
2878 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
2879 Perl does - treats the match as successful.
2880
2881 ****

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