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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 538 - (show annotations) (download)
Wed Jun 9 19:30:57 2010 UTC (3 years, 10 months ago) by ph10
File size: 189009 byte(s)
Fix pcre_study() problem with non-C-locale chartables in UTF-8 mode.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.10 03-Jun-2010
5 ------------------------
6
7 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
8 THEN.
9
10 2. (*ACCEPT) was not working when inside an atomic group.
11
12 3. Inside a character class, \B is treated as a literal by default, but
13 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
14 causes the error). The code is unchanged, but I tidied the documentation.
15
16 4. Inside a character class, PCRE always treated \R and \X as literals,
17 whereas Perl faults them if its -w option is set. I have changed PCRE so
18 that it faults them when PCRE_EXTRA is set.
19
20 5. Added support for \N, which always matches any character other than
21 newline. (It is the same as "." when PCRE_DOTALL is not set.)
22
23 6. When compiling pcregrep with newer versions of gcc which may have
24 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
25 declared with attribute warn_unused_result" were given. Just casting the
26 result to (void) does not stop the warnings; a more elaborate fudge is
27 needed. I've used a macro to implement this.
28
29 7. Minor change to pcretest.c to avoid a compiler warning.
30
31 8. Added four artifical Unicode properties to help with an option to make
32 \s etc use properties (see next item). The new properties are: Xan
33 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
34
35 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
36 use Unicode properties. (*UCP) at the start of a pattern can be used to set
37 this option. Modified pcretest to add /W to test this facility. Added
38 REG_UCP to make it available via the POSIX interface.
39
40 10. Added --line-buffered to pcregrep.
41
42 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
43 studied, and the match started with a letter with a code point greater than
44 127 whose first byte was different to the first byte of the other case of
45 the letter, the other case of this starting letter was not recognized
46 (#976).
47
48 12. If a pattern that was studied started with a repeated Unicode property
49 test, for example, \p{Nd}+, there was the theoretical possibility of
50 setting up an incorrect bitmap of starting bytes, but fortunately it could
51 not have actually happened in practice until change 8 above was made (it
52 added property types that matched character-matching opcodes).
53
54 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
55 possible starting bytes for non-anchored patterns.
56
57 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
58 \R, and also a number of cases that involve Unicode properties, both
59 explicit and implicit when PCRE_UCP is set.
60
61 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
62 input, it could crash or give wrong results if characters with values
63 greater than 0xc0 were present in the subject string. (Detail: it assumed
64 UTF-8 input when processing these items.)
65
66 16. Added a lot of (int) casts to avoid compiler warnings in systems where
67 size_t is 64-bit (#991).
68
69 17. Added a check for running out of memory when PCRE is compiled with
70 --disable-stack-for-recursion (#990).
71
72 18. If the last data line in a file for pcretest does not have a newline on
73 the end, a newline was missing in the output.
74
75 19. The default pcre_chartables.c file recognizes only ASCII characters (values
76 less than 128) in its various bitmaps. However, there is a facility for
77 generating tables according to the current locale when PCRE is compiled. It
78 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
79 characters, are recognized by isspace() and therefore were getting set in
80 these tables. This caused a problem in UTF-8 mode when pcre_study() was
81 used to create a list of bytes that can start a match. For \s, it was
82 including 0x85 and 0xa0, which of course cannot start UTF-8 characters. I
83 have changed the code so that only real ASCII characters (less than 128)
84 are set in this case because the \s etc escapes are documented as
85 recognizing only ASCII characters. (When PCRE_UCP is set - see 9 above -
86 the code is different altogether.)
87
88
89 Version 8.02 19-Mar-2010
90 ------------------------
91
92 1. The Unicode data tables have been updated to Unicode 5.2.0.
93
94 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
95 configured.
96
97 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
98 original author of that file, following a query about its status.
99
100 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
101 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
102
103 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
104 quantifier applied to a forward-referencing subroutine call, could compile
105 incorrect code or give the error "internal error: previously-checked
106 referenced subpattern not found".
107
108 6. Both MS Visual Studio and Symbian OS have problems with initializing
109 variables to point to external functions. For these systems, therefore,
110 pcre_malloc etc. are now initialized to local functions that call the
111 relevant global functions.
112
113 7. There were two entries missing in the vectors called coptable and poptable
114 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
115 I've fixed the data, and added a kludgy way of testing at compile time that
116 the lengths are correct (equal to the number of opcodes).
117
118 8. Following on from 7, I added a similar kludge to check the length of the
119 eint vector in pcreposix.c.
120
121 9. Error texts for pcre_compile() are held as one long string to avoid too
122 much relocation at load time. To find a text, the string is searched,
123 counting zeros. There was no check for running off the end of the string,
124 which could happen if a new error number was added without updating the
125 string.
126
127 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
128
129 11. \K was not working if it appeared in an atomic group or in a group that
130 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
131 \K is "not well defined" if used in an assertion. PCRE now accepts it if
132 the assertion is positive, but not if it is negative.
133
134 12. Change 11 fortuitously reduced the size of the stack frame used in the
135 "match()" function of pcre_exec.c by one pointer. Forthcoming
136 implementation of support for (*MARK) will need an extra pointer on the
137 stack; I have reserved it now, so that the stack frame size does not
138 decrease.
139
140 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
141 item in branch that calls a recursion is a subroutine call - as in the
142 second branch in the above example - was incorrectly given the compile-
143 time error "recursive call could loop indefinitely" because pcre_compile()
144 was not correctly checking the subroutine for matching a non-empty string.
145
146 14. The checks for overrunning compiling workspace could trigger after an
147 overrun had occurred. This is a "should never occur" error, but it can be
148 triggered by pathological patterns such as hundreds of nested parentheses.
149 The checks now trigger 100 bytes before the end of the workspace.
150
151 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
152
153
154 Version 8.01 19-Jan-2010
155 ------------------------
156
157 1. If a pattern contained a conditional subpattern with only one branch (in
158 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
159 computed the wrong minimum data length (which is of course zero for such
160 subpatterns). This could cause incorrect "no match" results.
161
162 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
163 the pattern is reset in the first branch, pcre_compile() failed with
164 "internal error: code overflow at offset...". This happened only when
165 the reset was to the original external option setting. (An optimization
166 abstracts leading options settings into an external setting, which was the
167 cause of this.)
168
169 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
170 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
171 assertion pattern did not match (meaning that the assertion was true), it
172 was incorrectly treated as false if the SKIP had been reached during the
173 matching. This also applied to assertions used as conditions.
174
175 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
176 assertion subpattern, including such a pattern used as a condition,
177 unpredictable results occurred, instead of the error return
178 PCRE_ERROR_DFA_UITEM.
179
180 5. The C++ GlobalReplace function was not working like Perl for the special
181 situation when an empty string is matched. It now does the fancy magic
182 stuff that is necessary.
183
184 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
185 removed. (These were left over from very, very early versions of PCRE.)
186
187 7. Some cosmetic changes to the code to make life easier when compiling it
188 as part of something else:
189
190 (a) Change DEBUG to PCRE_DEBUG.
191
192 (b) In pcre_compile(), rename the member of the "branch_chain" structure
193 called "current" as "current_branch", to prevent a collision with the
194 Linux macro when compiled as a kernel module.
195
196 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
197 prevent a collision with the Linux macro when compiled as a kernel
198 module.
199
200 8. In pcre_compile() there are some checks for integer overflows that used to
201 cast potentially large values to (double). This has been changed to that
202 when building, a check for int64_t is made, and if it is found, it is used
203 instead, thus avoiding the use of floating point arithmetic. (There is no
204 other use of FP in PCRE.) If int64_t is not found, the fallback is to
205 double.
206
207 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
208 2005 (difference between two addresses compared to an unsigned value).
209
210 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
211 custom one, because of the following reported problem in Windows:
212
213 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
214 under Win32.
215 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
216 therefore missing the function definition.
217 - The compiler thus generates a "C" signature for the test function.
218 - The linker fails to find the "C" function.
219 - PCRE fails to configure if asked to do so against libbz2.
220
221 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
222 messages were output:
223
224 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
225 rerunning libtoolize, to keep the correct libtool macros in-tree.
226 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
227
228 I have done both of these things.
229
230 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
231 most of the time, it *can* run out if it is given a pattern that contains a
232 runaway infinite recursion. I updated the discussion in the pcrestack man
233 page.
234
235 13. Now that we have gone to the x.xx style of version numbers, the minor
236 version may start with zero. Using 08 or 09 is a bad idea because users
237 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
238 interpreted as invalid octal numbers. I've updated the previous comment in
239 configure.ac, and also added a check that gives an error if 08 or 09 are
240 used.
241
242 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
243 causing partial matching to fail when the end of the subject matched \W
244 in a UTF-8 pattern where \W was quantified with a minimum of 3.
245
246 15. There were some discrepancies between the declarations in pcre_internal.h
247 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
248 their definitions. The declarations used "const uschar *" and the
249 definitions used USPTR. Even though USPTR is normally defined as "const
250 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
251 reported that: "This difference in casting confuses some C++ compilers, for
252 example, SunCC recognizes above declarations as different functions and
253 generates broken code for hbpcre." I have changed the declarations to use
254 USPTR.
255
256 16. GNU libtool is named differently on some systems. The autogen.sh script now
257 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
258 (FreeBSD).
259
260 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
261 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
262 comment: "Figure out how to create a longlong from a string: strtoll and
263 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
264 instance, but it only takes 2 args instead of 3!"
265
266 18. A subtle bug concerned with back references has been fixed by a change of
267 specification, with a corresponding code fix. A pattern such as
268 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
269 refers, was giving matches when it shouldn't. For example, xa=xaaa would
270 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
271 same bug. Such groups have to be quantified to be useful, or contained
272 inside another quantified group. (If there's no repetition, the reference
273 can never match.) The problem arises because, having left the group and
274 moved on to the rest of the pattern, a later failure that backtracks into
275 the group uses the captured value from the final iteration of the group
276 rather than the correct earlier one. I have fixed this in PCRE by forcing
277 any group that contains a reference to itself to be an atomic group; that
278 is, there cannot be any backtracking into it once it has completed. This is
279 similar to recursive and subroutine calls.
280
281
282 Version 8.00 19-Oct-09
283 ----------------------
284
285 1. The table for translating pcre_compile() error codes into POSIX error codes
286 was out-of-date, and there was no check on the pcre_compile() error code
287 being within the table. This could lead to an OK return being given in
288 error.
289
290 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
291 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
292 in a Windows environment.
293
294 3. The pcregrep --count option prints the count for each file even when it is
295 zero, as does GNU grep. However, pcregrep was also printing all files when
296 --files-with-matches was added. Now, when both options are given, it prints
297 counts only for those files that have at least one match. (GNU grep just
298 prints the file name in this circumstance, but including the count seems
299 more useful - otherwise, why use --count?) Also ensured that the
300 combination -clh just lists non-zero counts, with no names.
301
302 4. The long form of the pcregrep -F option was incorrectly implemented as
303 --fixed_strings instead of --fixed-strings. This is an incompatible change,
304 but it seems right to fix it, and I didn't think it was worth preserving
305 the old behaviour.
306
307 5. The command line items --regex=pattern and --regexp=pattern were not
308 recognized by pcregrep, which required --regex pattern or --regexp pattern
309 (with a space rather than an '='). The man page documented the '=' forms,
310 which are compatible with GNU grep; these now work.
311
312 6. No libpcreposix.pc file was created for pkg-config; there was just
313 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
314
315 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
316 when UCP support is not needed, by modifying the Python script that
317 generates it from Unicode data files. This should not matter if the module
318 is correctly used as a library, but I received one complaint about 50K of
319 unwanted data. My guess is that the person linked everything into his
320 program rather than using a library. Anyway, it does no harm.
321
322 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
323 was a minimum greater than 1 for a wide character in a possessive
324 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
325 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
326 character. Chaos in the form of incorrect output or a compiling loop could
327 result.
328
329 9. The restrictions on what a pattern can contain when partial matching is
330 requested for pcre_exec() have been removed. All patterns can now be
331 partially matched by this function. In addition, if there are at least two
332 slots in the offset vector, the offset of the earliest inspected character
333 for the match and the offset of the end of the subject are set in them when
334 PCRE_ERROR_PARTIAL is returned.
335
336 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
337 synonymous with PCRE_PARTIAL, for backwards compatibility, and
338 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
339 and may be more useful for multi-segment matching.
340
341 11. Partial matching with pcre_exec() is now more intuitive. A partial match
342 used to be given if ever the end of the subject was reached; now it is
343 given only if matching could not proceed because another character was
344 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
345 string "Z", which now yields "no match" instead of "partial match". In the
346 case of pcre_dfa_exec(), "no match" is given if every matching path for the
347 final character ended with (*FAIL).
348
349 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
350 if the pattern had a "must contain" character that was already found in the
351 earlier partial match, unless partial matching was again requested. For
352 example, with the pattern /dog.(body)?/, the "must contain" character is
353 "g". If the first part-match was for the string "dog", restarting with
354 "sbody" failed. This bug has been fixed.
355
356 13. The string returned by pcre_dfa_exec() after a partial match has been
357 changed so that it starts at the first inspected character rather than the
358 first character of the match. This makes a difference only if the pattern
359 starts with a lookbehind assertion or \b or \B (\K is not supported by
360 pcre_dfa_exec()). It's an incompatible change, but it makes the two
361 matching functions compatible, and I think it's the right thing to do.
362
363 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
364 so that the demonstration program is easily available in environments where
365 PCRE has not been installed from source.
366
367 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
368 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
369 library.
370
371 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
372 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
373 is not the first non-POSIX option to be added. Clearly some people find
374 these options useful.
375
376 17. If a caller to the POSIX matching function regexec() passes a non-zero
377 value for nmatch with a NULL value for pmatch, the value of
378 nmatch is forced to zero.
379
380 18. RunGrepTest did not have a test for the availability of the -u option of
381 the diff command, as RunTest does. It now checks in the same way as
382 RunTest, and also checks for the -b option.
383
384 19. If an odd number of negated classes containing just a single character
385 interposed, within parentheses, between a forward reference to a named
386 subpattern and the definition of the subpattern, compilation crashed with
387 an internal error, complaining that it could not find the referenced
388 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
389 [The bug was that it was starting one character too far in when skipping
390 over the character class, thus treating the ] as data rather than
391 terminating the class. This meant it could skip too much.]
392
393 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
394 /g option in pcretest when the pattern contains \K, which makes it possible
395 to have an empty string match not at the start, even when the pattern is
396 anchored. Updated pcretest and pcredemo to use this option.
397
398 21. If the maximum number of capturing subpatterns in a recursion was greater
399 than the maximum at the outer level, the higher number was returned, but
400 with unset values at the outer level. The correct (outer level) value is
401 now given.
402
403 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
404 PCRE did not set those parentheses (unlike Perl). I have now found a way to
405 make it do so. The string so far is captured, making this feature
406 compatible with Perl.
407
408 23. The tests have been re-organized, adding tests 11 and 12, to make it
409 possible to check the Perl 5.10 features against Perl 5.10.
410
411 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
412 pattern matches a fixed length string. PCRE did not allow this; now it
413 does. Neither allows recursion.
414
415 25. I finally figured out how to implement a request to provide the minimum
416 length of subject string that was needed in order to match a given pattern.
417 (It was back references and recursion that I had previously got hung up
418 on.) This code has now been added to pcre_study(); it finds a lower bound
419 to the length of subject needed. It is not necessarily the greatest lower
420 bound, but using it to avoid searching strings that are too short does give
421 some useful speed-ups. The value is available to calling programs via
422 pcre_fullinfo().
423
424 26. While implementing 25, I discovered to my embarrassment that pcretest had
425 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
426 study optimizations had never been tested with that matching function.
427 Oops. What is worse, even when it was passed study data, there was a bug in
428 pcre_dfa_exec() that meant it never actually used it. Double oops. There
429 were also very few tests of studied patterns with pcre_dfa_exec().
430
431 27. If (?| is used to create subpatterns with duplicate numbers, they are now
432 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
433 on the other side of the coin, they are no longer allowed to have different
434 names, because these cannot be distinguished in PCRE, and this has caused
435 confusion. (This is a difference from Perl.)
436
437 28. When duplicate subpattern names are present (necessarily with different
438 numbers, as required by 27 above), and a test is made by name in a
439 conditional pattern, either for a subpattern having been matched, or for
440 recursion in such a pattern, all the associated numbered subpatterns are
441 tested, and the overall condition is true if the condition is true for any
442 one of them. This is the way Perl works, and is also more like the way
443 testing by number works.
444
445
446 Version 7.9 11-Apr-09
447 ---------------------
448
449 1. When building with support for bzlib/zlib (pcregrep) and/or readline
450 (pcretest), all targets were linked against these libraries. This included
451 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
452 libraries. This caused unwanted dependencies to be created. This problem
453 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
454 pcretest is linked with readline.
455
456 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
457 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
458 moved outside it again, because FALSE and TRUE are already defined in AIX,
459 but BOOL is not.
460
461 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
462 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
463
464 4. The pcregrep documentation said spaces were inserted as well as colons (or
465 hyphens) following file names and line numbers when outputting matching
466 lines. This is not true; no spaces are inserted. I have also clarified the
467 wording for the --colour (or --color) option.
468
469 5. In pcregrep, when --colour was used with -o, the list of matching strings
470 was not coloured; this is different to GNU grep, so I have changed it to be
471 the same.
472
473 6. When --colo(u)r was used in pcregrep, only the first matching substring in
474 each matching line was coloured. Now it goes on to look for further matches
475 of any of the test patterns, which is the same behaviour as GNU grep.
476
477 7. A pattern that could match an empty string could cause pcregrep to loop; it
478 doesn't make sense to accept an empty string match in pcregrep, so I have
479 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
480 seems to be how GNU grep behaves.
481
482 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
483 start or after a newline", because the conditional assertion was not being
484 correctly handled. The rule now is that both the assertion and what follows
485 in the first alternative must satisfy the test.
486
487 9. If auto-callout was enabled in a pattern with a conditional group whose
488 condition was an assertion, PCRE could crash during matching, both with
489 pcre_exec() and pcre_dfa_exec().
490
491 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
492 used for matching.
493
494 11. Unicode property support in character classes was not working for
495 characters (bytes) greater than 127 when not in UTF-8 mode.
496
497 12. Added the -M command line option to pcretest.
498
499 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
500
501 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
502
503 16. Added comments and documentation about mis-use of no_arg in the C++
504 wrapper.
505
506 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
507 from Martin Jerabek that uses macro names for all relevant character and
508 string constants.
509
510 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
511 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
512 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
513 these, but not everybody uses configure.
514
515 19. A conditional group that had only one branch was not being correctly
516 recognized as an item that could match an empty string. This meant that an
517 enclosing group might also not be so recognized, causing infinite looping
518 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
519 with the subject "ab", where knowledge that the repeated group can match
520 nothing is needed in order to break the loop.
521
522 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
523 exec(), but without supplying a callout function, matching went wrong.
524
525 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
526 leak if the size of the offset vector was greater than 30. When the vector
527 is smaller, the saved offsets during recursion go onto a local stack
528 vector, but for larger vectors malloc() is used. It was failing to free
529 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
530 error, in fact).
531
532 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
533 heapframe that is used only when UTF-8 support is enabled. This caused no
534 problem, but was untidy.
535
536 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
537 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
538 included within another project.
539
540 24. Steven Van Ingelgem's patches to add more options to the CMake support,
541 slightly modified by me:
542
543 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
544 not building pcregrep.
545
546 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
547 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
548
549 25. Forward references, both numeric and by name, in patterns that made use of
550 duplicate group numbers, could behave incorrectly or give incorrect errors,
551 because when scanning forward to find the reference group, PCRE was not
552 taking into account the duplicate group numbers. A pattern such as
553 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
554
555 26. Changed a few more instances of "const unsigned char *" to USPTR, making
556 the feature of a custom pointer more persuasive (as requested by a user).
557
558 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
559 pcretest.c, inside #ifndefs, because it seems they are sometimes already
560 pre-defined.
561
562 28. Added support for (*UTF8) at the start of a pattern.
563
564 29. Arrange for flags added by the "release type" setting in CMake to be shown
565 in the configuration summary.
566
567
568 Version 7.8 05-Sep-08
569 ---------------------
570
571 1. Replaced UCP searching code with optimized version as implemented for Ad
572 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
573 stage table and inline lookup instead of a function, giving speed ups of 2
574 to 5 times on some simple patterns that I tested. Permission was given to
575 distribute the MultiStage2.py script that generates the tables (it's not in
576 the tarball, but is in the Subversion repository).
577
578 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
579 scripts.
580
581 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
582 a group with a zero qualifier. The result of the study could be incorrect,
583 or the function might crash, depending on the pattern.
584
585 4. Caseless matching was not working for non-ASCII characters in back
586 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
587 It now works when Unicode Property Support is available.
588
589 5. In pcretest, an escape such as \x{de} in the data was always generating
590 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
591 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
592 truncation.
593
594 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
595
596 7. Added two (int) casts to pcregrep when printing the difference of two
597 pointers, in case they are 64-bit values.
598
599 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
600 test 2 if it fails.
601
602 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
603 and a #define of that name to empty if it is not externally set. This is to
604 allow users of MSVC to set it if necessary.
605
606 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
607 the convenience functions in the pcre_get.c source file.
608
609 11. An option change at the start of a pattern that had top-level alternatives
610 could cause overwriting and/or a crash. This command provoked a crash in
611 some environments:
612
613 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
614
615 This potential security problem was recorded as CVE-2008-2371.
616
617 12. For a pattern where the match had to start at the beginning or immediately
618 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
619 pcre_dfa_exec() could read past the end of the passed subject if there was
620 no match. To help with detecting such bugs (e.g. with valgrind), I modified
621 pcretest so that it places the subject at the end of its malloc-ed buffer.
622
623 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
624 exec() might read past the end of the data buffer in UTF-8 mode.
625
626 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
627 the data contained the byte 0x85 as part of a UTF-8 character within its
628 first line. This applied both to normal and DFA matching.
629
630 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
631 /^[^d]*?$/8 failed to match "abc".
632
633 16. Added a missing copyright notice to pcrecpp_internal.h.
634
635 17. Make it more clear in the documentation that values returned from
636 pcre_exec() in ovector are byte offsets, not character counts.
637
638 18. Tidied a few places to stop certain compilers from issuing warnings.
639
640 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
641 supplied by Stefan Weber. I made a further small update for 7.8 because
642 there is a change of source arrangements: the pcre_searchfuncs.c module is
643 replaced by pcre_ucd.c.
644
645
646 Version 7.7 07-May-08
647 ---------------------
648
649 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
650 a string to a long long, pretend we don't even have a long long." This is
651 done by checking for the strtoq, strtoll, and _strtoi64 functions.
652
653 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
654 pre-7.6 versions, which defined a global no_arg variable instead of putting
655 it in the RE class. (See also #8 below.)
656
657 3. Remove a line of dead code, identified by coverity and reported by Nuno
658 Lopes.
659
660 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
661
662 (1) The include/exclude patterns were being applied to the whole pathnames
663 of files, instead of just to the final components.
664
665 (2) If there was more than one level of directory, the subdirectories were
666 skipped unless they satisfied the include/exclude conditions. This is
667 inconsistent with GNU grep (and could even be seen as contrary to the
668 pcregrep specification - which I improved to make it absolutely clear).
669 The action now is always to scan all levels of directory, and just
670 apply the include/exclude patterns to regular files.
671
672 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
673 --exclude_dir in the tests to avoid scanning .svn directories.
674
675 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
676 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
677 doesn't support NULs in patterns.
678
679 7. Added some missing "const"s to declarations of static tables in
680 pcre_compile.c and pcre_dfa_exec.c.
681
682 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
683 caused by fix #2 above. (Subsequently also a second patch to fix the
684 first patch. And a third patch - this was a messy problem.)
685
686 9. Applied Craig's patch to remove the use of push_back().
687
688 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
689 matching function regexec().
690
691 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
692 which, however, unlike Perl's \g{...}, are subroutine calls, not back
693 references. PCRE supports relative numbers with this syntax (I don't think
694 Oniguruma does).
695
696 12. Previously, a group with a zero repeat such as (...){0} was completely
697 omitted from the compiled regex. However, this means that if the group
698 was called as a subroutine from elsewhere in the pattern, things went wrong
699 (an internal error was given). Such groups are now left in the compiled
700 pattern, with a new opcode that causes them to be skipped at execution
701 time.
702
703 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
704 to the way PCRE behaves:
705
706 (a) A lone ] character is dis-allowed (Perl treats it as data).
707
708 (b) A back reference to an unmatched subpattern matches an empty string
709 (Perl fails the current match path).
710
711 (c) A data ] in a character class must be notated as \] because if the
712 first data character in a class is ], it defines an empty class. (In
713 Perl it is not possible to have an empty class.) The empty class []
714 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
715 The negative empty class [^] matches any one character, independently
716 of the DOTALL setting.
717
718 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
719 non-existent subpattern following a character class starting with ']' and
720 containing () gave an internal compiling error instead of "reference to
721 non-existent subpattern". Fortunately, when the pattern did exist, the
722 compiled code was correct. (When scanning forwards to check for the
723 existencd of the subpattern, it was treating the data ']' as terminating
724 the class, so got the count wrong. When actually compiling, the reference
725 was subsequently set up correctly.)
726
727 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
728 it was being rejected as not supported by pcre_dfa_exec(), even though
729 other assertions are supported. I have made pcre_dfa_exec() support
730 (*FAIL).
731
732 16. The implementation of 13c above involved the invention of a new opcode,
733 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
734 cannot be changed at match time, I realized I could make a small
735 improvement to matching performance by compiling OP_ALLANY instead of
736 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
737 on the OP_ANY path.
738
739 17. Compiling pcretest on Windows with readline support failed without the
740 following two fixes: (1) Make the unistd.h include conditional on
741 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
742
743 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
744 ncurses library to be included for pcretest when ReadLine support is
745 requested, but also to allow for it to be overridden. This patch came from
746 Daniel Bergström.
747
748 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
749 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
750 any errors with the current Unicode tables. Thanks to Peter Kankowski for
751 spotting this.
752
753
754 Version 7.6 28-Jan-08
755 ---------------------
756
757 1. A character class containing a very large number of characters with
758 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
759 overflow.
760
761 2. Patch to cut out the "long long" test in pcrecpp_unittest when
762 HAVE_LONG_LONG is not defined.
763
764 3. Applied Christian Ehrlicher's patch to update the CMake build files to
765 bring them up to date and include new features. This patch includes:
766
767 - Fixed PH's badly added libz and libbz2 support.
768 - Fixed a problem with static linking.
769 - Added pcredemo. [But later removed - see 7 below.]
770 - Fixed dftables problem and added an option.
771 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
772 HAVE_LONG_LONG.
773 - Added readline support for pcretest.
774 - Added an listing of the option settings after cmake has run.
775
776 4. A user submitted a patch to Makefile that makes it easy to create
777 "pcre.dll" under mingw when using Configure/Make. I added stuff to
778 Makefile.am that cause it to include this special target, without
779 affecting anything else. Note that the same mingw target plus all
780 the other distribution libraries and programs are now supported
781 when configuring with CMake (see 6 below) instead of with
782 Configure/Make.
783
784 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
785 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
786 exported in the Windows port". It has not yet been confirmed that the patch
787 solves the problem, but it does no harm.
788
789 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
790 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
791 with CMake, and also correct the comment about stack recursion.
792
793 7. Remove the automatic building of pcredemo from the ./configure system and
794 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
795 of a program that users should build themselves after PCRE is installed, so
796 building it automatically is not really right. What is more, it gave
797 trouble in some build environments.
798
799 8. Further tidies to CMakeLists.txt from Sheri and Christian.
800
801
802 Version 7.5 10-Jan-08
803 ---------------------
804
805 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
806 values in parens when parsing an RE using the C++ wrapper."
807
808 2. Negative specials like \S did not work in character classes in UTF-8 mode.
809 Characters greater than 255 were excluded from the class instead of being
810 included.
811
812 3. The same bug as (2) above applied to negated POSIX classes such as
813 [:^space:].
814
815 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
816 defined or documented. It seems to have been a typo for PCRE_STATIC, so
817 I have changed it.
818
819 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
820 first named subpattern) and a construct such as (?&a) would reference the
821 first named subpattern whose name started with "a" (in other words, the
822 length check was missing). Both these problems are fixed. "Subpattern name
823 expected" is now given for (?&) (a zero-length name), and this patch also
824 makes it give the same error for \k'' (previously it complained that that
825 was a reference to a non-existent subpattern).
826
827 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
828 this is right because (?- can be followed by option settings as well as by
829 digits. I have, however, made the messages clearer.
830
831 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
832 than the number used in the conditional) now cause a compile-time error.
833 This is actually not compatible with Perl, which accepts such patterns, but
834 treats the conditional as always being FALSE (as PCRE used to), but it
835 seems to me that giving a diagnostic is better.
836
837 8. Change "alphameric" to the more common word "alphanumeric" in comments
838 and messages.
839
840 9. Fix two occurrences of "backslash" in comments that should have been
841 "backspace".
842
843 10. Remove two redundant lines of code that can never be obeyed (their function
844 was moved elsewhere).
845
846 11. The program that makes PCRE's Unicode character property table had a bug
847 which caused it to generate incorrect table entries for sequences of
848 characters that have the same character type, but are in different scripts.
849 It amalgamated them into a single range, with the script of the first of
850 them. In other words, some characters were in the wrong script. There were
851 thirteen such cases, affecting characters in the following ranges:
852
853 U+002b0 - U+002c1
854 U+0060c - U+0060d
855 U+0061e - U+00612
856 U+0064b - U+0065e
857 U+0074d - U+0076d
858 U+01800 - U+01805
859 U+01d00 - U+01d77
860 U+01d9b - U+01dbf
861 U+0200b - U+0200f
862 U+030fc - U+030fe
863 U+03260 - U+0327f
864 U+0fb46 - U+0fbb1
865 U+10450 - U+1049d
866
867 12. The -o option (show only the matching part of a line) for pcregrep was not
868 compatible with GNU grep in that, if there was more than one match in a
869 line, it showed only the first of them. It now behaves in the same way as
870 GNU grep.
871
872 13. If the -o and -v options were combined for pcregrep, it printed a blank
873 line for every non-matching line. GNU grep prints nothing, and pcregrep now
874 does the same. The return code can be used to tell if there were any
875 non-matching lines.
876
877 14. Added --file-offsets and --line-offsets to pcregrep.
878
879 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
880 infinitely looping recursion. The bug was that positive lookaheads were not
881 being skipped when checking for a possible empty match (negative lookaheads
882 and both kinds of lookbehind were skipped).
883
884 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
885 inclusion of <windows.h> to before rather than after the definition of
886 INVALID_FILE_ATTRIBUTES (patch from David Byron).
887
888 17. Specifying a possessive quantifier with a specific limit for a Unicode
889 character property caused pcre_compile() to compile bad code, which led at
890 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
891 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
892 caused the error; without that there was no problem.
893
894 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
895
896 19. Added --enable-pcretest-libreadline.
897
898 20. In pcrecpp.cc, the variable 'count' was incremented twice in
899 RE::GlobalReplace(). As a result, the number of replacements returned was
900 double what it should be. I removed one of the increments, but Craig sent a
901 later patch that removed the other one (the right fix) and added unit tests
902 that check the return values (which was not done before).
903
904 21. Several CMake things:
905
906 (1) Arranged that, when cmake is used on Unix, the libraries end up with
907 the names libpcre and libpcreposix, not just pcre and pcreposix.
908
909 (2) The above change means that pcretest and pcregrep are now correctly
910 linked with the newly-built libraries, not previously installed ones.
911
912 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
913
914 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
915 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
916 UTF-8 newline character). The key issue is that the pattern starts .*;
917 this means that the match must be either at the beginning, or after a
918 newline. The bug was in the code for advancing after a failed match and
919 checking that the new position followed a newline. It was not taking
920 account of UTF-8 characters correctly.
921
922 23. PCRE was behaving differently from Perl in the way it recognized POSIX
923 character classes. PCRE was not treating the sequence [:...:] as a
924 character class unless the ... were all letters. Perl, however, seems to
925 allow any characters between [: and :], though of course it rejects as
926 unknown any "names" that contain non-letters, because all the known class
927 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
928 for example, whereas PCRE did not - it did not recognize a POSIX character
929 class. This seemed a bit dangerous, so the code has been changed to be
930 closer to Perl. The behaviour is not identical to Perl, because PCRE will
931 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
932 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
933 Perl does, and where it didn't before.
934
935 24. Rewrite so as to remove the single use of %n from pcregrep because in some
936 Windows environments %n is disabled by default.
937
938
939 Version 7.4 21-Sep-07
940 ---------------------
941
942 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
943 means that a class such as [\s] counted as "explicit reference to CR or
944 LF". That isn't really right - the whole point of the change was to try to
945 help when there was an actual mention of one of the two characters. So now
946 the change happens only if \r or \n (or a literal CR or LF) character is
947 encountered.
948
949 2. The 32-bit options word was also used for 6 internal flags, but the numbers
950 of both had grown to the point where there were only 3 bits left.
951 Fortunately, there was spare space in the data structure, and so I have
952 moved the internal flags into a new 16-bit field to free up more option
953 bits.
954
955 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
956 but did not set the internal JCHANGED flag - either of these is enough to
957 control the way the "get" function works - but the PCRE_INFO_JCHANGED
958 facility is supposed to tell if (?J) was ever used, so now (?J) at the
959 start sets both bits.
960
961 4. Added options (at build time, compile time, exec time) to change \R from
962 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
963
964 5. doc/pcresyntax.html was missing from the distribution.
965
966 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
967 compatibility, even though it is no longer used.
968
969 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
970 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
971 windows.h file is present (where different names are used). [This was
972 reversed later after testing - see 16 below.]
973
974 8. Changed all #include <config.h> to #include "config.h". There were also
975 some further <pcre.h> cases that I changed to "pcre.h".
976
977 9. When pcregrep was used with the --colour option, it missed the line ending
978 sequence off the lines that it output.
979
980 10. It was pointed out to me that arrays of string pointers cause lots of
981 relocations when a shared library is dynamically loaded. A technique of
982 using a single long string with a table of offsets can drastically reduce
983 these. I have refactored PCRE in four places to do this. The result is
984 dramatic:
985
986 Originally: 290
987 After changing UCP table: 187
988 After changing error message table: 43
989 After changing table of "verbs" 36
990 After changing table of Posix names 22
991
992 Thanks to the folks working on Gregex for glib for this insight.
993
994 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
995 unicode-properties was also set.
996
997 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
998
999 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1000 checked only for CRLF.
1001
1002 14. Added casts to pcretest.c to avoid compiler warnings.
1003
1004 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1005
1006 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1007 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1008 entirely. This removes changes made in 7 above.
1009
1010 17. The CMake files have been updated, and there is now more information about
1011 building with CMake in the NON-UNIX-USE document.
1012
1013
1014 Version 7.3 28-Aug-07
1015 ---------------------
1016
1017 1. In the rejigging of the build system that eventually resulted in 7.1, the
1018 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1019 brackets there is not right, since it causes compilers to look for an
1020 installed pcre.h, not the version that is in the source that is being
1021 compiled (which of course may be different). I have changed it back to:
1022
1023 #include "pcre.h"
1024
1025 I have a vague recollection that the change was concerned with compiling in
1026 different directories, but in the new build system, that is taken care of
1027 by the VPATH setting the Makefile.
1028
1029 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1030 when the subject happened to end in the byte 0x85 (e.g. if the last
1031 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1032 characters but of course it shouldn't be taken as a newline when it is part
1033 of another character. The bug was that, for an unlimited repeat of . in
1034 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1035 characters when looking for a newline.
1036
1037 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1038
1039 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1040 in debug output.
1041
1042 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1043 long printing in the pcrecpp unittest when running under MinGW.
1044
1045 6. ESC_K was left out of the EBCDIC table.
1046
1047 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1048 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1049 limit also applies to "virtual nesting" when a pattern is recursive, and in
1050 this case 1000 isn't so big. I have been able to remove this limit at the
1051 expense of backing off one optimization in certain circumstances. Normally,
1052 when pcre_exec() would call its internal match() function recursively and
1053 immediately return the result unconditionally, it uses a "tail recursion"
1054 feature to save stack. However, when a subpattern that can match an empty
1055 string has an unlimited repetition quantifier, it no longer makes this
1056 optimization. That gives it a stack frame in which to save the data for
1057 checking that an empty string has been matched. Previously this was taken
1058 from the 1000-entry workspace that had been reserved. So now there is no
1059 explicit limit, but more stack is used.
1060
1061 8. Applied Daniel's patches to solve problems with the import/export magic
1062 syntax that is required for Windows, and which was going wrong for the
1063 pcreposix and pcrecpp parts of the library. These were overlooked when this
1064 problem was solved for the main library.
1065
1066 9. There were some crude static tests to avoid integer overflow when computing
1067 the size of patterns that contain repeated groups with explicit upper
1068 limits. As the maximum quantifier is 65535, the maximum group length was
1069 set at 30,000 so that the product of these two numbers did not overflow a
1070 32-bit integer. However, it turns out that people want to use groups that
1071 are longer than 30,000 bytes (though not repeat them that many times).
1072 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1073 made it possible to implement the integer overflow checks in a much more
1074 dynamic way, which I have now done. The artificial limitation on group
1075 length has been removed - we now have only the limit on the total length of
1076 the compiled pattern, which depends on the LINK_SIZE setting.
1077
1078 10. Fixed a bug in the documentation for get/copy named substring when
1079 duplicate names are permitted. If none of the named substrings are set, the
1080 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1081 empty string.
1082
1083 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1084 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1085 because the ] is interpreted as the first data character and the
1086 terminating ] is not found. PCRE has been made compatible with Perl in this
1087 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1088 cause memory overwriting.
1089
1090 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1091 string has been matched (to stop an infinite loop). It was not recognizing
1092 a conditional subpattern that could match an empty string if that
1093 subpattern was within another subpattern. For example, it looped when
1094 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1095 condition was not nested. This bug has been fixed.
1096
1097 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1098 past the start of the subject in the presence of bytes with the top bit
1099 set, for example "\x8aBCD".
1100
1101 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1102 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1103
1104 14. Optimized (?!) to (*FAIL).
1105
1106 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1107 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1108 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1109 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1110 does: it's just the validity check that is more restrictive.
1111
1112 16. Inserted checks for integer overflows during escape sequence (backslash)
1113 processing, and also fixed erroneous offset values for syntax errors during
1114 backslash processing.
1115
1116 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1117 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1118
1119 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1120 caused an overrun.
1121
1122 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1123 something other than just ASCII characters) inside a group that had an
1124 unlimited repeat caused a loop at compile time (while checking to see
1125 whether the group could match an empty string).
1126
1127 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1128 [\P{Any}] did so. (Error in the code for printing property names.)
1129
1130 21. An orphan \E inside a character class could cause a crash.
1131
1132 22. A repeated capturing bracket such as (A)? could cause a wild memory
1133 reference during compilation.
1134
1135 23. There are several functions in pcre_compile() that scan along a compiled
1136 expression for various reasons (e.g. to see if it's fixed length for look
1137 behind). There were bugs in these functions when a repeated \p or \P was
1138 present in the pattern. These operators have additional parameters compared
1139 with \d, etc, and these were not being taken into account when moving along
1140 the compiled data. Specifically:
1141
1142 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1143 length.
1144
1145 (b) An item such as \pL+ within a repeated group could cause crashes or
1146 loops.
1147
1148 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1149 "reference to non-existent subpattern" error.
1150
1151 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1152
1153 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1154 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1155
1156 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1157
1158 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1159 character were causing crashes (broken optimization).
1160
1161 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1162 \p or \P) caused a compile-time loop.
1163
1164 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1165 break. For example, the unstudied pattern [\r\n]A does not match the string
1166 "\r\nA" because change 7.0/46 below moves the current point on by two
1167 characters after failing to match at the start. However, the pattern \nA
1168 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1169 the same is true. There doesn't seem any very clean way out of this, but
1170 what I have chosen to do makes the common cases work: PCRE now takes note
1171 of whether there can be an explicit match for \r or \n anywhere in the
1172 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1173 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1174 pattern has explicit CR or LF references.
1175
1176 29. Added (*CR) etc for changing newline setting at start of pattern.
1177
1178
1179 Version 7.2 19-Jun-07
1180 ---------------------
1181
1182 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1183 which is apparently normally available under Windows.
1184
1185 2. Re-jig the pcregrep tests with different newline settings in an attempt
1186 to make them independent of the local environment's newline setting.
1187
1188 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1189
1190 4. Some of the "internals" tests were previously cut out when the link size
1191 was not 2, because the output contained actual offsets. The recent new
1192 "Z" feature of pcretest means that these can be cut out, making the tests
1193 usable with all link sizes.
1194
1195 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1196 stack recursion. This gives a massive performance boost under BSD, but just
1197 a small improvement under Linux. However, it saves one field in the frame
1198 in all cases.
1199
1200 6. Added more features from the forthcoming Perl 5.10:
1201
1202 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1203 recursion call. It refers to the nth most recently opened parentheses.
1204
1205 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1206 to be opened parentheses.
1207
1208 (c) Conditions that refer to capturing parentheses can be specified
1209 relatively, for example, (?(-2)... or (?(+3)...
1210
1211 (d) \K resets the start of the current match so that everything before
1212 is not part of it.
1213
1214 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1215
1216 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1217 reference syntax.
1218
1219 (g) (?| introduces a group in which the numbering of parentheses in each
1220 alternative starts with the same number.
1221
1222 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1223
1224 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1225 PCRE_INFO_JCHANGED.
1226
1227 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1228 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1229 for detecting groups that can match an empty string.
1230
1231 9. A pattern with a very large number of alternatives (more than several
1232 hundred) was running out of internal workspace during the pre-compile
1233 phase, where pcre_compile() figures out how much memory will be needed. A
1234 bit of new cunning has reduced the workspace needed for groups with
1235 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1236 workspace instead of running out of the 4096 that are available.
1237
1238 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1239
1240 11. Applied patch from Google to remove an optimization that didn't quite work.
1241 The report of the bug said:
1242
1243 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1244 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1245 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1246
1247 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1248 it matched the wrong number of bytes.
1249
1250
1251 Version 7.1 24-Apr-07
1252 ---------------------
1253
1254 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1255 that is more "standard", making use of automake and other Autotools. There
1256 is some re-arrangement of the files and adjustment of comments consequent
1257 on this.
1258
1259 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1260 for recursive directory scanning broke on some systems because the files
1261 are not scanned in any specific order and on different systems the order
1262 was different. A call to "sort" has been inserted into RunGrepTest for the
1263 approprate test as a short-term fix. In the longer term there may be an
1264 alternative.
1265
1266 3. I had an email from Eric Raymond about problems translating some of PCRE's
1267 man pages to HTML (despite the fact that I distribute HTML pages, some
1268 people do their own conversions for various reasons). The problems
1269 concerned the use of low-level troff macros .br and .in. I have therefore
1270 removed all such uses from the man pages (some were redundant, some could
1271 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1272 HTML has been updated to handle .nf/.fi and to complain if it encounters
1273 .br or .in.
1274
1275 4. Updated comments in configure.ac that get placed in config.h.in and also
1276 arranged for config.h to be included in the distribution, with the name
1277 config.h.generic, for the benefit of those who have to compile without
1278 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1279
1280 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1281 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1282 makevp.bat for the current PCRE, using the additional files
1283 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1284
1285 6. A Windows user reported a minor discrepancy with test 2, which turned out
1286 to be caused by a trailing space on an input line that had got lost in his
1287 copy. The trailing space was an accident, so I've just removed it.
1288
1289 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1290 that is needed.
1291
1292 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1293 as "const" (a) because they are and (b) because it helps the PHP
1294 maintainers who have recently made a script to detect big data structures
1295 in the php code that should be moved to the .rodata section. I remembered
1296 to update Builducptable as well, so it won't revert if ucptable.h is ever
1297 re-created.
1298
1299 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1300 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1301 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1302 support is not required. This saves 1.5-2K of code, which is important in
1303 some applications.
1304
1305 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1306 so as not to refer to the tables, even though these functions will never be
1307 called when UTF-8 support is disabled. Otherwise there are problems with a
1308 shared library.
1309
1310 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1311
1312 (a) It was defining its arguments as char * instead of void *.
1313
1314 (b) It was assuming that all moves were upwards in memory; this was true
1315 a long time ago when I wrote it, but is no longer the case.
1316
1317 The emulated memove() is provided for those environments that have neither
1318 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1319 is clearly not the case, as these two bugs were recently reported.
1320
1321 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1322 and Detrail to create the HTML documentation, the .txt form of the man
1323 pages, and it removes trailing spaces from listed files. It also creates
1324 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1325 case, it wraps all the #defines with #ifndefs. This script should be run
1326 before "make dist".
1327
1328 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1329 with Unicode property support.
1330
1331 (a) For a maximizing quantifier, if the two different cases of the
1332 character were of different lengths in their UTF-8 codings (there are
1333 some cases like this - I found 11), and the matching function had to
1334 back up over a mixture of the two cases, it incorrectly assumed they
1335 were both the same length.
1336
1337 (b) When PCRE was configured to use the heap rather than the stack for
1338 recursion during matching, it was not correctly preserving the data for
1339 the other case of a UTF-8 character when checking ahead for a match
1340 while processing a minimizing repeat. If the check also involved
1341 matching a wide character, but failed, corruption could cause an
1342 erroneous result when trying to check for a repeat of the original
1343 character.
1344
1345 13. Some tidying changes to the testing mechanism:
1346
1347 (a) The RunTest script now detects the internal link size and whether there
1348 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1349 values substituted by "configure". (The RunGrepTest script already did
1350 this for UTF-8.) The configure.ac script no longer substitutes the
1351 relevant variables.
1352
1353 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1354 with length and offset values. This means that the output is different
1355 for different internal link sizes. Test 2 is skipped for link sizes
1356 other than 2 because of this, bypassing the problem. Unfortunately,
1357 there was also a test in test 3 (the locale tests) that used /B and
1358 failed for link sizes other than 2. Rather than cut the whole test out,
1359 I have added a new /Z option to pcretest that replaces the length and
1360 offset values with spaces. This is now used to make test 3 independent
1361 of link size. (Test 2 will be tidied up later.)
1362
1363 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1364 segmentation fault instead of returning the appropriate error message.
1365
1366 15. In multiline mode when the newline sequence was set to "any", the pattern
1367 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1368 This doesn't seem right; it now treats the CRLF combination as the line
1369 ending, and so does not match in that case. It's only a pattern such as ^$
1370 that would hit this one: something like ^ABC$ would have failed after \r
1371 and then tried again after \r\n.
1372
1373 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1374 in an attempt to make files that differ only in their line terminators
1375 compare equal. This works on Linux.
1376
1377 17. Under certain error circumstances pcregrep might try to free random memory
1378 as it exited. This is now fixed, thanks to valgrind.
1379
1380 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1381 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1382 was because its rules for how to advance for /g after matching an empty
1383 string at the end of a line did not allow for this case. They now check for
1384 it specially.
1385
1386 20. pcretest is supposed to handle patterns and data of any length, by
1387 extending its buffers when necessary. It was getting this wrong when the
1388 buffer for a data line had to be extended.
1389
1390 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1391 CRLF as a newline sequence.
1392
1393 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1394 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1395 I have nevertheless tidied it up.
1396
1397 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1398
1399 24. Added a man page for pcre-config.
1400
1401
1402 Version 7.0 19-Dec-06
1403 ---------------------
1404
1405 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1406 moving to gcc 4.1.1.
1407
1408 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1409 sys/time.h, which is documented as needed for this function. It doesn't
1410 seem to matter on Linux, but it showed up on some releases of OS X.
1411
1412 3. It seems that there are systems where bytes whose values are greater than
1413 127 match isprint() in the "C" locale. The "C" locale should be the
1414 default when a C program starts up. In most systems, only ASCII printing
1415 characters match isprint(). This difference caused the output from pcretest
1416 to vary, making some of the tests fail. I have changed pcretest so that:
1417
1418 (a) When it is outputting text in the compiled version of a pattern, bytes
1419 other than 32-126 are always shown as hex escapes.
1420
1421 (b) When it is outputting text that is a matched part of a subject string,
1422 it does the same, unless a different locale has been set for the match
1423 (using the /L modifier). In this case, it uses isprint() to decide.
1424
1425 4. Fixed a major bug that caused incorrect computation of the amount of memory
1426 required for a compiled pattern when options that changed within the
1427 pattern affected the logic of the preliminary scan that determines the
1428 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1429 that the computed length was too small. The symptoms of this bug were
1430 either the PCRE error "internal error: code overflow" from pcre_compile(),
1431 or a glibc crash with a message such as "pcretest: free(): invalid next
1432 size (fast)". Examples of patterns that provoked this bug (shown in
1433 pcretest format) are:
1434
1435 /(?-x: )/x
1436 /(?x)(?-x: \s*#\s*)/
1437 /((?i)[\x{c0}])/8
1438 /(?i:[\x{c0}])/8
1439
1440 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1441 is now done differently.
1442
1443 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1444 wrapper classes; (b) implement a new function in the C++ scanner that is
1445 more efficient than the old way of doing things because it avoids levels of
1446 recursion in the regex matching; (c) add a paragraph to the documentation
1447 for the FullMatch() function.
1448
1449 6. The escape sequence \n was being treated as whatever was defined as
1450 "newline". Not only was this contrary to the documentation, which states
1451 that \n is character 10 (hex 0A), but it also went horribly wrong when
1452 "newline" was defined as CRLF. This has been fixed.
1453
1454 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1455 was being set to -1 for the "end of line" case (supposedly a value that no
1456 character can have). Though this value is never used (the check for end of
1457 line is "zero bytes in current character"), it caused compiler complaints.
1458 I've changed it to 0xffffffff.
1459
1460 8. In pcre_version.c, the version string was being built by a sequence of
1461 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1462 string (as it is for production releases) called a macro with an empty
1463 argument. The C standard says the result of this is undefined. The gcc
1464 compiler treats it as an empty string (which was what was wanted) but it is
1465 reported that Visual C gives an error. The source has been hacked around to
1466 avoid this problem.
1467
1468 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1469 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1470 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1471 of them did).
1472
1473 10. Originally, pcretest opened its input and output without "b"; then I was
1474 told that "b" was needed in some environments, so it was added for release
1475 5.0 to both the input and output. (It makes no difference on Unix-like
1476 systems.) Later I was told that it is wrong for the input on Windows. I've
1477 now abstracted the modes into two macros, to make it easier to fiddle with
1478 them, and removed "b" from the input mode under Windows.
1479
1480 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1481
1482 12. Added -help and --help to pcretest as an official way of being reminded
1483 of the options.
1484
1485 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1486 and pcrecpp.cc because they annoy compilers at high warning levels.
1487
1488 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1489
1490 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1491 scripts are not C programs :-) and which was not noticed because it works
1492 on Linux.
1493
1494 16. pcretest is supposed to handle any length of pattern and data line (as one
1495 line or as a continued sequence of lines) by extending its input buffer if
1496 necessary. This feature was broken for very long pattern lines, leading to
1497 a string of junk being passed to pcre_compile() if the pattern was longer
1498 than about 50K.
1499
1500 17. I have done a major re-factoring of the way pcre_compile() computes the
1501 amount of memory needed for a compiled pattern. Previously, there was code
1502 that made a preliminary scan of the pattern in order to do this. That was
1503 OK when PCRE was new, but as the facilities have expanded, it has become
1504 harder and harder to keep it in step with the real compile phase, and there
1505 have been a number of bugs (see for example, 4 above). I have now found a
1506 cunning way of running the real compile function in a "fake" mode that
1507 enables it to compute how much memory it would need, while actually only
1508 ever using a few hundred bytes of working memory and without too many
1509 tests of the mode. This should make future maintenance and development
1510 easier. A side effect of this work is that the limit of 200 on the nesting
1511 depth of parentheses has been removed (though this was never a serious
1512 limitation, I suspect). However, there is a downside: pcre_compile() now
1513 runs more slowly than before (30% or more, depending on the pattern). I
1514 hope this isn't a big issue. There is no effect on runtime performance.
1515
1516 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1517 newline (only possible for the last line of a file) and it was a
1518 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1519
1520 19. Added additional timing features to pcretest. (1) The -tm option now times
1521 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1522 separate command line item, by a number that specifies the number of
1523 repeats to use when timing. The default is 50000; this gives better
1524 precision, but takes uncomfortably long for very large patterns.
1525
1526 20. Extended pcre_study() to be more clever in cases where a branch of a
1527 subpattern has no definite first character. For example, (a*|b*)[cd] would
1528 previously give no result from pcre_study(). Now it recognizes that the
1529 first character must be a, b, c, or d.
1530
1531 21. There was an incorrect error "recursive call could loop indefinitely" if
1532 a subpattern (or the entire pattern) that was being tested for matching an
1533 empty string contained only one non-empty item after a nested subpattern.
1534 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1535 incorrectly, because the \d was being skipped in the check.
1536
1537 22. The pcretest program now has a new pattern option /B and a command line
1538 option -b, which is equivalent to adding /B to every pattern. This causes
1539 it to show the compiled bytecode, without the additional information that
1540 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1541 is the same as /B/I).
1542
1543 23. A new optimization is now able automatically to treat some sequences such
1544 as a*b as a*+b. More specifically, if something simple (such as a character
1545 or a simple class like \d) has an unlimited quantifier, and is followed by
1546 something that cannot possibly match the quantified thing, the quantifier
1547 is automatically "possessified".
1548
1549 24. A recursive reference to a subpattern whose number was greater than 39
1550 went wrong under certain circumstances in UTF-8 mode. This bug could also
1551 have affected the operation of pcre_study().
1552
1553 25. Realized that a little bit of performance could be had by replacing
1554 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1555
1556 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1557
1558 27. Possessive quantifiers such as a++ were previously implemented by turning
1559 them into atomic groups such as ($>a+). Now they have their own opcodes,
1560 which improves performance. This includes the automatically created ones
1561 from 23 above.
1562
1563 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1564 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1565 the first matched character to be a colon. This applied both to named and
1566 numbered groups.
1567
1568 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1569
1570 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1571 building PCRE on Windows easier, so I have included it in the distribution.
1572
1573 31. There is now a check in pcretest against a ridiculously large number being
1574 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1575 loop, the loop is abandoned.
1576
1577 32. Forward references to subpatterns in conditions such as (?(2)...) where
1578 subpattern 2 is defined later cause pcre_compile() to search forwards in
1579 the pattern for the relevant set of parentheses. This search went wrong
1580 when there were unescaped parentheses in a character class, parentheses
1581 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1582
1583 33. "Subroutine" calls and backreferences were previously restricted to
1584 referencing subpatterns earlier in the regex. This restriction has now
1585 been removed.
1586
1587 34. Added a number of extra features that are going to be in Perl 5.10. On the
1588 whole, these are just syntactic alternatives for features that PCRE had
1589 previously implemented using the Python syntax or my own invention. The
1590 other formats are all retained for compatibility.
1591
1592 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1593 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1594 also .NET compatible.
1595
1596 (b) A recursion or subroutine call to a named group can now be defined as
1597 (?&name) as well as (?P>name).
1598
1599 (c) A backreference to a named group can now be defined as \k<name> or
1600 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1601 5.10, are also .NET compatible.
1602
1603 (d) A conditional reference to a named group can now use the syntax
1604 (?(<name>) or (?('name') as well as (?(name).
1605
1606 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1607 groups (named and numbered) that are never evaluated inline, but can be
1608 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1609 is always false. There may be only one alternative in such a group.
1610
1611 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1612 as the simple (?(R). The condition is true only if the most recent
1613 recursion is that of the given number or name. It does not search out
1614 through the entire recursion stack.
1615
1616 (g) The escape \gN or \g{N} has been added, where N is a positive or
1617 negative number, specifying an absolute or relative reference.
1618
1619 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1620 some "unreachable code" warnings.
1621
1622 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1623 things, this adds five new scripts.
1624
1625 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1626 There were also incompatibilities regarding the handling of \Q..\E inside
1627 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1628 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1629
1630 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1631 matches an empty string, and forcibly breaks the loop. There were bugs in
1632 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1633 against aaaa the result was just "a" rather than "aaaa", for example. Two
1634 separate and independent bugs (that affected different cases) have been
1635 fixed.
1636
1637 39. Refactored the code to abolish the use of different opcodes for small
1638 capturing bracket numbers. This is a tidy that I avoided doing when I
1639 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1640 The new approach is not only tidier, it makes it possible to reduce the
1641 memory needed to fix the previous bug (38).
1642
1643 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1644 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1645 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1646 mode.
1647
1648 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1649 report.
1650
1651 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1652 copy construction and assignment in the C++ wrapper.
1653
1654 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1655 couple of bugs that could have given wrong results in the "--newline=crlf"
1656 case.
1657
1658 44. Added a number of casts and did some reorganization of signed/unsigned int
1659 variables following suggestions from Dair Grant. Also renamed the variable
1660 "this" as "item" because it is a C++ keyword.
1661
1662 45. Arranged for dftables to add
1663
1664 #include "pcre_internal.h"
1665
1666 to pcre_chartables.c because without it, gcc 4.x may remove the array
1667 definition from the final binary if PCRE is built into a static library and
1668 dead code stripping is activated.
1669
1670 46. For an unanchored pattern, if a match attempt fails at the start of a
1671 newline sequence, and the newline setting is CRLF or ANY, and the next two
1672 characters are CRLF, advance by two characters instead of one.
1673
1674
1675 Version 6.7 04-Jul-06
1676 ---------------------
1677
1678 1. In order to handle tests when input lines are enormously long, pcretest has
1679 been re-factored so that it automatically extends its buffers when
1680 necessary. The code is crude, but this _is_ just a test program. The
1681 default size has been increased from 32K to 50K.
1682
1683 2. The code in pcre_study() was using the value of the re argument before
1684 testing it for NULL. (Of course, in any sensible call of the function, it
1685 won't be NULL.)
1686
1687 3. The memmove() emulation function in pcre_internal.h, which is used on
1688 systems that lack both memmove() and bcopy() - that is, hardly ever -
1689 was missing a "static" storage class specifier.
1690
1691 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1692 containing an extended class (one that cannot be represented by a bitmap
1693 because it contains high-valued characters or Unicode property items, e.g.
1694 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1695 pattern, but PCRE should not loop if you do not (it no longer does).
1696 [Detail: two cases were found: (a) a repeated subpattern containing an
1697 extended class; (b) a recursive reference to a subpattern that followed a
1698 previous extended class. It wasn't skipping over the extended class
1699 correctly when UTF-8 mode was not set.]
1700
1701 5. A negated single-character class was not being recognized as fixed-length
1702 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1703 compile error "lookbehind assertion is not fixed length".
1704
1705 6. The RunPerlTest auxiliary script was showing an unexpected difference
1706 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1707 write a Perl script that can interpret lines of an input file either as
1708 byte characters or as UTF-8, which is what "perltest" was being required to
1709 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1710 can't do is switch easily at run time between having the "use utf8;" pragma
1711 or not. In the end, I fudged it by using the RunPerlTest script to insert
1712 "use utf8;" explicitly for the UTF-8 tests.
1713
1714 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1715 the end of the subject string, contrary to the documentation and to what
1716 Perl does. This was true of both matching functions. Now it matches only at
1717 the start of the subject and immediately after *internal* newlines.
1718
1719 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1720 a pointer to an int instead of a pointer to an unsigned long int. This
1721 caused problems on 64-bit systems.
1722
1723 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1724 instance of the 'standard' template library not being so standard".
1725
1726 10. There was no check on the number of named subpatterns nor the maximum
1727 length of a subpattern name. The product of these values is used to compute
1728 the size of the memory block for a compiled pattern. By supplying a very
1729 long subpattern name and a large number of named subpatterns, the size
1730 computation could be caused to overflow. This is now prevented by limiting
1731 the length of names to 32 characters, and the number of named subpatterns
1732 to 10,000.
1733
1734 11. Subpatterns that are repeated with specific counts have to be replicated in
1735 the compiled pattern. The size of memory for this was computed from the
1736 length of the subpattern and the repeat count. The latter is limited to
1737 65535, but there was no limit on the former, meaning that integer overflow
1738 could in principle occur. The compiled length of a repeated subpattern is
1739 now limited to 30,000 bytes in order to prevent this.
1740
1741 12. Added the optional facility to have named substrings with the same name.
1742
1743 13. Added the ability to use a named substring as a condition, using the
1744 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1745 are numbers (not recommended). Forward references are permitted.
1746
1747 14. Added forward references in named backreferences (if you see what I mean).
1748
1749 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1750 pattern could run off the end of the subject. For example, the pattern
1751 "(?s)(.{1,5})"8 did this with the subject "ab".
1752
1753 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1754 PCRE_CASELESS was set when matching characters that were quantified with ?
1755 or *.
1756
1757 17. A character class other than a single negated character that had a minimum
1758 but no maximum quantifier - for example [ab]{6,} - was not handled
1759 correctly by pce_dfa_exec(). It would match only one character.
1760
1761 18. A valid (though odd) pattern that looked like a POSIX character
1762 class but used an invalid character after [ (for example [[,abc,]]) caused
1763 pcre_compile() to give the error "Failed: internal error: code overflow" or
1764 in some cases to crash with a glibc free() error. This could even happen if
1765 the pattern terminated after [[ but there just happened to be a sequence of
1766 letters, a binary zero, and a closing ] in the memory that followed.
1767
1768 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1769 over the years. Originally (before any Unicode support), just the bottom 8
1770 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1771 output from "man perlunicode" includes this:
1772
1773 The regular expression compiler produces polymorphic opcodes. That
1774 is, the pattern adapts to the data and automatically switches to
1775 the Unicode character scheme when presented with Unicode data--or
1776 instead uses a traditional byte scheme when presented with byte
1777 data.
1778
1779 Sadly, a wide octal escape does not cause a switch, and in a string with
1780 no other multibyte characters, these octal escapes are treated as before.
1781 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1782 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1783 Unicode string.
1784
1785 I have not perpetrated such confusion in PCRE. Up till now, it took just
1786 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1787 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1788 translate to the appropriate multibyte character.
1789
1790 29. Applied some refactoring to reduce the number of warnings from Microsoft
1791 and Borland compilers. This has included removing the fudge introduced
1792 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1793 a warning about an unused variable.
1794
1795 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1796 characters since release 4.0, because Perl (from release 5.004) does not.
1797 [Or at least, is documented not to: some releases seem to be in conflict
1798 with the documentation.] However, when a pattern was studied with
1799 pcre_study() and all its branches started with \s, PCRE still included VT
1800 as a possible starting character. Of course, this did no harm; it just
1801 caused an unnecessary match attempt.
1802
1803 22. Removed a now-redundant internal flag bit that recorded the fact that case
1804 dependency changed within the pattern. This was once needed for "required
1805 byte" processing, but is no longer used. This recovers a now-scarce options
1806 bit. Also moved the least significant internal flag bit to the most-
1807 significant bit of the word, which was not previously used (hangover from
1808 the days when it was an int rather than a uint) to free up another bit for
1809 the future.
1810
1811 23. Added support for CRLF line endings as well as CR and LF. As well as the
1812 default being selectable at build time, it can now be changed at runtime
1813 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1814 specify that it is scanning data with non-default line endings.
1815
1816 24. Changed the definition of CXXLINK to make it agree with the definition of
1817 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1818
1819 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1820 recursions. This makes a big different to stack usage for some patterns.
1821
1822 26. If a subpattern containing a named recursion or subroutine reference such
1823 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1824 the space required for the compiled pattern went wrong and gave too small a
1825 value. Depending on the environment, this could lead to "Failed: internal
1826 error: code overflow at offset 49" or "glibc detected double free or
1827 corruption" errors.
1828
1829 27. Applied patches from Google (a) to support the new newline modes and (b) to
1830 advance over multibyte UTF-8 characters in GlobalReplace.
1831
1832 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1833 difference for some implementation of PCRE in some Windows version.
1834
1835 29. Added some extra testing facilities to pcretest:
1836
1837 \q<number> in a data line sets the "match limit" value
1838 \Q<number> in a data line sets the "match recursion limt" value
1839 -S <number> sets the stack size, where <number> is in megabytes
1840
1841 The -S option isn't available for Windows.
1842
1843
1844 Version 6.6 06-Feb-06
1845 ---------------------
1846
1847 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1848 in pcreposix.h. I have copied the definition from pcre.h.
1849
1850 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1851 because pcre.h is no longer a built file.
1852
1853 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1854 not normally included in the compiled code.
1855
1856
1857 Version 6.5 01-Feb-06
1858 ---------------------
1859
1860 1. When using the partial match feature with pcre_dfa_exec(), it was not
1861 anchoring the second and subsequent partial matches at the new starting
1862 point. This could lead to incorrect results. For example, with the pattern
1863 /1234/, partially matching against "123" and then "a4" gave a match.
1864
1865 2. Changes to pcregrep:
1866
1867 (a) All non-match returns from pcre_exec() were being treated as failures
1868 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1869 error message is output. Some extra information is given for the
1870 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1871 probably the only errors that are likely to be caused by users (by
1872 specifying a regex that has nested indefinite repeats, for instance).
1873 If there are more than 20 of these errors, pcregrep is abandoned.
1874
1875 (b) A binary zero was treated as data while matching, but terminated the
1876 output line if it was written out. This has been fixed: binary zeroes
1877 are now no different to any other data bytes.
1878
1879 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1880 used to set a locale for matching. The --locale=xxxx long option has
1881 been added (no short equivalent) to specify a locale explicitly on the
1882 pcregrep command, overriding the environment variables.
1883
1884 (d) When -B was used with -n, some line numbers in the output were one less
1885 than they should have been.
1886
1887 (e) Added the -o (--only-matching) option.
1888
1889 (f) If -A or -C was used with -c (count only), some lines of context were
1890 accidentally printed for the final match.
1891
1892 (g) Added the -H (--with-filename) option.
1893
1894 (h) The combination of options -rh failed to suppress file names for files
1895 that were found from directory arguments.
1896
1897 (i) Added the -D (--devices) and -d (--directories) options.
1898
1899 (j) Added the -F (--fixed-strings) option.
1900
1901 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1902
1903 (l) Added the --colo(u)r option.
1904
1905 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1906 is not present by default.
1907
1908 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1909 items such as (?R) or (?1), when the recursion could match a number of
1910 alternatives. If it matched one of the alternatives, but subsequently,
1911 outside the recursion, there was a failure, the code tried to back up into
1912 the recursion. However, because of the way PCRE is implemented, this is not
1913 possible, and the result was an incorrect result from the match.
1914
1915 In order to prevent this happening, the specification of recursion has
1916 been changed so that all such subpatterns are automatically treated as
1917 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1918
1919 4. I had overlooked the fact that, in some locales, there are characters for
1920 which isalpha() is true but neither isupper() nor islower() are true. In
1921 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1922 and ordfeminine) are like this. This affected the treatment of \w and \W
1923 when they appeared in character classes, but not when they appeared outside
1924 a character class. The bit map for "word" characters is now created
1925 separately from the results of isalnum() instead of just taking it from the
1926 upper, lower, and digit maps. (Plus the underscore character, of course.)
1927
1928 5. The above bug also affected the handling of POSIX character classes such as
1929 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1930 permanent tables. Instead, the bit maps for such a class were previously
1931 created as the appropriate unions of the upper, lower, and digit bitmaps.
1932 Now they are created by subtraction from the [[:word:]] class, which has
1933 its own bitmap.
1934
1935 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1936 It is created by subtracting the vertical space characters (\x09, \x0a,
1937 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1938 subtraction was done in the overall bitmap for a character class, meaning
1939 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1940 be recognized. This bug has been fixed.
1941
1942 7. Patches from the folks at Google:
1943
1944 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1945 real life, but is still worth protecting against".
1946
1947 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1948 regular expressions".
1949
1950 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1951 have it.
1952
1953 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1954 "configure" and the latter not, in order to fix a problem somebody had
1955 with compiling the Arg class on HP-UX.
1956
1957 (e) Improve the error-handling of the C++ wrapper a little bit.
1958
1959 (f) New tests for checking recursion limiting.
1960
1961 8. The pcre_memmove() function, which is used only if the environment does not
1962 have a standard memmove() function (and is therefore rarely compiled),
1963 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1964 returning a result (though PCRE never actually uses the result).
1965
1966 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1967 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1968 returned instead of calling malloc() with an overflowing number that would
1969 most likely cause subsequent chaos.
1970
1971 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1972
1973 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1974 with this option is matched, the nmatch and pmatch options of regexec() are
1975 ignored.
1976
1977 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1978 provided in case anyone wants to the the POSIX interface with UTF-8
1979 strings.
1980
1981 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1982 C++ linking (needed for some HP-UX environments).
1983
1984 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1985 (unused parameter) and in the pcre_printint() function (omitted "default"
1986 switch label when the default is to do nothing).
1987
1988 15. Added some code to make it possible, when PCRE is compiled as a C++
1989 library, to replace subject pointers for pcre_exec() with a smart pointer
1990 class, thus making it possible to process discontinuous strings.
1991
1992 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1993 much the same function. They were added by different people who were trying
1994 to make PCRE easy to compile on non-Unix systems. It has been suggested
1995 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1996 for compiling on Windows systems. I have therefore replaced it with
1997 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1998 defaults to "extern" for C or "extern C" for C++, which works fine on
1999 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2000 SCOPE with something explicit in config.h. In addition:
2001
2002 (a) pcreposix.h still had just "extern" instead of either of these macros;
2003 I have replaced it with PCRE_DATA_SCOPE.
2004
2005 (b) Functions such as _pcre_xclass(), which are internal to the library,
2006 but external in the C sense, all had PCRE_EXPORT in their definitions.
2007 This is apparently wrong for the Windows case, so I have removed it.
2008 (It makes no difference on Unix-like systems.)
2009
2010 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2011 of recursive calls to match(). This is different to MATCH_LIMIT because
2012 that limits the total number of calls to match(), not all of which increase
2013 the depth of recursion. Limiting the recursion depth limits the amount of
2014 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2015 when PCRE is compiled, and changed at run time. A patch from Google adds
2016 this functionality to the C++ interface.
2017
2018 18. Changes to the handling of Unicode character properties:
2019
2020 (a) Updated the table to Unicode 4.1.0.
2021
2022 (b) Recognize characters that are not in the table as "Cn" (undefined).
2023
2024 (c) I revised the way the table is implemented to a much improved format
2025 which includes recognition of ranges. It now supports the ranges that
2026 are defined in UnicodeData.txt, and it also amalgamates other
2027 characters into ranges. This has reduced the number of entries in the
2028 table from around 16,000 to around 3,000, thus reducing its size
2029 considerably. I realized I did not need to use a tree structure after
2030 all - a binary chop search is just as efficient. Having reduced the
2031 number of entries, I extended their size from 6 bytes to 8 bytes to
2032 allow for more data.
2033
2034 (d) Added support for Unicode script names via properties such as \p{Han}.
2035
2036 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2037 matching that character.
2038
2039 20. When matching a repeated Unicode property with a minimum greater than zero,
2040 (for example \pL{2,}), PCRE could look past the end of the subject if it
2041 reached it while seeking the minimum number of characters. This could
2042 happen only if some of the characters were more than one byte long, because
2043 there is a check for at least the minimum number of bytes.
2044
2045 21. Refactored the implementation of \p and \P so as to be more general, to
2046 allow for more different types of property in future. This has changed the
2047 compiled form incompatibly. Anybody with saved compiled patterns that use
2048 \p or \P will have to recompile them.
2049
2050 22. Added "Any" and "L&" to the supported property types.
2051
2052 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2053 but give a compile time error if the value is greater than 0xff.
2054
2055 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2056 accidentally not being installed or uninstalled.
2057
2058 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2059 made were to insert the current release number. This seemed silly, because
2060 it made things harder for people building PCRE on systems that don't run
2061 "configure". I have turned pcre.h into a distributed file, no longer built
2062 by "configure", with the version identification directly included. There is
2063 no longer a pcre.h.in file.
2064
2065 However, this change necessitated a change to the pcre-config script as
2066 well. It is built from pcre-config.in, and one of the substitutions was the
2067 release number. I have updated configure.ac so that ./configure now finds
2068 the release number by grepping pcre.h.
2069
2070 26. Added the ability to run the tests under valgrind.
2071
2072
2073 Version 6.4 05-Sep-05
2074 ---------------------
2075
2076 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2077 "--" to be printed when multiple files were scanned, even when none of the
2078 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2079 consider it to be a bug, and have restored the previous behaviour.
2080
2081 2. A couple of code tidies to get rid of compiler warnings.
2082
2083 3. The pcretest program used to cheat by referring to symbols in the library
2084 whose names begin with _pcre_. These are internal symbols that are not
2085 really supposed to be visible externally, and in some environments it is
2086 possible to suppress them. The cheating is now confined to including
2087 certain files from the library's source, which is a bit cleaner.
2088
2089 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2090 file's purpose clearer.
2091
2092 5. Reorganized pcre_ucp_findchar().
2093
2094
2095 Version 6.3 15-Aug-05
2096 ---------------------
2097
2098 1. The file libpcre.pc.in did not have general read permission in the tarball.
2099
2100 2. There were some problems when building without C++ support:
2101
2102 (a) If C++ support was not built, "make install" and "make test" still
2103 tried to test it.
2104
2105 (b) There were problems when the value of CXX was explicitly set. Some
2106 changes have been made to try to fix these, and ...
2107
2108 (c) --disable-cpp can now be used to explicitly disable C++ support.
2109
2110 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2111 backslash in a target when C++ was disabled. This confuses some
2112 versions of "make", apparently. Using an intermediate variable solves
2113 this. (Same for CPP_LOBJ.)
2114
2115 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2116 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2117 necessary on certain architectures.
2118
2119 4. Added a setting of -export-symbols-regex to the link command to remove
2120 those symbols that are exported in the C sense, but actually are local
2121 within the library, and not documented. Their names all begin with
2122 "_pcre_". This is not a perfect job, because (a) we have to except some
2123 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2124 available (and never for static libraries). I have made a note to try to
2125 find a way round (a) in the future.
2126
2127
2128 Version 6.2 01-Aug-05
2129 ---------------------
2130
2131 1. There was no test for integer overflow of quantifier values. A construction
2132 such as {1111111111111111} would give undefined results. What is worse, if
2133 a minimum quantifier for a parenthesized subpattern overflowed and became
2134 negative, the calculation of the memory size went wrong. This could have
2135 led to memory overwriting.
2136
2137 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2138
2139 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2140 operating environments where this matters.
2141
2142 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2143 PCRE options from within the C++ wrapper.
2144
2145 5. Named capturing subpatterns were not being correctly counted when a pattern
2146 was compiled. This caused two problems: (a) If there were more than 100
2147 such subpatterns, the calculation of the memory needed for the whole
2148 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2149 back references of the form \12, where the number was greater than 9, were
2150 not recognized as back references, even though there were sufficient
2151 previous subpatterns.
2152
2153 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2154 versions of gcc, e.g. 2.95.4.
2155
2156
2157 Version 6.1 21-Jun-05
2158 ---------------------
2159
2160 1. There was one reference to the variable "posix" in pcretest.c that was not
2161 surrounded by "#if !defined NOPOSIX".
2162
2163 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2164 the cross-check on the old pcre_info() function, for the benefit of the
2165 cut-down version of PCRE that is currently imported into Exim.
2166
2167 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2168 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2169 patterns that start (?i) but I don't think that matters. The use of (?i) is
2170 just an example; this all applies to the other options as well.
2171
2172 4. Since libtool seems to echo the compile commands it is issuing, the output
2173 from "make" can be reduced a bit by putting "@" in front of each libtool
2174 compile command.
2175
2176 5. Patch from the folks at Google for configure.in to be a bit more thorough
2177 in checking for a suitable C++ installation before trying to compile the
2178 C++ stuff. This should fix a reported problem when a compiler was present,
2179 but no suitable headers.
2180
2181 6. The man pages all had just "PCRE" as their title. I have changed them to
2182 be the relevant file name. I have also arranged that these names are
2183 retained in the file doc/pcre.txt, which is a concatenation in text format
2184 of all the man pages except the little individual ones for each function.
2185
2186 7. The NON-UNIX-USE file had not been updated for the different set of source
2187 files that come with release 6. I also added a few comments about the C++
2188 wrapper.
2189
2190
2191 Version 6.0 07-Jun-05
2192 ---------------------
2193
2194 1. Some minor internal re-organization to help with my DFA experiments.
2195
2196 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2197 didn't matter for the library itself when fully configured, but did matter
2198 when compiling without UCP support, or within Exim, where the ucp files are
2199 not imported.
2200
2201 3. Refactoring of the library code to split up the various functions into
2202 different source modules. The addition of the new DFA matching code (see
2203 below) to a single monolithic source would have made it really too
2204 unwieldy, quite apart from causing all the code to be include in a
2205 statically linked application, when only some functions are used. This is
2206 relevant even without the DFA addition now that patterns can be compiled in
2207 one application and matched in another.
2208
2209 The downside of splitting up is that there have to be some external
2210 functions and data tables that are used internally in different modules of
2211 the library but which are not part of the API. These have all had their
2212 names changed to start with "_pcre_" so that they are unlikely to clash
2213 with other external names.
2214
2215 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2216 a different (DFA) algorithm. Although it is slower than the original
2217 function, it does have some advantages for certain types of matching
2218 problem.
2219
2220 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2221 including restarting after a partial match.
2222
2223 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2224 defined when compiling for Windows was sent to me. I have put it into the
2225 code, though I have no means of testing or verifying it.
2226
2227 7. Added the pcre_refcount() auxiliary function.
2228
2229 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2230 match before or at the first newline in the subject string. In pcretest,
2231 the /f option on a pattern can be used to set this.
2232
2233 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2234 would behave wrongly. This has been present in PCRE since release 4.0.
2235
2236 10. A number of changes to the pcregrep command:
2237
2238 (a) Refactored how -x works; insert ^(...)$ instead of setting
2239 PCRE_ANCHORED and checking the length, in preparation for adding
2240 something similar for -w.
2241
2242 (b) Added the -w (match as a word) option.
2243
2244 (c) Refactored the way lines are read and buffered so as to have more
2245 than one at a time available.
2246
2247 (d) Implemented a pcregrep test script.
2248
2249 (e) Added the -M (multiline match) option. This allows patterns to match
2250 over several lines of the subject. The buffering ensures that at least
2251 8K, or the rest of the document (whichever is the shorter) is available
2252 for matching (and similarly the previous 8K for lookbehind assertions).
2253
2254 (f) Changed the --help output so that it now says
2255
2256 -w, --word-regex(p)
2257
2258 instead of two lines, one with "regex" and the other with "regexp"
2259 because that confused at least one person since the short forms are the
2260 same. (This required a bit of code, as the output is generated
2261 automatically from a table. It wasn't just a text change.)
2262
2263 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2264 option but starts with a hyphen. Could be a pattern or a path name
2265 starting with a hyphen, for instance.
2266
2267 (h) "-" can be given as a file name to represent stdin.
2268
2269 (i) When file names are being printed, "(standard input)" is used for
2270 the standard input, for compatibility with GNU grep. Previously
2271 "<stdin>" was used.
2272
2273 (j) The option --label=xxx can be used to supply a name to be used for
2274 stdin when file names are being printed. There is no short form.
2275
2276 (k) Re-factored the options decoding logic because we are going to add
2277 two more options that take data. Such options can now be given in four
2278 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2279
2280 (l) Added the -A, -B, and -C options for requesting that lines of context
2281 around matches be printed.
2282
2283 (m) Added the -L option to print the names of files that do not contain
2284 any matching lines, that is, the complement of -l.
2285
2286 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2287 continue to scan other files.
2288
2289 (o) The -s option was incorrectly implemented. For compatibility with other
2290 greps, it now suppresses the error message for a non-existent or non-
2291 accessible file (but not the return code). There is a new option called
2292 -q that suppresses the output of matching lines, which was what -s was
2293 previously doing.
2294
2295 (p) Added --include and --exclude options to specify files for inclusion
2296 and exclusion when recursing.
2297
2298 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2299 Hopefully, it now does.
2300
2301 12. Missing cast in pcre_study().
2302
2303 13. Added an "uninstall" target to the makefile.
2304
2305 14. Replaced "extern" in the function prototypes in Makefile.in with
2306 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2307 world, but is set differently for Windows.
2308
2309 15. Added a second compiling function called pcre_compile2(). The only
2310 difference is that it has an extra argument, which is a pointer to an
2311 integer error code. When there is a compile-time failure, this is set
2312 non-zero, in addition to the error test pointer being set to point to an
2313 error message. The new argument may be NULL if no error number is required
2314 (but then you may as well call pcre_compile(), which is now just a
2315 wrapper). This facility is provided because some applications need a
2316 numeric error indication, but it has also enabled me to tidy up the way
2317 compile-time errors are handled in the POSIX wrapper.
2318
2319 16. Added VPATH=.libs to the makefile; this should help when building with one
2320 prefix path and installing with another. (Or so I'm told by someone who
2321 knows more about this stuff than I do.)
2322
2323 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2324 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2325 match everything, including newlines. This is not POSIX-compatible, but
2326 somebody wanted the feature. From pcretest it can be activated by using
2327 both the P and the s flags.
2328
2329 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2330
2331 19. libpcre.pc was being incorrectly installed as executable.
2332
2333 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2334 it now also looks for '\r' so that it will work unmodified on Windows.
2335
2336 21. Added Google's contributed C++ wrapper to the distribution.
2337
2338 22. Added some untidy missing memory free() calls in pcretest, to keep
2339 Electric Fence happy when testing.
2340
2341
2342
2343 Version 5.0 13-Sep-04
2344 ---------------------
2345
2346 1. Internal change: literal characters are no longer packed up into items
2347 containing multiple characters in a single byte-string. Each character
2348 is now matched using a separate opcode. However, there may be more than one
2349 byte in the character in UTF-8 mode.
2350
2351 2. The pcre_callout_block structure has two new fields: pattern_position and
2352 next_item_length. These contain the offset in the pattern to the next match
2353 item, and its length, respectively.
2354
2355 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2356 insertion of callouts before each pattern item. Added the /C option to
2357 pcretest to make use of this.
2358
2359 4. On the advice of a Windows user, the lines
2360
2361 #if defined(_WIN32) || defined(WIN32)
2362 _setmode( _fileno( stdout ), 0x8000 );
2363 #endif /* defined(_WIN32) || defined(WIN32) */
2364
2365 have been added to the source of pcretest. This apparently does useful
2366 magic in relation to line terminators.
2367
2368 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2369 for the benefit of those environments where the "b" makes a difference.
2370
2371 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2372 to know about it. I have put a hack into configure.in that adds in code
2373 to set GCC=yes if CC=icc. This seems to end up at a point in the
2374 generated configure script that is early enough to affect the setting of
2375 compiler options, which is what is needed, but I have no means of testing
2376 whether it really works. (The user who reported this had patched the
2377 generated configure script, which of course I cannot do.)
2378
2379 LATER: After change 22 below (new libtool files), the configure script
2380 seems to know about icc (and also ecc). Therefore, I have commented out
2381 this hack in configure.in.
2382
2383 7. Added support for pkg-config (2 patches were sent in).
2384
2385 8. Negated POSIX character classes that used a combination of internal tables
2386 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2387 [[:^ascii]]. Typically, they would match almost any characters. The other
2388 POSIX classes were not broken in this way.
2389
2390 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2391 to find the match, as PCRE was deluded into thinking that the match had to
2392 start at the start point or following a newline. The same bug applied to
2393 patterns with negative forward assertions or any backward assertions
2394 preceding ".*" at the start, unless the pattern required a fixed first
2395 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2396
2397 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2398 starting at the last subject character, bytes beyond the end of the subject
2399 string were read.
2400
2401 11. Renamed the variable "class" as "classbits" to make life easier for C++
2402 users. (Previously there was a macro definition, but it apparently wasn't
2403 enough.)
2404
2405 12. Added the new field "tables" to the extra data so that tables can be passed
2406 in at exec time, or the internal tables can be re-selected. This allows
2407 a compiled regex to be saved and re-used at a later time by a different
2408 program that might have everything at different addresses.
2409
2410 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2411 -R library as well as a -L library.
2412
2413 14. The debugging options of pcretest (-d on the command line or D on a
2414 pattern) showed incorrect output for anything following an extended class
2415 that contained multibyte characters and which was followed by a quantifier.
2416
2417 15. Added optional support for general category Unicode character properties
2418 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2419 support. It adds about 90K to the size of the library. The meanings of the
2420 inbuilt class escapes such as \d and \s have NOT been changed.
2421
2422 16. Updated pcredemo.c to include calls to free() to release the memory for the
2423 compiled pattern.
2424
2425 17. The generated file chartables.c was being created in the source directory
2426 instead of in the building directory. This caused the build to fail if the
2427 source directory was different from the building directory, and was
2428 read-only.
2429
2430 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2431 file. No doubt somebody will tell me if they don't make sense... Also added
2432 Dan Mooney's comments about building on OpenVMS.
2433
2434 19. Added support for partial matching via the PCRE_PARTIAL option for
2435 pcre_exec() and the \P data escape in pcretest.
2436
2437 20. Extended pcretest with 3 new pattern features:
2438
2439 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2440 write the compiled pattern to the file whose name is "rest-of-line".
2441 This is a straight binary dump of the data, with the saved pointer to
2442 the character tables forced to be NULL. The study data, if any, is
2443 written too. After writing, pcretest reads a new pattern.
2444
2445 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2446 compiled pattern from the given file. There must not be any
2447 occurrences of "<" in the file name (pretty unlikely); if there are,
2448 pcretest will instead treat the initial "<" as a pattern delimiter.
2449 After reading in the pattern, pcretest goes on to read data lines as
2450 usual.
2451
2452 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2453 and 16-bit fields in a compiled pattern, to simulate a pattern that
2454 was compiled on a host of opposite endianness.
2455
2456 21. The pcre-exec() function can now cope with patterns that were compiled on
2457 hosts of opposite endianness, with this restriction:
2458
2459 As for any compiled expression that is saved and used later, the tables
2460 pointer field cannot be preserved; the extra_data field in the arguments
2461 to pcre_exec() should be used to pass in a tables address if a value
2462 other than the default internal tables were used at compile time.
2463
2464 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2465 now diagnosed as an error. Previously, most of the time, a negative number
2466 would have been treated as zero, but if in addition "ovector" was passed as
2467 NULL, a crash could occur.
2468
2469 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2470 new versions from the libtool 1.5 distribution (the last one is a copy of
2471 a file called libtool.m4). This seems to have fixed the need to patch
2472 "configure" to support Darwin 1.3 (which I used to do). However, I still
2473 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2474 workstation).
2475
2476 24. Changed the PCRE licence to be the more standard "BSD" licence.
2477
2478
2479 Version 4.5 01-Dec-03
2480 ---------------------
2481
2482 1. There has been some re-arrangement of the code for the match() function so
2483 that it can be compiled in a version that does not call itself recursively.
2484 Instead, it keeps those local variables that need separate instances for
2485 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2486 needs to "recurse". Keeping track of where control must go is done by means
2487 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2488 hide most of the details from the main code, and operates only if
2489 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2490 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2491 operating.
2492
2493 To make it easier for callers to provide specially tailored get/free
2494 functions for this usage, two new functions, pcre_stack_malloc, and
2495 pcre_stack_free, are used. They are always called in strict stacking order,
2496 and the size of block requested is always the same.
2497
2498 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2499 PCRE has been compiled to use the stack or the heap for recursion. The
2500 -C option of pcretest uses this to show which version is compiled.
2501
2502 A new data escape \S, is added to pcretest; it causes the amounts of store
2503 obtained and freed by both kinds of malloc/free at match time to be added
2504 to the output.
2505
2506 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2507 what's available on my current Linux desktop machine.
2508
2509 3. When matching a UTF-8 string, the test for a valid string at the start has
2510 been extended. If start_offset is not zero, PCRE now checks that it points
2511 to a byte that is the start of a UTF-8 character. If not, it returns
2512 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2513 this is necessary because there may be backward assertions in the pattern.
2514 When matching the same subject several times, it may save resources to use
2515 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2516
2517 4. The code for checking the validity of UTF-8 strings has been tightened so
2518 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2519 containing "overlong sequences".
2520
2521 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2522 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2523 should have been "&", but it just so happened that all the cases this let
2524 through by mistake were picked up later in the function.
2525
2526 6. I had used a variable called "isblank" - this is a C99 function, causing
2527 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2528
2529 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2530 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2531 so the version is shown; (c) stop "make test" echoing "./RunTest".
2532
2533 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2534
2535 9. The prototype for memmove() for systems that don't have it was using
2536 size_t, but the inclusion of the header that defines size_t was later. I've
2537 moved the #includes for the C headers earlier to avoid this.
2538
2539 10. Added some adjustments to the code to make it easier to compiler on certain
2540 special systems:
2541
2542 (a) Some "const" qualifiers were missing.
2543 (b) Added the macro EXPORT before all exported functions; by default this
2544 is defined to be empty.
2545 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2546 that it reads its output file name as an argument instead of writing
2547 to the standard output and assuming this can be redirected.
2548
2549 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2550 class containing characters with values greater than 255, PCRE compilation
2551 went into a loop.
2552
2553 12. A recursive reference to a subpattern that was within another subpattern
2554 that had a minimum quantifier of zero caused PCRE to crash. For example,
2555 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2556 recursion. If the recursively-called subpattern itself had a zero repeat,
2557 that was OK.
2558
2559 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2560 buffer into which it was copied (for escape processing) was still set at
2561 1024, so long lines caused crashes.
2562
2563 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2564 "internal error: code overflow...". This applied to any character class
2565 that was followed by a possessive quantifier.
2566
2567 15. Modified the Makefile to add libpcre.la as a prerequisite for
2568 libpcreposix.la because I was told this is needed for a parallel build to
2569 work.
2570
2571 16. If a pattern that contained .* following optional items at the start was
2572 studied, the wrong optimizing data was generated, leading to matching
2573 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2574 matching string must start with a or b or c. The correct conclusion for
2575 this pattern is that a match can start with any character.
2576
2577
2578 Version 4.4 13-Aug-03
2579 ---------------------
2580
2581 1. In UTF-8 mode, a character class containing characters with values between
2582 127 and 255 was not handled correctly if the compiled pattern was studied.
2583 In fixing this, I have also improved the studying algorithm for such
2584 classes (slightly).
2585
2586 2. Three internal functions had redundant arguments passed to them. Removal
2587 might give a very teeny performance improvement.
2588
2589 3. Documentation bug: the value of the capture_top field in a callout is *one
2590 more than* the number of the hightest numbered captured substring.
2591
2592 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2593 in incorrectly linking with a previously installed version. They now link
2594 explicitly with libpcre.la.
2595
2596 5. configure.in no longer needs to recognize Cygwin specially.
2597
2598 6. A problem in pcre.in for Windows platforms is fixed.
2599
2600 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2601 pcretest, it used to include the size of the study block as part of its
2602 output. Unfortunately, the structure contains a field that has a different
2603 size on different hardware architectures. This meant that the tests that
2604 showed this size failed. As the block is currently always of a fixed size,
2605 this information isn't actually particularly useful in pcretest output, so
2606 I have just removed it.
2607
2608 8. Three pre-processor statements accidentally did not start in column 1.
2609 Sadly, there are *still* compilers around that complain, even though
2610 standard C has not required this for well over a decade. Sigh.
2611
2612 9. In pcretest, the code for checking callouts passed small integers in the
2613 callout_data field, which is a void * field. However, some picky compilers
2614 complained about the casts involved for this on 64-bit systems. Now
2615 pcretest passes the address of the small integer instead, which should get
2616 rid of the warnings.
2617
2618 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2619 both compile and run time, and gives an error if an invalid UTF-8 sequence
2620 is found. There is a option for disabling this check in cases where the
2621 string is known to be correct and/or the maximum performance is wanted.
2622
2623 11. In response to a bug report, I changed one line in Makefile.in from
2624
2625 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2626 to
2627 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2628
2629 to look similar to other lines, but I have no way of telling whether this
2630 is the right thing to do, as I do not use Windows. No doubt I'll get told
2631 if it's wrong...
2632
2633
2634 Version 4.3 21-May-03
2635 ---------------------
2636
2637 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2638 Makefile.
2639
2640 2. Some refactoring to improve the quality of the code:
2641
2642 (i) The utf8_table... variables are now declared "const".
2643
2644 (ii) The code for \cx, which used the "case flipping" table to upper case
2645 lower case letters, now just substracts 32. This is ASCII-specific,
2646 but the whole concept of \cx is ASCII-specific, so it seems
2647 reasonable.
2648
2649 (iii) PCRE was using its character types table to recognize decimal and
2650 hexadecimal digits in the pattern. This is silly, because it handles
2651 only 0-9, a-f, and A-F, but the character types table is locale-
2652 specific, which means strange things might happen. A private
2653 table is now used for this - though it costs 256 bytes, a table is
2654 much faster than multiple explicit tests. Of course, the standard
2655 character types table is still used for matching digits in subject
2656 strings against \d.
2657
2658 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2659 ending in _t are). So I've renamed it as ESC_tee.
2660
2661 3. The first argument for regexec() in the POSIX wrapper should have been
2662 defined as "const".
2663
2664 4. Changed pcretest to use malloc() for its buffers so that they can be
2665 Electric Fenced for debugging.
2666
2667 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2668 to read one or more bytes before the start of the subject string. Often this
2669 had no effect on PCRE's behaviour, but in some circumstances it could
2670 provoke a segmentation fault.
2671
2672 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2673 to try to read one or more bytes before the start of the subject string.
2674
2675 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2676 UTF-8 support could misbehave in various ways if the subject string
2677 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2678 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2679 back over UTF-8 characters.)
2680
2681
2682 Version 4.2 14-Apr-03
2683 ---------------------
2684
2685 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2686
2687 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2688 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2689 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2690 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2691 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2692 and BUILD_EXEEXT
2693 Note: automatic setting of the BUILD variables is not yet working
2694 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2695 compile-time but not at link-time
2696 [LINK]: use for linking executables only
2697 make different versions for Windows and non-Windows
2698 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2699 libraries
2700 [LINK_FOR_BUILD]: new variable
2701 [OBJEXT]: use throughout
2702 [EXEEXT]: use throughout
2703 <winshared>: new target
2704 <wininstall>: new target
2705 <dftables.o>: use native compiler
2706 <dftables>: use native linker
2707 <install>: handle Windows platform correctly
2708 <clean>: ditto
2709 <check>: ditto
2710 copy DLL to top builddir before testing
2711
2712 As part of these changes, -no-undefined was removed again. This was reported
2713 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2714 in any case.
2715
2716 3. Some tidies to get rid of compiler warnings:
2717
2718 . In the match_data structure, match_limit was an unsigned long int, whereas
2719 match_call_count was an int. I've made them both unsigned long ints.
2720
2721 . In pcretest the fact that a const uschar * doesn't automatically cast to
2722 a void * provoked a warning.
2723
2724 . Turning on some more compiler warnings threw up some "shadow" variables
2725 and a few more missing casts.
2726
2727 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2728 option, a class that contained a single character with a value between 128
2729 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2730
2731 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2732 option, a class that contained several characters, but with at least one
2733 whose value was between 128 and 255 caused PCRE to crash.
2734
2735
2736 Version 4.1 12-Mar-03
2737 ---------------------
2738
2739 1. Compiling with gcc -pedantic found a couple of places where casts were
2740 needed, and a string in dftables.c that was longer than standard compilers are
2741 required to support.
2742
2743 2. Compiling with Sun's compiler found a few more places where the code could
2744 be tidied up in order to avoid warnings.
2745
2746 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2747 first of these names is deprecated in the latest Autoconf in favour of the name
2748 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2749 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2750 analogy I have changed it to CFLAGS_FOR_BUILD.
2751
2752 4. Added -no-undefined to the linking command in the Makefile, because this is
2753 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2754 linking step for the pcreposix library.
2755
2756 5. PCRE was failing to diagnose the case of two named groups with the same
2757 name.
2758
2759 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2760 literal character that is needed in the subject for a match, and scans along to
2761 ensure that it is present before embarking on the full matching process. This
2762 saves time in cases of nested unlimited repeats that are never going to match.
2763 Problem: the scan can take a lot of time if the subject is very long (e.g.
2764 megabytes), thus penalizing straightforward matches. It is now done only if the
2765 amount of subject to be scanned is less than 1000 bytes.
2766
2767 7. A lesser problem with the same optimization is that it was recording the
2768 first character of an anchored pattern as "needed", thus provoking a search
2769 right along the subject, even when the first match of the pattern was going to
2770 fail. The "needed" character is now not set for anchored patterns, unless it
2771 follows something in the pattern that is of non-fixed length. Thus, it still
2772 fulfils its original purpose of finding quick non-matches in cases of nested
2773 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2774
2775
2776 Version 4.0 17-Feb-03
2777 ---------------------
2778
2779 1. If a comment in an extended regex that started immediately after a meta-item
2780 extended to the end of string, PCRE compiled incorrect data. This could lead to
2781 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2782
2783 2. Moved to autoconf 2.53 and libtool 1.4.2.
2784
2785 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2786 the special perltest8 script is no longer needed - all the tests can be run
2787 from a single perltest script.
2788
2789 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2790 by \s. It has now been removed in PCRE. This means it isn't recognized as
2791 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2792 class [:space:] *does* include VT, thereby creating a mess.
2793
2794 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2795 space and tab.
2796
2797 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2798 its new features into the main test script, reducing the number of scripts.
2799
2800 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2801 were backward compatible, and made the (?i) apply to the whole pattern, as if
2802 /i were given. Now it behaves more logically, and applies the option setting
2803 only to what follows. PCRE has been changed to follow suit. However, if it
2804 finds options settings right at the start of the pattern, it extracts them into
2805 the global options, as before. Thus, they show up in the info data.
2806
2807 8. Added support for the \Q...\E escape sequence. Characters in between are
2808 treated as literals. This is slightly different from Perl in that $ and @ are
2809 also handled as literals inside the quotes. In Perl, they will cause variable
2810 interpolation. Note the following examples:
2811
2812 Pattern PCRE matches Perl matches
2813
2814 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2815 \Qabc\$xyz\E abc\$xyz abc\$xyz
2816 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2817
2818 For compatibility with Perl, \Q...\E sequences are recognized inside character
2819 classes as well as outside them.
2820
2821 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2822 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2823 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2824 signed/unsigned warnings.
2825
2826 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2827 option for pcretest, so I've replaced it by a simple function that does just
2828 that job.
2829
2830 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2831 "pcregrep -".
2832
2833 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2834 Java package. This provides some syntactic sugar for simple cases of what my
2835 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2836 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2837 item, you can use this simplified notation. Note that only makes sense with
2838 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2839 greediness, whatever the setting of the PCRE_UNGREEDY option.
2840
2841 13. A change of greediness default within a pattern was not taking effect at
2842 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2843 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2844 was abstracted outside.
2845
2846 14. PCRE now supports the \G assertion. It is true when the current matching
2847 position is at the start point of the match. This differs from \A when the
2848 starting offset is non-zero. Used with the /g option of pcretest (or similar
2849 code), it works in the same way as it does for Perl's /g option. If all
2850 alternatives of a regex begin with \G, the expression is anchored to the start
2851 match position, and the "anchored" flag is set in the compiled expression.
2852
2853 15. Some bugs concerning the handling of certain option changes within patterns
2854 have been fixed. These applied to options other than (?ims). For example,
2855 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2856 been the other way round. Some of this was related to change 7 above.
2857
2858 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2859 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2860 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2861 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2862
2863 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2864 mode. Unlike ".", it always matches newline, whatever the setting of
2865 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2866 assertions. Perl allows it, but it doesn't (in general) work because it can't
2867 calculate the length of the lookbehind. At least, that's the case for Perl
2868 5.8.0 - I've been told they are going to document that it doesn't work in
2869 future.
2870
2871 18. Added an error diagnosis for escapes that PCRE does not support: these are
2872 \L, \l, \N, \P, \p, \U, \u, and \X.
2873
2874 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2875 reading past the end of the pattern in cases such as /[abcd/.
2876
2877 20. PCRE was getting more memory than necessary for patterns with classes that
2878 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2879
2880 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2881 compiling PCRE for use with Virtual Pascal.
2882
2883 22. Small fix to the Makefile to make it work properly if the build is done
2884 outside the source tree.
2885
2886 23. Added a new extension: a condition to go with recursion. If a conditional
2887 subpattern starts with (?(R) the "true" branch is used if recursion has
2888 happened, whereas the "false" branch is used only at the top level.
2889
2890 24. When there was a very long string of literal characters (over 255 bytes
2891 without UTF support, over 250 bytes with UTF support), the computation of how
2892 much memory was required could be incorrect, leading to segfaults or other
2893 strange effects.
2894
2895 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2896 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2897 there was a subsequent back reference to those brackets. This meant that, for
2898 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2899 possible to check for precisely this case. All we can do is abandon the
2900 optimization if .* occurs inside capturing brackets when there are any back
2901 references whatsoever. (See below for a better fix that came later.)
2902
2903 26. The handling of the optimization for finding the first character of a
2904 non-anchored pattern, and for finding a character that is required later in the
2905 match were failing in some cases. This didn't break the matching; it just
2906 failed to optimize when it could. The way this is done has been re-implemented.
2907
2908 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2909
2910 28. Added a new feature that provides some of the functionality that Perl
2911 provides with (?{...}). The facility is termed a "callout". The way it is done
2912 in PCRE is for the caller to provide an optional function, by setting
2913 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2914 global variable. By default it is unset, which disables all calling out. To get
2915 the function called, the regex must include (?C) at appropriate points. This
2916 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2917 This provides a means of identifying different callout points. When PCRE
2918 reaches such a point in the regex, if pcre_callout has been set, the external
2919 function is called. It is provided with data in a structure called
2920 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2921 matching continues; if it returns a non-zero value, the match at the current
2922 point fails. However, backtracking will occur if possible. [This was changed
2923 later and other features added - see item 49 below.]
2924
2925 29. pcretest is upgraded to test the callout functionality. It provides a
2926 callout function that displays information. By default, it shows the start of
2927 the match and the current position in the text. There are some new data escapes
2928 to vary what happens:
2929
2930 \C+ in addition, show current contents of captured substrings
2931 \C- do not supply a callout function
2932 \C!n return 1 when callout number n is reached
2933 \C!n!m return 1 when callout number n is reached for the mth time
2934
2935 30. If pcregrep was called with the -l option and just a single file name, it
2936 output "<stdin>" if a match was found, instead of the file name.
2937
2938 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2939 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2940 pcre_exec(). This saves a malloc/free per call. The default value of
2941 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2942 when configuring.
2943
2944 32. The default maximum size of a compiled pattern is 64K. There have been a
2945 few cases of people hitting this limit. The code now uses macros to handle the
2946 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2947 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2948 configuring. Tests 2 and 5 work only with 2-byte links because they output
2949 debugging information about compiled patterns.
2950
2951 33. Internal code re-arrangements:
2952
2953 (a) Moved the debugging function for printing out a compiled regex into
2954 its own source file (printint.c) and used #include to pull it into
2955 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2956 separate copies.
2957
2958 (b) Defined the list of op-code names for debugging as a macro in
2959 internal.h so that it is next to the definition of the opcodes.
2960
2961 (c) Defined a table of op-code lengths for simpler skipping along compiled
2962 code. This is again a macro in internal.h so that it is next to the
2963 definition of the opcodes.
2964
2965 34. Added support for recursive calls to individual subpatterns, along the
2966 lines of Robin Houston's patch (but implemented somewhat differently).
2967
2968 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2969 allow it to read and process whole directories in Win32. This code was
2970 contributed by Lionel Fourquaux; it has not been tested by me.
2971
2972 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2973 used to name a group. Names consist of alphanumerics and underscores, and must
2974 be unique. Back references use the syntax (?P=name) and recursive calls use
2975 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2976 numbers. The function pcre_fullinfo() can be used after compilation to extract
2977 a name/number map. There are three relevant calls:
2978
2979 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2980 PCRE_INFO_NAMECOUNT yields the number of entries
2981 PCRE_INFO_NAMETABLE yields a pointer to the map.
2982
2983 The map is a vector of fixed-size entries. The size of each entry depends on
2984 the length of the longest name used. The first two bytes of each entry are the
2985 group number, most significant byte first. There follows the corresponding
2986 name, zero terminated. The names are in alphabetical order.
2987
2988 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2989 case instead of 255. Making it the same both with and without UTF-8 support
2990 means that the same test output works with both.
2991
2992 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2993 calling malloc() with a zero argument.
2994
2995 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2996 optimization. I've improved things by keeping a bitmap of backreferences with
2997 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2998 fact referenced, the optimization can be applied. It is unlikely that a
2999 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3000 the match to follow \n) will appear inside brackets with a number greater than
3001 31, but if it does, any back reference > 31 suppresses the optimization.
3002
3003 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3004 of disabling numbered capturing parentheses. Any opening parenthesis that is
3005 not followed by ? behaves as if it were followed by ?: but named parentheses
3006 can still be used for capturing (and they will acquire numbers in the usual
3007 way).
3008
3009 41. Redesigned the return codes from the match() function into yes/no/error so
3010 that errors can be passed back from deep inside the nested calls. A malloc
3011 failure while inside a recursive subpattern call now causes the
3012 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3013
3014 42. It is now possible to set a limit on the number of times the match()
3015 function is called in a call to pcre_exec(). This facility makes it possible to
3016 limit the amount of recursion and backtracking, though not in a directly
3017 obvious way, because the match() function is used in a number of different
3018 circumstances. The count starts from zero for each position in the subject
3019 string (for non-anchored patterns). The default limit is, for compatibility, a
3020 large number, namely 10 000 000. You can change this in two ways:
3021
3022 (a) When configuring PCRE before making, you can use --with-match-limit=n
3023 to set a default value for the compiled library.
3024
3025 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3026 a different value is set. See 45 below.
3027
3028 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3029
3030 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3031 of things that can be changed at compile time. The first argument specifies
3032 what is wanted and the second points to where the information is to be placed.
3033 The current list of available information is:
3034
3035 PCRE_CONFIG_UTF8
3036
3037 The output is an integer that is set to one if UTF-8 support is available;
3038 otherwise it is set to zero.
3039
3040 PCRE_CONFIG_NEWLINE
3041
3042 The output is an integer that it set to the value of the code that is used for
3043 newline. It is either LF (10) or CR (13).
3044
3045 PCRE_CONFIG_LINK_SIZE
3046
3047 The output is an integer that contains the number of bytes used for internal
3048 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3049
3050 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3051
3052 The output is an integer that contains the threshold above which the POSIX
3053 interface uses malloc() for output vectors. See item 31 above.
3054
3055 PCRE_CONFIG_MATCH_LIMIT
3056
3057 The output is an unsigned integer that contains the default limit of the number
3058 of match() calls in a pcre_exec() execution. See 42 above.
3059
3060 44. pcretest has been upgraded by the addition of the -C option. This causes it
3061 to extract all the available output from the new pcre_config() function, and to
3062 output it. The program then exits immediately.
3063
3064 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3065 order to support additional features. One way would have been to define
3066 pcre_exec2() (for example) with extra arguments, but this would not have been
3067 extensible, and would also have required all calls to the original function to
3068 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3069 is used for passing in "extra" data from pcre_study().
3070
3071 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3072 contains the following fields:
3073
3074 flags a bitmap indicating which of the following fields are set
3075 study_data opaque data from pcre_study()
3076 match_limit a way of specifying a limit on match() calls for a specific
3077 call to pcre_exec()
3078 callout_data data for callouts (see 49 below)
3079
3080 The flag bits are also defined in pcre.h, and are
3081
3082 PCRE_EXTRA_STUDY_DATA
3083 PCRE_EXTRA_MATCH_LIMIT
3084 PCRE_EXTRA_CALLOUT_DATA
3085
3086 The pcre_study() function now returns one of these new pcre_extra blocks, with
3087 the actual study data pointed to by the study_data field, and the
3088 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3089 before. That is, this change is entirely upwards-compatible and requires no
3090 change to existing code.
3091
3092 If you want to pass in additional data to pcre_exec(), you can either place it
3093 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3094 block.
3095
3096 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3097 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3098 times with different match limits, until it finds the minimum value needed for
3099 pcre_exec() to complete. The value is then output. This can be instructive; for
3100 most simple matches the number is quite small, but for pathological cases it
3101 gets very large very quickly.
3102
3103 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3104 returns the size of the data block pointed to by the study_data field in a
3105 pcre_extra block, that is, the value that was passed as the argument to
3106 pcre_malloc() when PCRE was getting memory in which to place the information
3107 created by pcre_study(). The fourth argument should point to a size_t variable.
3108 pcretest has been extended so that this information is shown after a successful
3109 pcre_study() call when information about the compiled regex is being displayed.
3110
3111 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3112 because what follows is always an absolute path. (Later: it turns out that this
3113 is more than cosmetic for MinGW, because it doesn't like empty path
3114 components.)
3115
3116 49. Some changes have been made to the callout feature (see 28 above):
3117
3118 (i) A callout function now has three choices for what it returns:
3119
3120 0 => success, carry on matching
3121 > 0 => failure at this point, but backtrack if possible
3122 < 0 => serious error, return this value from pcre_exec()
3123
3124 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3125 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3126 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3127 use by callout functions. It will never be used by PCRE itself.
3128
3129 (ii) The pcre_extra structure (see 45 above) has a void * field called
3130 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3131 pcre_callout_block structure has a field of the same name. The contents of
3132 the field passed in the pcre_extra structure are passed to the callout
3133 function in the corresponding field in the callout block. This makes it
3134 easier to use the same callout-containing regex from multiple threads. For
3135 testing, the pcretest program has a new data escape
3136
3137 \C*n pass the number n (may be negative) as callout_data
3138
3139 If the callout function in pcretest receives a non-zero value as
3140 callout_data, it returns that value.
3141
3142 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3143 there were some redundant $(CFLAGS) in commands that are now specified as
3144 $(LINK), which already includes $(CFLAGS).
3145
3146 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3147 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3148 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3149 one-byte characters throughout. Note that case-insensitive matching applies
3150 only to characters whose values are less than 256. PCRE doesn't support the
3151 notion of cases for higher-valued characters.
3152
3153 (i) A character class whose characters are all within 0-255 is handled as
3154 a bit map, and the map is inverted for negative classes. Previously, a
3155 character > 255 always failed to match such a class; however it should
3156 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3157
3158 (ii) A negated character class with a single character < 255 is coded as
3159 "not this character" (OP_NOT). This wasn't working properly when the test
3160 character was multibyte, either singly or repeated.
3161
3162 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3163 mode, for example: \x{100}{2,3}.
3164
3165 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3166 singly or repeated) now correctly test multibyte characters. However,
3167 PCRE doesn't recognize any characters with values greater than 255 as
3168 digits, spaces, or word characters. Such characters always match \D, \S,
3169 and \W, and never match \d, \s, or \w.
3170
3171 (v) Classes may now contain characters and character ranges with values
3172 greater than 255. For example: [ab\x{100}-\x{400}].
3173
3174 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3175 PCRE in UTF-8 mode.
3176
3177 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3178 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3179 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3180 value.)
3181
3182 53. The single man page has become too large. I have therefore split it up into
3183 a number of separate man pages. These also give rise to individual HTML pages;
3184 these are now put in a separate directory, and there is an index.html page that
3185 lists them all. Some hyperlinking between the pages has been installed.
3186
3187 54. Added convenience functions for handling named capturing parentheses.
3188
3189 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3190 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3191 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3192 are faulted.
3193
3194 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3195 calling configure. These values are used when compiling the dftables.c program
3196 which is run to generate the source of the default character tables. They
3197 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3198 you will need to set these values.
3199
3200 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3201
3202
3203 Version 3.9 02-Jan-02
3204 ---------------------
3205
3206 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3207
3208 2. If --disable-static was given, the building process failed when trying to
3209 build pcretest and pcregrep. (For some reason it was using libtool to compile
3210 them, which is not right, as they aren't part of the library.)
3211
3212
3213 Version 3.8 18-Dec-01
3214 ---------------------
3215
3216 1. The experimental UTF-8 code was completely screwed up. It was packing the
3217 bytes in the wrong order. How dumb can you get?
3218
3219
3220 Version 3.7 29-Oct-01
3221 ---------------------
3222
3223 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3224 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3225 this didn't happen under Solaris 8, where I normally test things.
3226
3227 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3228 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3229 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3230 this fix an hour or so after the initial 3.7 release.)
3231
3232
3233 Version 3.6 23-Oct-01
3234 ---------------------
3235
3236 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3237 offsets passed as NULL with zero offset count.
3238
3239 2. The config.guess and config.sub files had not been updated when I moved to
3240 the latest autoconf.
3241
3242
3243 Version 3.5 15-Aug-01
3244 ---------------------
3245
3246 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3247 had been forgotten.
3248
3249 2. By using declared but undefined structures, we can avoid using "void"
3250 definitions in pcre.h while keeping the internal definitions of the structures
3251 private.
3252
3253 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3254 user point of view, this means that both static and shared libraries are built
3255 by default, but this can be individually controlled. More of the work of
3256 handling this static/shared cases is now inside libtool instead of PCRE's make
3257 file.
3258
3259 4. The pcretest utility is now installed along with pcregrep because it is
3260 useful for users (to test regexs) and by doing this, it automatically gets
3261 relinked by libtool. The documentation has been turned into a man page, so
3262 there are now .1, .txt, and .html versions in /doc.
3263
3264 5. Upgrades to pcregrep:
3265 (i) Added long-form option names like gnu grep.
3266 (ii) Added --help to list all options with an explanatory phrase.
3267 (iii) Added -r, --recursive to recurse into sub-directories.
3268 (iv) Added -f, --file to read patterns from a file.
3269
3270 6. pcre_exec() was referring to its "code" argument before testing that
3271 argument for NULL (and giving an error if it was NULL).
3272
3273 7. Upgraded Makefile.in to allow for compiling in a different directory from
3274 the source directory.
3275
3276 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3277 options bits, the pointer it was passed was to an int instead of to an unsigned
3278 long int. This mattered only on 64-bit systems.
3279
3280 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3281 generated) instead of pcre.in, which it its source. Also made the same change
3282 in several of the .c files.
3283
3284 10. A new release of gcc defines printf() as a macro, which broke pcretest
3285 because it had an ifdef in the middle of a string argument for printf(). Fixed
3286 by using separate calls to printf().
3287
3288 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3289 script, to force use of CR or LF instead of \n in the source. On non-Unix
3290 systems, the value can be set in config.h.
3291
3292 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3293 absolute limit. Changed the text of the error message to make this clear, and
3294 likewise updated the man page.
3295
3296 13. The limit of 99 on the number of capturing subpatterns has been removed.
3297 The new limit is 65535, which I hope will not be a "real" limit.
3298
3299
3300 Version 3.4 22-Aug-00
3301 ---------------------
3302
3303 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3304
3305 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3306
3307
3308 Version 3.3 01-Aug-00
3309 ---------------------
3310
3311 1. If an octal character was given, but the value was greater than \377, it
3312 was not getting masked to the least significant bits, as documented. This could
3313 lead to crashes in some systems.
3314
3315 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3316 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3317
3318 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3319 These just pass their arguments on to (pcre_free)(), but they are provided
3320 because some uses of PCRE bind it to non-C systems that can call its functions,
3321 but cannot call free() or pcre_free() directly.
3322
3323 4. Add "make test" as a synonym for "make check". Corrected some comments in
3324 the Makefile.
3325
3326 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3327 Makefile.
3328
3329 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3330 command called pgrep for grepping around the active processes.
3331
3332 7. Added the beginnings of support for UTF-8 character strings.
3333
3334 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3335 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3336 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3337 out for the ar command.)
3338
3339
3340 Version 3.2 12-May-00
3341 ---------------------
3342
3343 This is purely a bug fixing release.
3344
3345 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3346 of ZA. This was just one example of several cases that could provoke this bug,
3347 which was introduced by change 9 of version 2.00. The code for breaking
3348 infinite loops after an iteration that matches an empty string was't working
3349 correctly.
3350
3351 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3352 when matched against abbab (for example). After matching an empty string, it
3353 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3354 caused it to match further down the string than it should.
3355
3356 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3357 was there because it doesn't seem to be needed, and it causes trouble on some
3358 systems, as it is not a Standard C header. It has been removed.
3359
3360 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3361 were reported on the Macintosh. The changes were from
3362
3363 while ((c = *(++ptr)) != 0 && c != '\n');
3364 to
3365 while ((c = *(++ptr)) != 0 && c != '\n') ;
3366
3367 Totally extraordinary, but if that's what it takes...
3368
3369 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3370 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3371 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3372 assumes the way PCRE uses memmove() (always moving upwards).
3373
3374 6. PCRE is being used in one environment where strchr() is not available. There
3375 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3376 faster code anyway.
3377
3378
3379 Version 3.1 09-Feb-00
3380 ---------------------
3381
3382 The only change in this release is the fixing of some bugs in Makefile.in for
3383 the "install" target:
3384
3385 (1) It was failing to install pcreposix.h.
3386
3387 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3388
3389
3390 Version 3.0 01-Feb-00
3391 ---------------------
3392
3393 1. Add support for the /+ modifier to perltest (to output $` like it does in
3394 pcretest).
3395
3396 2. Add support for the /g modifier to perltest.
3397
3398 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3399 matches null strings.
3400
3401 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3402 pattern. Perl treats empty patterns specially - it reuses the most recent
3403 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3404 effect.
3405
3406 5. The POSIX interface was broken in that it was just handing over the POSIX
3407 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3408 required a bigger vector, with some working space on the end. This means that
3409 the POSIX wrapper now has to get and free some memory, and copy the results.
3410
3411 6. Added some simple autoconf support, placing the test data and the
3412 documentation in separate directories, re-organizing some of the
3413 information files, and making it build pcre-config (a GNU standard). Also added
3414 libtool support for building PCRE as a shared library, which is now the
3415 default.
3416
3417 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3418 09 are not valid octal constants. Single digits will be used for minor values
3419 less than 10.
3420
3421 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3422 existing programs that set these in the POSIX interface can use PCRE without
3423 modification.
3424
3425 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3426 return all that pcre_info() returns, plus additional data. The pcre_info()
3427 function is retained for compatibility, but is considered to be obsolete.
3428
3429 10. Added experimental recursion feature (?R) to handle one common case that
3430 Perl 5.6 will be able to do with (?p{...}).
3431
3432 11. Added support for POSIX character classes like [:alpha:], which Perl is
3433 adopting.
3434
3435
3436 Version 2.08 31-Aug-99
3437 ----------------------
3438
3439 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3440 trying to match at the startoffset position, but instead was moving forward to
3441 the next newline as if a previous match had failed.
3442
3443 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3444 and could get into a loop if a null string was matched other than at the start
3445 of the subject.
3446
3447 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3448 be distinguished at compile time, and for completeness also added PCRE_DATE.
3449
3450 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3451 in GnuWin32 environments.
3452
3453
3454 Version 2.07 29-Jul-99
3455 ----------------------
3456
3457 1. The documentation is now supplied in plain text form and HTML as well as in
3458 the form of man page sources.
3459
3460 2. C++ compilers don't like assigning (void *) values to other pointer types.
3461 In particular this affects malloc(). Although there is no problem in Standard
3462 C, I've put in casts to keep C++ compilers happy.
3463
3464 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3465 should be (const char *).
3466
3467 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3468 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3469 However, I haven't made this a standard facility. The documentation doesn't
3470 mention it, and the Makefile doesn't support it.
3471
3472 5. The Makefile now contains an "install" target, with editable destinations at
3473 the top of the file. The pcretest program is not installed.
3474
3475 6. pgrep -V now gives the PCRE version number and date.
3476
3477 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3478 causing the entire string to be ignored, instead of just the last character.
3479
3480 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3481 non-matching string, it can take a very, very long time, even for strings of
3482 quite modest length, because of the nested recursion. PCRE now does better in
3483 some of these cases. It does this by remembering the last required literal
3484 character in the pattern, and pre-searching the subject to ensure it is present
3485 before running the real match. In other words, it applies a heuristic to detect
3486 some types of certain failure quickly, and in the above example, if presented
3487 with a string that has no trailing " it gives "no match" very quickly.
3488
3489 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3490 other alternatives are tried instead.
3491
3492
3493 Version 2.06 09-Jun-99
3494 ----------------------
3495
3496 1. Change pcretest's output for amount of store used to show just the code
3497 space, because the remainder (the data block) varies in size between 32-bit and
3498 64-bit systems.
3499
3500 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3501 start matching at. This allows lookbehinds to work when searching for multiple
3502 occurrences in a string.
3503
3504 3. Added additional options to pcretest for testing multiple occurrences:
3505
3506 /+ outputs the rest of the string that follows a match
3507 /g loops for multiple occurrences, using the new startoffset argument
3508 /G loops for multiple occurrences by passing an incremented pointer
3509
3510 4. PCRE wasn't doing the "first character" optimization for patterns starting
3511 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3512 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3513 the letter 'x'. On long subject strings, this gives a significant speed-up.
3514
3515
3516 Version 2.05 21-Apr-99
3517 ----------------------
3518
3519 1. Changed the type of magic_number from int to long int so that it works
3520 properly on 16-bit systems.
3521
3522 2. Fixed a bug which caused patterns starting with .* not to work correctly
3523 when the subject string contained newline characters. PCRE was assuming
3524 anchoring for such patterns in all cases, which is not correct because .* will
3525 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3526 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3527 must be retried after every newline in the subject.
3528
3529
3530 Version 2.04 18-Feb-99
3531 ----------------------
3532
3533 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3534 computation of the store needed to hold the pattern was incorrect (too large).
3535 If such patterns were nested a few deep, this could multiply and become a real
3536 problem.
3537
3538 2. Added /M option to pcretest to show the memory requirement of a specific
3539 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3540
3541 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3542 compiled in such a way that the backtracking after subsequent failure was
3543 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3544 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3545
3546
3547 Version 2.03 02-Feb-99
3548 ----------------------
3549
3550 1. Fixed typo and small mistake in man page.
3551
3552 2. Added 4th condition (GPL supersedes if conflict) and created separate
3553 LICENCE file containing the conditions.
3554
3555 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3556 Perl, that is the internal \ allows the delimiter to be included in the
3557 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3558 the final delimiter, add \ to the end of the pattern (to test the error).
3559
3560 4. Added the convenience functions for extracting substrings after a successful
3561 match. Updated pcretest to make it able to test these functions.
3562
3563
3564 Version 2.02 14-Jan-99
3565 ----------------------
3566
3567 1. Initialized the working variables associated with each extraction so that
3568 their saving and restoring doesn't refer to uninitialized store.
3569
3570 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3571 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3572 fix the problem.
3573
3574 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3575 calls, and wasn't printing the correct value for compiling calls. Increased the
3576 default value of LOOPREPEAT, and the number of significant figures in the
3577 times.
3578
3579 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3580
3581 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3582 a building problem on Windows NT with a FAT file system.
3583
3584
3585 Version 2.01 21-Oct-98
3586 ----------------------
3587
3588 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3589 to character tables built by pcre_maketables() in the current locale. If NULL
3590 is passed, the default tables are used.
3591
3592
3593 Version 2.00 24-Sep-98
3594 ----------------------
3595
3596 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3597 it any more.
3598
3599 2. Allow quantification of (?>) groups, and make it work correctly.
3600
3601 3. The first character computation wasn't working for (?>) groups.
3602
3603 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3604 end of the subject) and add 5.005's \z, which really does match only at the
3605 very end of the subject.
3606
3607 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3608
3609 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3610 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3611 localized options. All options to pcre_study() were also removed.
3612
3613 7. Add other new features from 5.005:
3614
3615 $(?<= positive lookbehind
3616 $(?<! negative lookbehind
3617 (?imsx-imsx) added the unsetting capability
3618 such a setting is global if at outer level; local otherwise
3619 (?imsx-imsx:) non-capturing groups with option setting
3620 (?(cond)re|re) conditional pattern matching
3621
3622 A backreference to itself in a repeated group matches the previous
3623 captured string.
3624
3625 8. General tidying up of studying (both automatic and via "study")
3626 consequential on the addition of new assertions.
3627
3628 9. As in 5.005, unlimited repeated groups that could match an empty substring
3629 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3630 runtime if any iteration does actually match an empty substring.
3631
3632 10. Include the RunTest script in the distribution.
3633
3634 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3635 discrepancies, some of which were old and were also with respect to 5.004. They
3636 have now been fixed.
3637
3638
3639 Version 1.09 28-Apr-98
3640 ----------------------
3641
3642 1. A negated single character class followed by a quantifier with a minimum
3643 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3644 program crashes, or just wrong answers. This did not apply to negated classes
3645 containing more than one character, or to minima other than one.
3646
3647
3648 Version 1.08 27-Mar-98
3649 ----------------------
3650
3651 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3652
3653 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3654 latter must appear before anything that relies on it in the pattern.
3655
3656
3657 Version 1.07 16-Feb-98
3658 ----------------------
3659
3660 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3661 repeat of a potentially empty string).
3662
3663
3664 Version 1.06 23-Jan-98
3665 ----------------------
3666
3667 1. Added Markus Oberhumer's little patches for C++.
3668
3669 2. Literal strings longer than 255 characters were broken.
3670
3671
3672 Version 1.05 23-Dec-97
3673 ----------------------
3674
3675 1. Negated character classes containing more than one character were failing if
3676 PCRE_CASELESS was set at run time.
3677
3678
3679 Version 1.04 19-Dec-97
3680 ----------------------
3681
3682 1. Corrected the man page, where some "const" qualifiers had been omitted.
3683
3684 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3685 input syntax.
3686
3687 3. Fixed memory leak which occurred when a regex with back references was
3688 matched with an offsets vector that wasn't big enough. The temporary memory
3689 that is used in this case wasn't being freed if the match failed.
3690
3691 4. Tidied pcretest to ensure it frees memory that it gets.
3692
3693 5. Temporary memory was being obtained in the case where the passed offsets
3694 vector was exactly big enough.
3695
3696 6. Corrected definition of offsetof() from change 5 below.
3697
3698 7. I had screwed up change 6 below and broken the rules for the use of
3699 setjmp(). Now fixed.
3700
3701
3702 Version 1.03 18-Dec-97
3703 ----------------------
3704
3705 1. A erroneous regex with a missing opening parenthesis was correctly
3706 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3707 on some systems.
3708
3709 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3710 it was reported that one broken compiler failed on the former because "code" is
3711 also an independent variable.
3712
3713 3. The erroneous regex a[]b caused an array overrun reference.
3714
3715 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3716 fail on data ending with that character. (It was going on too far, and checking
3717 the next character, typically a binary zero.) This was specific to the
3718 optimized code for single-character negative classes.
3719
3720 5. Added a contributed patch from the TIN world which does the following:
3721
3722 + Add an undef for memmove, in case the the system defines a macro for it.
3723
3724 + Add a definition of offsetof(), in case there isn't one. (I don't know
3725 the reason behind this - offsetof() is part of the ANSI standard - but
3726 it does no harm).
3727
3728 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3729 most of the places where whitespace preceded '#'. I have given up and
3730 allowed the remaining 2 cases to be at the margin.
3731
3732 + Rename some variables in pcre to eliminate shadowing. This seems very
3733 pedantic, but does no harm, of course.
3734
3735 6. Moved the call to setjmp() into its own function, to get rid of warnings
3736 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3737
3738 7. Constructs such as \d{8,} were compiling into the equivalent of
3739 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3740 outcome, but in this particular case used more store than had been allocated,
3741 which caused the bug to be discovered because it threw up an internal error.
3742
3743 8. The debugging code in both pcre and pcretest for outputting the compiled
3744 form of a regex was going wrong in the case of back references followed by
3745 curly-bracketed repeats.
3746
3747
3748 Version 1.02 12-Dec-97
3749 ----------------------
3750
3751 1. Typos in pcre.3 and comments in the source fixed.
3752
3753 2. Applied a contributed patch to get rid of places where it used to remove
3754 'const' from variables, and fixed some signed/unsigned and uninitialized
3755 variable warnings.
3756
3757 3. Added the "runtest" target to Makefile.
3758
3759 4. Set default compiler flag to -O2 rather than just -O.
3760
3761
3762 Version 1.01 19-Nov-97
3763 ----------------------
3764
3765 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3766 like /([ab]*)*/, that is, for classes with more than one character in them.
3767
3768 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3769 as /((?>a*))*/ (a PCRE_EXTRA facility).
3770
3771
3772 Version 1.00 18-Nov-97
3773 ----------------------
3774
3775 1. Added compile-time macros to support systems such as SunOS4 which don't have
3776 memmove() or strerror() but have other things that can be used instead.
3777
3778 2. Arranged that "make clean" removes the executables.
3779
3780
3781 Version 0.99 27-Oct-97
3782 ----------------------
3783
3784 1. Fixed bug in code for optimizing classes with only one character. It was
3785 initializing a 32-byte map regardless, which could cause it to run off the end
3786 of the memory it had got.
3787
3788 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3789
3790
3791 Version 0.98 22-Oct-97
3792 ----------------------
3793
3794 1. Fixed bug in code for handling temporary memory usage when there are more
3795 back references than supplied space in the ovector. This could cause segfaults.
3796
3797
3798 Version 0.97 21-Oct-97
3799 ----------------------
3800
3801 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3802
3803 2. Optimized negated single characters not to use a bit map.
3804
3805 3. Brought error texts together as macro definitions; clarified some of them;
3806 fixed one that was wrong - it said "range out of order" when it meant "invalid
3807 escape sequence".
3808
3809 4. Changed some char * arguments to const char *.
3810
3811 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3812
3813 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3814 pcretest.
3815
3816
3817 Version 0.96 16-Oct-97
3818 ----------------------
3819
3820 1. Added a simple "pgrep" utility to the distribution.
3821
3822 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3823 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3824 where "ddd" means "one or more decimal digits".
3825
3826 3. Fixed serious bug. If a pattern had a back reference, but the call to
3827 pcre_exec() didn't supply a large enough ovector to record the related
3828 identifying subpattern, the match always failed. PCRE now remembers the number
3829 of the largest back reference, and gets some temporary memory in which to save
3830 the offsets during matching if necessary, in order to ensure that
3831 backreferences always work.
3832
3833 4. Increased the compatibility with Perl in a number of ways:
3834
3835 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3836 to request this handling. The option can be set at compile or exec time.
3837
3838 (b) $ matches before a terminating newline by default; an option
3839 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3840 mode). The option can be set at compile or exec time.
3841
3842 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3843 the same as Perl's. If the decimal number it represents is less than 10
3844 or there aren't that many previous left capturing parentheses, an octal
3845 escape is read. Inside a character class, it's always an octal escape,
3846 even if it is a single digit.
3847
3848 (d) An escaped but undefined alphabetic character is taken as a literal,
3849 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3850 escapes.
3851
3852 (e) {0} is now permitted. (The previous item is removed from the compiled
3853 pattern).
3854
3855 5. Changed all the names of code files so that the basic parts are no longer
3856 than 10 characters, and abolished the teeny "globals.c" file.
3857
3858 6. Changed the handling of character classes; they are now done with a 32-byte
3859 bit map always.
3860
3861 7. Added the -d and /D options to pcretest to make it possible to look at the
3862 internals of compilation without having to recompile pcre.
3863
3864
3865 Version 0.95 23-Sep-97
3866 ----------------------
3867
3868 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3869 \x20 at the start of a run of normal characters. These were being treated as
3870 real characters, instead of the source characters being re-checked.
3871
3872
3873 Version 0.94 18-Sep-97
3874 ----------------------
3875
3876 1. The functions are now thread-safe, with the caveat that the global variables
3877 containing pointers to malloc() and free() or alternative functions are the
3878 same for all threads.
3879
3880 2. Get pcre_study() to generate a bitmap of initial characters for non-
3881 anchored patterns when this is possible, and use it if passed to pcre_exec().
3882
3883
3884 Version 0.93 15-Sep-97
3885 ----------------------
3886
3887 1. /(b)|(:+)/ was computing an incorrect first character.
3888
3889 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3890 but not actually doing anything yet.
3891
3892 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3893 as Perl does (e.g. [-az] or [az-]).
3894
3895 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3896 all possible positions.
3897
3898 5. Split up into different modules to avoid including unneeded functions in a
3899 compiled binary. However, compile and exec are still in one module. The "study"
3900 function is split off.
3901
3902 6. The character tables are now in a separate module whose source is generated
3903 by an auxiliary program - but can then be edited by hand if required. There are
3904 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3905 toupper() in the code.
3906
3907 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3908 make them global. Abolish the function for setting them, as the caller can now
3909 set them directly.
3910
3911
3912 Version 0.92 11-Sep-97
3913 ----------------------
3914
3915 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3916 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3917
3918 2. Caseless matching was not working in character classes if the characters in
3919 the pattern were in upper case.
3920
3921 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3922
3923 4. Make PCRE_ANCHORED public and accept as a compile option.
3924
3925 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3926 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3927 pass them.
3928
3929 6. Give an error if bad option bits passed at compile or run time.
3930
3931 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3932 pcretest to cause it to pass that flag.
3933
3934 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3935 options, and the first character, if set.
3936
3937 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3938
3939
3940 Version 0.91 10-Sep-97
3941 ----------------------
3942
3943 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3944 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3945
3946 2. PCRE was looping on encountering an indefinitely repeated back reference to
3947 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3948 Perl does - treats the match as successful.
3949
3950 ****

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