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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 550 - (show annotations) (download)
Sun Oct 10 16:24:11 2010 UTC (4 years, 1 month ago) by ph10
File size: 190018 byte(s)
Fix problem with (*THEN) not backing up far enough.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.11 10-Oct-2010
5 ------------------------
6
7 1. (*THEN) was not working properly if there were untried alternatives prior
8 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
9 backtracked to try for "b" instead of moving to the next alternative branch
10 at the same level (in this case, to look for "c"). The Perl documentation
11 is clear that when (*THEN) is backtracked onto, it goes to the "next
12 alternative in the innermost enclosing group".
13
14
15 Version 8.10 25-Jun-2010
16 ------------------------
17
18 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
19 THEN.
20
21 2. (*ACCEPT) was not working when inside an atomic group.
22
23 3. Inside a character class, \B is treated as a literal by default, but
24 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
25 causes the error). The code is unchanged, but I tidied the documentation.
26
27 4. Inside a character class, PCRE always treated \R and \X as literals,
28 whereas Perl faults them if its -w option is set. I have changed PCRE so
29 that it faults them when PCRE_EXTRA is set.
30
31 5. Added support for \N, which always matches any character other than
32 newline. (It is the same as "." when PCRE_DOTALL is not set.)
33
34 6. When compiling pcregrep with newer versions of gcc which may have
35 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
36 declared with attribute warn_unused_result" were given. Just casting the
37 result to (void) does not stop the warnings; a more elaborate fudge is
38 needed. I've used a macro to implement this.
39
40 7. Minor change to pcretest.c to avoid a compiler warning.
41
42 8. Added four artifical Unicode properties to help with an option to make
43 \s etc use properties (see next item). The new properties are: Xan
44 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
45
46 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
47 use Unicode properties. (*UCP) at the start of a pattern can be used to set
48 this option. Modified pcretest to add /W to test this facility. Added
49 REG_UCP to make it available via the POSIX interface.
50
51 10. Added --line-buffered to pcregrep.
52
53 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
54 studied, and the match started with a letter with a code point greater than
55 127 whose first byte was different to the first byte of the other case of
56 the letter, the other case of this starting letter was not recognized
57 (#976).
58
59 12. If a pattern that was studied started with a repeated Unicode property
60 test, for example, \p{Nd}+, there was the theoretical possibility of
61 setting up an incorrect bitmap of starting bytes, but fortunately it could
62 not have actually happened in practice until change 8 above was made (it
63 added property types that matched character-matching opcodes).
64
65 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
66 possible starting bytes for non-anchored patterns.
67
68 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
69 \R, and also a number of cases that involve Unicode properties, both
70 explicit and implicit when PCRE_UCP is set.
71
72 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
73 input, it could crash or give wrong results if characters with values
74 greater than 0xc0 were present in the subject string. (Detail: it assumed
75 UTF-8 input when processing these items.)
76
77 16. Added a lot of (int) casts to avoid compiler warnings in systems where
78 size_t is 64-bit (#991).
79
80 17. Added a check for running out of memory when PCRE is compiled with
81 --disable-stack-for-recursion (#990).
82
83 18. If the last data line in a file for pcretest does not have a newline on
84 the end, a newline was missing in the output.
85
86 19. The default pcre_chartables.c file recognizes only ASCII characters (values
87 less than 128) in its various bitmaps. However, there is a facility for
88 generating tables according to the current locale when PCRE is compiled. It
89 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
90 characters, are recognized by isspace() and therefore were getting set in
91 these tables, and indeed these tables seem to approximate to ISO 8859. This
92 caused a problem in UTF-8 mode when pcre_study() was used to create a list
93 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
94 which of course cannot start UTF-8 characters. I have changed the code so
95 that only real ASCII characters (less than 128) and the correct starting
96 bytes for UTF-8 encodings are set for characters greater than 127 when in
97 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
98 altogether.)
99
100 20. Added the /T option to pcretest so as to be able to run tests with non-
101 standard character tables, thus making it possible to include the tests
102 used for 19 above in the standard set of tests.
103
104 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
105 reference to a subpattern the other side of a comment that contains an
106 opening parenthesis caused either an internal compiling error, or a
107 reference to the wrong subpattern.
108
109
110 Version 8.02 19-Mar-2010
111 ------------------------
112
113 1. The Unicode data tables have been updated to Unicode 5.2.0.
114
115 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
116 configured.
117
118 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
119 original author of that file, following a query about its status.
120
121 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
122 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
123
124 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
125 quantifier applied to a forward-referencing subroutine call, could compile
126 incorrect code or give the error "internal error: previously-checked
127 referenced subpattern not found".
128
129 6. Both MS Visual Studio and Symbian OS have problems with initializing
130 variables to point to external functions. For these systems, therefore,
131 pcre_malloc etc. are now initialized to local functions that call the
132 relevant global functions.
133
134 7. There were two entries missing in the vectors called coptable and poptable
135 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
136 I've fixed the data, and added a kludgy way of testing at compile time that
137 the lengths are correct (equal to the number of opcodes).
138
139 8. Following on from 7, I added a similar kludge to check the length of the
140 eint vector in pcreposix.c.
141
142 9. Error texts for pcre_compile() are held as one long string to avoid too
143 much relocation at load time. To find a text, the string is searched,
144 counting zeros. There was no check for running off the end of the string,
145 which could happen if a new error number was added without updating the
146 string.
147
148 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
149
150 11. \K was not working if it appeared in an atomic group or in a group that
151 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
152 \K is "not well defined" if used in an assertion. PCRE now accepts it if
153 the assertion is positive, but not if it is negative.
154
155 12. Change 11 fortuitously reduced the size of the stack frame used in the
156 "match()" function of pcre_exec.c by one pointer. Forthcoming
157 implementation of support for (*MARK) will need an extra pointer on the
158 stack; I have reserved it now, so that the stack frame size does not
159 decrease.
160
161 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
162 item in branch that calls a recursion is a subroutine call - as in the
163 second branch in the above example - was incorrectly given the compile-
164 time error "recursive call could loop indefinitely" because pcre_compile()
165 was not correctly checking the subroutine for matching a non-empty string.
166
167 14. The checks for overrunning compiling workspace could trigger after an
168 overrun had occurred. This is a "should never occur" error, but it can be
169 triggered by pathological patterns such as hundreds of nested parentheses.
170 The checks now trigger 100 bytes before the end of the workspace.
171
172 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
173
174
175 Version 8.01 19-Jan-2010
176 ------------------------
177
178 1. If a pattern contained a conditional subpattern with only one branch (in
179 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
180 computed the wrong minimum data length (which is of course zero for such
181 subpatterns). This could cause incorrect "no match" results.
182
183 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
184 the pattern is reset in the first branch, pcre_compile() failed with
185 "internal error: code overflow at offset...". This happened only when
186 the reset was to the original external option setting. (An optimization
187 abstracts leading options settings into an external setting, which was the
188 cause of this.)
189
190 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
191 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
192 assertion pattern did not match (meaning that the assertion was true), it
193 was incorrectly treated as false if the SKIP had been reached during the
194 matching. This also applied to assertions used as conditions.
195
196 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
197 assertion subpattern, including such a pattern used as a condition,
198 unpredictable results occurred, instead of the error return
199 PCRE_ERROR_DFA_UITEM.
200
201 5. The C++ GlobalReplace function was not working like Perl for the special
202 situation when an empty string is matched. It now does the fancy magic
203 stuff that is necessary.
204
205 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
206 removed. (These were left over from very, very early versions of PCRE.)
207
208 7. Some cosmetic changes to the code to make life easier when compiling it
209 as part of something else:
210
211 (a) Change DEBUG to PCRE_DEBUG.
212
213 (b) In pcre_compile(), rename the member of the "branch_chain" structure
214 called "current" as "current_branch", to prevent a collision with the
215 Linux macro when compiled as a kernel module.
216
217 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
218 prevent a collision with the Linux macro when compiled as a kernel
219 module.
220
221 8. In pcre_compile() there are some checks for integer overflows that used to
222 cast potentially large values to (double). This has been changed to that
223 when building, a check for int64_t is made, and if it is found, it is used
224 instead, thus avoiding the use of floating point arithmetic. (There is no
225 other use of FP in PCRE.) If int64_t is not found, the fallback is to
226 double.
227
228 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
229 2005 (difference between two addresses compared to an unsigned value).
230
231 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
232 custom one, because of the following reported problem in Windows:
233
234 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
235 under Win32.
236 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
237 therefore missing the function definition.
238 - The compiler thus generates a "C" signature for the test function.
239 - The linker fails to find the "C" function.
240 - PCRE fails to configure if asked to do so against libbz2.
241
242 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
243 messages were output:
244
245 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
246 rerunning libtoolize, to keep the correct libtool macros in-tree.
247 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
248
249 I have done both of these things.
250
251 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
252 most of the time, it *can* run out if it is given a pattern that contains a
253 runaway infinite recursion. I updated the discussion in the pcrestack man
254 page.
255
256 13. Now that we have gone to the x.xx style of version numbers, the minor
257 version may start with zero. Using 08 or 09 is a bad idea because users
258 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
259 interpreted as invalid octal numbers. I've updated the previous comment in
260 configure.ac, and also added a check that gives an error if 08 or 09 are
261 used.
262
263 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
264 causing partial matching to fail when the end of the subject matched \W
265 in a UTF-8 pattern where \W was quantified with a minimum of 3.
266
267 15. There were some discrepancies between the declarations in pcre_internal.h
268 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
269 their definitions. The declarations used "const uschar *" and the
270 definitions used USPTR. Even though USPTR is normally defined as "const
271 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
272 reported that: "This difference in casting confuses some C++ compilers, for
273 example, SunCC recognizes above declarations as different functions and
274 generates broken code for hbpcre." I have changed the declarations to use
275 USPTR.
276
277 16. GNU libtool is named differently on some systems. The autogen.sh script now
278 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
279 (FreeBSD).
280
281 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
282 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
283 comment: "Figure out how to create a longlong from a string: strtoll and
284 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
285 instance, but it only takes 2 args instead of 3!"
286
287 18. A subtle bug concerned with back references has been fixed by a change of
288 specification, with a corresponding code fix. A pattern such as
289 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
290 refers, was giving matches when it shouldn't. For example, xa=xaaa would
291 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
292 same bug. Such groups have to be quantified to be useful, or contained
293 inside another quantified group. (If there's no repetition, the reference
294 can never match.) The problem arises because, having left the group and
295 moved on to the rest of the pattern, a later failure that backtracks into
296 the group uses the captured value from the final iteration of the group
297 rather than the correct earlier one. I have fixed this in PCRE by forcing
298 any group that contains a reference to itself to be an atomic group; that
299 is, there cannot be any backtracking into it once it has completed. This is
300 similar to recursive and subroutine calls.
301
302
303 Version 8.00 19-Oct-09
304 ----------------------
305
306 1. The table for translating pcre_compile() error codes into POSIX error codes
307 was out-of-date, and there was no check on the pcre_compile() error code
308 being within the table. This could lead to an OK return being given in
309 error.
310
311 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
312 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
313 in a Windows environment.
314
315 3. The pcregrep --count option prints the count for each file even when it is
316 zero, as does GNU grep. However, pcregrep was also printing all files when
317 --files-with-matches was added. Now, when both options are given, it prints
318 counts only for those files that have at least one match. (GNU grep just
319 prints the file name in this circumstance, but including the count seems
320 more useful - otherwise, why use --count?) Also ensured that the
321 combination -clh just lists non-zero counts, with no names.
322
323 4. The long form of the pcregrep -F option was incorrectly implemented as
324 --fixed_strings instead of --fixed-strings. This is an incompatible change,
325 but it seems right to fix it, and I didn't think it was worth preserving
326 the old behaviour.
327
328 5. The command line items --regex=pattern and --regexp=pattern were not
329 recognized by pcregrep, which required --regex pattern or --regexp pattern
330 (with a space rather than an '='). The man page documented the '=' forms,
331 which are compatible with GNU grep; these now work.
332
333 6. No libpcreposix.pc file was created for pkg-config; there was just
334 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
335
336 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
337 when UCP support is not needed, by modifying the Python script that
338 generates it from Unicode data files. This should not matter if the module
339 is correctly used as a library, but I received one complaint about 50K of
340 unwanted data. My guess is that the person linked everything into his
341 program rather than using a library. Anyway, it does no harm.
342
343 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
344 was a minimum greater than 1 for a wide character in a possessive
345 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
346 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
347 character. Chaos in the form of incorrect output or a compiling loop could
348 result.
349
350 9. The restrictions on what a pattern can contain when partial matching is
351 requested for pcre_exec() have been removed. All patterns can now be
352 partially matched by this function. In addition, if there are at least two
353 slots in the offset vector, the offset of the earliest inspected character
354 for the match and the offset of the end of the subject are set in them when
355 PCRE_ERROR_PARTIAL is returned.
356
357 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
358 synonymous with PCRE_PARTIAL, for backwards compatibility, and
359 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
360 and may be more useful for multi-segment matching.
361
362 11. Partial matching with pcre_exec() is now more intuitive. A partial match
363 used to be given if ever the end of the subject was reached; now it is
364 given only if matching could not proceed because another character was
365 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
366 string "Z", which now yields "no match" instead of "partial match". In the
367 case of pcre_dfa_exec(), "no match" is given if every matching path for the
368 final character ended with (*FAIL).
369
370 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
371 if the pattern had a "must contain" character that was already found in the
372 earlier partial match, unless partial matching was again requested. For
373 example, with the pattern /dog.(body)?/, the "must contain" character is
374 "g". If the first part-match was for the string "dog", restarting with
375 "sbody" failed. This bug has been fixed.
376
377 13. The string returned by pcre_dfa_exec() after a partial match has been
378 changed so that it starts at the first inspected character rather than the
379 first character of the match. This makes a difference only if the pattern
380 starts with a lookbehind assertion or \b or \B (\K is not supported by
381 pcre_dfa_exec()). It's an incompatible change, but it makes the two
382 matching functions compatible, and I think it's the right thing to do.
383
384 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
385 so that the demonstration program is easily available in environments where
386 PCRE has not been installed from source.
387
388 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
389 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
390 library.
391
392 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
393 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
394 is not the first non-POSIX option to be added. Clearly some people find
395 these options useful.
396
397 17. If a caller to the POSIX matching function regexec() passes a non-zero
398 value for nmatch with a NULL value for pmatch, the value of
399 nmatch is forced to zero.
400
401 18. RunGrepTest did not have a test for the availability of the -u option of
402 the diff command, as RunTest does. It now checks in the same way as
403 RunTest, and also checks for the -b option.
404
405 19. If an odd number of negated classes containing just a single character
406 interposed, within parentheses, between a forward reference to a named
407 subpattern and the definition of the subpattern, compilation crashed with
408 an internal error, complaining that it could not find the referenced
409 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
410 [The bug was that it was starting one character too far in when skipping
411 over the character class, thus treating the ] as data rather than
412 terminating the class. This meant it could skip too much.]
413
414 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
415 /g option in pcretest when the pattern contains \K, which makes it possible
416 to have an empty string match not at the start, even when the pattern is
417 anchored. Updated pcretest and pcredemo to use this option.
418
419 21. If the maximum number of capturing subpatterns in a recursion was greater
420 than the maximum at the outer level, the higher number was returned, but
421 with unset values at the outer level. The correct (outer level) value is
422 now given.
423
424 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
425 PCRE did not set those parentheses (unlike Perl). I have now found a way to
426 make it do so. The string so far is captured, making this feature
427 compatible with Perl.
428
429 23. The tests have been re-organized, adding tests 11 and 12, to make it
430 possible to check the Perl 5.10 features against Perl 5.10.
431
432 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
433 pattern matches a fixed length string. PCRE did not allow this; now it
434 does. Neither allows recursion.
435
436 25. I finally figured out how to implement a request to provide the minimum
437 length of subject string that was needed in order to match a given pattern.
438 (It was back references and recursion that I had previously got hung up
439 on.) This code has now been added to pcre_study(); it finds a lower bound
440 to the length of subject needed. It is not necessarily the greatest lower
441 bound, but using it to avoid searching strings that are too short does give
442 some useful speed-ups. The value is available to calling programs via
443 pcre_fullinfo().
444
445 26. While implementing 25, I discovered to my embarrassment that pcretest had
446 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
447 study optimizations had never been tested with that matching function.
448 Oops. What is worse, even when it was passed study data, there was a bug in
449 pcre_dfa_exec() that meant it never actually used it. Double oops. There
450 were also very few tests of studied patterns with pcre_dfa_exec().
451
452 27. If (?| is used to create subpatterns with duplicate numbers, they are now
453 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
454 on the other side of the coin, they are no longer allowed to have different
455 names, because these cannot be distinguished in PCRE, and this has caused
456 confusion. (This is a difference from Perl.)
457
458 28. When duplicate subpattern names are present (necessarily with different
459 numbers, as required by 27 above), and a test is made by name in a
460 conditional pattern, either for a subpattern having been matched, or for
461 recursion in such a pattern, all the associated numbered subpatterns are
462 tested, and the overall condition is true if the condition is true for any
463 one of them. This is the way Perl works, and is also more like the way
464 testing by number works.
465
466
467 Version 7.9 11-Apr-09
468 ---------------------
469
470 1. When building with support for bzlib/zlib (pcregrep) and/or readline
471 (pcretest), all targets were linked against these libraries. This included
472 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
473 libraries. This caused unwanted dependencies to be created. This problem
474 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
475 pcretest is linked with readline.
476
477 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
478 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
479 moved outside it again, because FALSE and TRUE are already defined in AIX,
480 but BOOL is not.
481
482 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
483 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
484
485 4. The pcregrep documentation said spaces were inserted as well as colons (or
486 hyphens) following file names and line numbers when outputting matching
487 lines. This is not true; no spaces are inserted. I have also clarified the
488 wording for the --colour (or --color) option.
489
490 5. In pcregrep, when --colour was used with -o, the list of matching strings
491 was not coloured; this is different to GNU grep, so I have changed it to be
492 the same.
493
494 6. When --colo(u)r was used in pcregrep, only the first matching substring in
495 each matching line was coloured. Now it goes on to look for further matches
496 of any of the test patterns, which is the same behaviour as GNU grep.
497
498 7. A pattern that could match an empty string could cause pcregrep to loop; it
499 doesn't make sense to accept an empty string match in pcregrep, so I have
500 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
501 seems to be how GNU grep behaves.
502
503 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
504 start or after a newline", because the conditional assertion was not being
505 correctly handled. The rule now is that both the assertion and what follows
506 in the first alternative must satisfy the test.
507
508 9. If auto-callout was enabled in a pattern with a conditional group whose
509 condition was an assertion, PCRE could crash during matching, both with
510 pcre_exec() and pcre_dfa_exec().
511
512 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
513 used for matching.
514
515 11. Unicode property support in character classes was not working for
516 characters (bytes) greater than 127 when not in UTF-8 mode.
517
518 12. Added the -M command line option to pcretest.
519
520 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
521
522 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
523
524 16. Added comments and documentation about mis-use of no_arg in the C++
525 wrapper.
526
527 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
528 from Martin Jerabek that uses macro names for all relevant character and
529 string constants.
530
531 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
532 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
533 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
534 these, but not everybody uses configure.
535
536 19. A conditional group that had only one branch was not being correctly
537 recognized as an item that could match an empty string. This meant that an
538 enclosing group might also not be so recognized, causing infinite looping
539 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
540 with the subject "ab", where knowledge that the repeated group can match
541 nothing is needed in order to break the loop.
542
543 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
544 exec(), but without supplying a callout function, matching went wrong.
545
546 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
547 leak if the size of the offset vector was greater than 30. When the vector
548 is smaller, the saved offsets during recursion go onto a local stack
549 vector, but for larger vectors malloc() is used. It was failing to free
550 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
551 error, in fact).
552
553 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
554 heapframe that is used only when UTF-8 support is enabled. This caused no
555 problem, but was untidy.
556
557 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
558 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
559 included within another project.
560
561 24. Steven Van Ingelgem's patches to add more options to the CMake support,
562 slightly modified by me:
563
564 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
565 not building pcregrep.
566
567 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
568 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
569
570 25. Forward references, both numeric and by name, in patterns that made use of
571 duplicate group numbers, could behave incorrectly or give incorrect errors,
572 because when scanning forward to find the reference group, PCRE was not
573 taking into account the duplicate group numbers. A pattern such as
574 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
575
576 26. Changed a few more instances of "const unsigned char *" to USPTR, making
577 the feature of a custom pointer more persuasive (as requested by a user).
578
579 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
580 pcretest.c, inside #ifndefs, because it seems they are sometimes already
581 pre-defined.
582
583 28. Added support for (*UTF8) at the start of a pattern.
584
585 29. Arrange for flags added by the "release type" setting in CMake to be shown
586 in the configuration summary.
587
588
589 Version 7.8 05-Sep-08
590 ---------------------
591
592 1. Replaced UCP searching code with optimized version as implemented for Ad
593 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
594 stage table and inline lookup instead of a function, giving speed ups of 2
595 to 5 times on some simple patterns that I tested. Permission was given to
596 distribute the MultiStage2.py script that generates the tables (it's not in
597 the tarball, but is in the Subversion repository).
598
599 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
600 scripts.
601
602 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
603 a group with a zero qualifier. The result of the study could be incorrect,
604 or the function might crash, depending on the pattern.
605
606 4. Caseless matching was not working for non-ASCII characters in back
607 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
608 It now works when Unicode Property Support is available.
609
610 5. In pcretest, an escape such as \x{de} in the data was always generating
611 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
612 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
613 truncation.
614
615 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
616
617 7. Added two (int) casts to pcregrep when printing the difference of two
618 pointers, in case they are 64-bit values.
619
620 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
621 test 2 if it fails.
622
623 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
624 and a #define of that name to empty if it is not externally set. This is to
625 allow users of MSVC to set it if necessary.
626
627 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
628 the convenience functions in the pcre_get.c source file.
629
630 11. An option change at the start of a pattern that had top-level alternatives
631 could cause overwriting and/or a crash. This command provoked a crash in
632 some environments:
633
634 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
635
636 This potential security problem was recorded as CVE-2008-2371.
637
638 12. For a pattern where the match had to start at the beginning or immediately
639 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
640 pcre_dfa_exec() could read past the end of the passed subject if there was
641 no match. To help with detecting such bugs (e.g. with valgrind), I modified
642 pcretest so that it places the subject at the end of its malloc-ed buffer.
643
644 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
645 exec() might read past the end of the data buffer in UTF-8 mode.
646
647 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
648 the data contained the byte 0x85 as part of a UTF-8 character within its
649 first line. This applied both to normal and DFA matching.
650
651 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
652 /^[^d]*?$/8 failed to match "abc".
653
654 16. Added a missing copyright notice to pcrecpp_internal.h.
655
656 17. Make it more clear in the documentation that values returned from
657 pcre_exec() in ovector are byte offsets, not character counts.
658
659 18. Tidied a few places to stop certain compilers from issuing warnings.
660
661 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
662 supplied by Stefan Weber. I made a further small update for 7.8 because
663 there is a change of source arrangements: the pcre_searchfuncs.c module is
664 replaced by pcre_ucd.c.
665
666
667 Version 7.7 07-May-08
668 ---------------------
669
670 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
671 a string to a long long, pretend we don't even have a long long." This is
672 done by checking for the strtoq, strtoll, and _strtoi64 functions.
673
674 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
675 pre-7.6 versions, which defined a global no_arg variable instead of putting
676 it in the RE class. (See also #8 below.)
677
678 3. Remove a line of dead code, identified by coverity and reported by Nuno
679 Lopes.
680
681 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
682
683 (1) The include/exclude patterns were being applied to the whole pathnames
684 of files, instead of just to the final components.
685
686 (2) If there was more than one level of directory, the subdirectories were
687 skipped unless they satisfied the include/exclude conditions. This is
688 inconsistent with GNU grep (and could even be seen as contrary to the
689 pcregrep specification - which I improved to make it absolutely clear).
690 The action now is always to scan all levels of directory, and just
691 apply the include/exclude patterns to regular files.
692
693 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
694 --exclude_dir in the tests to avoid scanning .svn directories.
695
696 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
697 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
698 doesn't support NULs in patterns.
699
700 7. Added some missing "const"s to declarations of static tables in
701 pcre_compile.c and pcre_dfa_exec.c.
702
703 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
704 caused by fix #2 above. (Subsequently also a second patch to fix the
705 first patch. And a third patch - this was a messy problem.)
706
707 9. Applied Craig's patch to remove the use of push_back().
708
709 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
710 matching function regexec().
711
712 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
713 which, however, unlike Perl's \g{...}, are subroutine calls, not back
714 references. PCRE supports relative numbers with this syntax (I don't think
715 Oniguruma does).
716
717 12. Previously, a group with a zero repeat such as (...){0} was completely
718 omitted from the compiled regex. However, this means that if the group
719 was called as a subroutine from elsewhere in the pattern, things went wrong
720 (an internal error was given). Such groups are now left in the compiled
721 pattern, with a new opcode that causes them to be skipped at execution
722 time.
723
724 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
725 to the way PCRE behaves:
726
727 (a) A lone ] character is dis-allowed (Perl treats it as data).
728
729 (b) A back reference to an unmatched subpattern matches an empty string
730 (Perl fails the current match path).
731
732 (c) A data ] in a character class must be notated as \] because if the
733 first data character in a class is ], it defines an empty class. (In
734 Perl it is not possible to have an empty class.) The empty class []
735 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
736 The negative empty class [^] matches any one character, independently
737 of the DOTALL setting.
738
739 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
740 non-existent subpattern following a character class starting with ']' and
741 containing () gave an internal compiling error instead of "reference to
742 non-existent subpattern". Fortunately, when the pattern did exist, the
743 compiled code was correct. (When scanning forwards to check for the
744 existencd of the subpattern, it was treating the data ']' as terminating
745 the class, so got the count wrong. When actually compiling, the reference
746 was subsequently set up correctly.)
747
748 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
749 it was being rejected as not supported by pcre_dfa_exec(), even though
750 other assertions are supported. I have made pcre_dfa_exec() support
751 (*FAIL).
752
753 16. The implementation of 13c above involved the invention of a new opcode,
754 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
755 cannot be changed at match time, I realized I could make a small
756 improvement to matching performance by compiling OP_ALLANY instead of
757 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
758 on the OP_ANY path.
759
760 17. Compiling pcretest on Windows with readline support failed without the
761 following two fixes: (1) Make the unistd.h include conditional on
762 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
763
764 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
765 ncurses library to be included for pcretest when ReadLine support is
766 requested, but also to allow for it to be overridden. This patch came from
767 Daniel Bergström.
768
769 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
770 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
771 any errors with the current Unicode tables. Thanks to Peter Kankowski for
772 spotting this.
773
774
775 Version 7.6 28-Jan-08
776 ---------------------
777
778 1. A character class containing a very large number of characters with
779 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
780 overflow.
781
782 2. Patch to cut out the "long long" test in pcrecpp_unittest when
783 HAVE_LONG_LONG is not defined.
784
785 3. Applied Christian Ehrlicher's patch to update the CMake build files to
786 bring them up to date and include new features. This patch includes:
787
788 - Fixed PH's badly added libz and libbz2 support.
789 - Fixed a problem with static linking.
790 - Added pcredemo. [But later removed - see 7 below.]
791 - Fixed dftables problem and added an option.
792 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
793 HAVE_LONG_LONG.
794 - Added readline support for pcretest.
795 - Added an listing of the option settings after cmake has run.
796
797 4. A user submitted a patch to Makefile that makes it easy to create
798 "pcre.dll" under mingw when using Configure/Make. I added stuff to
799 Makefile.am that cause it to include this special target, without
800 affecting anything else. Note that the same mingw target plus all
801 the other distribution libraries and programs are now supported
802 when configuring with CMake (see 6 below) instead of with
803 Configure/Make.
804
805 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
806 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
807 exported in the Windows port". It has not yet been confirmed that the patch
808 solves the problem, but it does no harm.
809
810 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
811 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
812 with CMake, and also correct the comment about stack recursion.
813
814 7. Remove the automatic building of pcredemo from the ./configure system and
815 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
816 of a program that users should build themselves after PCRE is installed, so
817 building it automatically is not really right. What is more, it gave
818 trouble in some build environments.
819
820 8. Further tidies to CMakeLists.txt from Sheri and Christian.
821
822
823 Version 7.5 10-Jan-08
824 ---------------------
825
826 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
827 values in parens when parsing an RE using the C++ wrapper."
828
829 2. Negative specials like \S did not work in character classes in UTF-8 mode.
830 Characters greater than 255 were excluded from the class instead of being
831 included.
832
833 3. The same bug as (2) above applied to negated POSIX classes such as
834 [:^space:].
835
836 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
837 defined or documented. It seems to have been a typo for PCRE_STATIC, so
838 I have changed it.
839
840 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
841 first named subpattern) and a construct such as (?&a) would reference the
842 first named subpattern whose name started with "a" (in other words, the
843 length check was missing). Both these problems are fixed. "Subpattern name
844 expected" is now given for (?&) (a zero-length name), and this patch also
845 makes it give the same error for \k'' (previously it complained that that
846 was a reference to a non-existent subpattern).
847
848 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
849 this is right because (?- can be followed by option settings as well as by
850 digits. I have, however, made the messages clearer.
851
852 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
853 than the number used in the conditional) now cause a compile-time error.
854 This is actually not compatible with Perl, which accepts such patterns, but
855 treats the conditional as always being FALSE (as PCRE used to), but it
856 seems to me that giving a diagnostic is better.
857
858 8. Change "alphameric" to the more common word "alphanumeric" in comments
859 and messages.
860
861 9. Fix two occurrences of "backslash" in comments that should have been
862 "backspace".
863
864 10. Remove two redundant lines of code that can never be obeyed (their function
865 was moved elsewhere).
866
867 11. The program that makes PCRE's Unicode character property table had a bug
868 which caused it to generate incorrect table entries for sequences of
869 characters that have the same character type, but are in different scripts.
870 It amalgamated them into a single range, with the script of the first of
871 them. In other words, some characters were in the wrong script. There were
872 thirteen such cases, affecting characters in the following ranges:
873
874 U+002b0 - U+002c1
875 U+0060c - U+0060d
876 U+0061e - U+00612
877 U+0064b - U+0065e
878 U+0074d - U+0076d
879 U+01800 - U+01805
880 U+01d00 - U+01d77
881 U+01d9b - U+01dbf
882 U+0200b - U+0200f
883 U+030fc - U+030fe
884 U+03260 - U+0327f
885 U+0fb46 - U+0fbb1
886 U+10450 - U+1049d
887
888 12. The -o option (show only the matching part of a line) for pcregrep was not
889 compatible with GNU grep in that, if there was more than one match in a
890 line, it showed only the first of them. It now behaves in the same way as
891 GNU grep.
892
893 13. If the -o and -v options were combined for pcregrep, it printed a blank
894 line for every non-matching line. GNU grep prints nothing, and pcregrep now
895 does the same. The return code can be used to tell if there were any
896 non-matching lines.
897
898 14. Added --file-offsets and --line-offsets to pcregrep.
899
900 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
901 infinitely looping recursion. The bug was that positive lookaheads were not
902 being skipped when checking for a possible empty match (negative lookaheads
903 and both kinds of lookbehind were skipped).
904
905 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
906 inclusion of <windows.h> to before rather than after the definition of
907 INVALID_FILE_ATTRIBUTES (patch from David Byron).
908
909 17. Specifying a possessive quantifier with a specific limit for a Unicode
910 character property caused pcre_compile() to compile bad code, which led at
911 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
912 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
913 caused the error; without that there was no problem.
914
915 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
916
917 19. Added --enable-pcretest-libreadline.
918
919 20. In pcrecpp.cc, the variable 'count' was incremented twice in
920 RE::GlobalReplace(). As a result, the number of replacements returned was
921 double what it should be. I removed one of the increments, but Craig sent a
922 later patch that removed the other one (the right fix) and added unit tests
923 that check the return values (which was not done before).
924
925 21. Several CMake things:
926
927 (1) Arranged that, when cmake is used on Unix, the libraries end up with
928 the names libpcre and libpcreposix, not just pcre and pcreposix.
929
930 (2) The above change means that pcretest and pcregrep are now correctly
931 linked with the newly-built libraries, not previously installed ones.
932
933 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
934
935 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
936 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
937 UTF-8 newline character). The key issue is that the pattern starts .*;
938 this means that the match must be either at the beginning, or after a
939 newline. The bug was in the code for advancing after a failed match and
940 checking that the new position followed a newline. It was not taking
941 account of UTF-8 characters correctly.
942
943 23. PCRE was behaving differently from Perl in the way it recognized POSIX
944 character classes. PCRE was not treating the sequence [:...:] as a
945 character class unless the ... were all letters. Perl, however, seems to
946 allow any characters between [: and :], though of course it rejects as
947 unknown any "names" that contain non-letters, because all the known class
948 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
949 for example, whereas PCRE did not - it did not recognize a POSIX character
950 class. This seemed a bit dangerous, so the code has been changed to be
951 closer to Perl. The behaviour is not identical to Perl, because PCRE will
952 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
953 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
954 Perl does, and where it didn't before.
955
956 24. Rewrite so as to remove the single use of %n from pcregrep because in some
957 Windows environments %n is disabled by default.
958
959
960 Version 7.4 21-Sep-07
961 ---------------------
962
963 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
964 means that a class such as [\s] counted as "explicit reference to CR or
965 LF". That isn't really right - the whole point of the change was to try to
966 help when there was an actual mention of one of the two characters. So now
967 the change happens only if \r or \n (or a literal CR or LF) character is
968 encountered.
969
970 2. The 32-bit options word was also used for 6 internal flags, but the numbers
971 of both had grown to the point where there were only 3 bits left.
972 Fortunately, there was spare space in the data structure, and so I have
973 moved the internal flags into a new 16-bit field to free up more option
974 bits.
975
976 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
977 but did not set the internal JCHANGED flag - either of these is enough to
978 control the way the "get" function works - but the PCRE_INFO_JCHANGED
979 facility is supposed to tell if (?J) was ever used, so now (?J) at the
980 start sets both bits.
981
982 4. Added options (at build time, compile time, exec time) to change \R from
983 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
984
985 5. doc/pcresyntax.html was missing from the distribution.
986
987 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
988 compatibility, even though it is no longer used.
989
990 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
991 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
992 windows.h file is present (where different names are used). [This was
993 reversed later after testing - see 16 below.]
994
995 8. Changed all #include <config.h> to #include "config.h". There were also
996 some further <pcre.h> cases that I changed to "pcre.h".
997
998 9. When pcregrep was used with the --colour option, it missed the line ending
999 sequence off the lines that it output.
1000
1001 10. It was pointed out to me that arrays of string pointers cause lots of
1002 relocations when a shared library is dynamically loaded. A technique of
1003 using a single long string with a table of offsets can drastically reduce
1004 these. I have refactored PCRE in four places to do this. The result is
1005 dramatic:
1006
1007 Originally: 290
1008 After changing UCP table: 187
1009 After changing error message table: 43
1010 After changing table of "verbs" 36
1011 After changing table of Posix names 22
1012
1013 Thanks to the folks working on Gregex for glib for this insight.
1014
1015 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1016 unicode-properties was also set.
1017
1018 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1019
1020 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1021 checked only for CRLF.
1022
1023 14. Added casts to pcretest.c to avoid compiler warnings.
1024
1025 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1026
1027 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1028 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1029 entirely. This removes changes made in 7 above.
1030
1031 17. The CMake files have been updated, and there is now more information about
1032 building with CMake in the NON-UNIX-USE document.
1033
1034
1035 Version 7.3 28-Aug-07
1036 ---------------------
1037
1038 1. In the rejigging of the build system that eventually resulted in 7.1, the
1039 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1040 brackets there is not right, since it causes compilers to look for an
1041 installed pcre.h, not the version that is in the source that is being
1042 compiled (which of course may be different). I have changed it back to:
1043
1044 #include "pcre.h"
1045
1046 I have a vague recollection that the change was concerned with compiling in
1047 different directories, but in the new build system, that is taken care of
1048 by the VPATH setting the Makefile.
1049
1050 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1051 when the subject happened to end in the byte 0x85 (e.g. if the last
1052 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1053 characters but of course it shouldn't be taken as a newline when it is part
1054 of another character. The bug was that, for an unlimited repeat of . in
1055 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1056 characters when looking for a newline.
1057
1058 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1059
1060 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1061 in debug output.
1062
1063 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1064 long printing in the pcrecpp unittest when running under MinGW.
1065
1066 6. ESC_K was left out of the EBCDIC table.
1067
1068 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1069 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1070 limit also applies to "virtual nesting" when a pattern is recursive, and in
1071 this case 1000 isn't so big. I have been able to remove this limit at the
1072 expense of backing off one optimization in certain circumstances. Normally,
1073 when pcre_exec() would call its internal match() function recursively and
1074 immediately return the result unconditionally, it uses a "tail recursion"
1075 feature to save stack. However, when a subpattern that can match an empty
1076 string has an unlimited repetition quantifier, it no longer makes this
1077 optimization. That gives it a stack frame in which to save the data for
1078 checking that an empty string has been matched. Previously this was taken
1079 from the 1000-entry workspace that had been reserved. So now there is no
1080 explicit limit, but more stack is used.
1081
1082 8. Applied Daniel's patches to solve problems with the import/export magic
1083 syntax that is required for Windows, and which was going wrong for the
1084 pcreposix and pcrecpp parts of the library. These were overlooked when this
1085 problem was solved for the main library.
1086
1087 9. There were some crude static tests to avoid integer overflow when computing
1088 the size of patterns that contain repeated groups with explicit upper
1089 limits. As the maximum quantifier is 65535, the maximum group length was
1090 set at 30,000 so that the product of these two numbers did not overflow a
1091 32-bit integer. However, it turns out that people want to use groups that
1092 are longer than 30,000 bytes (though not repeat them that many times).
1093 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1094 made it possible to implement the integer overflow checks in a much more
1095 dynamic way, which I have now done. The artificial limitation on group
1096 length has been removed - we now have only the limit on the total length of
1097 the compiled pattern, which depends on the LINK_SIZE setting.
1098
1099 10. Fixed a bug in the documentation for get/copy named substring when
1100 duplicate names are permitted. If none of the named substrings are set, the
1101 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1102 empty string.
1103
1104 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1105 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1106 because the ] is interpreted as the first data character and the
1107 terminating ] is not found. PCRE has been made compatible with Perl in this
1108 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1109 cause memory overwriting.
1110
1111 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1112 string has been matched (to stop an infinite loop). It was not recognizing
1113 a conditional subpattern that could match an empty string if that
1114 subpattern was within another subpattern. For example, it looped when
1115 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1116 condition was not nested. This bug has been fixed.
1117
1118 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1119 past the start of the subject in the presence of bytes with the top bit
1120 set, for example "\x8aBCD".
1121
1122 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1123 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1124
1125 14. Optimized (?!) to (*FAIL).
1126
1127 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1128 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1129 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1130 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1131 does: it's just the validity check that is more restrictive.
1132
1133 16. Inserted checks for integer overflows during escape sequence (backslash)
1134 processing, and also fixed erroneous offset values for syntax errors during
1135 backslash processing.
1136
1137 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1138 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1139
1140 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1141 caused an overrun.
1142
1143 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1144 something other than just ASCII characters) inside a group that had an
1145 unlimited repeat caused a loop at compile time (while checking to see
1146 whether the group could match an empty string).
1147
1148 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1149 [\P{Any}] did so. (Error in the code for printing property names.)
1150
1151 21. An orphan \E inside a character class could cause a crash.
1152
1153 22. A repeated capturing bracket such as (A)? could cause a wild memory
1154 reference during compilation.
1155
1156 23. There are several functions in pcre_compile() that scan along a compiled
1157 expression for various reasons (e.g. to see if it's fixed length for look
1158 behind). There were bugs in these functions when a repeated \p or \P was
1159 present in the pattern. These operators have additional parameters compared
1160 with \d, etc, and these were not being taken into account when moving along
1161 the compiled data. Specifically:
1162
1163 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1164 length.
1165
1166 (b) An item such as \pL+ within a repeated group could cause crashes or
1167 loops.
1168
1169 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1170 "reference to non-existent subpattern" error.
1171
1172 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1173
1174 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1175 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1176
1177 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1178
1179 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1180 character were causing crashes (broken optimization).
1181
1182 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1183 \p or \P) caused a compile-time loop.
1184
1185 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1186 break. For example, the unstudied pattern [\r\n]A does not match the string
1187 "\r\nA" because change 7.0/46 below moves the current point on by two
1188 characters after failing to match at the start. However, the pattern \nA
1189 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1190 the same is true. There doesn't seem any very clean way out of this, but
1191 what I have chosen to do makes the common cases work: PCRE now takes note
1192 of whether there can be an explicit match for \r or \n anywhere in the
1193 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1194 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1195 pattern has explicit CR or LF references.
1196
1197 29. Added (*CR) etc for changing newline setting at start of pattern.
1198
1199
1200 Version 7.2 19-Jun-07
1201 ---------------------
1202
1203 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1204 which is apparently normally available under Windows.
1205
1206 2. Re-jig the pcregrep tests with different newline settings in an attempt
1207 to make them independent of the local environment's newline setting.
1208
1209 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1210
1211 4. Some of the "internals" tests were previously cut out when the link size
1212 was not 2, because the output contained actual offsets. The recent new
1213 "Z" feature of pcretest means that these can be cut out, making the tests
1214 usable with all link sizes.
1215
1216 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1217 stack recursion. This gives a massive performance boost under BSD, but just
1218 a small improvement under Linux. However, it saves one field in the frame
1219 in all cases.
1220
1221 6. Added more features from the forthcoming Perl 5.10:
1222
1223 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1224 recursion call. It refers to the nth most recently opened parentheses.
1225
1226 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1227 to be opened parentheses.
1228
1229 (c) Conditions that refer to capturing parentheses can be specified
1230 relatively, for example, (?(-2)... or (?(+3)...
1231
1232 (d) \K resets the start of the current match so that everything before
1233 is not part of it.
1234
1235 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1236
1237 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1238 reference syntax.
1239
1240 (g) (?| introduces a group in which the numbering of parentheses in each
1241 alternative starts with the same number.
1242
1243 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1244
1245 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1246 PCRE_INFO_JCHANGED.
1247
1248 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1249 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1250 for detecting groups that can match an empty string.
1251
1252 9. A pattern with a very large number of alternatives (more than several
1253 hundred) was running out of internal workspace during the pre-compile
1254 phase, where pcre_compile() figures out how much memory will be needed. A
1255 bit of new cunning has reduced the workspace needed for groups with
1256 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1257 workspace instead of running out of the 4096 that are available.
1258
1259 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1260
1261 11. Applied patch from Google to remove an optimization that didn't quite work.
1262 The report of the bug said:
1263
1264 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1265 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1266 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1267
1268 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1269 it matched the wrong number of bytes.
1270
1271
1272 Version 7.1 24-Apr-07
1273 ---------------------
1274
1275 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1276 that is more "standard", making use of automake and other Autotools. There
1277 is some re-arrangement of the files and adjustment of comments consequent
1278 on this.
1279
1280 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1281 for recursive directory scanning broke on some systems because the files
1282 are not scanned in any specific order and on different systems the order
1283 was different. A call to "sort" has been inserted into RunGrepTest for the
1284 approprate test as a short-term fix. In the longer term there may be an
1285 alternative.
1286
1287 3. I had an email from Eric Raymond about problems translating some of PCRE's
1288 man pages to HTML (despite the fact that I distribute HTML pages, some
1289 people do their own conversions for various reasons). The problems
1290 concerned the use of low-level troff macros .br and .in. I have therefore
1291 removed all such uses from the man pages (some were redundant, some could
1292 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1293 HTML has been updated to handle .nf/.fi and to complain if it encounters
1294 .br or .in.
1295
1296 4. Updated comments in configure.ac that get placed in config.h.in and also
1297 arranged for config.h to be included in the distribution, with the name
1298 config.h.generic, for the benefit of those who have to compile without
1299 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1300
1301 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1302 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1303 makevp.bat for the current PCRE, using the additional files
1304 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1305
1306 6. A Windows user reported a minor discrepancy with test 2, which turned out
1307 to be caused by a trailing space on an input line that had got lost in his
1308 copy. The trailing space was an accident, so I've just removed it.
1309
1310 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1311 that is needed.
1312
1313 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1314 as "const" (a) because they are and (b) because it helps the PHP
1315 maintainers who have recently made a script to detect big data structures
1316 in the php code that should be moved to the .rodata section. I remembered
1317 to update Builducptable as well, so it won't revert if ucptable.h is ever
1318 re-created.
1319
1320 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1321 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1322 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1323 support is not required. This saves 1.5-2K of code, which is important in
1324 some applications.
1325
1326 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1327 so as not to refer to the tables, even though these functions will never be
1328 called when UTF-8 support is disabled. Otherwise there are problems with a
1329 shared library.
1330
1331 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1332
1333 (a) It was defining its arguments as char * instead of void *.
1334
1335 (b) It was assuming that all moves were upwards in memory; this was true
1336 a long time ago when I wrote it, but is no longer the case.
1337
1338 The emulated memove() is provided for those environments that have neither
1339 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1340 is clearly not the case, as these two bugs were recently reported.
1341
1342 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1343 and Detrail to create the HTML documentation, the .txt form of the man
1344 pages, and it removes trailing spaces from listed files. It also creates
1345 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1346 case, it wraps all the #defines with #ifndefs. This script should be run
1347 before "make dist".
1348
1349 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1350 with Unicode property support.
1351
1352 (a) For a maximizing quantifier, if the two different cases of the
1353 character were of different lengths in their UTF-8 codings (there are
1354 some cases like this - I found 11), and the matching function had to
1355 back up over a mixture of the two cases, it incorrectly assumed they
1356 were both the same length.
1357
1358 (b) When PCRE was configured to use the heap rather than the stack for
1359 recursion during matching, it was not correctly preserving the data for
1360 the other case of a UTF-8 character when checking ahead for a match
1361 while processing a minimizing repeat. If the check also involved
1362 matching a wide character, but failed, corruption could cause an
1363 erroneous result when trying to check for a repeat of the original
1364 character.
1365
1366 13. Some tidying changes to the testing mechanism:
1367
1368 (a) The RunTest script now detects the internal link size and whether there
1369 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1370 values substituted by "configure". (The RunGrepTest script already did
1371 this for UTF-8.) The configure.ac script no longer substitutes the
1372 relevant variables.
1373
1374 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1375 with length and offset values. This means that the output is different
1376 for different internal link sizes. Test 2 is skipped for link sizes
1377 other than 2 because of this, bypassing the problem. Unfortunately,
1378 there was also a test in test 3 (the locale tests) that used /B and
1379 failed for link sizes other than 2. Rather than cut the whole test out,
1380 I have added a new /Z option to pcretest that replaces the length and
1381 offset values with spaces. This is now used to make test 3 independent
1382 of link size. (Test 2 will be tidied up later.)
1383
1384 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1385 segmentation fault instead of returning the appropriate error message.
1386
1387 15. In multiline mode when the newline sequence was set to "any", the pattern
1388 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1389 This doesn't seem right; it now treats the CRLF combination as the line
1390 ending, and so does not match in that case. It's only a pattern such as ^$
1391 that would hit this one: something like ^ABC$ would have failed after \r
1392 and then tried again after \r\n.
1393
1394 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1395 in an attempt to make files that differ only in their line terminators
1396 compare equal. This works on Linux.
1397
1398 17. Under certain error circumstances pcregrep might try to free random memory
1399 as it exited. This is now fixed, thanks to valgrind.
1400
1401 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1402 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1403 was because its rules for how to advance for /g after matching an empty
1404 string at the end of a line did not allow for this case. They now check for
1405 it specially.
1406
1407 20. pcretest is supposed to handle patterns and data of any length, by
1408 extending its buffers when necessary. It was getting this wrong when the
1409 buffer for a data line had to be extended.
1410
1411 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1412 CRLF as a newline sequence.
1413
1414 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1415 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1416 I have nevertheless tidied it up.
1417
1418 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1419
1420 24. Added a man page for pcre-config.
1421
1422
1423 Version 7.0 19-Dec-06
1424 ---------------------
1425
1426 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1427 moving to gcc 4.1.1.
1428
1429 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1430 sys/time.h, which is documented as needed for this function. It doesn't
1431 seem to matter on Linux, but it showed up on some releases of OS X.
1432
1433 3. It seems that there are systems where bytes whose values are greater than
1434 127 match isprint() in the "C" locale. The "C" locale should be the
1435 default when a C program starts up. In most systems, only ASCII printing
1436 characters match isprint(). This difference caused the output from pcretest
1437 to vary, making some of the tests fail. I have changed pcretest so that:
1438
1439 (a) When it is outputting text in the compiled version of a pattern, bytes
1440 other than 32-126 are always shown as hex escapes.
1441
1442 (b) When it is outputting text that is a matched part of a subject string,
1443 it does the same, unless a different locale has been set for the match
1444 (using the /L modifier). In this case, it uses isprint() to decide.
1445
1446 4. Fixed a major bug that caused incorrect computation of the amount of memory
1447 required for a compiled pattern when options that changed within the
1448 pattern affected the logic of the preliminary scan that determines the
1449 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1450 that the computed length was too small. The symptoms of this bug were
1451 either the PCRE error "internal error: code overflow" from pcre_compile(),
1452 or a glibc crash with a message such as "pcretest: free(): invalid next
1453 size (fast)". Examples of patterns that provoked this bug (shown in
1454 pcretest format) are:
1455
1456 /(?-x: )/x
1457 /(?x)(?-x: \s*#\s*)/
1458 /((?i)[\x{c0}])/8
1459 /(?i:[\x{c0}])/8
1460
1461 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1462 is now done differently.
1463
1464 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1465 wrapper classes; (b) implement a new function in the C++ scanner that is
1466 more efficient than the old way of doing things because it avoids levels of
1467 recursion in the regex matching; (c) add a paragraph to the documentation
1468 for the FullMatch() function.
1469
1470 6. The escape sequence \n was being treated as whatever was defined as
1471 "newline". Not only was this contrary to the documentation, which states
1472 that \n is character 10 (hex 0A), but it also went horribly wrong when
1473 "newline" was defined as CRLF. This has been fixed.
1474
1475 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1476 was being set to -1 for the "end of line" case (supposedly a value that no
1477 character can have). Though this value is never used (the check for end of
1478 line is "zero bytes in current character"), it caused compiler complaints.
1479 I've changed it to 0xffffffff.
1480
1481 8. In pcre_version.c, the version string was being built by a sequence of
1482 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1483 string (as it is for production releases) called a macro with an empty
1484 argument. The C standard says the result of this is undefined. The gcc
1485 compiler treats it as an empty string (which was what was wanted) but it is
1486 reported that Visual C gives an error. The source has been hacked around to
1487 avoid this problem.
1488
1489 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1490 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1491 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1492 of them did).
1493
1494 10. Originally, pcretest opened its input and output without "b"; then I was
1495 told that "b" was needed in some environments, so it was added for release
1496 5.0 to both the input and output. (It makes no difference on Unix-like
1497 systems.) Later I was told that it is wrong for the input on Windows. I've
1498 now abstracted the modes into two macros, to make it easier to fiddle with
1499 them, and removed "b" from the input mode under Windows.
1500
1501 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1502
1503 12. Added -help and --help to pcretest as an official way of being reminded
1504 of the options.
1505
1506 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1507 and pcrecpp.cc because they annoy compilers at high warning levels.
1508
1509 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1510
1511 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1512 scripts are not C programs :-) and which was not noticed because it works
1513 on Linux.
1514
1515 16. pcretest is supposed to handle any length of pattern and data line (as one
1516 line or as a continued sequence of lines) by extending its input buffer if
1517 necessary. This feature was broken for very long pattern lines, leading to
1518 a string of junk being passed to pcre_compile() if the pattern was longer
1519 than about 50K.
1520
1521 17. I have done a major re-factoring of the way pcre_compile() computes the
1522 amount of memory needed for a compiled pattern. Previously, there was code
1523 that made a preliminary scan of the pattern in order to do this. That was
1524 OK when PCRE was new, but as the facilities have expanded, it has become
1525 harder and harder to keep it in step with the real compile phase, and there
1526 have been a number of bugs (see for example, 4 above). I have now found a
1527 cunning way of running the real compile function in a "fake" mode that
1528 enables it to compute how much memory it would need, while actually only
1529 ever using a few hundred bytes of working memory and without too many
1530 tests of the mode. This should make future maintenance and development
1531 easier. A side effect of this work is that the limit of 200 on the nesting
1532 depth of parentheses has been removed (though this was never a serious
1533 limitation, I suspect). However, there is a downside: pcre_compile() now
1534 runs more slowly than before (30% or more, depending on the pattern). I
1535 hope this isn't a big issue. There is no effect on runtime performance.
1536
1537 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1538 newline (only possible for the last line of a file) and it was a
1539 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1540
1541 19. Added additional timing features to pcretest. (1) The -tm option now times
1542 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1543 separate command line item, by a number that specifies the number of
1544 repeats to use when timing. The default is 50000; this gives better
1545 precision, but takes uncomfortably long for very large patterns.
1546
1547 20. Extended pcre_study() to be more clever in cases where a branch of a
1548 subpattern has no definite first character. For example, (a*|b*)[cd] would
1549 previously give no result from pcre_study(). Now it recognizes that the
1550 first character must be a, b, c, or d.
1551
1552 21. There was an incorrect error "recursive call could loop indefinitely" if
1553 a subpattern (or the entire pattern) that was being tested for matching an
1554 empty string contained only one non-empty item after a nested subpattern.
1555 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1556 incorrectly, because the \d was being skipped in the check.
1557
1558 22. The pcretest program now has a new pattern option /B and a command line
1559 option -b, which is equivalent to adding /B to every pattern. This causes
1560 it to show the compiled bytecode, without the additional information that
1561 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1562 is the same as /B/I).
1563
1564 23. A new optimization is now able automatically to treat some sequences such
1565 as a*b as a*+b. More specifically, if something simple (such as a character
1566 or a simple class like \d) has an unlimited quantifier, and is followed by
1567 something that cannot possibly match the quantified thing, the quantifier
1568 is automatically "possessified".
1569
1570 24. A recursive reference to a subpattern whose number was greater than 39
1571 went wrong under certain circumstances in UTF-8 mode. This bug could also
1572 have affected the operation of pcre_study().
1573
1574 25. Realized that a little bit of performance could be had by replacing
1575 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1576
1577 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1578
1579 27. Possessive quantifiers such as a++ were previously implemented by turning
1580 them into atomic groups such as ($>a+). Now they have their own opcodes,
1581 which improves performance. This includes the automatically created ones
1582 from 23 above.
1583
1584 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1585 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1586 the first matched character to be a colon. This applied both to named and
1587 numbered groups.
1588
1589 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1590
1591 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1592 building PCRE on Windows easier, so I have included it in the distribution.
1593
1594 31. There is now a check in pcretest against a ridiculously large number being
1595 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1596 loop, the loop is abandoned.
1597
1598 32. Forward references to subpatterns in conditions such as (?(2)...) where
1599 subpattern 2 is defined later cause pcre_compile() to search forwards in
1600 the pattern for the relevant set of parentheses. This search went wrong
1601 when there were unescaped parentheses in a character class, parentheses
1602 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1603
1604 33. "Subroutine" calls and backreferences were previously restricted to
1605 referencing subpatterns earlier in the regex. This restriction has now
1606 been removed.
1607
1608 34. Added a number of extra features that are going to be in Perl 5.10. On the
1609 whole, these are just syntactic alternatives for features that PCRE had
1610 previously implemented using the Python syntax or my own invention. The
1611 other formats are all retained for compatibility.
1612
1613 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1614 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1615 also .NET compatible.
1616
1617 (b) A recursion or subroutine call to a named group can now be defined as
1618 (?&name) as well as (?P>name).
1619
1620 (c) A backreference to a named group can now be defined as \k<name> or
1621 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1622 5.10, are also .NET compatible.
1623
1624 (d) A conditional reference to a named group can now use the syntax
1625 (?(<name>) or (?('name') as well as (?(name).
1626
1627 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1628 groups (named and numbered) that are never evaluated inline, but can be
1629 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1630 is always false. There may be only one alternative in such a group.
1631
1632 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1633 as the simple (?(R). The condition is true only if the most recent
1634 recursion is that of the given number or name. It does not search out
1635 through the entire recursion stack.
1636
1637 (g) The escape \gN or \g{N} has been added, where N is a positive or
1638 negative number, specifying an absolute or relative reference.
1639
1640 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1641 some "unreachable code" warnings.
1642
1643 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1644 things, this adds five new scripts.
1645
1646 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1647 There were also incompatibilities regarding the handling of \Q..\E inside
1648 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1649 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1650
1651 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1652 matches an empty string, and forcibly breaks the loop. There were bugs in
1653 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1654 against aaaa the result was just "a" rather than "aaaa", for example. Two
1655 separate and independent bugs (that affected different cases) have been
1656 fixed.
1657
1658 39. Refactored the code to abolish the use of different opcodes for small
1659 capturing bracket numbers. This is a tidy that I avoided doing when I
1660 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1661 The new approach is not only tidier, it makes it possible to reduce the
1662 memory needed to fix the previous bug (38).
1663
1664 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1665 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1666 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1667 mode.
1668
1669 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1670 report.
1671
1672 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1673 copy construction and assignment in the C++ wrapper.
1674
1675 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1676 couple of bugs that could have given wrong results in the "--newline=crlf"
1677 case.
1678
1679 44. Added a number of casts and did some reorganization of signed/unsigned int
1680 variables following suggestions from Dair Grant. Also renamed the variable
1681 "this" as "item" because it is a C++ keyword.
1682
1683 45. Arranged for dftables to add
1684
1685 #include "pcre_internal.h"
1686
1687 to pcre_chartables.c because without it, gcc 4.x may remove the array
1688 definition from the final binary if PCRE is built into a static library and
1689 dead code stripping is activated.
1690
1691 46. For an unanchored pattern, if a match attempt fails at the start of a
1692 newline sequence, and the newline setting is CRLF or ANY, and the next two
1693 characters are CRLF, advance by two characters instead of one.
1694
1695
1696 Version 6.7 04-Jul-06
1697 ---------------------
1698
1699 1. In order to handle tests when input lines are enormously long, pcretest has
1700 been re-factored so that it automatically extends its buffers when
1701 necessary. The code is crude, but this _is_ just a test program. The
1702 default size has been increased from 32K to 50K.
1703
1704 2. The code in pcre_study() was using the value of the re argument before
1705 testing it for NULL. (Of course, in any sensible call of the function, it
1706 won't be NULL.)
1707
1708 3. The memmove() emulation function in pcre_internal.h, which is used on
1709 systems that lack both memmove() and bcopy() - that is, hardly ever -
1710 was missing a "static" storage class specifier.
1711
1712 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1713 containing an extended class (one that cannot be represented by a bitmap
1714 because it contains high-valued characters or Unicode property items, e.g.
1715 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1716 pattern, but PCRE should not loop if you do not (it no longer does).
1717 [Detail: two cases were found: (a) a repeated subpattern containing an
1718 extended class; (b) a recursive reference to a subpattern that followed a
1719 previous extended class. It wasn't skipping over the extended class
1720 correctly when UTF-8 mode was not set.]
1721
1722 5. A negated single-character class was not being recognized as fixed-length
1723 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1724 compile error "lookbehind assertion is not fixed length".
1725
1726 6. The RunPerlTest auxiliary script was showing an unexpected difference
1727 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1728 write a Perl script that can interpret lines of an input file either as
1729 byte characters or as UTF-8, which is what "perltest" was being required to
1730 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1731 can't do is switch easily at run time between having the "use utf8;" pragma
1732 or not. In the end, I fudged it by using the RunPerlTest script to insert
1733 "use utf8;" explicitly for the UTF-8 tests.
1734
1735 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1736 the end of the subject string, contrary to the documentation and to what
1737 Perl does. This was true of both matching functions. Now it matches only at
1738 the start of the subject and immediately after *internal* newlines.
1739
1740 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1741 a pointer to an int instead of a pointer to an unsigned long int. This
1742 caused problems on 64-bit systems.
1743
1744 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1745 instance of the 'standard' template library not being so standard".
1746
1747 10. There was no check on the number of named subpatterns nor the maximum
1748 length of a subpattern name. The product of these values is used to compute
1749 the size of the memory block for a compiled pattern. By supplying a very
1750 long subpattern name and a large number of named subpatterns, the size
1751 computation could be caused to overflow. This is now prevented by limiting
1752 the length of names to 32 characters, and the number of named subpatterns
1753 to 10,000.
1754
1755 11. Subpatterns that are repeated with specific counts have to be replicated in
1756 the compiled pattern. The size of memory for this was computed from the
1757 length of the subpattern and the repeat count. The latter is limited to
1758 65535, but there was no limit on the former, meaning that integer overflow
1759 could in principle occur. The compiled length of a repeated subpattern is
1760 now limited to 30,000 bytes in order to prevent this.
1761
1762 12. Added the optional facility to have named substrings with the same name.
1763
1764 13. Added the ability to use a named substring as a condition, using the
1765 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1766 are numbers (not recommended). Forward references are permitted.
1767
1768 14. Added forward references in named backreferences (if you see what I mean).
1769
1770 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1771 pattern could run off the end of the subject. For example, the pattern
1772 "(?s)(.{1,5})"8 did this with the subject "ab".
1773
1774 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1775 PCRE_CASELESS was set when matching characters that were quantified with ?
1776 or *.
1777
1778 17. A character class other than a single negated character that had a minimum
1779 but no maximum quantifier - for example [ab]{6,} - was not handled
1780 correctly by pce_dfa_exec(). It would match only one character.
1781
1782 18. A valid (though odd) pattern that looked like a POSIX character
1783 class but used an invalid character after [ (for example [[,abc,]]) caused
1784 pcre_compile() to give the error "Failed: internal error: code overflow" or
1785 in some cases to crash with a glibc free() error. This could even happen if
1786 the pattern terminated after [[ but there just happened to be a sequence of
1787 letters, a binary zero, and a closing ] in the memory that followed.
1788
1789 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1790 over the years. Originally (before any Unicode support), just the bottom 8
1791 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1792 output from "man perlunicode" includes this:
1793
1794 The regular expression compiler produces polymorphic opcodes. That
1795 is, the pattern adapts to the data and automatically switches to
1796 the Unicode character scheme when presented with Unicode data--or
1797 instead uses a traditional byte scheme when presented with byte
1798 data.
1799
1800 Sadly, a wide octal escape does not cause a switch, and in a string with
1801 no other multibyte characters, these octal escapes are treated as before.
1802 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1803 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1804 Unicode string.
1805
1806 I have not perpetrated such confusion in PCRE. Up till now, it took just
1807 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1808 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1809 translate to the appropriate multibyte character.
1810
1811 29. Applied some refactoring to reduce the number of warnings from Microsoft
1812 and Borland compilers. This has included removing the fudge introduced
1813 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1814 a warning about an unused variable.
1815
1816 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1817 characters since release 4.0, because Perl (from release 5.004) does not.
1818 [Or at least, is documented not to: some releases seem to be in conflict
1819 with the documentation.] However, when a pattern was studied with
1820 pcre_study() and all its branches started with \s, PCRE still included VT
1821 as a possible starting character. Of course, this did no harm; it just
1822 caused an unnecessary match attempt.
1823
1824 22. Removed a now-redundant internal flag bit that recorded the fact that case
1825 dependency changed within the pattern. This was once needed for "required
1826 byte" processing, but is no longer used. This recovers a now-scarce options
1827 bit. Also moved the least significant internal flag bit to the most-
1828 significant bit of the word, which was not previously used (hangover from
1829 the days when it was an int rather than a uint) to free up another bit for
1830 the future.
1831
1832 23. Added support for CRLF line endings as well as CR and LF. As well as the
1833 default being selectable at build time, it can now be changed at runtime
1834 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1835 specify that it is scanning data with non-default line endings.
1836
1837 24. Changed the definition of CXXLINK to make it agree with the definition of
1838 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1839
1840 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1841 recursions. This makes a big different to stack usage for some patterns.
1842
1843 26. If a subpattern containing a named recursion or subroutine reference such
1844 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1845 the space required for the compiled pattern went wrong and gave too small a
1846 value. Depending on the environment, this could lead to "Failed: internal
1847 error: code overflow at offset 49" or "glibc detected double free or
1848 corruption" errors.
1849
1850 27. Applied patches from Google (a) to support the new newline modes and (b) to
1851 advance over multibyte UTF-8 characters in GlobalReplace.
1852
1853 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1854 difference for some implementation of PCRE in some Windows version.
1855
1856 29. Added some extra testing facilities to pcretest:
1857
1858 \q<number> in a data line sets the "match limit" value
1859 \Q<number> in a data line sets the "match recursion limt" value
1860 -S <number> sets the stack size, where <number> is in megabytes
1861
1862 The -S option isn't available for Windows.
1863
1864
1865 Version 6.6 06-Feb-06
1866 ---------------------
1867
1868 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1869 in pcreposix.h. I have copied the definition from pcre.h.
1870
1871 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1872 because pcre.h is no longer a built file.
1873
1874 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1875 not normally included in the compiled code.
1876
1877
1878 Version 6.5 01-Feb-06
1879 ---------------------
1880
1881 1. When using the partial match feature with pcre_dfa_exec(), it was not
1882 anchoring the second and subsequent partial matches at the new starting
1883 point. This could lead to incorrect results. For example, with the pattern
1884 /1234/, partially matching against "123" and then "a4" gave a match.
1885
1886 2. Changes to pcregrep:
1887
1888 (a) All non-match returns from pcre_exec() were being treated as failures
1889 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1890 error message is output. Some extra information is given for the
1891 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1892 probably the only errors that are likely to be caused by users (by
1893 specifying a regex that has nested indefinite repeats, for instance).
1894 If there are more than 20 of these errors, pcregrep is abandoned.
1895
1896 (b) A binary zero was treated as data while matching, but terminated the
1897 output line if it was written out. This has been fixed: binary zeroes
1898 are now no different to any other data bytes.
1899
1900 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1901 used to set a locale for matching. The --locale=xxxx long option has
1902 been added (no short equivalent) to specify a locale explicitly on the
1903 pcregrep command, overriding the environment variables.
1904
1905 (d) When -B was used with -n, some line numbers in the output were one less
1906 than they should have been.
1907
1908 (e) Added the -o (--only-matching) option.
1909
1910 (f) If -A or -C was used with -c (count only), some lines of context were
1911 accidentally printed for the final match.
1912
1913 (g) Added the -H (--with-filename) option.
1914
1915 (h) The combination of options -rh failed to suppress file names for files
1916 that were found from directory arguments.
1917
1918 (i) Added the -D (--devices) and -d (--directories) options.
1919
1920 (j) Added the -F (--fixed-strings) option.
1921
1922 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1923
1924 (l) Added the --colo(u)r option.
1925
1926 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1927 is not present by default.
1928
1929 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1930 items such as (?R) or (?1), when the recursion could match a number of
1931 alternatives. If it matched one of the alternatives, but subsequently,
1932 outside the recursion, there was a failure, the code tried to back up into
1933 the recursion. However, because of the way PCRE is implemented, this is not
1934 possible, and the result was an incorrect result from the match.
1935
1936 In order to prevent this happening, the specification of recursion has
1937 been changed so that all such subpatterns are automatically treated as
1938 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1939
1940 4. I had overlooked the fact that, in some locales, there are characters for
1941 which isalpha() is true but neither isupper() nor islower() are true. In
1942 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1943 and ordfeminine) are like this. This affected the treatment of \w and \W
1944 when they appeared in character classes, but not when they appeared outside
1945 a character class. The bit map for "word" characters is now created
1946 separately from the results of isalnum() instead of just taking it from the
1947 upper, lower, and digit maps. (Plus the underscore character, of course.)
1948
1949 5. The above bug also affected the handling of POSIX character classes such as
1950 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1951 permanent tables. Instead, the bit maps for such a class were previously
1952 created as the appropriate unions of the upper, lower, and digit bitmaps.
1953 Now they are created by subtraction from the [[:word:]] class, which has
1954 its own bitmap.
1955
1956 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1957 It is created by subtracting the vertical space characters (\x09, \x0a,
1958 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1959 subtraction was done in the overall bitmap for a character class, meaning
1960 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1961 be recognized. This bug has been fixed.
1962
1963 7. Patches from the folks at Google:
1964
1965 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1966 real life, but is still worth protecting against".
1967
1968 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1969 regular expressions".
1970
1971 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1972 have it.
1973
1974 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1975 "configure" and the latter not, in order to fix a problem somebody had
1976 with compiling the Arg class on HP-UX.
1977
1978 (e) Improve the error-handling of the C++ wrapper a little bit.
1979
1980 (f) New tests for checking recursion limiting.
1981
1982 8. The pcre_memmove() function, which is used only if the environment does not
1983 have a standard memmove() function (and is therefore rarely compiled),
1984 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1985 returning a result (though PCRE never actually uses the result).
1986
1987 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1988 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1989 returned instead of calling malloc() with an overflowing number that would
1990 most likely cause subsequent chaos.
1991
1992 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1993
1994 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1995 with this option is matched, the nmatch and pmatch options of regexec() are
1996 ignored.
1997
1998 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1999 provided in case anyone wants to the the POSIX interface with UTF-8
2000 strings.
2001
2002 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2003 C++ linking (needed for some HP-UX environments).
2004
2005 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2006 (unused parameter) and in the pcre_printint() function (omitted "default"
2007 switch label when the default is to do nothing).
2008
2009 15. Added some code to make it possible, when PCRE is compiled as a C++
2010 library, to replace subject pointers for pcre_exec() with a smart pointer
2011 class, thus making it possible to process discontinuous strings.
2012
2013 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2014 much the same function. They were added by different people who were trying
2015 to make PCRE easy to compile on non-Unix systems. It has been suggested
2016 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2017 for compiling on Windows systems. I have therefore replaced it with
2018 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2019 defaults to "extern" for C or "extern C" for C++, which works fine on
2020 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2021 SCOPE with something explicit in config.h. In addition:
2022
2023 (a) pcreposix.h still had just "extern" instead of either of these macros;
2024 I have replaced it with PCRE_DATA_SCOPE.
2025
2026 (b) Functions such as _pcre_xclass(), which are internal to the library,
2027 but external in the C sense, all had PCRE_EXPORT in their definitions.
2028 This is apparently wrong for the Windows case, so I have removed it.
2029 (It makes no difference on Unix-like systems.)
2030
2031 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2032 of recursive calls to match(). This is different to MATCH_LIMIT because
2033 that limits the total number of calls to match(), not all of which increase
2034 the depth of recursion. Limiting the recursion depth limits the amount of
2035 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2036 when PCRE is compiled, and changed at run time. A patch from Google adds
2037 this functionality to the C++ interface.
2038
2039 18. Changes to the handling of Unicode character properties:
2040
2041 (a) Updated the table to Unicode 4.1.0.
2042
2043 (b) Recognize characters that are not in the table as "Cn" (undefined).
2044
2045 (c) I revised the way the table is implemented to a much improved format
2046 which includes recognition of ranges. It now supports the ranges that
2047 are defined in UnicodeData.txt, and it also amalgamates other
2048 characters into ranges. This has reduced the number of entries in the
2049 table from around 16,000 to around 3,000, thus reducing its size
2050 considerably. I realized I did not need to use a tree structure after
2051 all - a binary chop search is just as efficient. Having reduced the
2052 number of entries, I extended their size from 6 bytes to 8 bytes to
2053 allow for more data.
2054
2055 (d) Added support for Unicode script names via properties such as \p{Han}.
2056
2057 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2058 matching that character.
2059
2060 20. When matching a repeated Unicode property with a minimum greater than zero,
2061 (for example \pL{2,}), PCRE could look past the end of the subject if it
2062 reached it while seeking the minimum number of characters. This could
2063 happen only if some of the characters were more than one byte long, because
2064 there is a check for at least the minimum number of bytes.
2065
2066 21. Refactored the implementation of \p and \P so as to be more general, to
2067 allow for more different types of property in future. This has changed the
2068 compiled form incompatibly. Anybody with saved compiled patterns that use
2069 \p or \P will have to recompile them.
2070
2071 22. Added "Any" and "L&" to the supported property types.
2072
2073 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2074 but give a compile time error if the value is greater than 0xff.
2075
2076 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2077 accidentally not being installed or uninstalled.
2078
2079 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2080 made were to insert the current release number. This seemed silly, because
2081 it made things harder for people building PCRE on systems that don't run
2082 "configure". I have turned pcre.h into a distributed file, no longer built
2083 by "configure", with the version identification directly included. There is
2084 no longer a pcre.h.in file.
2085
2086 However, this change necessitated a change to the pcre-config script as
2087 well. It is built from pcre-config.in, and one of the substitutions was the
2088 release number. I have updated configure.ac so that ./configure now finds
2089 the release number by grepping pcre.h.
2090
2091 26. Added the ability to run the tests under valgrind.
2092
2093
2094 Version 6.4 05-Sep-05
2095 ---------------------
2096
2097 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2098 "--" to be printed when multiple files were scanned, even when none of the
2099 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2100 consider it to be a bug, and have restored the previous behaviour.
2101
2102 2. A couple of code tidies to get rid of compiler warnings.
2103
2104 3. The pcretest program used to cheat by referring to symbols in the library
2105 whose names begin with _pcre_. These are internal symbols that are not
2106 really supposed to be visible externally, and in some environments it is
2107 possible to suppress them. The cheating is now confined to including
2108 certain files from the library's source, which is a bit cleaner.
2109
2110 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2111 file's purpose clearer.
2112
2113 5. Reorganized pcre_ucp_findchar().
2114
2115
2116 Version 6.3 15-Aug-05
2117 ---------------------
2118
2119 1. The file libpcre.pc.in did not have general read permission in the tarball.
2120
2121 2. There were some problems when building without C++ support:
2122
2123 (a) If C++ support was not built, "make install" and "make test" still
2124 tried to test it.
2125
2126 (b) There were problems when the value of CXX was explicitly set. Some
2127 changes have been made to try to fix these, and ...
2128
2129 (c) --disable-cpp can now be used to explicitly disable C++ support.
2130
2131 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2132 backslash in a target when C++ was disabled. This confuses some
2133 versions of "make", apparently. Using an intermediate variable solves
2134 this. (Same for CPP_LOBJ.)
2135
2136 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2137 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2138 necessary on certain architectures.
2139
2140 4. Added a setting of -export-symbols-regex to the link command to remove
2141 those symbols that are exported in the C sense, but actually are local
2142 within the library, and not documented. Their names all begin with
2143 "_pcre_". This is not a perfect job, because (a) we have to except some
2144 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2145 available (and never for static libraries). I have made a note to try to
2146 find a way round (a) in the future.
2147
2148
2149 Version 6.2 01-Aug-05
2150 ---------------------
2151
2152 1. There was no test for integer overflow of quantifier values. A construction
2153 such as {1111111111111111} would give undefined results. What is worse, if
2154 a minimum quantifier for a parenthesized subpattern overflowed and became
2155 negative, the calculation of the memory size went wrong. This could have
2156 led to memory overwriting.
2157
2158 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2159
2160 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2161 operating environments where this matters.
2162
2163 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2164 PCRE options from within the C++ wrapper.
2165
2166 5. Named capturing subpatterns were not being correctly counted when a pattern
2167 was compiled. This caused two problems: (a) If there were more than 100
2168 such subpatterns, the calculation of the memory needed for the whole
2169 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2170 back references of the form \12, where the number was greater than 9, were
2171 not recognized as back references, even though there were sufficient
2172 previous subpatterns.
2173
2174 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2175 versions of gcc, e.g. 2.95.4.
2176
2177
2178 Version 6.1 21-Jun-05
2179 ---------------------
2180
2181 1. There was one reference to the variable "posix" in pcretest.c that was not
2182 surrounded by "#if !defined NOPOSIX".
2183
2184 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2185 the cross-check on the old pcre_info() function, for the benefit of the
2186 cut-down version of PCRE that is currently imported into Exim.
2187
2188 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2189 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2190 patterns that start (?i) but I don't think that matters. The use of (?i) is
2191 just an example; this all applies to the other options as well.
2192
2193 4. Since libtool seems to echo the compile commands it is issuing, the output
2194 from "make" can be reduced a bit by putting "@" in front of each libtool
2195 compile command.
2196
2197 5. Patch from the folks at Google for configure.in to be a bit more thorough
2198 in checking for a suitable C++ installation before trying to compile the
2199 C++ stuff. This should fix a reported problem when a compiler was present,
2200 but no suitable headers.
2201
2202 6. The man pages all had just "PCRE" as their title. I have changed them to
2203 be the relevant file name. I have also arranged that these names are
2204 retained in the file doc/pcre.txt, which is a concatenation in text format
2205 of all the man pages except the little individual ones for each function.
2206
2207 7. The NON-UNIX-USE file had not been updated for the different set of source
2208 files that come with release 6. I also added a few comments about the C++
2209 wrapper.
2210
2211
2212 Version 6.0 07-Jun-05
2213 ---------------------
2214
2215 1. Some minor internal re-organization to help with my DFA experiments.
2216
2217 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2218 didn't matter for the library itself when fully configured, but did matter
2219 when compiling without UCP support, or within Exim, where the ucp files are
2220 not imported.
2221
2222 3. Refactoring of the library code to split up the various functions into
2223 different source modules. The addition of the new DFA matching code (see
2224 below) to a single monolithic source would have made it really too
2225 unwieldy, quite apart from causing all the code to be include in a
2226 statically linked application, when only some functions are used. This is
2227 relevant even without the DFA addition now that patterns can be compiled in
2228 one application and matched in another.
2229
2230 The downside of splitting up is that there have to be some external
2231 functions and data tables that are used internally in different modules of
2232 the library but which are not part of the API. These have all had their
2233 names changed to start with "_pcre_" so that they are unlikely to clash
2234 with other external names.
2235
2236 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2237 a different (DFA) algorithm. Although it is slower than the original
2238 function, it does have some advantages for certain types of matching
2239 problem.
2240
2241 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2242 including restarting after a partial match.
2243
2244 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2245 defined when compiling for Windows was sent to me. I have put it into the
2246 code, though I have no means of testing or verifying it.
2247
2248 7. Added the pcre_refcount() auxiliary function.
2249
2250 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2251 match before or at the first newline in the subject string. In pcretest,
2252 the /f option on a pattern can be used to set this.
2253
2254 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2255 would behave wrongly. This has been present in PCRE since release 4.0.
2256
2257 10. A number of changes to the pcregrep command:
2258
2259 (a) Refactored how -x works; insert ^(...)$ instead of setting
2260 PCRE_ANCHORED and checking the length, in preparation for adding
2261 something similar for -w.
2262
2263 (b) Added the -w (match as a word) option.
2264
2265 (c) Refactored the way lines are read and buffered so as to have more
2266 than one at a time available.
2267
2268 (d) Implemented a pcregrep test script.
2269
2270 (e) Added the -M (multiline match) option. This allows patterns to match
2271 over several lines of the subject. The buffering ensures that at least
2272 8K, or the rest of the document (whichever is the shorter) is available
2273 for matching (and similarly the previous 8K for lookbehind assertions).
2274
2275 (f) Changed the --help output so that it now says
2276
2277 -w, --word-regex(p)
2278
2279 instead of two lines, one with "regex" and the other with "regexp"
2280 because that confused at least one person since the short forms are the
2281 same. (This required a bit of code, as the output is generated
2282 automatically from a table. It wasn't just a text change.)
2283
2284 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2285 option but starts with a hyphen. Could be a pattern or a path name
2286 starting with a hyphen, for instance.
2287
2288 (h) "-" can be given as a file name to represent stdin.
2289
2290 (i) When file names are being printed, "(standard input)" is used for
2291 the standard input, for compatibility with GNU grep. Previously
2292 "<stdin>" was used.
2293
2294 (j) The option --label=xxx can be used to supply a name to be used for
2295 stdin when file names are being printed. There is no short form.
2296
2297 (k) Re-factored the options decoding logic because we are going to add
2298 two more options that take data. Such options can now be given in four
2299 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2300
2301 (l) Added the -A, -B, and -C options for requesting that lines of context
2302 around matches be printed.
2303
2304 (m) Added the -L option to print the names of files that do not contain
2305 any matching lines, that is, the complement of -l.
2306
2307 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2308 continue to scan other files.
2309
2310 (o) The -s option was incorrectly implemented. For compatibility with other
2311 greps, it now suppresses the error message for a non-existent or non-
2312 accessible file (but not the return code). There is a new option called
2313 -q that suppresses the output of matching lines, which was what -s was
2314 previously doing.
2315
2316 (p) Added --include and --exclude options to specify files for inclusion
2317 and exclusion when recursing.
2318
2319 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2320 Hopefully, it now does.
2321
2322 12. Missing cast in pcre_study().
2323
2324 13. Added an "uninstall" target to the makefile.
2325
2326 14. Replaced "extern" in the function prototypes in Makefile.in with
2327 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2328 world, but is set differently for Windows.
2329
2330 15. Added a second compiling function called pcre_compile2(). The only
2331 difference is that it has an extra argument, which is a pointer to an
2332 integer error code. When there is a compile-time failure, this is set
2333 non-zero, in addition to the error test pointer being set to point to an
2334 error message. The new argument may be NULL if no error number is required
2335 (but then you may as well call pcre_compile(), which is now just a
2336 wrapper). This facility is provided because some applications need a
2337 numeric error indication, but it has also enabled me to tidy up the way
2338 compile-time errors are handled in the POSIX wrapper.
2339
2340 16. Added VPATH=.libs to the makefile; this should help when building with one
2341 prefix path and installing with another. (Or so I'm told by someone who
2342 knows more about this stuff than I do.)
2343
2344 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2345 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2346 match everything, including newlines. This is not POSIX-compatible, but
2347 somebody wanted the feature. From pcretest it can be activated by using
2348 both the P and the s flags.
2349
2350 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2351
2352 19. libpcre.pc was being incorrectly installed as executable.
2353
2354 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2355 it now also looks for '\r' so that it will work unmodified on Windows.
2356
2357 21. Added Google's contributed C++ wrapper to the distribution.
2358
2359 22. Added some untidy missing memory free() calls in pcretest, to keep
2360 Electric Fence happy when testing.
2361
2362
2363
2364 Version 5.0 13-Sep-04
2365 ---------------------
2366
2367 1. Internal change: literal characters are no longer packed up into items
2368 containing multiple characters in a single byte-string. Each character
2369 is now matched using a separate opcode. However, there may be more than one
2370 byte in the character in UTF-8 mode.
2371
2372 2. The pcre_callout_block structure has two new fields: pattern_position and
2373 next_item_length. These contain the offset in the pattern to the next match
2374 item, and its length, respectively.
2375
2376 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2377 insertion of callouts before each pattern item. Added the /C option to
2378 pcretest to make use of this.
2379
2380 4. On the advice of a Windows user, the lines
2381
2382 #if defined(_WIN32) || defined(WIN32)
2383 _setmode( _fileno( stdout ), 0x8000 );
2384 #endif /* defined(_WIN32) || defined(WIN32) */
2385
2386 have been added to the source of pcretest. This apparently does useful
2387 magic in relation to line terminators.
2388
2389 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2390 for the benefit of those environments where the "b" makes a difference.
2391
2392 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2393 to know about it. I have put a hack into configure.in that adds in code
2394 to set GCC=yes if CC=icc. This seems to end up at a point in the
2395 generated configure script that is early enough to affect the setting of
2396 compiler options, which is what is needed, but I have no means of testing
2397 whether it really works. (The user who reported this had patched the
2398 generated configure script, which of course I cannot do.)
2399
2400 LATER: After change 22 below (new libtool files), the configure script
2401 seems to know about icc (and also ecc). Therefore, I have commented out
2402 this hack in configure.in.
2403
2404 7. Added support for pkg-config (2 patches were sent in).
2405
2406 8. Negated POSIX character classes that used a combination of internal tables
2407 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2408 [[:^ascii]]. Typically, they would match almost any characters. The other
2409 POSIX classes were not broken in this way.
2410
2411 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2412 to find the match, as PCRE was deluded into thinking that the match had to
2413 start at the start point or following a newline. The same bug applied to
2414 patterns with negative forward assertions or any backward assertions
2415 preceding ".*" at the start, unless the pattern required a fixed first
2416 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2417
2418 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2419 starting at the last subject character, bytes beyond the end of the subject
2420 string were read.
2421
2422 11. Renamed the variable "class" as "classbits" to make life easier for C++
2423 users. (Previously there was a macro definition, but it apparently wasn't
2424 enough.)
2425
2426 12. Added the new field "tables" to the extra data so that tables can be passed
2427 in at exec time, or the internal tables can be re-selected. This allows
2428 a compiled regex to be saved and re-used at a later time by a different
2429 program that might have everything at different addresses.
2430
2431 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2432 -R library as well as a -L library.
2433
2434 14. The debugging options of pcretest (-d on the command line or D on a
2435 pattern) showed incorrect output for anything following an extended class
2436 that contained multibyte characters and which was followed by a quantifier.
2437
2438 15. Added optional support for general category Unicode character properties
2439 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2440 support. It adds about 90K to the size of the library. The meanings of the
2441 inbuilt class escapes such as \d and \s have NOT been changed.
2442
2443 16. Updated pcredemo.c to include calls to free() to release the memory for the
2444 compiled pattern.
2445
2446 17. The generated file chartables.c was being created in the source directory
2447 instead of in the building directory. This caused the build to fail if the
2448 source directory was different from the building directory, and was
2449 read-only.
2450
2451 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2452 file. No doubt somebody will tell me if they don't make sense... Also added
2453 Dan Mooney's comments about building on OpenVMS.
2454
2455 19. Added support for partial matching via the PCRE_PARTIAL option for
2456 pcre_exec() and the \P data escape in pcretest.
2457
2458 20. Extended pcretest with 3 new pattern features:
2459
2460 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2461 write the compiled pattern to the file whose name is "rest-of-line".
2462 This is a straight binary dump of the data, with the saved pointer to
2463 the character tables forced to be NULL. The study data, if any, is
2464 written too. After writing, pcretest reads a new pattern.
2465
2466 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2467 compiled pattern from the given file. There must not be any
2468 occurrences of "<" in the file name (pretty unlikely); if there are,
2469 pcretest will instead treat the initial "<" as a pattern delimiter.
2470 After reading in the pattern, pcretest goes on to read data lines as
2471 usual.
2472
2473 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2474 and 16-bit fields in a compiled pattern, to simulate a pattern that
2475 was compiled on a host of opposite endianness.
2476
2477 21. The pcre-exec() function can now cope with patterns that were compiled on
2478 hosts of opposite endianness, with this restriction:
2479
2480 As for any compiled expression that is saved and used later, the tables
2481 pointer field cannot be preserved; the extra_data field in the arguments
2482 to pcre_exec() should be used to pass in a tables address if a value
2483 other than the default internal tables were used at compile time.
2484
2485 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2486 now diagnosed as an error. Previously, most of the time, a negative number
2487 would have been treated as zero, but if in addition "ovector" was passed as
2488 NULL, a crash could occur.
2489
2490 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2491 new versions from the libtool 1.5 distribution (the last one is a copy of
2492 a file called libtool.m4). This seems to have fixed the need to patch
2493 "configure" to support Darwin 1.3 (which I used to do). However, I still
2494 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2495 workstation).
2496
2497 24. Changed the PCRE licence to be the more standard "BSD" licence.
2498
2499
2500 Version 4.5 01-Dec-03
2501 ---------------------
2502
2503 1. There has been some re-arrangement of the code for the match() function so
2504 that it can be compiled in a version that does not call itself recursively.
2505 Instead, it keeps those local variables that need separate instances for
2506 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2507 needs to "recurse". Keeping track of where control must go is done by means
2508 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2509 hide most of the details from the main code, and operates only if
2510 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2511 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2512 operating.
2513
2514 To make it easier for callers to provide specially tailored get/free
2515 functions for this usage, two new functions, pcre_stack_malloc, and
2516 pcre_stack_free, are used. They are always called in strict stacking order,
2517 and the size of block requested is always the same.
2518
2519 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2520 PCRE has been compiled to use the stack or the heap for recursion. The
2521 -C option of pcretest uses this to show which version is compiled.
2522
2523 A new data escape \S, is added to pcretest; it causes the amounts of store
2524 obtained and freed by both kinds of malloc/free at match time to be added
2525 to the output.
2526
2527 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2528 what's available on my current Linux desktop machine.
2529
2530 3. When matching a UTF-8 string, the test for a valid string at the start has
2531 been extended. If start_offset is not zero, PCRE now checks that it points
2532 to a byte that is the start of a UTF-8 character. If not, it returns
2533 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2534 this is necessary because there may be backward assertions in the pattern.
2535 When matching the same subject several times, it may save resources to use
2536 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2537
2538 4. The code for checking the validity of UTF-8 strings has been tightened so
2539 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2540 containing "overlong sequences".
2541
2542 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2543 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2544 should have been "&", but it just so happened that all the cases this let
2545 through by mistake were picked up later in the function.
2546
2547 6. I had used a variable called "isblank" - this is a C99 function, causing
2548 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2549
2550 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2551 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2552 so the version is shown; (c) stop "make test" echoing "./RunTest".
2553
2554 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2555
2556 9. The prototype for memmove() for systems that don't have it was using
2557 size_t, but the inclusion of the header that defines size_t was later. I've
2558 moved the #includes for the C headers earlier to avoid this.
2559
2560 10. Added some adjustments to the code to make it easier to compiler on certain
2561 special systems:
2562
2563 (a) Some "const" qualifiers were missing.
2564 (b) Added the macro EXPORT before all exported functions; by default this
2565 is defined to be empty.
2566 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2567 that it reads its output file name as an argument instead of writing
2568 to the standard output and assuming this can be redirected.
2569
2570 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2571 class containing characters with values greater than 255, PCRE compilation
2572 went into a loop.
2573
2574 12. A recursive reference to a subpattern that was within another subpattern
2575 that had a minimum quantifier of zero caused PCRE to crash. For example,
2576 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2577 recursion. If the recursively-called subpattern itself had a zero repeat,
2578 that was OK.
2579
2580 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2581 buffer into which it was copied (for escape processing) was still set at
2582 1024, so long lines caused crashes.
2583
2584 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2585 "internal error: code overflow...". This applied to any character class
2586 that was followed by a possessive quantifier.
2587
2588 15. Modified the Makefile to add libpcre.la as a prerequisite for
2589 libpcreposix.la because I was told this is needed for a parallel build to
2590 work.
2591
2592 16. If a pattern that contained .* following optional items at the start was
2593 studied, the wrong optimizing data was generated, leading to matching
2594 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2595 matching string must start with a or b or c. The correct conclusion for
2596 this pattern is that a match can start with any character.
2597
2598
2599 Version 4.4 13-Aug-03
2600 ---------------------
2601
2602 1. In UTF-8 mode, a character class containing characters with values between
2603 127 and 255 was not handled correctly if the compiled pattern was studied.
2604 In fixing this, I have also improved the studying algorithm for such
2605 classes (slightly).
2606
2607 2. Three internal functions had redundant arguments passed to them. Removal
2608 might give a very teeny performance improvement.
2609
2610 3. Documentation bug: the value of the capture_top field in a callout is *one
2611 more than* the number of the hightest numbered captured substring.
2612
2613 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2614 in incorrectly linking with a previously installed version. They now link
2615 explicitly with libpcre.la.
2616
2617 5. configure.in no longer needs to recognize Cygwin specially.
2618
2619 6. A problem in pcre.in for Windows platforms is fixed.
2620
2621 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2622 pcretest, it used to include the size of the study block as part of its
2623 output. Unfortunately, the structure contains a field that has a different
2624 size on different hardware architectures. This meant that the tests that
2625 showed this size failed. As the block is currently always of a fixed size,
2626 this information isn't actually particularly useful in pcretest output, so
2627 I have just removed it.
2628
2629 8. Three pre-processor statements accidentally did not start in column 1.
2630 Sadly, there are *still* compilers around that complain, even though
2631 standard C has not required this for well over a decade. Sigh.
2632
2633 9. In pcretest, the code for checking callouts passed small integers in the
2634 callout_data field, which is a void * field. However, some picky compilers
2635 complained about the casts involved for this on 64-bit systems. Now
2636 pcretest passes the address of the small integer instead, which should get
2637 rid of the warnings.
2638
2639 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2640 both compile and run time, and gives an error if an invalid UTF-8 sequence
2641 is found. There is a option for disabling this check in cases where the
2642 string is known to be correct and/or the maximum performance is wanted.
2643
2644 11. In response to a bug report, I changed one line in Makefile.in from
2645
2646 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2647 to
2648 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2649
2650 to look similar to other lines, but I have no way of telling whether this
2651 is the right thing to do, as I do not use Windows. No doubt I'll get told
2652 if it's wrong...
2653
2654
2655 Version 4.3 21-May-03
2656 ---------------------
2657
2658 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2659 Makefile.
2660
2661 2. Some refactoring to improve the quality of the code:
2662
2663 (i) The utf8_table... variables are now declared "const".
2664
2665 (ii) The code for \cx, which used the "case flipping" table to upper case
2666 lower case letters, now just substracts 32. This is ASCII-specific,
2667 but the whole concept of \cx is ASCII-specific, so it seems
2668 reasonable.
2669
2670 (iii) PCRE was using its character types table to recognize decimal and
2671 hexadecimal digits in the pattern. This is silly, because it handles
2672 only 0-9, a-f, and A-F, but the character types table is locale-
2673 specific, which means strange things might happen. A private
2674 table is now used for this - though it costs 256 bytes, a table is
2675 much faster than multiple explicit tests. Of course, the standard
2676 character types table is still used for matching digits in subject
2677 strings against \d.
2678
2679 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2680 ending in _t are). So I've renamed it as ESC_tee.
2681
2682 3. The first argument for regexec() in the POSIX wrapper should have been
2683 defined as "const".
2684
2685 4. Changed pcretest to use malloc() for its buffers so that they can be
2686 Electric Fenced for debugging.
2687
2688 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2689 to read one or more bytes before the start of the subject string. Often this
2690 had no effect on PCRE's behaviour, but in some circumstances it could
2691 provoke a segmentation fault.
2692
2693 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2694 to try to read one or more bytes before the start of the subject string.
2695
2696 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2697 UTF-8 support could misbehave in various ways if the subject string
2698 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2699 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2700 back over UTF-8 characters.)
2701
2702
2703 Version 4.2 14-Apr-03
2704 ---------------------
2705
2706 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2707
2708 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2709 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2710 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2711 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2712 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2713 and BUILD_EXEEXT
2714 Note: automatic setting of the BUILD variables is not yet working
2715 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2716 compile-time but not at link-time
2717 [LINK]: use for linking executables only
2718 make different versions for Windows and non-Windows
2719 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2720 libraries
2721 [LINK_FOR_BUILD]: new variable
2722 [OBJEXT]: use throughout
2723 [EXEEXT]: use throughout
2724 <winshared>: new target
2725 <wininstall>: new target
2726 <dftables.o>: use native compiler
2727 <dftables>: use native linker
2728 <install>: handle Windows platform correctly
2729 <clean>: ditto
2730 <check>: ditto
2731 copy DLL to top builddir before testing
2732
2733 As part of these changes, -no-undefined was removed again. This was reported
2734 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2735 in any case.
2736
2737 3. Some tidies to get rid of compiler warnings:
2738
2739 . In the match_data structure, match_limit was an unsigned long int, whereas
2740 match_call_count was an int. I've made them both unsigned long ints.
2741
2742 . In pcretest the fact that a const uschar * doesn't automatically cast to
2743 a void * provoked a warning.
2744
2745 . Turning on some more compiler warnings threw up some "shadow" variables
2746 and a few more missing casts.
2747
2748 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2749 option, a class that contained a single character with a value between 128
2750 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2751
2752 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2753 option, a class that contained several characters, but with at least one
2754 whose value was between 128 and 255 caused PCRE to crash.
2755
2756
2757 Version 4.1 12-Mar-03
2758 ---------------------
2759
2760 1. Compiling with gcc -pedantic found a couple of places where casts were
2761 needed, and a string in dftables.c that was longer than standard compilers are
2762 required to support.
2763
2764 2. Compiling with Sun's compiler found a few more places where the code could
2765 be tidied up in order to avoid warnings.
2766
2767 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2768 first of these names is deprecated in the latest Autoconf in favour of the name
2769 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2770 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2771 analogy I have changed it to CFLAGS_FOR_BUILD.
2772
2773 4. Added -no-undefined to the linking command in the Makefile, because this is
2774 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2775 linking step for the pcreposix library.
2776
2777 5. PCRE was failing to diagnose the case of two named groups with the same
2778 name.
2779
2780 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2781 literal character that is needed in the subject for a match, and scans along to
2782 ensure that it is present before embarking on the full matching process. This
2783 saves time in cases of nested unlimited repeats that are never going to match.
2784 Problem: the scan can take a lot of time if the subject is very long (e.g.
2785 megabytes), thus penalizing straightforward matches. It is now done only if the
2786 amount of subject to be scanned is less than 1000 bytes.
2787
2788 7. A lesser problem with the same optimization is that it was recording the
2789 first character of an anchored pattern as "needed", thus provoking a search
2790 right along the subject, even when the first match of the pattern was going to
2791 fail. The "needed" character is now not set for anchored patterns, unless it
2792 follows something in the pattern that is of non-fixed length. Thus, it still
2793 fulfils its original purpose of finding quick non-matches in cases of nested
2794 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2795
2796
2797 Version 4.0 17-Feb-03
2798 ---------------------
2799
2800 1. If a comment in an extended regex that started immediately after a meta-item
2801 extended to the end of string, PCRE compiled incorrect data. This could lead to
2802 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2803
2804 2. Moved to autoconf 2.53 and libtool 1.4.2.
2805
2806 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2807 the special perltest8 script is no longer needed - all the tests can be run
2808 from a single perltest script.
2809
2810 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2811 by \s. It has now been removed in PCRE. This means it isn't recognized as
2812 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2813 class [:space:] *does* include VT, thereby creating a mess.
2814
2815 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2816 space and tab.
2817
2818 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2819 its new features into the main test script, reducing the number of scripts.
2820
2821 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2822 were backward compatible, and made the (?i) apply to the whole pattern, as if
2823 /i were given. Now it behaves more logically, and applies the option setting
2824 only to what follows. PCRE has been changed to follow suit. However, if it
2825 finds options settings right at the start of the pattern, it extracts them into
2826 the global options, as before. Thus, they show up in the info data.
2827
2828 8. Added support for the \Q...\E escape sequence. Characters in between are
2829 treated as literals. This is slightly different from Perl in that $ and @ are
2830 also handled as literals inside the quotes. In Perl, they will cause variable
2831 interpolation. Note the following examples:
2832
2833 Pattern PCRE matches Perl matches
2834
2835 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2836 \Qabc\$xyz\E abc\$xyz abc\$xyz
2837 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2838
2839 For compatibility with Perl, \Q...\E sequences are recognized inside character
2840 classes as well as outside them.
2841
2842 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2843 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2844 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2845 signed/unsigned warnings.
2846
2847 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2848 option for pcretest, so I've replaced it by a simple function that does just
2849 that job.
2850
2851 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2852 "pcregrep -".
2853
2854 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2855 Java package. This provides some syntactic sugar for simple cases of what my
2856 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2857 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2858 item, you can use this simplified notation. Note that only makes sense with
2859 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2860 greediness, whatever the setting of the PCRE_UNGREEDY option.
2861
2862 13. A change of greediness default within a pattern was not taking effect at
2863 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2864 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2865 was abstracted outside.
2866
2867 14. PCRE now supports the \G assertion. It is true when the current matching
2868 position is at the start point of the match. This differs from \A when the
2869 starting offset is non-zero. Used with the /g option of pcretest (or similar
2870 code), it works in the same way as it does for Perl's /g option. If all
2871 alternatives of a regex begin with \G, the expression is anchored to the start
2872 match position, and the "anchored" flag is set in the compiled expression.
2873
2874 15. Some bugs concerning the handling of certain option changes within patterns
2875 have been fixed. These applied to options other than (?ims). For example,
2876 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2877 been the other way round. Some of this was related to change 7 above.
2878
2879 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2880 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2881 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2882 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2883
2884 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2885 mode. Unlike ".", it always matches newline, whatever the setting of
2886 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2887 assertions. Perl allows it, but it doesn't (in general) work because it can't
2888 calculate the length of the lookbehind. At least, that's the case for Perl
2889 5.8.0 - I've been told they are going to document that it doesn't work in
2890 future.
2891
2892 18. Added an error diagnosis for escapes that PCRE does not support: these are
2893 \L, \l, \N, \P, \p, \U, \u, and \X.
2894
2895 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2896 reading past the end of the pattern in cases such as /[abcd/.
2897
2898 20. PCRE was getting more memory than necessary for patterns with classes that
2899 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2900
2901 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2902 compiling PCRE for use with Virtual Pascal.
2903
2904 22. Small fix to the Makefile to make it work properly if the build is done
2905 outside the source tree.
2906
2907 23. Added a new extension: a condition to go with recursion. If a conditional
2908 subpattern starts with (?(R) the "true" branch is used if recursion has
2909 happened, whereas the "false" branch is used only at the top level.
2910
2911 24. When there was a very long string of literal characters (over 255 bytes
2912 without UTF support, over 250 bytes with UTF support), the computation of how
2913 much memory was required could be incorrect, leading to segfaults or other
2914 strange effects.
2915
2916 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2917 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2918 there was a subsequent back reference to those brackets. This meant that, for
2919 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2920 possible to check for precisely this case. All we can do is abandon the
2921 optimization if .* occurs inside capturing brackets when there are any back
2922 references whatsoever. (See below for a better fix that came later.)
2923
2924 26. The handling of the optimization for finding the first character of a
2925 non-anchored pattern, and for finding a character that is required later in the
2926 match were failing in some cases. This didn't break the matching; it just
2927 failed to optimize when it could. The way this is done has been re-implemented.
2928
2929 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2930
2931 28. Added a new feature that provides some of the functionality that Perl
2932 provides with (?{...}). The facility is termed a "callout". The way it is done
2933 in PCRE is for the caller to provide an optional function, by setting
2934 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2935 global variable. By default it is unset, which disables all calling out. To get
2936 the function called, the regex must include (?C) at appropriate points. This
2937 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2938 This provides a means of identifying different callout points. When PCRE
2939 reaches such a point in the regex, if pcre_callout has been set, the external
2940 function is called. It is provided with data in a structure called
2941 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2942 matching continues; if it returns a non-zero value, the match at the current
2943 point fails. However, backtracking will occur if possible. [This was changed
2944 later and other features added - see item 49 below.]
2945
2946 29. pcretest is upgraded to test the callout functionality. It provides a
2947 callout function that displays information. By default, it shows the start of
2948 the match and the current position in the text. There are some new data escapes
2949 to vary what happens:
2950
2951 \C+ in addition, show current contents of captured substrings
2952 \C- do not supply a callout function
2953 \C!n return 1 when callout number n is reached
2954 \C!n!m return 1 when callout number n is reached for the mth time
2955
2956 30. If pcregrep was called with the -l option and just a single file name, it
2957 output "<stdin>" if a match was found, instead of the file name.
2958
2959 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2960 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2961 pcre_exec(). This saves a malloc/free per call. The default value of
2962 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2963 when configuring.
2964
2965 32. The default maximum size of a compiled pattern is 64K. There have been a
2966 few cases of people hitting this limit. The code now uses macros to handle the
2967 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2968 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2969 configuring. Tests 2 and 5 work only with 2-byte links because they output
2970 debugging information about compiled patterns.
2971
2972 33. Internal code re-arrangements:
2973
2974 (a) Moved the debugging function for printing out a compiled regex into
2975 its own source file (printint.c) and used #include to pull it into
2976 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2977 separate copies.
2978
2979 (b) Defined the list of op-code names for debugging as a macro in
2980 internal.h so that it is next to the definition of the opcodes.
2981
2982 (c) Defined a table of op-code lengths for simpler skipping along compiled
2983 code. This is again a macro in internal.h so that it is next to the
2984 definition of the opcodes.
2985
2986 34. Added support for recursive calls to individual subpatterns, along the
2987 lines of Robin Houston's patch (but implemented somewhat differently).
2988
2989 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2990 allow it to read and process whole directories in Win32. This code was
2991 contributed by Lionel Fourquaux; it has not been tested by me.
2992
2993 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2994 used to name a group. Names consist of alphanumerics and underscores, and must
2995 be unique. Back references use the syntax (?P=name) and recursive calls use
2996 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2997 numbers. The function pcre_fullinfo() can be used after compilation to extract
2998 a name/number map. There are three relevant calls:
2999
3000 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3001 PCRE_INFO_NAMECOUNT yields the number of entries
3002 PCRE_INFO_NAMETABLE yields a pointer to the map.
3003
3004 The map is a vector of fixed-size entries. The size of each entry depends on
3005 the length of the longest name used. The first two bytes of each entry are the
3006 group number, most significant byte first. There follows the corresponding
3007 name, zero terminated. The names are in alphabetical order.
3008
3009 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3010 case instead of 255. Making it the same both with and without UTF-8 support
3011 means that the same test output works with both.
3012
3013 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3014 calling malloc() with a zero argument.
3015
3016 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3017 optimization. I've improved things by keeping a bitmap of backreferences with
3018 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3019 fact referenced, the optimization can be applied. It is unlikely that a
3020 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3021 the match to follow \n) will appear inside brackets with a number greater than
3022 31, but if it does, any back reference > 31 suppresses the optimization.
3023
3024 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3025 of disabling numbered capturing parentheses. Any opening parenthesis that is
3026 not followed by ? behaves as if it were followed by ?: but named parentheses
3027 can still be used for capturing (and they will acquire numbers in the usual
3028 way).
3029
3030 41. Redesigned the return codes from the match() function into yes/no/error so
3031 that errors can be passed back from deep inside the nested calls. A malloc
3032 failure while inside a recursive subpattern call now causes the
3033 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3034
3035 42. It is now possible to set a limit on the number of times the match()
3036 function is called in a call to pcre_exec(). This facility makes it possible to
3037 limit the amount of recursion and backtracking, though not in a directly
3038 obvious way, because the match() function is used in a number of different
3039 circumstances. The count starts from zero for each position in the subject
3040 string (for non-anchored patterns). The default limit is, for compatibility, a
3041 large number, namely 10 000 000. You can change this in two ways:
3042
3043 (a) When configuring PCRE before making, you can use --with-match-limit=n
3044 to set a default value for the compiled library.
3045
3046 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3047 a different value is set. See 45 below.
3048
3049 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3050
3051 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3052 of things that can be changed at compile time. The first argument specifies
3053 what is wanted and the second points to where the information is to be placed.
3054 The current list of available information is:
3055
3056 PCRE_CONFIG_UTF8
3057
3058 The output is an integer that is set to one if UTF-8 support is available;
3059 otherwise it is set to zero.
3060
3061 PCRE_CONFIG_NEWLINE
3062
3063 The output is an integer that it set to the value of the code that is used for
3064 newline. It is either LF (10) or CR (13).
3065
3066 PCRE_CONFIG_LINK_SIZE
3067
3068 The output is an integer that contains the number of bytes used for internal
3069 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3070
3071 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3072
3073 The output is an integer that contains the threshold above which the POSIX
3074 interface uses malloc() for output vectors. See item 31 above.
3075
3076 PCRE_CONFIG_MATCH_LIMIT
3077
3078 The output is an unsigned integer that contains the default limit of the number
3079 of match() calls in a pcre_exec() execution. See 42 above.
3080
3081 44. pcretest has been upgraded by the addition of the -C option. This causes it
3082 to extract all the available output from the new pcre_config() function, and to
3083 output it. The program then exits immediately.
3084
3085 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3086 order to support additional features. One way would have been to define
3087 pcre_exec2() (for example) with extra arguments, but this would not have been
3088 extensible, and would also have required all calls to the original function to
3089 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3090 is used for passing in "extra" data from pcre_study().
3091
3092 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3093 contains the following fields:
3094
3095 flags a bitmap indicating which of the following fields are set
3096 study_data opaque data from pcre_study()
3097 match_limit a way of specifying a limit on match() calls for a specific
3098 call to pcre_exec()
3099 callout_data data for callouts (see 49 below)
3100
3101 The flag bits are also defined in pcre.h, and are
3102
3103 PCRE_EXTRA_STUDY_DATA
3104 PCRE_EXTRA_MATCH_LIMIT
3105 PCRE_EXTRA_CALLOUT_DATA
3106
3107 The pcre_study() function now returns one of these new pcre_extra blocks, with
3108 the actual study data pointed to by the study_data field, and the
3109 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3110 before. That is, this change is entirely upwards-compatible and requires no
3111 change to existing code.
3112
3113 If you want to pass in additional data to pcre_exec(), you can either place it
3114 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3115 block.
3116
3117 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3118 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3119 times with different match limits, until it finds the minimum value needed for
3120 pcre_exec() to complete. The value is then output. This can be instructive; for
3121 most simple matches the number is quite small, but for pathological cases it
3122 gets very large very quickly.
3123
3124 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3125 returns the size of the data block pointed to by the study_data field in a
3126 pcre_extra block, that is, the value that was passed as the argument to
3127 pcre_malloc() when PCRE was getting memory in which to place the information
3128 created by pcre_study(). The fourth argument should point to a size_t variable.
3129 pcretest has been extended so that this information is shown after a successful
3130 pcre_study() call when information about the compiled regex is being displayed.
3131
3132 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3133 because what follows is always an absolute path. (Later: it turns out that this
3134 is more than cosmetic for MinGW, because it doesn't like empty path
3135 components.)
3136
3137 49. Some changes have been made to the callout feature (see 28 above):
3138
3139 (i) A callout function now has three choices for what it returns:
3140
3141 0 => success, carry on matching
3142 > 0 => failure at this point, but backtrack if possible
3143 < 0 => serious error, return this value from pcre_exec()
3144
3145 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3146 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3147 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3148 use by callout functions. It will never be used by PCRE itself.
3149
3150 (ii) The pcre_extra structure (see 45 above) has a void * field called
3151 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3152 pcre_callout_block structure has a field of the same name. The contents of
3153 the field passed in the pcre_extra structure are passed to the callout
3154 function in the corresponding field in the callout block. This makes it
3155 easier to use the same callout-containing regex from multiple threads. For
3156 testing, the pcretest program has a new data escape
3157
3158 \C*n pass the number n (may be negative) as callout_data
3159
3160 If the callout function in pcretest receives a non-zero value as
3161 callout_data, it returns that value.
3162
3163 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3164 there were some redundant $(CFLAGS) in commands that are now specified as
3165 $(LINK), which already includes $(CFLAGS).
3166
3167 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3168 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3169 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3170 one-byte characters throughout. Note that case-insensitive matching applies
3171 only to characters whose values are less than 256. PCRE doesn't support the
3172 notion of cases for higher-valued characters.
3173
3174 (i) A character class whose characters are all within 0-255 is handled as
3175 a bit map, and the map is inverted for negative classes. Previously, a
3176 character > 255 always failed to match such a class; however it should
3177 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3178
3179 (ii) A negated character class with a single character < 255 is coded as
3180 "not this character" (OP_NOT). This wasn't working properly when the test
3181 character was multibyte, either singly or repeated.
3182
3183 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3184 mode, for example: \x{100}{2,3}.
3185
3186 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3187 singly or repeated) now correctly test multibyte characters. However,
3188 PCRE doesn't recognize any characters with values greater than 255 as
3189 digits, spaces, or word characters. Such characters always match \D, \S,
3190 and \W, and never match \d, \s, or \w.
3191
3192 (v) Classes may now contain characters and character ranges with values
3193 greater than 255. For example: [ab\x{100}-\x{400}].
3194
3195 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3196 PCRE in UTF-8 mode.
3197
3198 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3199 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3200 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3201 value.)
3202
3203 53. The single man page has become too large. I have therefore split it up into
3204 a number of separate man pages. These also give rise to individual HTML pages;
3205 these are now put in a separate directory, and there is an index.html page that
3206 lists them all. Some hyperlinking between the pages has been installed.
3207
3208 54. Added convenience functions for handling named capturing parentheses.
3209
3210 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3211 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3212 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3213 are faulted.
3214
3215 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3216 calling configure. These values are used when compiling the dftables.c program
3217 which is run to generate the source of the default character tables. They
3218 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3219 you will need to set these values.
3220
3221 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3222
3223
3224 Version 3.9 02-Jan-02
3225 ---------------------
3226
3227 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3228
3229 2. If --disable-static was given, the building process failed when trying to
3230 build pcretest and pcregrep. (For some reason it was using libtool to compile
3231 them, which is not right, as they aren't part of the library.)
3232
3233
3234 Version 3.8 18-Dec-01
3235 ---------------------
3236
3237 1. The experimental UTF-8 code was completely screwed up. It was packing the
3238 bytes in the wrong order. How dumb can you get?
3239
3240
3241 Version 3.7 29-Oct-01
3242 ---------------------
3243
3244 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3245 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3246 this didn't happen under Solaris 8, where I normally test things.
3247
3248 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3249 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3250 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3251 this fix an hour or so after the initial 3.7 release.)
3252
3253
3254 Version 3.6 23-Oct-01
3255 ---------------------
3256
3257 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3258 offsets passed as NULL with zero offset count.
3259
3260 2. The config.guess and config.sub files had not been updated when I moved to
3261 the latest autoconf.
3262
3263
3264 Version 3.5 15-Aug-01
3265 ---------------------
3266
3267 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3268 had been forgotten.
3269
3270 2. By using declared but undefined structures, we can avoid using "void"
3271 definitions in pcre.h while keeping the internal definitions of the structures
3272 private.
3273
3274 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3275 user point of view, this means that both static and shared libraries are built
3276 by default, but this can be individually controlled. More of the work of
3277 handling this static/shared cases is now inside libtool instead of PCRE's make
3278 file.
3279
3280 4. The pcretest utility is now installed along with pcregrep because it is
3281 useful for users (to test regexs) and by doing this, it automatically gets
3282 relinked by libtool. The documentation has been turned into a man page, so
3283 there are now .1, .txt, and .html versions in /doc.
3284
3285 5. Upgrades to pcregrep:
3286 (i) Added long-form option names like gnu grep.
3287 (ii) Added --help to list all options with an explanatory phrase.
3288 (iii) Added -r, --recursive to recurse into sub-directories.
3289 (iv) Added -f, --file to read patterns from a file.
3290
3291 6. pcre_exec() was referring to its "code" argument before testing that
3292 argument for NULL (and giving an error if it was NULL).
3293
3294 7. Upgraded Makefile.in to allow for compiling in a different directory from
3295 the source directory.
3296
3297 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3298 options bits, the pointer it was passed was to an int instead of to an unsigned
3299 long int. This mattered only on 64-bit systems.
3300
3301 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3302 generated) instead of pcre.in, which it its source. Also made the same change
3303 in several of the .c files.
3304
3305 10. A new release of gcc defines printf() as a macro, which broke pcretest
3306 because it had an ifdef in the middle of a string argument for printf(). Fixed
3307 by using separate calls to printf().
3308
3309 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3310 script, to force use of CR or LF instead of \n in the source. On non-Unix
3311 systems, the value can be set in config.h.
3312
3313 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3314 absolute limit. Changed the text of the error message to make this clear, and
3315 likewise updated the man page.
3316
3317 13. The limit of 99 on the number of capturing subpatterns has been removed.
3318 The new limit is 65535, which I hope will not be a "real" limit.
3319
3320
3321 Version 3.4 22-Aug-00
3322 ---------------------
3323
3324 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3325
3326 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3327
3328
3329 Version 3.3 01-Aug-00
3330 ---------------------
3331
3332 1. If an octal character was given, but the value was greater than \377, it
3333 was not getting masked to the least significant bits, as documented. This could
3334 lead to crashes in some systems.
3335
3336 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3337 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3338
3339 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3340 These just pass their arguments on to (pcre_free)(), but they are provided
3341 because some uses of PCRE bind it to non-C systems that can call its functions,
3342 but cannot call free() or pcre_free() directly.
3343
3344 4. Add "make test" as a synonym for "make check". Corrected some comments in
3345 the Makefile.
3346
3347 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3348 Makefile.
3349
3350 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3351 command called pgrep for grepping around the active processes.
3352
3353 7. Added the beginnings of support for UTF-8 character strings.
3354
3355 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3356 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3357 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3358 out for the ar command.)
3359
3360
3361 Version 3.2 12-May-00
3362 ---------------------
3363
3364 This is purely a bug fixing release.
3365
3366 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3367 of ZA. This was just one example of several cases that could provoke this bug,
3368 which was introduced by change 9 of version 2.00. The code for breaking
3369 infinite loops after an iteration that matches an empty string was't working
3370 correctly.
3371
3372 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3373 when matched against abbab (for example). After matching an empty string, it
3374 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3375 caused it to match further down the string than it should.
3376
3377 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3378 was there because it doesn't seem to be needed, and it causes trouble on some
3379 systems, as it is not a Standard C header. It has been removed.
3380
3381 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3382 were reported on the Macintosh. The changes were from
3383
3384 while ((c = *(++ptr)) != 0 && c != '\n');
3385 to
3386 while ((c = *(++ptr)) != 0 && c != '\n') ;
3387
3388 Totally extraordinary, but if that's what it takes...
3389
3390 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3391 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3392 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3393 assumes the way PCRE uses memmove() (always moving upwards).
3394
3395 6. PCRE is being used in one environment where strchr() is not available. There
3396 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3397 faster code anyway.
3398
3399
3400 Version 3.1 09-Feb-00
3401 ---------------------
3402
3403 The only change in this release is the fixing of some bugs in Makefile.in for
3404 the "install" target:
3405
3406 (1) It was failing to install pcreposix.h.
3407
3408 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3409
3410
3411 Version 3.0 01-Feb-00
3412 ---------------------
3413
3414 1. Add support for the /+ modifier to perltest (to output $` like it does in
3415 pcretest).
3416
3417 2. Add support for the /g modifier to perltest.
3418
3419 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3420 matches null strings.
3421
3422 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3423 pattern. Perl treats empty patterns specially - it reuses the most recent
3424 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3425 effect.
3426
3427 5. The POSIX interface was broken in that it was just handing over the POSIX
3428 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3429 required a bigger vector, with some working space on the end. This means that
3430 the POSIX wrapper now has to get and free some memory, and copy the results.
3431
3432 6. Added some simple autoconf support, placing the test data and the
3433 documentation in separate directories, re-organizing some of the
3434 information files, and making it build pcre-config (a GNU standard). Also added
3435 libtool support for building PCRE as a shared library, which is now the
3436 default.
3437
3438 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3439 09 are not valid octal constants. Single digits will be used for minor values
3440 less than 10.
3441
3442 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3443 existing programs that set these in the POSIX interface can use PCRE without
3444 modification.
3445
3446 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3447 return all that pcre_info() returns, plus additional data. The pcre_info()
3448 function is retained for compatibility, but is considered to be obsolete.
3449
3450 10. Added experimental recursion feature (?R) to handle one common case that
3451 Perl 5.6 will be able to do with (?p{...}).
3452
3453 11. Added support for POSIX character classes like [:alpha:], which Perl is
3454 adopting.
3455
3456
3457 Version 2.08 31-Aug-99
3458 ----------------------
3459
3460 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3461 trying to match at the startoffset position, but instead was moving forward to
3462 the next newline as if a previous match had failed.
3463
3464 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3465 and could get into a loop if a null string was matched other than at the start
3466 of the subject.
3467
3468 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3469 be distinguished at compile time, and for completeness also added PCRE_DATE.
3470
3471 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3472 in GnuWin32 environments.
3473
3474
3475 Version 2.07 29-Jul-99
3476 ----------------------
3477
3478 1. The documentation is now supplied in plain text form and HTML as well as in
3479 the form of man page sources.
3480
3481 2. C++ compilers don't like assigning (void *) values to other pointer types.
3482 In particular this affects malloc(). Although there is no problem in Standard
3483 C, I've put in casts to keep C++ compilers happy.
3484
3485 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3486 should be (const char *).
3487
3488 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3489 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3490 However, I haven't made this a standard facility. The documentation doesn't
3491 mention it, and the Makefile doesn't support it.
3492
3493 5. The Makefile now contains an "install" target, with editable destinations at
3494 the top of the file. The pcretest program is not installed.
3495
3496 6. pgrep -V now gives the PCRE version number and date.
3497
3498 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3499 causing the entire string to be ignored, instead of just the last character.
3500
3501 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3502 non-matching string, it can take a very, very long time, even for strings of
3503 quite modest length, because of the nested recursion. PCRE now does better in
3504 some of these cases. It does this by remembering the last required literal
3505 character in the pattern, and pre-searching the subject to ensure it is present
3506 before running the real match. In other words, it applies a heuristic to detect
3507 some types of certain failure quickly, and in the above example, if presented
3508 with a string that has no trailing " it gives "no match" very quickly.
3509
3510 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3511 other alternatives are tried instead.
3512
3513
3514 Version 2.06 09-Jun-99
3515 ----------------------
3516
3517 1. Change pcretest's output for amount of store used to show just the code
3518 space, because the remainder (the data block) varies in size between 32-bit and
3519 64-bit systems.
3520
3521 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3522 start matching at. This allows lookbehinds to work when searching for multiple
3523 occurrences in a string.
3524
3525 3. Added additional options to pcretest for testing multiple occurrences:
3526
3527 /+ outputs the rest of the string that follows a match
3528 /g loops for multiple occurrences, using the new startoffset argument
3529 /G loops for multiple occurrences by passing an incremented pointer
3530
3531 4. PCRE wasn't doing the "first character" optimization for patterns starting
3532 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3533 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3534 the letter 'x'. On long subject strings, this gives a significant speed-up.
3535
3536
3537 Version 2.05 21-Apr-99
3538 ----------------------
3539
3540 1. Changed the type of magic_number from int to long int so that it works
3541 properly on 16-bit systems.
3542
3543 2. Fixed a bug which caused patterns starting with .* not to work correctly
3544 when the subject string contained newline characters. PCRE was assuming
3545 anchoring for such patterns in all cases, which is not correct because .* will
3546 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3547 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3548 must be retried after every newline in the subject.
3549
3550
3551 Version 2.04 18-Feb-99
3552 ----------------------
3553
3554 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3555 computation of the store needed to hold the pattern was incorrect (too large).
3556 If such patterns were nested a few deep, this could multiply and become a real
3557 problem.
3558
3559 2. Added /M option to pcretest to show the memory requirement of a specific
3560 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3561
3562 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3563 compiled in such a way that the backtracking after subsequent failure was
3564 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3565 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3566
3567
3568 Version 2.03 02-Feb-99
3569 ----------------------
3570
3571 1. Fixed typo and small mistake in man page.
3572
3573 2. Added 4th condition (GPL supersedes if conflict) and created separate
3574 LICENCE file containing the conditions.
3575
3576 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3577 Perl, that is the internal \ allows the delimiter to be included in the
3578 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3579 the final delimiter, add \ to the end of the pattern (to test the error).
3580
3581 4. Added the convenience functions for extracting substrings after a successful
3582 match. Updated pcretest to make it able to test these functions.
3583
3584
3585 Version 2.02 14-Jan-99
3586 ----------------------
3587
3588 1. Initialized the working variables associated with each extraction so that
3589 their saving and restoring doesn't refer to uninitialized store.
3590
3591 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3592 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3593 fix the problem.
3594
3595 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3596 calls, and wasn't printing the correct value for compiling calls. Increased the
3597 default value of LOOPREPEAT, and the number of significant figures in the
3598 times.
3599
3600 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3601
3602 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3603 a building problem on Windows NT with a FAT file system.
3604
3605
3606 Version 2.01 21-Oct-98
3607 ----------------------
3608
3609 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3610 to character tables built by pcre_maketables() in the current locale. If NULL
3611 is passed, the default tables are used.
3612
3613
3614 Version 2.00 24-Sep-98
3615 ----------------------
3616
3617 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3618 it any more.
3619
3620 2. Allow quantification of (?>) groups, and make it work correctly.
3621
3622 3. The first character computation wasn't working for (?>) groups.
3623
3624 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3625 end of the subject) and add 5.005's \z, which really does match only at the
3626 very end of the subject.
3627
3628 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3629
3630 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3631 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3632 localized options. All options to pcre_study() were also removed.
3633
3634 7. Add other new features from 5.005:
3635
3636 $(?<= positive lookbehind
3637 $(?<! negative lookbehind
3638 (?imsx-imsx) added the unsetting capability
3639 such a setting is global if at outer level; local otherwise
3640 (?imsx-imsx:) non-capturing groups with option setting
3641 (?(cond)re|re) conditional pattern matching
3642
3643 A backreference to itself in a repeated group matches the previous
3644 captured string.
3645
3646 8. General tidying up of studying (both automatic and via "study")
3647 consequential on the addition of new assertions.
3648
3649 9. As in 5.005, unlimited repeated groups that could match an empty substring
3650 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3651 runtime if any iteration does actually match an empty substring.
3652
3653 10. Include the RunTest script in the distribution.
3654
3655 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3656 discrepancies, some of which were old and were also with respect to 5.004. They
3657 have now been fixed.
3658
3659
3660 Version 1.09 28-Apr-98
3661 ----------------------
3662
3663 1. A negated single character class followed by a quantifier with a minimum
3664 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3665 program crashes, or just wrong answers. This did not apply to negated classes
3666 containing more than one character, or to minima other than one.
3667
3668
3669 Version 1.08 27-Mar-98
3670 ----------------------
3671
3672 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3673
3674 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3675 latter must appear before anything that relies on it in the pattern.
3676
3677
3678 Version 1.07 16-Feb-98
3679 ----------------------
3680
3681 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3682 repeat of a potentially empty string).
3683
3684
3685 Version 1.06 23-Jan-98
3686 ----------------------
3687
3688 1. Added Markus Oberhumer's little patches for C++.
3689
3690 2. Literal strings longer than 255 characters were broken.
3691
3692
3693 Version 1.05 23-Dec-97
3694 ----------------------
3695
3696 1. Negated character classes containing more than one character were failing if
3697 PCRE_CASELESS was set at run time.
3698
3699
3700 Version 1.04 19-Dec-97
3701 ----------------------
3702
3703 1. Corrected the man page, where some "const" qualifiers had been omitted.
3704
3705 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3706 input syntax.
3707
3708 3. Fixed memory leak which occurred when a regex with back references was
3709 matched with an offsets vector that wasn't big enough. The temporary memory
3710 that is used in this case wasn't being freed if the match failed.
3711
3712 4. Tidied pcretest to ensure it frees memory that it gets.
3713
3714 5. Temporary memory was being obtained in the case where the passed offsets
3715 vector was exactly big enough.
3716
3717 6. Corrected definition of offsetof() from change 5 below.
3718
3719 7. I had screwed up change 6 below and broken the rules for the use of
3720 setjmp(). Now fixed.
3721
3722
3723 Version 1.03 18-Dec-97
3724 ----------------------
3725
3726 1. A erroneous regex with a missing opening parenthesis was correctly
3727 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3728 on some systems.
3729
3730 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3731 it was reported that one broken compiler failed on the former because "code" is
3732 also an independent variable.
3733
3734 3. The erroneous regex a[]b caused an array overrun reference.
3735
3736 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3737 fail on data ending with that character. (It was going on too far, and checking
3738 the next character, typically a binary zero.) This was specific to the
3739 optimized code for single-character negative classes.
3740
3741 5. Added a contributed patch from the TIN world which does the following:
3742
3743 + Add an undef for memmove, in case the the system defines a macro for it.
3744
3745 + Add a definition of offsetof(), in case there isn't one. (I don't know
3746 the reason behind this - offsetof() is part of the ANSI standard - but
3747 it does no harm).
3748
3749 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3750 most of the places where whitespace preceded '#'. I have given up and
3751 allowed the remaining 2 cases to be at the margin.
3752
3753 + Rename some variables in pcre to eliminate shadowing. This seems very
3754 pedantic, but does no harm, of course.
3755
3756 6. Moved the call to setjmp() into its own function, to get rid of warnings
3757 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3758
3759 7. Constructs such as \d{8,} were compiling into the equivalent of
3760 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3761 outcome, but in this particular case used more store than had been allocated,
3762 which caused the bug to be discovered because it threw up an internal error.
3763
3764 8. The debugging code in both pcre and pcretest for outputting the compiled
3765 form of a regex was going wrong in the case of back references followed by
3766 curly-bracketed repeats.
3767
3768
3769 Version 1.02 12-Dec-97
3770 ----------------------
3771
3772 1. Typos in pcre.3 and comments in the source fixed.
3773
3774 2. Applied a contributed patch to get rid of places where it used to remove
3775 'const' from variables, and fixed some signed/unsigned and uninitialized
3776 variable warnings.
3777
3778 3. Added the "runtest" target to Makefile.
3779
3780 4. Set default compiler flag to -O2 rather than just -O.
3781
3782
3783 Version 1.01 19-Nov-97
3784 ----------------------
3785
3786 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3787 like /([ab]*)*/, that is, for classes with more than one character in them.
3788
3789 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3790 as /((?>a*))*/ (a PCRE_EXTRA facility).
3791
3792
3793 Version 1.00 18-Nov-97
3794 ----------------------
3795
3796 1. Added compile-time macros to support systems such as SunOS4 which don't have
3797 memmove() or strerror() but have other things that can be used instead.
3798
3799 2. Arranged that "make clean" removes the executables.
3800
3801
3802 Version 0.99 27-Oct-97
3803 ----------------------
3804
3805 1. Fixed bug in code for optimizing classes with only one character. It was
3806 initializing a 32-byte map regardless, which could cause it to run off the end
3807 of the memory it had got.
3808
3809 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3810
3811
3812 Version 0.98 22-Oct-97
3813 ----------------------
3814
3815 1. Fixed bug in code for handling temporary memory usage when there are more
3816 back references than supplied space in the ovector. This could cause segfaults.
3817
3818
3819 Version 0.97 21-Oct-97
3820 ----------------------
3821
3822 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3823
3824 2. Optimized negated single characters not to use a bit map.
3825
3826 3. Brought error texts together as macro definitions; clarified some of them;
3827 fixed one that was wrong - it said "range out of order" when it meant "invalid
3828 escape sequence".
3829
3830 4. Changed some char * arguments to const char *.
3831
3832 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3833
3834 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3835 pcretest.
3836
3837
3838 Version 0.96 16-Oct-97
3839 ----------------------
3840
3841 1. Added a simple "pgrep" utility to the distribution.
3842
3843 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3844 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3845 where "ddd" means "one or more decimal digits".
3846
3847 3. Fixed serious bug. If a pattern had a back reference, but the call to
3848 pcre_exec() didn't supply a large enough ovector to record the related
3849 identifying subpattern, the match always failed. PCRE now remembers the number
3850 of the largest back reference, and gets some temporary memory in which to save
3851 the offsets during matching if necessary, in order to ensure that
3852 backreferences always work.
3853
3854 4. Increased the compatibility with Perl in a number of ways:
3855
3856 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3857 to request this handling. The option can be set at compile or exec time.
3858
3859 (b) $ matches before a terminating newline by default; an option
3860 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3861 mode). The option can be set at compile or exec time.
3862
3863 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3864 the same as Perl's. If the decimal number it represents is less than 10
3865 or there aren't that many previous left capturing parentheses, an octal
3866 escape is read. Inside a character class, it's always an octal escape,
3867 even if it is a single digit.
3868
3869 (d) An escaped but undefined alphabetic character is taken as a literal,
3870 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3871 escapes.
3872
3873 (e) {0} is now permitted. (The previous item is removed from the compiled
3874 pattern).
3875
3876 5. Changed all the names of code files so that the basic parts are no longer
3877 than 10 characters, and abolished the teeny "globals.c" file.
3878
3879 6. Changed the handling of character classes; they are now done with a 32-byte
3880 bit map always.
3881
3882 7. Added the -d and /D options to pcretest to make it possible to look at the
3883 internals of compilation without having to recompile pcre.
3884
3885
3886 Version 0.95 23-Sep-97
3887 ----------------------
3888
3889 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3890 \x20 at the start of a run of normal characters. These were being treated as
3891 real characters, instead of the source characters being re-checked.
3892
3893
3894 Version 0.94 18-Sep-97
3895 ----------------------
3896
3897 1. The functions are now thread-safe, with the caveat that the global variables
3898 containing pointers to malloc() and free() or alternative functions are the
3899 same for all threads.
3900
3901 2. Get pcre_study() to generate a bitmap of initial characters for non-
3902 anchored patterns when this is possible, and use it if passed to pcre_exec().
3903
3904
3905 Version 0.93 15-Sep-97
3906 ----------------------
3907
3908 1. /(b)|(:+)/ was computing an incorrect first character.
3909
3910 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3911 but not actually doing anything yet.
3912
3913 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3914 as Perl does (e.g. [-az] or [az-]).
3915
3916 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3917 all possible positions.
3918
3919 5. Split up into different modules to avoid including unneeded functions in a
3920 compiled binary. However, compile and exec are still in one module. The "study"
3921 function is split off.
3922
3923 6. The character tables are now in a separate module whose source is generated
3924 by an auxiliary program - but can then be edited by hand if required. There are
3925 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3926 toupper() in the code.
3927
3928 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3929 make them global. Abolish the function for setting them, as the caller can now
3930 set them directly.
3931
3932
3933 Version 0.92 11-Sep-97
3934 ----------------------
3935
3936 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3937 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3938
3939 2. Caseless matching was not working in character classes if the characters in
3940 the pattern were in upper case.
3941
3942 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3943
3944 4. Make PCRE_ANCHORED public and accept as a compile option.
3945
3946 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3947 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3948 pass them.
3949
3950 6. Give an error if bad option bits passed at compile or run time.
3951
3952 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3953 pcretest to cause it to pass that flag.
3954
3955 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3956 options, and the first character, if set.
3957
3958 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3959
3960
3961 Version 0.91 10-Sep-97
3962 ----------------------
3963
3964 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3965 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3966
3967 2. PCRE was looping on encountering an indefinitely repeated back reference to
3968 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3969 Perl does - treats the match as successful.
3970
3971 ****

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