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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 552 - (show annotations) (download)
Wed Oct 13 10:15:41 2010 UTC (3 years, 6 months ago) by ph10
File size: 190635 byte(s)
Fix \s bug in character classes (always removing VT).

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 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
15 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
16 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
17 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
18 (*THEN).
19
20 3. If \s appeared in a character class, it removed the VT character from
21 the class, even if it had been included by some previous item, for example
22 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
23 of \s, but is part of the POSIX "space" class.)
24
25
26 Version 8.10 25-Jun-2010
27 ------------------------
28
29 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
30 THEN.
31
32 2. (*ACCEPT) was not working when inside an atomic group.
33
34 3. Inside a character class, \B is treated as a literal by default, but
35 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
36 causes the error). The code is unchanged, but I tidied the documentation.
37
38 4. Inside a character class, PCRE always treated \R and \X as literals,
39 whereas Perl faults them if its -w option is set. I have changed PCRE so
40 that it faults them when PCRE_EXTRA is set.
41
42 5. Added support for \N, which always matches any character other than
43 newline. (It is the same as "." when PCRE_DOTALL is not set.)
44
45 6. When compiling pcregrep with newer versions of gcc which may have
46 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
47 declared with attribute warn_unused_result" were given. Just casting the
48 result to (void) does not stop the warnings; a more elaborate fudge is
49 needed. I've used a macro to implement this.
50
51 7. Minor change to pcretest.c to avoid a compiler warning.
52
53 8. Added four artifical Unicode properties to help with an option to make
54 \s etc use properties (see next item). The new properties are: Xan
55 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
56
57 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
58 use Unicode properties. (*UCP) at the start of a pattern can be used to set
59 this option. Modified pcretest to add /W to test this facility. Added
60 REG_UCP to make it available via the POSIX interface.
61
62 10. Added --line-buffered to pcregrep.
63
64 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
65 studied, and the match started with a letter with a code point greater than
66 127 whose first byte was different to the first byte of the other case of
67 the letter, the other case of this starting letter was not recognized
68 (#976).
69
70 12. If a pattern that was studied started with a repeated Unicode property
71 test, for example, \p{Nd}+, there was the theoretical possibility of
72 setting up an incorrect bitmap of starting bytes, but fortunately it could
73 not have actually happened in practice until change 8 above was made (it
74 added property types that matched character-matching opcodes).
75
76 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
77 possible starting bytes for non-anchored patterns.
78
79 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
80 \R, and also a number of cases that involve Unicode properties, both
81 explicit and implicit when PCRE_UCP is set.
82
83 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
84 input, it could crash or give wrong results if characters with values
85 greater than 0xc0 were present in the subject string. (Detail: it assumed
86 UTF-8 input when processing these items.)
87
88 16. Added a lot of (int) casts to avoid compiler warnings in systems where
89 size_t is 64-bit (#991).
90
91 17. Added a check for running out of memory when PCRE is compiled with
92 --disable-stack-for-recursion (#990).
93
94 18. If the last data line in a file for pcretest does not have a newline on
95 the end, a newline was missing in the output.
96
97 19. The default pcre_chartables.c file recognizes only ASCII characters (values
98 less than 128) in its various bitmaps. However, there is a facility for
99 generating tables according to the current locale when PCRE is compiled. It
100 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
101 characters, are recognized by isspace() and therefore were getting set in
102 these tables, and indeed these tables seem to approximate to ISO 8859. This
103 caused a problem in UTF-8 mode when pcre_study() was used to create a list
104 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
105 which of course cannot start UTF-8 characters. I have changed the code so
106 that only real ASCII characters (less than 128) and the correct starting
107 bytes for UTF-8 encodings are set for characters greater than 127 when in
108 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
109 altogether.)
110
111 20. Added the /T option to pcretest so as to be able to run tests with non-
112 standard character tables, thus making it possible to include the tests
113 used for 19 above in the standard set of tests.
114
115 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
116 reference to a subpattern the other side of a comment that contains an
117 opening parenthesis caused either an internal compiling error, or a
118 reference to the wrong subpattern.
119
120
121 Version 8.02 19-Mar-2010
122 ------------------------
123
124 1. The Unicode data tables have been updated to Unicode 5.2.0.
125
126 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
127 configured.
128
129 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
130 original author of that file, following a query about its status.
131
132 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
133 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
134
135 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
136 quantifier applied to a forward-referencing subroutine call, could compile
137 incorrect code or give the error "internal error: previously-checked
138 referenced subpattern not found".
139
140 6. Both MS Visual Studio and Symbian OS have problems with initializing
141 variables to point to external functions. For these systems, therefore,
142 pcre_malloc etc. are now initialized to local functions that call the
143 relevant global functions.
144
145 7. There were two entries missing in the vectors called coptable and poptable
146 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
147 I've fixed the data, and added a kludgy way of testing at compile time that
148 the lengths are correct (equal to the number of opcodes).
149
150 8. Following on from 7, I added a similar kludge to check the length of the
151 eint vector in pcreposix.c.
152
153 9. Error texts for pcre_compile() are held as one long string to avoid too
154 much relocation at load time. To find a text, the string is searched,
155 counting zeros. There was no check for running off the end of the string,
156 which could happen if a new error number was added without updating the
157 string.
158
159 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
160
161 11. \K was not working if it appeared in an atomic group or in a group that
162 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
163 \K is "not well defined" if used in an assertion. PCRE now accepts it if
164 the assertion is positive, but not if it is negative.
165
166 12. Change 11 fortuitously reduced the size of the stack frame used in the
167 "match()" function of pcre_exec.c by one pointer. Forthcoming
168 implementation of support for (*MARK) will need an extra pointer on the
169 stack; I have reserved it now, so that the stack frame size does not
170 decrease.
171
172 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
173 item in branch that calls a recursion is a subroutine call - as in the
174 second branch in the above example - was incorrectly given the compile-
175 time error "recursive call could loop indefinitely" because pcre_compile()
176 was not correctly checking the subroutine for matching a non-empty string.
177
178 14. The checks for overrunning compiling workspace could trigger after an
179 overrun had occurred. This is a "should never occur" error, but it can be
180 triggered by pathological patterns such as hundreds of nested parentheses.
181 The checks now trigger 100 bytes before the end of the workspace.
182
183 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
184
185
186 Version 8.01 19-Jan-2010
187 ------------------------
188
189 1. If a pattern contained a conditional subpattern with only one branch (in
190 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
191 computed the wrong minimum data length (which is of course zero for such
192 subpatterns). This could cause incorrect "no match" results.
193
194 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
195 the pattern is reset in the first branch, pcre_compile() failed with
196 "internal error: code overflow at offset...". This happened only when
197 the reset was to the original external option setting. (An optimization
198 abstracts leading options settings into an external setting, which was the
199 cause of this.)
200
201 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
202 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
203 assertion pattern did not match (meaning that the assertion was true), it
204 was incorrectly treated as false if the SKIP had been reached during the
205 matching. This also applied to assertions used as conditions.
206
207 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
208 assertion subpattern, including such a pattern used as a condition,
209 unpredictable results occurred, instead of the error return
210 PCRE_ERROR_DFA_UITEM.
211
212 5. The C++ GlobalReplace function was not working like Perl for the special
213 situation when an empty string is matched. It now does the fancy magic
214 stuff that is necessary.
215
216 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
217 removed. (These were left over from very, very early versions of PCRE.)
218
219 7. Some cosmetic changes to the code to make life easier when compiling it
220 as part of something else:
221
222 (a) Change DEBUG to PCRE_DEBUG.
223
224 (b) In pcre_compile(), rename the member of the "branch_chain" structure
225 called "current" as "current_branch", to prevent a collision with the
226 Linux macro when compiled as a kernel module.
227
228 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
229 prevent a collision with the Linux macro when compiled as a kernel
230 module.
231
232 8. In pcre_compile() there are some checks for integer overflows that used to
233 cast potentially large values to (double). This has been changed to that
234 when building, a check for int64_t is made, and if it is found, it is used
235 instead, thus avoiding the use of floating point arithmetic. (There is no
236 other use of FP in PCRE.) If int64_t is not found, the fallback is to
237 double.
238
239 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
240 2005 (difference between two addresses compared to an unsigned value).
241
242 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
243 custom one, because of the following reported problem in Windows:
244
245 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
246 under Win32.
247 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
248 therefore missing the function definition.
249 - The compiler thus generates a "C" signature for the test function.
250 - The linker fails to find the "C" function.
251 - PCRE fails to configure if asked to do so against libbz2.
252
253 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
254 messages were output:
255
256 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
257 rerunning libtoolize, to keep the correct libtool macros in-tree.
258 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
259
260 I have done both of these things.
261
262 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
263 most of the time, it *can* run out if it is given a pattern that contains a
264 runaway infinite recursion. I updated the discussion in the pcrestack man
265 page.
266
267 13. Now that we have gone to the x.xx style of version numbers, the minor
268 version may start with zero. Using 08 or 09 is a bad idea because users
269 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
270 interpreted as invalid octal numbers. I've updated the previous comment in
271 configure.ac, and also added a check that gives an error if 08 or 09 are
272 used.
273
274 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
275 causing partial matching to fail when the end of the subject matched \W
276 in a UTF-8 pattern where \W was quantified with a minimum of 3.
277
278 15. There were some discrepancies between the declarations in pcre_internal.h
279 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
280 their definitions. The declarations used "const uschar *" and the
281 definitions used USPTR. Even though USPTR is normally defined as "const
282 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
283 reported that: "This difference in casting confuses some C++ compilers, for
284 example, SunCC recognizes above declarations as different functions and
285 generates broken code for hbpcre." I have changed the declarations to use
286 USPTR.
287
288 16. GNU libtool is named differently on some systems. The autogen.sh script now
289 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
290 (FreeBSD).
291
292 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
293 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
294 comment: "Figure out how to create a longlong from a string: strtoll and
295 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
296 instance, but it only takes 2 args instead of 3!"
297
298 18. A subtle bug concerned with back references has been fixed by a change of
299 specification, with a corresponding code fix. A pattern such as
300 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
301 refers, was giving matches when it shouldn't. For example, xa=xaaa would
302 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
303 same bug. Such groups have to be quantified to be useful, or contained
304 inside another quantified group. (If there's no repetition, the reference
305 can never match.) The problem arises because, having left the group and
306 moved on to the rest of the pattern, a later failure that backtracks into
307 the group uses the captured value from the final iteration of the group
308 rather than the correct earlier one. I have fixed this in PCRE by forcing
309 any group that contains a reference to itself to be an atomic group; that
310 is, there cannot be any backtracking into it once it has completed. This is
311 similar to recursive and subroutine calls.
312
313
314 Version 8.00 19-Oct-09
315 ----------------------
316
317 1. The table for translating pcre_compile() error codes into POSIX error codes
318 was out-of-date, and there was no check on the pcre_compile() error code
319 being within the table. This could lead to an OK return being given in
320 error.
321
322 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
323 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
324 in a Windows environment.
325
326 3. The pcregrep --count option prints the count for each file even when it is
327 zero, as does GNU grep. However, pcregrep was also printing all files when
328 --files-with-matches was added. Now, when both options are given, it prints
329 counts only for those files that have at least one match. (GNU grep just
330 prints the file name in this circumstance, but including the count seems
331 more useful - otherwise, why use --count?) Also ensured that the
332 combination -clh just lists non-zero counts, with no names.
333
334 4. The long form of the pcregrep -F option was incorrectly implemented as
335 --fixed_strings instead of --fixed-strings. This is an incompatible change,
336 but it seems right to fix it, and I didn't think it was worth preserving
337 the old behaviour.
338
339 5. The command line items --regex=pattern and --regexp=pattern were not
340 recognized by pcregrep, which required --regex pattern or --regexp pattern
341 (with a space rather than an '='). The man page documented the '=' forms,
342 which are compatible with GNU grep; these now work.
343
344 6. No libpcreposix.pc file was created for pkg-config; there was just
345 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
346
347 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
348 when UCP support is not needed, by modifying the Python script that
349 generates it from Unicode data files. This should not matter if the module
350 is correctly used as a library, but I received one complaint about 50K of
351 unwanted data. My guess is that the person linked everything into his
352 program rather than using a library. Anyway, it does no harm.
353
354 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
355 was a minimum greater than 1 for a wide character in a possessive
356 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
357 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
358 character. Chaos in the form of incorrect output or a compiling loop could
359 result.
360
361 9. The restrictions on what a pattern can contain when partial matching is
362 requested for pcre_exec() have been removed. All patterns can now be
363 partially matched by this function. In addition, if there are at least two
364 slots in the offset vector, the offset of the earliest inspected character
365 for the match and the offset of the end of the subject are set in them when
366 PCRE_ERROR_PARTIAL is returned.
367
368 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
369 synonymous with PCRE_PARTIAL, for backwards compatibility, and
370 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
371 and may be more useful for multi-segment matching.
372
373 11. Partial matching with pcre_exec() is now more intuitive. A partial match
374 used to be given if ever the end of the subject was reached; now it is
375 given only if matching could not proceed because another character was
376 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
377 string "Z", which now yields "no match" instead of "partial match". In the
378 case of pcre_dfa_exec(), "no match" is given if every matching path for the
379 final character ended with (*FAIL).
380
381 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
382 if the pattern had a "must contain" character that was already found in the
383 earlier partial match, unless partial matching was again requested. For
384 example, with the pattern /dog.(body)?/, the "must contain" character is
385 "g". If the first part-match was for the string "dog", restarting with
386 "sbody" failed. This bug has been fixed.
387
388 13. The string returned by pcre_dfa_exec() after a partial match has been
389 changed so that it starts at the first inspected character rather than the
390 first character of the match. This makes a difference only if the pattern
391 starts with a lookbehind assertion or \b or \B (\K is not supported by
392 pcre_dfa_exec()). It's an incompatible change, but it makes the two
393 matching functions compatible, and I think it's the right thing to do.
394
395 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
396 so that the demonstration program is easily available in environments where
397 PCRE has not been installed from source.
398
399 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
400 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
401 library.
402
403 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
404 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
405 is not the first non-POSIX option to be added. Clearly some people find
406 these options useful.
407
408 17. If a caller to the POSIX matching function regexec() passes a non-zero
409 value for nmatch with a NULL value for pmatch, the value of
410 nmatch is forced to zero.
411
412 18. RunGrepTest did not have a test for the availability of the -u option of
413 the diff command, as RunTest does. It now checks in the same way as
414 RunTest, and also checks for the -b option.
415
416 19. If an odd number of negated classes containing just a single character
417 interposed, within parentheses, between a forward reference to a named
418 subpattern and the definition of the subpattern, compilation crashed with
419 an internal error, complaining that it could not find the referenced
420 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
421 [The bug was that it was starting one character too far in when skipping
422 over the character class, thus treating the ] as data rather than
423 terminating the class. This meant it could skip too much.]
424
425 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
426 /g option in pcretest when the pattern contains \K, which makes it possible
427 to have an empty string match not at the start, even when the pattern is
428 anchored. Updated pcretest and pcredemo to use this option.
429
430 21. If the maximum number of capturing subpatterns in a recursion was greater
431 than the maximum at the outer level, the higher number was returned, but
432 with unset values at the outer level. The correct (outer level) value is
433 now given.
434
435 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
436 PCRE did not set those parentheses (unlike Perl). I have now found a way to
437 make it do so. The string so far is captured, making this feature
438 compatible with Perl.
439
440 23. The tests have been re-organized, adding tests 11 and 12, to make it
441 possible to check the Perl 5.10 features against Perl 5.10.
442
443 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
444 pattern matches a fixed length string. PCRE did not allow this; now it
445 does. Neither allows recursion.
446
447 25. I finally figured out how to implement a request to provide the minimum
448 length of subject string that was needed in order to match a given pattern.
449 (It was back references and recursion that I had previously got hung up
450 on.) This code has now been added to pcre_study(); it finds a lower bound
451 to the length of subject needed. It is not necessarily the greatest lower
452 bound, but using it to avoid searching strings that are too short does give
453 some useful speed-ups. The value is available to calling programs via
454 pcre_fullinfo().
455
456 26. While implementing 25, I discovered to my embarrassment that pcretest had
457 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
458 study optimizations had never been tested with that matching function.
459 Oops. What is worse, even when it was passed study data, there was a bug in
460 pcre_dfa_exec() that meant it never actually used it. Double oops. There
461 were also very few tests of studied patterns with pcre_dfa_exec().
462
463 27. If (?| is used to create subpatterns with duplicate numbers, they are now
464 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
465 on the other side of the coin, they are no longer allowed to have different
466 names, because these cannot be distinguished in PCRE, and this has caused
467 confusion. (This is a difference from Perl.)
468
469 28. When duplicate subpattern names are present (necessarily with different
470 numbers, as required by 27 above), and a test is made by name in a
471 conditional pattern, either for a subpattern having been matched, or for
472 recursion in such a pattern, all the associated numbered subpatterns are
473 tested, and the overall condition is true if the condition is true for any
474 one of them. This is the way Perl works, and is also more like the way
475 testing by number works.
476
477
478 Version 7.9 11-Apr-09
479 ---------------------
480
481 1. When building with support for bzlib/zlib (pcregrep) and/or readline
482 (pcretest), all targets were linked against these libraries. This included
483 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
484 libraries. This caused unwanted dependencies to be created. This problem
485 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
486 pcretest is linked with readline.
487
488 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
489 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
490 moved outside it again, because FALSE and TRUE are already defined in AIX,
491 but BOOL is not.
492
493 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
494 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
495
496 4. The pcregrep documentation said spaces were inserted as well as colons (or
497 hyphens) following file names and line numbers when outputting matching
498 lines. This is not true; no spaces are inserted. I have also clarified the
499 wording for the --colour (or --color) option.
500
501 5. In pcregrep, when --colour was used with -o, the list of matching strings
502 was not coloured; this is different to GNU grep, so I have changed it to be
503 the same.
504
505 6. When --colo(u)r was used in pcregrep, only the first matching substring in
506 each matching line was coloured. Now it goes on to look for further matches
507 of any of the test patterns, which is the same behaviour as GNU grep.
508
509 7. A pattern that could match an empty string could cause pcregrep to loop; it
510 doesn't make sense to accept an empty string match in pcregrep, so I have
511 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
512 seems to be how GNU grep behaves.
513
514 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
515 start or after a newline", because the conditional assertion was not being
516 correctly handled. The rule now is that both the assertion and what follows
517 in the first alternative must satisfy the test.
518
519 9. If auto-callout was enabled in a pattern with a conditional group whose
520 condition was an assertion, PCRE could crash during matching, both with
521 pcre_exec() and pcre_dfa_exec().
522
523 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
524 used for matching.
525
526 11. Unicode property support in character classes was not working for
527 characters (bytes) greater than 127 when not in UTF-8 mode.
528
529 12. Added the -M command line option to pcretest.
530
531 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
532
533 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
534
535 16. Added comments and documentation about mis-use of no_arg in the C++
536 wrapper.
537
538 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
539 from Martin Jerabek that uses macro names for all relevant character and
540 string constants.
541
542 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
543 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
544 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
545 these, but not everybody uses configure.
546
547 19. A conditional group that had only one branch was not being correctly
548 recognized as an item that could match an empty string. This meant that an
549 enclosing group might also not be so recognized, causing infinite looping
550 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
551 with the subject "ab", where knowledge that the repeated group can match
552 nothing is needed in order to break the loop.
553
554 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
555 exec(), but without supplying a callout function, matching went wrong.
556
557 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
558 leak if the size of the offset vector was greater than 30. When the vector
559 is smaller, the saved offsets during recursion go onto a local stack
560 vector, but for larger vectors malloc() is used. It was failing to free
561 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
562 error, in fact).
563
564 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
565 heapframe that is used only when UTF-8 support is enabled. This caused no
566 problem, but was untidy.
567
568 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
569 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
570 included within another project.
571
572 24. Steven Van Ingelgem's patches to add more options to the CMake support,
573 slightly modified by me:
574
575 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
576 not building pcregrep.
577
578 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
579 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
580
581 25. Forward references, both numeric and by name, in patterns that made use of
582 duplicate group numbers, could behave incorrectly or give incorrect errors,
583 because when scanning forward to find the reference group, PCRE was not
584 taking into account the duplicate group numbers. A pattern such as
585 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
586
587 26. Changed a few more instances of "const unsigned char *" to USPTR, making
588 the feature of a custom pointer more persuasive (as requested by a user).
589
590 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
591 pcretest.c, inside #ifndefs, because it seems they are sometimes already
592 pre-defined.
593
594 28. Added support for (*UTF8) at the start of a pattern.
595
596 29. Arrange for flags added by the "release type" setting in CMake to be shown
597 in the configuration summary.
598
599
600 Version 7.8 05-Sep-08
601 ---------------------
602
603 1. Replaced UCP searching code with optimized version as implemented for Ad
604 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
605 stage table and inline lookup instead of a function, giving speed ups of 2
606 to 5 times on some simple patterns that I tested. Permission was given to
607 distribute the MultiStage2.py script that generates the tables (it's not in
608 the tarball, but is in the Subversion repository).
609
610 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
611 scripts.
612
613 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
614 a group with a zero qualifier. The result of the study could be incorrect,
615 or the function might crash, depending on the pattern.
616
617 4. Caseless matching was not working for non-ASCII characters in back
618 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
619 It now works when Unicode Property Support is available.
620
621 5. In pcretest, an escape such as \x{de} in the data was always generating
622 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
623 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
624 truncation.
625
626 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
627
628 7. Added two (int) casts to pcregrep when printing the difference of two
629 pointers, in case they are 64-bit values.
630
631 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
632 test 2 if it fails.
633
634 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
635 and a #define of that name to empty if it is not externally set. This is to
636 allow users of MSVC to set it if necessary.
637
638 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
639 the convenience functions in the pcre_get.c source file.
640
641 11. An option change at the start of a pattern that had top-level alternatives
642 could cause overwriting and/or a crash. This command provoked a crash in
643 some environments:
644
645 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
646
647 This potential security problem was recorded as CVE-2008-2371.
648
649 12. For a pattern where the match had to start at the beginning or immediately
650 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
651 pcre_dfa_exec() could read past the end of the passed subject if there was
652 no match. To help with detecting such bugs (e.g. with valgrind), I modified
653 pcretest so that it places the subject at the end of its malloc-ed buffer.
654
655 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
656 exec() might read past the end of the data buffer in UTF-8 mode.
657
658 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
659 the data contained the byte 0x85 as part of a UTF-8 character within its
660 first line. This applied both to normal and DFA matching.
661
662 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
663 /^[^d]*?$/8 failed to match "abc".
664
665 16. Added a missing copyright notice to pcrecpp_internal.h.
666
667 17. Make it more clear in the documentation that values returned from
668 pcre_exec() in ovector are byte offsets, not character counts.
669
670 18. Tidied a few places to stop certain compilers from issuing warnings.
671
672 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
673 supplied by Stefan Weber. I made a further small update for 7.8 because
674 there is a change of source arrangements: the pcre_searchfuncs.c module is
675 replaced by pcre_ucd.c.
676
677
678 Version 7.7 07-May-08
679 ---------------------
680
681 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
682 a string to a long long, pretend we don't even have a long long." This is
683 done by checking for the strtoq, strtoll, and _strtoi64 functions.
684
685 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
686 pre-7.6 versions, which defined a global no_arg variable instead of putting
687 it in the RE class. (See also #8 below.)
688
689 3. Remove a line of dead code, identified by coverity and reported by Nuno
690 Lopes.
691
692 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
693
694 (1) The include/exclude patterns were being applied to the whole pathnames
695 of files, instead of just to the final components.
696
697 (2) If there was more than one level of directory, the subdirectories were
698 skipped unless they satisfied the include/exclude conditions. This is
699 inconsistent with GNU grep (and could even be seen as contrary to the
700 pcregrep specification - which I improved to make it absolutely clear).
701 The action now is always to scan all levels of directory, and just
702 apply the include/exclude patterns to regular files.
703
704 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
705 --exclude_dir in the tests to avoid scanning .svn directories.
706
707 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
708 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
709 doesn't support NULs in patterns.
710
711 7. Added some missing "const"s to declarations of static tables in
712 pcre_compile.c and pcre_dfa_exec.c.
713
714 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
715 caused by fix #2 above. (Subsequently also a second patch to fix the
716 first patch. And a third patch - this was a messy problem.)
717
718 9. Applied Craig's patch to remove the use of push_back().
719
720 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
721 matching function regexec().
722
723 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
724 which, however, unlike Perl's \g{...}, are subroutine calls, not back
725 references. PCRE supports relative numbers with this syntax (I don't think
726 Oniguruma does).
727
728 12. Previously, a group with a zero repeat such as (...){0} was completely
729 omitted from the compiled regex. However, this means that if the group
730 was called as a subroutine from elsewhere in the pattern, things went wrong
731 (an internal error was given). Such groups are now left in the compiled
732 pattern, with a new opcode that causes them to be skipped at execution
733 time.
734
735 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
736 to the way PCRE behaves:
737
738 (a) A lone ] character is dis-allowed (Perl treats it as data).
739
740 (b) A back reference to an unmatched subpattern matches an empty string
741 (Perl fails the current match path).
742
743 (c) A data ] in a character class must be notated as \] because if the
744 first data character in a class is ], it defines an empty class. (In
745 Perl it is not possible to have an empty class.) The empty class []
746 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
747 The negative empty class [^] matches any one character, independently
748 of the DOTALL setting.
749
750 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
751 non-existent subpattern following a character class starting with ']' and
752 containing () gave an internal compiling error instead of "reference to
753 non-existent subpattern". Fortunately, when the pattern did exist, the
754 compiled code was correct. (When scanning forwards to check for the
755 existencd of the subpattern, it was treating the data ']' as terminating
756 the class, so got the count wrong. When actually compiling, the reference
757 was subsequently set up correctly.)
758
759 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
760 it was being rejected as not supported by pcre_dfa_exec(), even though
761 other assertions are supported. I have made pcre_dfa_exec() support
762 (*FAIL).
763
764 16. The implementation of 13c above involved the invention of a new opcode,
765 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
766 cannot be changed at match time, I realized I could make a small
767 improvement to matching performance by compiling OP_ALLANY instead of
768 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
769 on the OP_ANY path.
770
771 17. Compiling pcretest on Windows with readline support failed without the
772 following two fixes: (1) Make the unistd.h include conditional on
773 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
774
775 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
776 ncurses library to be included for pcretest when ReadLine support is
777 requested, but also to allow for it to be overridden. This patch came from
778 Daniel Bergström.
779
780 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
781 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
782 any errors with the current Unicode tables. Thanks to Peter Kankowski for
783 spotting this.
784
785
786 Version 7.6 28-Jan-08
787 ---------------------
788
789 1. A character class containing a very large number of characters with
790 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
791 overflow.
792
793 2. Patch to cut out the "long long" test in pcrecpp_unittest when
794 HAVE_LONG_LONG is not defined.
795
796 3. Applied Christian Ehrlicher's patch to update the CMake build files to
797 bring them up to date and include new features. This patch includes:
798
799 - Fixed PH's badly added libz and libbz2 support.
800 - Fixed a problem with static linking.
801 - Added pcredemo. [But later removed - see 7 below.]
802 - Fixed dftables problem and added an option.
803 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
804 HAVE_LONG_LONG.
805 - Added readline support for pcretest.
806 - Added an listing of the option settings after cmake has run.
807
808 4. A user submitted a patch to Makefile that makes it easy to create
809 "pcre.dll" under mingw when using Configure/Make. I added stuff to
810 Makefile.am that cause it to include this special target, without
811 affecting anything else. Note that the same mingw target plus all
812 the other distribution libraries and programs are now supported
813 when configuring with CMake (see 6 below) instead of with
814 Configure/Make.
815
816 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
817 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
818 exported in the Windows port". It has not yet been confirmed that the patch
819 solves the problem, but it does no harm.
820
821 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
822 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
823 with CMake, and also correct the comment about stack recursion.
824
825 7. Remove the automatic building of pcredemo from the ./configure system and
826 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
827 of a program that users should build themselves after PCRE is installed, so
828 building it automatically is not really right. What is more, it gave
829 trouble in some build environments.
830
831 8. Further tidies to CMakeLists.txt from Sheri and Christian.
832
833
834 Version 7.5 10-Jan-08
835 ---------------------
836
837 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
838 values in parens when parsing an RE using the C++ wrapper."
839
840 2. Negative specials like \S did not work in character classes in UTF-8 mode.
841 Characters greater than 255 were excluded from the class instead of being
842 included.
843
844 3. The same bug as (2) above applied to negated POSIX classes such as
845 [:^space:].
846
847 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
848 defined or documented. It seems to have been a typo for PCRE_STATIC, so
849 I have changed it.
850
851 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
852 first named subpattern) and a construct such as (?&a) would reference the
853 first named subpattern whose name started with "a" (in other words, the
854 length check was missing). Both these problems are fixed. "Subpattern name
855 expected" is now given for (?&) (a zero-length name), and this patch also
856 makes it give the same error for \k'' (previously it complained that that
857 was a reference to a non-existent subpattern).
858
859 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
860 this is right because (?- can be followed by option settings as well as by
861 digits. I have, however, made the messages clearer.
862
863 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
864 than the number used in the conditional) now cause a compile-time error.
865 This is actually not compatible with Perl, which accepts such patterns, but
866 treats the conditional as always being FALSE (as PCRE used to), but it
867 seems to me that giving a diagnostic is better.
868
869 8. Change "alphameric" to the more common word "alphanumeric" in comments
870 and messages.
871
872 9. Fix two occurrences of "backslash" in comments that should have been
873 "backspace".
874
875 10. Remove two redundant lines of code that can never be obeyed (their function
876 was moved elsewhere).
877
878 11. The program that makes PCRE's Unicode character property table had a bug
879 which caused it to generate incorrect table entries for sequences of
880 characters that have the same character type, but are in different scripts.
881 It amalgamated them into a single range, with the script of the first of
882 them. In other words, some characters were in the wrong script. There were
883 thirteen such cases, affecting characters in the following ranges:
884
885 U+002b0 - U+002c1
886 U+0060c - U+0060d
887 U+0061e - U+00612
888 U+0064b - U+0065e
889 U+0074d - U+0076d
890 U+01800 - U+01805
891 U+01d00 - U+01d77
892 U+01d9b - U+01dbf
893 U+0200b - U+0200f
894 U+030fc - U+030fe
895 U+03260 - U+0327f
896 U+0fb46 - U+0fbb1
897 U+10450 - U+1049d
898
899 12. The -o option (show only the matching part of a line) for pcregrep was not
900 compatible with GNU grep in that, if there was more than one match in a
901 line, it showed only the first of them. It now behaves in the same way as
902 GNU grep.
903
904 13. If the -o and -v options were combined for pcregrep, it printed a blank
905 line for every non-matching line. GNU grep prints nothing, and pcregrep now
906 does the same. The return code can be used to tell if there were any
907 non-matching lines.
908
909 14. Added --file-offsets and --line-offsets to pcregrep.
910
911 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
912 infinitely looping recursion. The bug was that positive lookaheads were not
913 being skipped when checking for a possible empty match (negative lookaheads
914 and both kinds of lookbehind were skipped).
915
916 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
917 inclusion of <windows.h> to before rather than after the definition of
918 INVALID_FILE_ATTRIBUTES (patch from David Byron).
919
920 17. Specifying a possessive quantifier with a specific limit for a Unicode
921 character property caused pcre_compile() to compile bad code, which led at
922 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
923 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
924 caused the error; without that there was no problem.
925
926 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
927
928 19. Added --enable-pcretest-libreadline.
929
930 20. In pcrecpp.cc, the variable 'count' was incremented twice in
931 RE::GlobalReplace(). As a result, the number of replacements returned was
932 double what it should be. I removed one of the increments, but Craig sent a
933 later patch that removed the other one (the right fix) and added unit tests
934 that check the return values (which was not done before).
935
936 21. Several CMake things:
937
938 (1) Arranged that, when cmake is used on Unix, the libraries end up with
939 the names libpcre and libpcreposix, not just pcre and pcreposix.
940
941 (2) The above change means that pcretest and pcregrep are now correctly
942 linked with the newly-built libraries, not previously installed ones.
943
944 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
945
946 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
947 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
948 UTF-8 newline character). The key issue is that the pattern starts .*;
949 this means that the match must be either at the beginning, or after a
950 newline. The bug was in the code for advancing after a failed match and
951 checking that the new position followed a newline. It was not taking
952 account of UTF-8 characters correctly.
953
954 23. PCRE was behaving differently from Perl in the way it recognized POSIX
955 character classes. PCRE was not treating the sequence [:...:] as a
956 character class unless the ... were all letters. Perl, however, seems to
957 allow any characters between [: and :], though of course it rejects as
958 unknown any "names" that contain non-letters, because all the known class
959 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
960 for example, whereas PCRE did not - it did not recognize a POSIX character
961 class. This seemed a bit dangerous, so the code has been changed to be
962 closer to Perl. The behaviour is not identical to Perl, because PCRE will
963 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
964 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
965 Perl does, and where it didn't before.
966
967 24. Rewrite so as to remove the single use of %n from pcregrep because in some
968 Windows environments %n is disabled by default.
969
970
971 Version 7.4 21-Sep-07
972 ---------------------
973
974 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
975 means that a class such as [\s] counted as "explicit reference to CR or
976 LF". That isn't really right - the whole point of the change was to try to
977 help when there was an actual mention of one of the two characters. So now
978 the change happens only if \r or \n (or a literal CR or LF) character is
979 encountered.
980
981 2. The 32-bit options word was also used for 6 internal flags, but the numbers
982 of both had grown to the point where there were only 3 bits left.
983 Fortunately, there was spare space in the data structure, and so I have
984 moved the internal flags into a new 16-bit field to free up more option
985 bits.
986
987 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
988 but did not set the internal JCHANGED flag - either of these is enough to
989 control the way the "get" function works - but the PCRE_INFO_JCHANGED
990 facility is supposed to tell if (?J) was ever used, so now (?J) at the
991 start sets both bits.
992
993 4. Added options (at build time, compile time, exec time) to change \R from
994 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
995
996 5. doc/pcresyntax.html was missing from the distribution.
997
998 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
999 compatibility, even though it is no longer used.
1000
1001 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1002 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1003 windows.h file is present (where different names are used). [This was
1004 reversed later after testing - see 16 below.]
1005
1006 8. Changed all #include <config.h> to #include "config.h". There were also
1007 some further <pcre.h> cases that I changed to "pcre.h".
1008
1009 9. When pcregrep was used with the --colour option, it missed the line ending
1010 sequence off the lines that it output.
1011
1012 10. It was pointed out to me that arrays of string pointers cause lots of
1013 relocations when a shared library is dynamically loaded. A technique of
1014 using a single long string with a table of offsets can drastically reduce
1015 these. I have refactored PCRE in four places to do this. The result is
1016 dramatic:
1017
1018 Originally: 290
1019 After changing UCP table: 187
1020 After changing error message table: 43
1021 After changing table of "verbs" 36
1022 After changing table of Posix names 22
1023
1024 Thanks to the folks working on Gregex for glib for this insight.
1025
1026 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1027 unicode-properties was also set.
1028
1029 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1030
1031 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1032 checked only for CRLF.
1033
1034 14. Added casts to pcretest.c to avoid compiler warnings.
1035
1036 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1037
1038 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1039 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1040 entirely. This removes changes made in 7 above.
1041
1042 17. The CMake files have been updated, and there is now more information about
1043 building with CMake in the NON-UNIX-USE document.
1044
1045
1046 Version 7.3 28-Aug-07
1047 ---------------------
1048
1049 1. In the rejigging of the build system that eventually resulted in 7.1, the
1050 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1051 brackets there is not right, since it causes compilers to look for an
1052 installed pcre.h, not the version that is in the source that is being
1053 compiled (which of course may be different). I have changed it back to:
1054
1055 #include "pcre.h"
1056
1057 I have a vague recollection that the change was concerned with compiling in
1058 different directories, but in the new build system, that is taken care of
1059 by the VPATH setting the Makefile.
1060
1061 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1062 when the subject happened to end in the byte 0x85 (e.g. if the last
1063 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1064 characters but of course it shouldn't be taken as a newline when it is part
1065 of another character. The bug was that, for an unlimited repeat of . in
1066 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1067 characters when looking for a newline.
1068
1069 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1070
1071 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1072 in debug output.
1073
1074 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1075 long printing in the pcrecpp unittest when running under MinGW.
1076
1077 6. ESC_K was left out of the EBCDIC table.
1078
1079 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1080 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1081 limit also applies to "virtual nesting" when a pattern is recursive, and in
1082 this case 1000 isn't so big. I have been able to remove this limit at the
1083 expense of backing off one optimization in certain circumstances. Normally,
1084 when pcre_exec() would call its internal match() function recursively and
1085 immediately return the result unconditionally, it uses a "tail recursion"
1086 feature to save stack. However, when a subpattern that can match an empty
1087 string has an unlimited repetition quantifier, it no longer makes this
1088 optimization. That gives it a stack frame in which to save the data for
1089 checking that an empty string has been matched. Previously this was taken
1090 from the 1000-entry workspace that had been reserved. So now there is no
1091 explicit limit, but more stack is used.
1092
1093 8. Applied Daniel's patches to solve problems with the import/export magic
1094 syntax that is required for Windows, and which was going wrong for the
1095 pcreposix and pcrecpp parts of the library. These were overlooked when this
1096 problem was solved for the main library.
1097
1098 9. There were some crude static tests to avoid integer overflow when computing
1099 the size of patterns that contain repeated groups with explicit upper
1100 limits. As the maximum quantifier is 65535, the maximum group length was
1101 set at 30,000 so that the product of these two numbers did not overflow a
1102 32-bit integer. However, it turns out that people want to use groups that
1103 are longer than 30,000 bytes (though not repeat them that many times).
1104 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1105 made it possible to implement the integer overflow checks in a much more
1106 dynamic way, which I have now done. The artificial limitation on group
1107 length has been removed - we now have only the limit on the total length of
1108 the compiled pattern, which depends on the LINK_SIZE setting.
1109
1110 10. Fixed a bug in the documentation for get/copy named substring when
1111 duplicate names are permitted. If none of the named substrings are set, the
1112 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1113 empty string.
1114
1115 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1116 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1117 because the ] is interpreted as the first data character and the
1118 terminating ] is not found. PCRE has been made compatible with Perl in this
1119 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1120 cause memory overwriting.
1121
1122 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1123 string has been matched (to stop an infinite loop). It was not recognizing
1124 a conditional subpattern that could match an empty string if that
1125 subpattern was within another subpattern. For example, it looped when
1126 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1127 condition was not nested. This bug has been fixed.
1128
1129 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1130 past the start of the subject in the presence of bytes with the top bit
1131 set, for example "\x8aBCD".
1132
1133 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1134 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1135
1136 14. Optimized (?!) to (*FAIL).
1137
1138 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1139 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1140 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1141 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1142 does: it's just the validity check that is more restrictive.
1143
1144 16. Inserted checks for integer overflows during escape sequence (backslash)
1145 processing, and also fixed erroneous offset values for syntax errors during
1146 backslash processing.
1147
1148 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1149 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1150
1151 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1152 caused an overrun.
1153
1154 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1155 something other than just ASCII characters) inside a group that had an
1156 unlimited repeat caused a loop at compile time (while checking to see
1157 whether the group could match an empty string).
1158
1159 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1160 [\P{Any}] did so. (Error in the code for printing property names.)
1161
1162 21. An orphan \E inside a character class could cause a crash.
1163
1164 22. A repeated capturing bracket such as (A)? could cause a wild memory
1165 reference during compilation.
1166
1167 23. There are several functions in pcre_compile() that scan along a compiled
1168 expression for various reasons (e.g. to see if it's fixed length for look
1169 behind). There were bugs in these functions when a repeated \p or \P was
1170 present in the pattern. These operators have additional parameters compared
1171 with \d, etc, and these were not being taken into account when moving along
1172 the compiled data. Specifically:
1173
1174 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1175 length.
1176
1177 (b) An item such as \pL+ within a repeated group could cause crashes or
1178 loops.
1179
1180 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1181 "reference to non-existent subpattern" error.
1182
1183 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1184
1185 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1186 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1187
1188 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1189
1190 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1191 character were causing crashes (broken optimization).
1192
1193 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1194 \p or \P) caused a compile-time loop.
1195
1196 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1197 break. For example, the unstudied pattern [\r\n]A does not match the string
1198 "\r\nA" because change 7.0/46 below moves the current point on by two
1199 characters after failing to match at the start. However, the pattern \nA
1200 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1201 the same is true. There doesn't seem any very clean way out of this, but
1202 what I have chosen to do makes the common cases work: PCRE now takes note
1203 of whether there can be an explicit match for \r or \n anywhere in the
1204 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1205 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1206 pattern has explicit CR or LF references.
1207
1208 29. Added (*CR) etc for changing newline setting at start of pattern.
1209
1210
1211 Version 7.2 19-Jun-07
1212 ---------------------
1213
1214 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1215 which is apparently normally available under Windows.
1216
1217 2. Re-jig the pcregrep tests with different newline settings in an attempt
1218 to make them independent of the local environment's newline setting.
1219
1220 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1221
1222 4. Some of the "internals" tests were previously cut out when the link size
1223 was not 2, because the output contained actual offsets. The recent new
1224 "Z" feature of pcretest means that these can be cut out, making the tests
1225 usable with all link sizes.
1226
1227 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1228 stack recursion. This gives a massive performance boost under BSD, but just
1229 a small improvement under Linux. However, it saves one field in the frame
1230 in all cases.
1231
1232 6. Added more features from the forthcoming Perl 5.10:
1233
1234 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1235 recursion call. It refers to the nth most recently opened parentheses.
1236
1237 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1238 to be opened parentheses.
1239
1240 (c) Conditions that refer to capturing parentheses can be specified
1241 relatively, for example, (?(-2)... or (?(+3)...
1242
1243 (d) \K resets the start of the current match so that everything before
1244 is not part of it.
1245
1246 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1247
1248 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1249 reference syntax.
1250
1251 (g) (?| introduces a group in which the numbering of parentheses in each
1252 alternative starts with the same number.
1253
1254 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1255
1256 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1257 PCRE_INFO_JCHANGED.
1258
1259 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1260 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1261 for detecting groups that can match an empty string.
1262
1263 9. A pattern with a very large number of alternatives (more than several
1264 hundred) was running out of internal workspace during the pre-compile
1265 phase, where pcre_compile() figures out how much memory will be needed. A
1266 bit of new cunning has reduced the workspace needed for groups with
1267 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1268 workspace instead of running out of the 4096 that are available.
1269
1270 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1271
1272 11. Applied patch from Google to remove an optimization that didn't quite work.
1273 The report of the bug said:
1274
1275 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1276 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1277 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1278
1279 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1280 it matched the wrong number of bytes.
1281
1282
1283 Version 7.1 24-Apr-07
1284 ---------------------
1285
1286 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1287 that is more "standard", making use of automake and other Autotools. There
1288 is some re-arrangement of the files and adjustment of comments consequent
1289 on this.
1290
1291 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1292 for recursive directory scanning broke on some systems because the files
1293 are not scanned in any specific order and on different systems the order
1294 was different. A call to "sort" has been inserted into RunGrepTest for the
1295 approprate test as a short-term fix. In the longer term there may be an
1296 alternative.
1297
1298 3. I had an email from Eric Raymond about problems translating some of PCRE's
1299 man pages to HTML (despite the fact that I distribute HTML pages, some
1300 people do their own conversions for various reasons). The problems
1301 concerned the use of low-level troff macros .br and .in. I have therefore
1302 removed all such uses from the man pages (some were redundant, some could
1303 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1304 HTML has been updated to handle .nf/.fi and to complain if it encounters
1305 .br or .in.
1306
1307 4. Updated comments in configure.ac that get placed in config.h.in and also
1308 arranged for config.h to be included in the distribution, with the name
1309 config.h.generic, for the benefit of those who have to compile without
1310 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1311
1312 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1313 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1314 makevp.bat for the current PCRE, using the additional files
1315 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1316
1317 6. A Windows user reported a minor discrepancy with test 2, which turned out
1318 to be caused by a trailing space on an input line that had got lost in his
1319 copy. The trailing space was an accident, so I've just removed it.
1320
1321 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1322 that is needed.
1323
1324 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1325 as "const" (a) because they are and (b) because it helps the PHP
1326 maintainers who have recently made a script to detect big data structures
1327 in the php code that should be moved to the .rodata section. I remembered
1328 to update Builducptable as well, so it won't revert if ucptable.h is ever
1329 re-created.
1330
1331 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1332 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1333 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1334 support is not required. This saves 1.5-2K of code, which is important in
1335 some applications.
1336
1337 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1338 so as not to refer to the tables, even though these functions will never be
1339 called when UTF-8 support is disabled. Otherwise there are problems with a
1340 shared library.
1341
1342 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1343
1344 (a) It was defining its arguments as char * instead of void *.
1345
1346 (b) It was assuming that all moves were upwards in memory; this was true
1347 a long time ago when I wrote it, but is no longer the case.
1348
1349 The emulated memove() is provided for those environments that have neither
1350 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1351 is clearly not the case, as these two bugs were recently reported.
1352
1353 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1354 and Detrail to create the HTML documentation, the .txt form of the man
1355 pages, and it removes trailing spaces from listed files. It also creates
1356 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1357 case, it wraps all the #defines with #ifndefs. This script should be run
1358 before "make dist".
1359
1360 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1361 with Unicode property support.
1362
1363 (a) For a maximizing quantifier, if the two different cases of the
1364 character were of different lengths in their UTF-8 codings (there are
1365 some cases like this - I found 11), and the matching function had to
1366 back up over a mixture of the two cases, it incorrectly assumed they
1367 were both the same length.
1368
1369 (b) When PCRE was configured to use the heap rather than the stack for
1370 recursion during matching, it was not correctly preserving the data for
1371 the other case of a UTF-8 character when checking ahead for a match
1372 while processing a minimizing repeat. If the check also involved
1373 matching a wide character, but failed, corruption could cause an
1374 erroneous result when trying to check for a repeat of the original
1375 character.
1376
1377 13. Some tidying changes to the testing mechanism:
1378
1379 (a) The RunTest script now detects the internal link size and whether there
1380 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1381 values substituted by "configure". (The RunGrepTest script already did
1382 this for UTF-8.) The configure.ac script no longer substitutes the
1383 relevant variables.
1384
1385 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1386 with length and offset values. This means that the output is different
1387 for different internal link sizes. Test 2 is skipped for link sizes
1388 other than 2 because of this, bypassing the problem. Unfortunately,
1389 there was also a test in test 3 (the locale tests) that used /B and
1390 failed for link sizes other than 2. Rather than cut the whole test out,
1391 I have added a new /Z option to pcretest that replaces the length and
1392 offset values with spaces. This is now used to make test 3 independent
1393 of link size. (Test 2 will be tidied up later.)
1394
1395 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1396 segmentation fault instead of returning the appropriate error message.
1397
1398 15. In multiline mode when the newline sequence was set to "any", the pattern
1399 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1400 This doesn't seem right; it now treats the CRLF combination as the line
1401 ending, and so does not match in that case. It's only a pattern such as ^$
1402 that would hit this one: something like ^ABC$ would have failed after \r
1403 and then tried again after \r\n.
1404
1405 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1406 in an attempt to make files that differ only in their line terminators
1407 compare equal. This works on Linux.
1408
1409 17. Under certain error circumstances pcregrep might try to free random memory
1410 as it exited. This is now fixed, thanks to valgrind.
1411
1412 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1413 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1414 was because its rules for how to advance for /g after matching an empty
1415 string at the end of a line did not allow for this case. They now check for
1416 it specially.
1417
1418 20. pcretest is supposed to handle patterns and data of any length, by
1419 extending its buffers when necessary. It was getting this wrong when the
1420 buffer for a data line had to be extended.
1421
1422 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1423 CRLF as a newline sequence.
1424
1425 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1426 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1427 I have nevertheless tidied it up.
1428
1429 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1430
1431 24. Added a man page for pcre-config.
1432
1433
1434 Version 7.0 19-Dec-06
1435 ---------------------
1436
1437 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1438 moving to gcc 4.1.1.
1439
1440 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1441 sys/time.h, which is documented as needed for this function. It doesn't
1442 seem to matter on Linux, but it showed up on some releases of OS X.
1443
1444 3. It seems that there are systems where bytes whose values are greater than
1445 127 match isprint() in the "C" locale. The "C" locale should be the
1446 default when a C program starts up. In most systems, only ASCII printing
1447 characters match isprint(). This difference caused the output from pcretest
1448 to vary, making some of the tests fail. I have changed pcretest so that:
1449
1450 (a) When it is outputting text in the compiled version of a pattern, bytes
1451 other than 32-126 are always shown as hex escapes.
1452
1453 (b) When it is outputting text that is a matched part of a subject string,
1454 it does the same, unless a different locale has been set for the match
1455 (using the /L modifier). In this case, it uses isprint() to decide.
1456
1457 4. Fixed a major bug that caused incorrect computation of the amount of memory
1458 required for a compiled pattern when options that changed within the
1459 pattern affected the logic of the preliminary scan that determines the
1460 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1461 that the computed length was too small. The symptoms of this bug were
1462 either the PCRE error "internal error: code overflow" from pcre_compile(),
1463 or a glibc crash with a message such as "pcretest: free(): invalid next
1464 size (fast)". Examples of patterns that provoked this bug (shown in
1465 pcretest format) are:
1466
1467 /(?-x: )/x
1468 /(?x)(?-x: \s*#\s*)/
1469 /((?i)[\x{c0}])/8
1470 /(?i:[\x{c0}])/8
1471
1472 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1473 is now done differently.
1474
1475 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1476 wrapper classes; (b) implement a new function in the C++ scanner that is
1477 more efficient than the old way of doing things because it avoids levels of
1478 recursion in the regex matching; (c) add a paragraph to the documentation
1479 for the FullMatch() function.
1480
1481 6. The escape sequence \n was being treated as whatever was defined as
1482 "newline". Not only was this contrary to the documentation, which states
1483 that \n is character 10 (hex 0A), but it also went horribly wrong when
1484 "newline" was defined as CRLF. This has been fixed.
1485
1486 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1487 was being set to -1 for the "end of line" case (supposedly a value that no
1488 character can have). Though this value is never used (the check for end of
1489 line is "zero bytes in current character"), it caused compiler complaints.
1490 I've changed it to 0xffffffff.
1491
1492 8. In pcre_version.c, the version string was being built by a sequence of
1493 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1494 string (as it is for production releases) called a macro with an empty
1495 argument. The C standard says the result of this is undefined. The gcc
1496 compiler treats it as an empty string (which was what was wanted) but it is
1497 reported that Visual C gives an error. The source has been hacked around to
1498 avoid this problem.
1499
1500 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1501 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1502 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1503 of them did).
1504
1505 10. Originally, pcretest opened its input and output without "b"; then I was
1506 told that "b" was needed in some environments, so it was added for release
1507 5.0 to both the input and output. (It makes no difference on Unix-like
1508 systems.) Later I was told that it is wrong for the input on Windows. I've
1509 now abstracted the modes into two macros, to make it easier to fiddle with
1510 them, and removed "b" from the input mode under Windows.
1511
1512 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1513
1514 12. Added -help and --help to pcretest as an official way of being reminded
1515 of the options.
1516
1517 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1518 and pcrecpp.cc because they annoy compilers at high warning levels.
1519
1520 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1521
1522 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1523 scripts are not C programs :-) and which was not noticed because it works
1524 on Linux.
1525
1526 16. pcretest is supposed to handle any length of pattern and data line (as one
1527 line or as a continued sequence of lines) by extending its input buffer if
1528 necessary. This feature was broken for very long pattern lines, leading to
1529 a string of junk being passed to pcre_compile() if the pattern was longer
1530 than about 50K.
1531
1532 17. I have done a major re-factoring of the way pcre_compile() computes the
1533 amount of memory needed for a compiled pattern. Previously, there was code
1534 that made a preliminary scan of the pattern in order to do this. That was
1535 OK when PCRE was new, but as the facilities have expanded, it has become
1536 harder and harder to keep it in step with the real compile phase, and there
1537 have been a number of bugs (see for example, 4 above). I have now found a
1538 cunning way of running the real compile function in a "fake" mode that
1539 enables it to compute how much memory it would need, while actually only
1540 ever using a few hundred bytes of working memory and without too many
1541 tests of the mode. This should make future maintenance and development
1542 easier. A side effect of this work is that the limit of 200 on the nesting
1543 depth of parentheses has been removed (though this was never a serious
1544 limitation, I suspect). However, there is a downside: pcre_compile() now
1545 runs more slowly than before (30% or more, depending on the pattern). I
1546 hope this isn't a big issue. There is no effect on runtime performance.
1547
1548 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1549 newline (only possible for the last line of a file) and it was a
1550 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1551
1552 19. Added additional timing features to pcretest. (1) The -tm option now times
1553 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1554 separate command line item, by a number that specifies the number of
1555 repeats to use when timing. The default is 50000; this gives better
1556 precision, but takes uncomfortably long for very large patterns.
1557
1558 20. Extended pcre_study() to be more clever in cases where a branch of a
1559 subpattern has no definite first character. For example, (a*|b*)[cd] would
1560 previously give no result from pcre_study(). Now it recognizes that the
1561 first character must be a, b, c, or d.
1562
1563 21. There was an incorrect error "recursive call could loop indefinitely" if
1564 a subpattern (or the entire pattern) that was being tested for matching an
1565 empty string contained only one non-empty item after a nested subpattern.
1566 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1567 incorrectly, because the \d was being skipped in the check.
1568
1569 22. The pcretest program now has a new pattern option /B and a command line
1570 option -b, which is equivalent to adding /B to every pattern. This causes
1571 it to show the compiled bytecode, without the additional information that
1572 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1573 is the same as /B/I).
1574
1575 23. A new optimization is now able automatically to treat some sequences such
1576 as a*b as a*+b. More specifically, if something simple (such as a character
1577 or a simple class like \d) has an unlimited quantifier, and is followed by
1578 something that cannot possibly match the quantified thing, the quantifier
1579 is automatically "possessified".
1580
1581 24. A recursive reference to a subpattern whose number was greater than 39
1582 went wrong under certain circumstances in UTF-8 mode. This bug could also
1583 have affected the operation of pcre_study().
1584
1585 25. Realized that a little bit of performance could be had by replacing
1586 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1587
1588 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1589
1590 27. Possessive quantifiers such as a++ were previously implemented by turning
1591 them into atomic groups such as ($>a+). Now they have their own opcodes,
1592 which improves performance. This includes the automatically created ones
1593 from 23 above.
1594
1595 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1596 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1597 the first matched character to be a colon. This applied both to named and
1598 numbered groups.
1599
1600 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1601
1602 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1603 building PCRE on Windows easier, so I have included it in the distribution.
1604
1605 31. There is now a check in pcretest against a ridiculously large number being
1606 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1607 loop, the loop is abandoned.
1608
1609 32. Forward references to subpatterns in conditions such as (?(2)...) where
1610 subpattern 2 is defined later cause pcre_compile() to search forwards in
1611 the pattern for the relevant set of parentheses. This search went wrong
1612 when there were unescaped parentheses in a character class, parentheses
1613 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1614
1615 33. "Subroutine" calls and backreferences were previously restricted to
1616 referencing subpatterns earlier in the regex. This restriction has now
1617 been removed.
1618
1619 34. Added a number of extra features that are going to be in Perl 5.10. On the
1620 whole, these are just syntactic alternatives for features that PCRE had
1621 previously implemented using the Python syntax or my own invention. The
1622 other formats are all retained for compatibility.
1623
1624 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1625 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1626 also .NET compatible.
1627
1628 (b) A recursion or subroutine call to a named group can now be defined as
1629 (?&name) as well as (?P>name).
1630
1631 (c) A backreference to a named group can now be defined as \k<name> or
1632 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1633 5.10, are also .NET compatible.
1634
1635 (d) A conditional reference to a named group can now use the syntax
1636 (?(<name>) or (?('name') as well as (?(name).
1637
1638 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1639 groups (named and numbered) that are never evaluated inline, but can be
1640 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1641 is always false. There may be only one alternative in such a group.
1642
1643 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1644 as the simple (?(R). The condition is true only if the most recent
1645 recursion is that of the given number or name. It does not search out
1646 through the entire recursion stack.
1647
1648 (g) The escape \gN or \g{N} has been added, where N is a positive or
1649 negative number, specifying an absolute or relative reference.
1650
1651 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1652 some "unreachable code" warnings.
1653
1654 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1655 things, this adds five new scripts.
1656
1657 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1658 There were also incompatibilities regarding the handling of \Q..\E inside
1659 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1660 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1661
1662 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1663 matches an empty string, and forcibly breaks the loop. There were bugs in
1664 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1665 against aaaa the result was just "a" rather than "aaaa", for example. Two
1666 separate and independent bugs (that affected different cases) have been
1667 fixed.
1668
1669 39. Refactored the code to abolish the use of different opcodes for small
1670 capturing bracket numbers. This is a tidy that I avoided doing when I
1671 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1672 The new approach is not only tidier, it makes it possible to reduce the
1673 memory needed to fix the previous bug (38).
1674
1675 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1676 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1677 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1678 mode.
1679
1680 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1681 report.
1682
1683 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1684 copy construction and assignment in the C++ wrapper.
1685
1686 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1687 couple of bugs that could have given wrong results in the "--newline=crlf"
1688 case.
1689
1690 44. Added a number of casts and did some reorganization of signed/unsigned int
1691 variables following suggestions from Dair Grant. Also renamed the variable
1692 "this" as "item" because it is a C++ keyword.
1693
1694 45. Arranged for dftables to add
1695
1696 #include "pcre_internal.h"
1697
1698 to pcre_chartables.c because without it, gcc 4.x may remove the array
1699 definition from the final binary if PCRE is built into a static library and
1700 dead code stripping is activated.
1701
1702 46. For an unanchored pattern, if a match attempt fails at the start of a
1703 newline sequence, and the newline setting is CRLF or ANY, and the next two
1704 characters are CRLF, advance by two characters instead of one.
1705
1706
1707 Version 6.7 04-Jul-06
1708 ---------------------
1709
1710 1. In order to handle tests when input lines are enormously long, pcretest has
1711 been re-factored so that it automatically extends its buffers when
1712 necessary. The code is crude, but this _is_ just a test program. The
1713 default size has been increased from 32K to 50K.
1714
1715 2. The code in pcre_study() was using the value of the re argument before
1716 testing it for NULL. (Of course, in any sensible call of the function, it
1717 won't be NULL.)
1718
1719 3. The memmove() emulation function in pcre_internal.h, which is used on
1720 systems that lack both memmove() and bcopy() - that is, hardly ever -
1721 was missing a "static" storage class specifier.
1722
1723 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1724 containing an extended class (one that cannot be represented by a bitmap
1725 because it contains high-valued characters or Unicode property items, e.g.
1726 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1727 pattern, but PCRE should not loop if you do not (it no longer does).
1728 [Detail: two cases were found: (a) a repeated subpattern containing an
1729 extended class; (b) a recursive reference to a subpattern that followed a
1730 previous extended class. It wasn't skipping over the extended class
1731 correctly when UTF-8 mode was not set.]
1732
1733 5. A negated single-character class was not being recognized as fixed-length
1734 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1735 compile error "lookbehind assertion is not fixed length".
1736
1737 6. The RunPerlTest auxiliary script was showing an unexpected difference
1738 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1739 write a Perl script that can interpret lines of an input file either as
1740 byte characters or as UTF-8, which is what "perltest" was being required to
1741 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1742 can't do is switch easily at run time between having the "use utf8;" pragma
1743 or not. In the end, I fudged it by using the RunPerlTest script to insert
1744 "use utf8;" explicitly for the UTF-8 tests.
1745
1746 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1747 the end of the subject string, contrary to the documentation and to what
1748 Perl does. This was true of both matching functions. Now it matches only at
1749 the start of the subject and immediately after *internal* newlines.
1750
1751 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1752 a pointer to an int instead of a pointer to an unsigned long int. This
1753 caused problems on 64-bit systems.
1754
1755 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1756 instance of the 'standard' template library not being so standard".
1757
1758 10. There was no check on the number of named subpatterns nor the maximum
1759 length of a subpattern name. The product of these values is used to compute
1760 the size of the memory block for a compiled pattern. By supplying a very
1761 long subpattern name and a large number of named subpatterns, the size
1762 computation could be caused to overflow. This is now prevented by limiting
1763 the length of names to 32 characters, and the number of named subpatterns
1764 to 10,000.
1765
1766 11. Subpatterns that are repeated with specific counts have to be replicated in
1767 the compiled pattern. The size of memory for this was computed from the
1768 length of the subpattern and the repeat count. The latter is limited to
1769 65535, but there was no limit on the former, meaning that integer overflow
1770 could in principle occur. The compiled length of a repeated subpattern is
1771 now limited to 30,000 bytes in order to prevent this.
1772
1773 12. Added the optional facility to have named substrings with the same name.
1774
1775 13. Added the ability to use a named substring as a condition, using the
1776 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1777 are numbers (not recommended). Forward references are permitted.
1778
1779 14. Added forward references in named backreferences (if you see what I mean).
1780
1781 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1782 pattern could run off the end of the subject. For example, the pattern
1783 "(?s)(.{1,5})"8 did this with the subject "ab".
1784
1785 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1786 PCRE_CASELESS was set when matching characters that were quantified with ?
1787 or *.
1788
1789 17. A character class other than a single negated character that had a minimum
1790 but no maximum quantifier - for example [ab]{6,} - was not handled
1791 correctly by pce_dfa_exec(). It would match only one character.
1792
1793 18. A valid (though odd) pattern that looked like a POSIX character
1794 class but used an invalid character after [ (for example [[,abc,]]) caused
1795 pcre_compile() to give the error "Failed: internal error: code overflow" or
1796 in some cases to crash with a glibc free() error. This could even happen if
1797 the pattern terminated after [[ but there just happened to be a sequence of
1798 letters, a binary zero, and a closing ] in the memory that followed.
1799
1800 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1801 over the years. Originally (before any Unicode support), just the bottom 8
1802 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1803 output from "man perlunicode" includes this:
1804
1805 The regular expression compiler produces polymorphic opcodes. That
1806 is, the pattern adapts to the data and automatically switches to
1807 the Unicode character scheme when presented with Unicode data--or
1808 instead uses a traditional byte scheme when presented with byte
1809 data.
1810
1811 Sadly, a wide octal escape does not cause a switch, and in a string with
1812 no other multibyte characters, these octal escapes are treated as before.
1813 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1814 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1815 Unicode string.
1816
1817 I have not perpetrated such confusion in PCRE. Up till now, it took just
1818 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1819 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1820 translate to the appropriate multibyte character.
1821
1822 29. Applied some refactoring to reduce the number of warnings from Microsoft
1823 and Borland compilers. This has included removing the fudge introduced
1824 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1825 a warning about an unused variable.
1826
1827 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1828 characters since release 4.0, because Perl (from release 5.004) does not.
1829 [Or at least, is documented not to: some releases seem to be in conflict
1830 with the documentation.] However, when a pattern was studied with
1831 pcre_study() and all its branches started with \s, PCRE still included VT
1832 as a possible starting character. Of course, this did no harm; it just
1833 caused an unnecessary match attempt.
1834
1835 22. Removed a now-redundant internal flag bit that recorded the fact that case
1836 dependency changed within the pattern. This was once needed for "required
1837 byte" processing, but is no longer used. This recovers a now-scarce options
1838 bit. Also moved the least significant internal flag bit to the most-
1839 significant bit of the word, which was not previously used (hangover from
1840 the days when it was an int rather than a uint) to free up another bit for
1841 the future.
1842
1843 23. Added support for CRLF line endings as well as CR and LF. As well as the
1844 default being selectable at build time, it can now be changed at runtime
1845 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1846 specify that it is scanning data with non-default line endings.
1847
1848 24. Changed the definition of CXXLINK to make it agree with the definition of
1849 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1850
1851 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1852 recursions. This makes a big different to stack usage for some patterns.
1853
1854 26. If a subpattern containing a named recursion or subroutine reference such
1855 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1856 the space required for the compiled pattern went wrong and gave too small a
1857 value. Depending on the environment, this could lead to "Failed: internal
1858 error: code overflow at offset 49" or "glibc detected double free or
1859 corruption" errors.
1860
1861 27. Applied patches from Google (a) to support the new newline modes and (b) to
1862 advance over multibyte UTF-8 characters in GlobalReplace.
1863
1864 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1865 difference for some implementation of PCRE in some Windows version.
1866
1867 29. Added some extra testing facilities to pcretest:
1868
1869 \q<number> in a data line sets the "match limit" value
1870 \Q<number> in a data line sets the "match recursion limt" value
1871 -S <number> sets the stack size, where <number> is in megabytes
1872
1873 The -S option isn't available for Windows.
1874
1875
1876 Version 6.6 06-Feb-06
1877 ---------------------
1878
1879 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1880 in pcreposix.h. I have copied the definition from pcre.h.
1881
1882 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1883 because pcre.h is no longer a built file.
1884
1885 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1886 not normally included in the compiled code.
1887
1888
1889 Version 6.5 01-Feb-06
1890 ---------------------
1891
1892 1. When using the partial match feature with pcre_dfa_exec(), it was not
1893 anchoring the second and subsequent partial matches at the new starting
1894 point. This could lead to incorrect results. For example, with the pattern
1895 /1234/, partially matching against "123" and then "a4" gave a match.
1896
1897 2. Changes to pcregrep:
1898
1899 (a) All non-match returns from pcre_exec() were being treated as failures
1900 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1901 error message is output. Some extra information is given for the
1902 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1903 probably the only errors that are likely to be caused by users (by
1904 specifying a regex that has nested indefinite repeats, for instance).
1905 If there are more than 20 of these errors, pcregrep is abandoned.
1906
1907 (b) A binary zero was treated as data while matching, but terminated the
1908 output line if it was written out. This has been fixed: binary zeroes
1909 are now no different to any other data bytes.
1910
1911 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1912 used to set a locale for matching. The --locale=xxxx long option has
1913 been added (no short equivalent) to specify a locale explicitly on the
1914 pcregrep command, overriding the environment variables.
1915
1916 (d) When -B was used with -n, some line numbers in the output were one less
1917 than they should have been.
1918
1919 (e) Added the -o (--only-matching) option.
1920
1921 (f) If -A or -C was used with -c (count only), some lines of context were
1922 accidentally printed for the final match.
1923
1924 (g) Added the -H (--with-filename) option.
1925
1926 (h) The combination of options -rh failed to suppress file names for files
1927 that were found from directory arguments.
1928
1929 (i) Added the -D (--devices) and -d (--directories) options.
1930
1931 (j) Added the -F (--fixed-strings) option.
1932
1933 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1934
1935 (l) Added the --colo(u)r option.
1936
1937 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1938 is not present by default.
1939
1940 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1941 items such as (?R) or (?1), when the recursion could match a number of
1942 alternatives. If it matched one of the alternatives, but subsequently,
1943 outside the recursion, there was a failure, the code tried to back up into
1944 the recursion. However, because of the way PCRE is implemented, this is not
1945 possible, and the result was an incorrect result from the match.
1946
1947 In order to prevent this happening, the specification of recursion has
1948 been changed so that all such subpatterns are automatically treated as
1949 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1950
1951 4. I had overlooked the fact that, in some locales, there are characters for
1952 which isalpha() is true but neither isupper() nor islower() are true. In
1953 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1954 and ordfeminine) are like this. This affected the treatment of \w and \W
1955 when they appeared in character classes, but not when they appeared outside
1956 a character class. The bit map for "word" characters is now created
1957 separately from the results of isalnum() instead of just taking it from the
1958 upper, lower, and digit maps. (Plus the underscore character, of course.)
1959
1960 5. The above bug also affected the handling of POSIX character classes such as
1961 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1962 permanent tables. Instead, the bit maps for such a class were previously
1963 created as the appropriate unions of the upper, lower, and digit bitmaps.
1964 Now they are created by subtraction from the [[:word:]] class, which has
1965 its own bitmap.
1966
1967 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1968 It is created by subtracting the vertical space characters (\x09, \x0a,
1969 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1970 subtraction was done in the overall bitmap for a character class, meaning
1971 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1972 be recognized. This bug has been fixed.
1973
1974 7. Patches from the folks at Google:
1975
1976 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1977 real life, but is still worth protecting against".
1978
1979 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1980 regular expressions".
1981
1982 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1983 have it.
1984
1985 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1986 "configure" and the latter not, in order to fix a problem somebody had
1987 with compiling the Arg class on HP-UX.
1988
1989 (e) Improve the error-handling of the C++ wrapper a little bit.
1990
1991 (f) New tests for checking recursion limiting.
1992
1993 8. The pcre_memmove() function, which is used only if the environment does not
1994 have a standard memmove() function (and is therefore rarely compiled),
1995 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1996 returning a result (though PCRE never actually uses the result).
1997
1998 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1999 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2000 returned instead of calling malloc() with an overflowing number that would
2001 most likely cause subsequent chaos.
2002
2003 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2004
2005 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2006 with this option is matched, the nmatch and pmatch options of regexec() are
2007 ignored.
2008
2009 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2010 provided in case anyone wants to the the POSIX interface with UTF-8
2011 strings.
2012
2013 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2014 C++ linking (needed for some HP-UX environments).
2015
2016 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2017 (unused parameter) and in the pcre_printint() function (omitted "default"
2018 switch label when the default is to do nothing).
2019
2020 15. Added some code to make it possible, when PCRE is compiled as a C++
2021 library, to replace subject pointers for pcre_exec() with a smart pointer
2022 class, thus making it possible to process discontinuous strings.
2023
2024 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2025 much the same function. They were added by different people who were trying
2026 to make PCRE easy to compile on non-Unix systems. It has been suggested
2027 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2028 for compiling on Windows systems. I have therefore replaced it with
2029 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2030 defaults to "extern" for C or "extern C" for C++, which works fine on
2031 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2032 SCOPE with something explicit in config.h. In addition:
2033
2034 (a) pcreposix.h still had just "extern" instead of either of these macros;
2035 I have replaced it with PCRE_DATA_SCOPE.
2036
2037 (b) Functions such as _pcre_xclass(), which are internal to the library,
2038 but external in the C sense, all had PCRE_EXPORT in their definitions.
2039 This is apparently wrong for the Windows case, so I have removed it.
2040 (It makes no difference on Unix-like systems.)
2041
2042 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2043 of recursive calls to match(). This is different to MATCH_LIMIT because
2044 that limits the total number of calls to match(), not all of which increase
2045 the depth of recursion. Limiting the recursion depth limits the amount of
2046 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2047 when PCRE is compiled, and changed at run time. A patch from Google adds
2048 this functionality to the C++ interface.
2049
2050 18. Changes to the handling of Unicode character properties:
2051
2052 (a) Updated the table to Unicode 4.1.0.
2053
2054 (b) Recognize characters that are not in the table as "Cn" (undefined).
2055
2056 (c) I revised the way the table is implemented to a much improved format
2057 which includes recognition of ranges. It now supports the ranges that
2058 are defined in UnicodeData.txt, and it also amalgamates other
2059 characters into ranges. This has reduced the number of entries in the
2060 table from around 16,000 to around 3,000, thus reducing its size
2061 considerably. I realized I did not need to use a tree structure after
2062 all - a binary chop search is just as efficient. Having reduced the
2063 number of entries, I extended their size from 6 bytes to 8 bytes to
2064 allow for more data.
2065
2066 (d) Added support for Unicode script names via properties such as \p{Han}.
2067
2068 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2069 matching that character.
2070
2071 20. When matching a repeated Unicode property with a minimum greater than zero,
2072 (for example \pL{2,}), PCRE could look past the end of the subject if it
2073 reached it while seeking the minimum number of characters. This could
2074 happen only if some of the characters were more than one byte long, because
2075 there is a check for at least the minimum number of bytes.
2076
2077 21. Refactored the implementation of \p and \P so as to be more general, to
2078 allow for more different types of property in future. This has changed the
2079 compiled form incompatibly. Anybody with saved compiled patterns that use
2080 \p or \P will have to recompile them.
2081
2082 22. Added "Any" and "L&" to the supported property types.
2083
2084 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2085 but give a compile time error if the value is greater than 0xff.
2086
2087 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2088 accidentally not being installed or uninstalled.
2089
2090 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2091 made were to insert the current release number. This seemed silly, because
2092 it made things harder for people building PCRE on systems that don't run
2093 "configure". I have turned pcre.h into a distributed file, no longer built
2094 by "configure", with the version identification directly included. There is
2095 no longer a pcre.h.in file.
2096
2097 However, this change necessitated a change to the pcre-config script as
2098 well. It is built from pcre-config.in, and one of the substitutions was the
2099 release number. I have updated configure.ac so that ./configure now finds
2100 the release number by grepping pcre.h.
2101
2102 26. Added the ability to run the tests under valgrind.
2103
2104
2105 Version 6.4 05-Sep-05
2106 ---------------------
2107
2108 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2109 "--" to be printed when multiple files were scanned, even when none of the
2110 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2111 consider it to be a bug, and have restored the previous behaviour.
2112
2113 2. A couple of code tidies to get rid of compiler warnings.
2114
2115 3. The pcretest program used to cheat by referring to symbols in the library
2116 whose names begin with _pcre_. These are internal symbols that are not
2117 really supposed to be visible externally, and in some environments it is
2118 possible to suppress them. The cheating is now confined to including
2119 certain files from the library's source, which is a bit cleaner.
2120
2121 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2122 file's purpose clearer.
2123
2124 5. Reorganized pcre_ucp_findchar().
2125
2126
2127 Version 6.3 15-Aug-05
2128 ---------------------
2129
2130 1. The file libpcre.pc.in did not have general read permission in the tarball.
2131
2132 2. There were some problems when building without C++ support:
2133
2134 (a) If C++ support was not built, "make install" and "make test" still
2135 tried to test it.
2136
2137 (b) There were problems when the value of CXX was explicitly set. Some
2138 changes have been made to try to fix these, and ...
2139
2140 (c) --disable-cpp can now be used to explicitly disable C++ support.
2141
2142 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2143 backslash in a target when C++ was disabled. This confuses some
2144 versions of "make", apparently. Using an intermediate variable solves
2145 this. (Same for CPP_LOBJ.)
2146
2147 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2148 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2149 necessary on certain architectures.
2150
2151 4. Added a setting of -export-symbols-regex to the link command to remove
2152 those symbols that are exported in the C sense, but actually are local
2153 within the library, and not documented. Their names all begin with
2154 "_pcre_". This is not a perfect job, because (a) we have to except some
2155 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2156 available (and never for static libraries). I have made a note to try to
2157 find a way round (a) in the future.
2158
2159
2160 Version 6.2 01-Aug-05
2161 ---------------------
2162
2163 1. There was no test for integer overflow of quantifier values. A construction
2164 such as {1111111111111111} would give undefined results. What is worse, if
2165 a minimum quantifier for a parenthesized subpattern overflowed and became
2166 negative, the calculation of the memory size went wrong. This could have
2167 led to memory overwriting.
2168
2169 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2170
2171 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2172 operating environments where this matters.
2173
2174 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2175 PCRE options from within the C++ wrapper.
2176
2177 5. Named capturing subpatterns were not being correctly counted when a pattern
2178 was compiled. This caused two problems: (a) If there were more than 100
2179 such subpatterns, the calculation of the memory needed for the whole
2180 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2181 back references of the form \12, where the number was greater than 9, were
2182 not recognized as back references, even though there were sufficient
2183 previous subpatterns.
2184
2185 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2186 versions of gcc, e.g. 2.95.4.
2187
2188
2189 Version 6.1 21-Jun-05
2190 ---------------------
2191
2192 1. There was one reference to the variable "posix" in pcretest.c that was not
2193 surrounded by "#if !defined NOPOSIX".
2194
2195 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2196 the cross-check on the old pcre_info() function, for the benefit of the
2197 cut-down version of PCRE that is currently imported into Exim.
2198
2199 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2200 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2201 patterns that start (?i) but I don't think that matters. The use of (?i) is
2202 just an example; this all applies to the other options as well.
2203
2204 4. Since libtool seems to echo the compile commands it is issuing, the output
2205 from "make" can be reduced a bit by putting "@" in front of each libtool
2206 compile command.
2207
2208 5. Patch from the folks at Google for configure.in to be a bit more thorough
2209 in checking for a suitable C++ installation before trying to compile the
2210 C++ stuff. This should fix a reported problem when a compiler was present,
2211 but no suitable headers.
2212
2213 6. The man pages all had just "PCRE" as their title. I have changed them to
2214 be the relevant file name. I have also arranged that these names are
2215 retained in the file doc/pcre.txt, which is a concatenation in text format
2216 of all the man pages except the little individual ones for each function.
2217
2218 7. The NON-UNIX-USE file had not been updated for the different set of source
2219 files that come with release 6. I also added a few comments about the C++
2220 wrapper.
2221
2222
2223 Version 6.0 07-Jun-05
2224 ---------------------
2225
2226 1. Some minor internal re-organization to help with my DFA experiments.
2227
2228 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2229 didn't matter for the library itself when fully configured, but did matter
2230 when compiling without UCP support, or within Exim, where the ucp files are
2231 not imported.
2232
2233 3. Refactoring of the library code to split up the various functions into
2234 different source modules. The addition of the new DFA matching code (see
2235 below) to a single monolithic source would have made it really too
2236 unwieldy, quite apart from causing all the code to be include in a
2237 statically linked application, when only some functions are used. This is
2238 relevant even without the DFA addition now that patterns can be compiled in
2239 one application and matched in another.
2240
2241 The downside of splitting up is that there have to be some external
2242 functions and data tables that are used internally in different modules of
2243 the library but which are not part of the API. These have all had their
2244 names changed to start with "_pcre_" so that they are unlikely to clash
2245 with other external names.
2246
2247 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2248 a different (DFA) algorithm. Although it is slower than the original
2249 function, it does have some advantages for certain types of matching
2250 problem.
2251
2252 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2253 including restarting after a partial match.
2254
2255 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2256 defined when compiling for Windows was sent to me. I have put it into the
2257 code, though I have no means of testing or verifying it.
2258
2259 7. Added the pcre_refcount() auxiliary function.
2260
2261 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2262 match before or at the first newline in the subject string. In pcretest,
2263 the /f option on a pattern can be used to set this.
2264
2265 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2266 would behave wrongly. This has been present in PCRE since release 4.0.
2267
2268 10. A number of changes to the pcregrep command:
2269
2270 (a) Refactored how -x works; insert ^(...)$ instead of setting
2271 PCRE_ANCHORED and checking the length, in preparation for adding
2272 something similar for -w.
2273
2274 (b) Added the -w (match as a word) option.
2275
2276 (c) Refactored the way lines are read and buffered so as to have more
2277 than one at a time available.
2278
2279 (d) Implemented a pcregrep test script.
2280
2281 (e) Added the -M (multiline match) option. This allows patterns to match
2282 over several lines of the subject. The buffering ensures that at least
2283 8K, or the rest of the document (whichever is the shorter) is available
2284 for matching (and similarly the previous 8K for lookbehind assertions).
2285
2286 (f) Changed the --help output so that it now says
2287
2288 -w, --word-regex(p)
2289
2290 instead of two lines, one with "regex" and the other with "regexp"
2291 because that confused at least one person since the short forms are the
2292 same. (This required a bit of code, as the output is generated
2293 automatically from a table. It wasn't just a text change.)
2294
2295 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2296 option but starts with a hyphen. Could be a pattern or a path name
2297 starting with a hyphen, for instance.
2298
2299 (h) "-" can be given as a file name to represent stdin.
2300
2301 (i) When file names are being printed, "(standard input)" is used for
2302 the standard input, for compatibility with GNU grep. Previously
2303 "<stdin>" was used.
2304
2305 (j) The option --label=xxx can be used to supply a name to be used for
2306 stdin when file names are being printed. There is no short form.
2307
2308 (k) Re-factored the options decoding logic because we are going to add
2309 two more options that take data. Such options can now be given in four
2310 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2311
2312 (l) Added the -A, -B, and -C options for requesting that lines of context
2313 around matches be printed.
2314
2315 (m) Added the -L option to print the names of files that do not contain
2316 any matching lines, that is, the complement of -l.
2317
2318 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2319 continue to scan other files.
2320
2321 (o) The -s option was incorrectly implemented. For compatibility with other
2322 greps, it now suppresses the error message for a non-existent or non-
2323 accessible file (but not the return code). There is a new option called
2324 -q that suppresses the output of matching lines, which was what -s was
2325 previously doing.
2326
2327 (p) Added --include and --exclude options to specify files for inclusion
2328 and exclusion when recursing.
2329
2330 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2331 Hopefully, it now does.
2332
2333 12. Missing cast in pcre_study().
2334
2335 13. Added an "uninstall" target to the makefile.
2336
2337 14. Replaced "extern" in the function prototypes in Makefile.in with
2338 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2339 world, but is set differently for Windows.
2340
2341 15. Added a second compiling function called pcre_compile2(). The only
2342 difference is that it has an extra argument, which is a pointer to an
2343 integer error code. When there is a compile-time failure, this is set
2344 non-zero, in addition to the error test pointer being set to point to an
2345 error message. The new argument may be NULL if no error number is required
2346 (but then you may as well call pcre_compile(), which is now just a
2347 wrapper). This facility is provided because some applications need a
2348 numeric error indication, but it has also enabled me to tidy up the way
2349 compile-time errors are handled in the POSIX wrapper.
2350
2351 16. Added VPATH=.libs to the makefile; this should help when building with one
2352 prefix path and installing with another. (Or so I'm told by someone who
2353 knows more about this stuff than I do.)
2354
2355 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2356 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2357 match everything, including newlines. This is not POSIX-compatible, but
2358 somebody wanted the feature. From pcretest it can be activated by using
2359 both the P and the s flags.
2360
2361 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2362
2363 19. libpcre.pc was being incorrectly installed as executable.
2364
2365 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2366 it now also looks for '\r' so that it will work unmodified on Windows.
2367
2368 21. Added Google's contributed C++ wrapper to the distribution.
2369
2370 22. Added some untidy missing memory free() calls in pcretest, to keep
2371 Electric Fence happy when testing.
2372
2373
2374
2375 Version 5.0 13-Sep-04
2376 ---------------------
2377
2378 1. Internal change: literal characters are no longer packed up into items
2379 containing multiple characters in a single byte-string. Each character
2380 is now matched using a separate opcode. However, there may be more than one
2381 byte in the character in UTF-8 mode.
2382
2383 2. The pcre_callout_block structure has two new fields: pattern_position and
2384 next_item_length. These contain the offset in the pattern to the next match
2385 item, and its length, respectively.
2386
2387 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2388 insertion of callouts before each pattern item. Added the /C option to
2389 pcretest to make use of this.
2390
2391 4. On the advice of a Windows user, the lines
2392
2393 #if defined(_WIN32) || defined(WIN32)
2394 _setmode( _fileno( stdout ), 0x8000 );
2395 #endif /* defined(_WIN32) || defined(WIN32) */
2396
2397 have been added to the source of pcretest. This apparently does useful
2398 magic in relation to line terminators.
2399
2400 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2401 for the benefit of those environments where the "b" makes a difference.
2402
2403 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2404 to know about it. I have put a hack into configure.in that adds in code
2405 to set GCC=yes if CC=icc. This seems to end up at a point in the
2406 generated configure script that is early enough to affect the setting of
2407 compiler options, which is what is needed, but I have no means of testing
2408 whether it really works. (The user who reported this had patched the
2409 generated configure script, which of course I cannot do.)
2410
2411 LATER: After change 22 below (new libtool files), the configure script
2412 seems to know about icc (and also ecc). Therefore, I have commented out
2413 this hack in configure.in.
2414
2415 7. Added support for pkg-config (2 patches were sent in).
2416
2417 8. Negated POSIX character classes that used a combination of internal tables
2418 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2419 [[:^ascii]]. Typically, they would match almost any characters. The other
2420 POSIX classes were not broken in this way.
2421
2422 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2423 to find the match, as PCRE was deluded into thinking that the match had to
2424 start at the start point or following a newline. The same bug applied to
2425 patterns with negative forward assertions or any backward assertions
2426 preceding ".*" at the start, unless the pattern required a fixed first
2427 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2428
2429 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2430 starting at the last subject character, bytes beyond the end of the subject
2431 string were read.
2432
2433 11. Renamed the variable "class" as "classbits" to make life easier for C++
2434 users. (Previously there was a macro definition, but it apparently wasn't
2435 enough.)
2436
2437 12. Added the new field "tables" to the extra data so that tables can be passed
2438 in at exec time, or the internal tables can be re-selected. This allows
2439 a compiled regex to be saved and re-used at a later time by a different
2440 program that might have everything at different addresses.
2441
2442 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2443 -R library as well as a -L library.
2444
2445 14. The debugging options of pcretest (-d on the command line or D on a
2446 pattern) showed incorrect output for anything following an extended class
2447 that contained multibyte characters and which was followed by a quantifier.
2448
2449 15. Added optional support for general category Unicode character properties
2450 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2451 support. It adds about 90K to the size of the library. The meanings of the
2452 inbuilt class escapes such as \d and \s have NOT been changed.
2453
2454 16. Updated pcredemo.c to include calls to free() to release the memory for the
2455 compiled pattern.
2456
2457 17. The generated file chartables.c was being created in the source directory
2458 instead of in the building directory. This caused the build to fail if the
2459 source directory was different from the building directory, and was
2460 read-only.
2461
2462 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2463 file. No doubt somebody will tell me if they don't make sense... Also added
2464 Dan Mooney's comments about building on OpenVMS.
2465
2466 19. Added support for partial matching via the PCRE_PARTIAL option for
2467 pcre_exec() and the \P data escape in pcretest.
2468
2469 20. Extended pcretest with 3 new pattern features:
2470
2471 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2472 write the compiled pattern to the file whose name is "rest-of-line".
2473 This is a straight binary dump of the data, with the saved pointer to
2474 the character tables forced to be NULL. The study data, if any, is
2475 written too. After writing, pcretest reads a new pattern.
2476
2477 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2478 compiled pattern from the given file. There must not be any
2479 occurrences of "<" in the file name (pretty unlikely); if there are,
2480 pcretest will instead treat the initial "<" as a pattern delimiter.
2481 After reading in the pattern, pcretest goes on to read data lines as
2482 usual.
2483
2484 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2485 and 16-bit fields in a compiled pattern, to simulate a pattern that
2486 was compiled on a host of opposite endianness.
2487
2488 21. The pcre-exec() function can now cope with patterns that were compiled on
2489 hosts of opposite endianness, with this restriction:
2490
2491 As for any compiled expression that is saved and used later, the tables
2492 pointer field cannot be preserved; the extra_data field in the arguments
2493 to pcre_exec() should be used to pass in a tables address if a value
2494 other than the default internal tables were used at compile time.
2495
2496 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2497 now diagnosed as an error. Previously, most of the time, a negative number
2498 would have been treated as zero, but if in addition "ovector" was passed as
2499 NULL, a crash could occur.
2500
2501 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2502 new versions from the libtool 1.5 distribution (the last one is a copy of
2503 a file called libtool.m4). This seems to have fixed the need to patch
2504 "configure" to support Darwin 1.3 (which I used to do). However, I still
2505 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2506 workstation).
2507
2508 24. Changed the PCRE licence to be the more standard "BSD" licence.
2509
2510
2511 Version 4.5 01-Dec-03
2512 ---------------------
2513
2514 1. There has been some re-arrangement of the code for the match() function so
2515 that it can be compiled in a version that does not call itself recursively.
2516 Instead, it keeps those local variables that need separate instances for
2517 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2518 needs to "recurse". Keeping track of where control must go is done by means
2519 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2520 hide most of the details from the main code, and operates only if
2521 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2522 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2523 operating.
2524
2525 To make it easier for callers to provide specially tailored get/free
2526 functions for this usage, two new functions, pcre_stack_malloc, and
2527 pcre_stack_free, are used. They are always called in strict stacking order,
2528 and the size of block requested is always the same.
2529
2530 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2531 PCRE has been compiled to use the stack or the heap for recursion. The
2532 -C option of pcretest uses this to show which version is compiled.
2533
2534 A new data escape \S, is added to pcretest; it causes the amounts of store
2535 obtained and freed by both kinds of malloc/free at match time to be added
2536 to the output.
2537
2538 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2539 what's available on my current Linux desktop machine.
2540
2541 3. When matching a UTF-8 string, the test for a valid string at the start has
2542 been extended. If start_offset is not zero, PCRE now checks that it points
2543 to a byte that is the start of a UTF-8 character. If not, it returns
2544 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2545 this is necessary because there may be backward assertions in the pattern.
2546 When matching the same subject several times, it may save resources to use
2547 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2548
2549 4. The code for checking the validity of UTF-8 strings has been tightened so
2550 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2551 containing "overlong sequences".
2552
2553 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2554 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2555 should have been "&", but it just so happened that all the cases this let
2556 through by mistake were picked up later in the function.
2557
2558 6. I had used a variable called "isblank" - this is a C99 function, causing
2559 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2560
2561 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2562 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2563 so the version is shown; (c) stop "make test" echoing "./RunTest".
2564
2565 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2566
2567 9. The prototype for memmove() for systems that don't have it was using
2568 size_t, but the inclusion of the header that defines size_t was later. I've
2569 moved the #includes for the C headers earlier to avoid this.
2570
2571 10. Added some adjustments to the code to make it easier to compiler on certain
2572 special systems:
2573
2574 (a) Some "const" qualifiers were missing.
2575 (b) Added the macro EXPORT before all exported functions; by default this
2576 is defined to be empty.
2577 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2578 that it reads its output file name as an argument instead of writing
2579 to the standard output and assuming this can be redirected.
2580
2581 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2582 class containing characters with values greater than 255, PCRE compilation
2583 went into a loop.
2584
2585 12. A recursive reference to a subpattern that was within another subpattern
2586 that had a minimum quantifier of zero caused PCRE to crash. For example,
2587 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2588 recursion. If the recursively-called subpattern itself had a zero repeat,
2589 that was OK.
2590
2591 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2592 buffer into which it was copied (for escape processing) was still set at
2593 1024, so long lines caused crashes.
2594
2595 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2596 "internal error: code overflow...". This applied to any character class
2597 that was followed by a possessive quantifier.
2598
2599 15. Modified the Makefile to add libpcre.la as a prerequisite for
2600 libpcreposix.la because I was told this is needed for a parallel build to
2601 work.
2602
2603 16. If a pattern that contained .* following optional items at the start was
2604 studied, the wrong optimizing data was generated, leading to matching
2605 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2606 matching string must start with a or b or c. The correct conclusion for
2607 this pattern is that a match can start with any character.
2608
2609
2610 Version 4.4 13-Aug-03
2611 ---------------------
2612
2613 1. In UTF-8 mode, a character class containing characters with values between
2614 127 and 255 was not handled correctly if the compiled pattern was studied.
2615 In fixing this, I have also improved the studying algorithm for such
2616 classes (slightly).
2617
2618 2. Three internal functions had redundant arguments passed to them. Removal
2619 might give a very teeny performance improvement.
2620
2621 3. Documentation bug: the value of the capture_top field in a callout is *one
2622 more than* the number of the hightest numbered captured substring.
2623
2624 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2625 in incorrectly linking with a previously installed version. They now link
2626 explicitly with libpcre.la.
2627
2628 5. configure.in no longer needs to recognize Cygwin specially.
2629
2630 6. A problem in pcre.in for Windows platforms is fixed.
2631
2632 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2633 pcretest, it used to include the size of the study block as part of its
2634 output. Unfortunately, the structure contains a field that has a different
2635 size on different hardware architectures. This meant that the tests that
2636 showed this size failed. As the block is currently always of a fixed size,
2637 this information isn't actually particularly useful in pcretest output, so
2638 I have just removed it.
2639
2640 8. Three pre-processor statements accidentally did not start in column 1.
2641 Sadly, there are *still* compilers around that complain, even though
2642 standard C has not required this for well over a decade. Sigh.
2643
2644 9. In pcretest, the code for checking callouts passed small integers in the
2645 callout_data field, which is a void * field. However, some picky compilers
2646 complained about the casts involved for this on 64-bit systems. Now
2647 pcretest passes the address of the small integer instead, which should get
2648 rid of the warnings.
2649
2650 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2651 both compile and run time, and gives an error if an invalid UTF-8 sequence
2652 is found. There is a option for disabling this check in cases where the
2653 string is known to be correct and/or the maximum performance is wanted.
2654
2655 11. In response to a bug report, I changed one line in Makefile.in from
2656
2657 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2658 to
2659 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2660
2661 to look similar to other lines, but I have no way of telling whether this
2662 is the right thing to do, as I do not use Windows. No doubt I'll get told
2663 if it's wrong...
2664
2665
2666 Version 4.3 21-May-03
2667 ---------------------
2668
2669 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2670 Makefile.
2671
2672 2. Some refactoring to improve the quality of the code:
2673
2674 (i) The utf8_table... variables are now declared "const".
2675
2676 (ii) The code for \cx, which used the "case flipping" table to upper case
2677 lower case letters, now just substracts 32. This is ASCII-specific,
2678 but the whole concept of \cx is ASCII-specific, so it seems
2679 reasonable.
2680
2681 (iii) PCRE was using its character types table to recognize decimal and
2682 hexadecimal digits in the pattern. This is silly, because it handles
2683 only 0-9, a-f, and A-F, but the character types table is locale-
2684 specific, which means strange things might happen. A private
2685 table is now used for this - though it costs 256 bytes, a table is
2686 much faster than multiple explicit tests. Of course, the standard
2687 character types table is still used for matching digits in subject
2688 strings against \d.
2689
2690 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2691 ending in _t are). So I've renamed it as ESC_tee.
2692
2693 3. The first argument for regexec() in the POSIX wrapper should have been
2694 defined as "const".
2695
2696 4. Changed pcretest to use malloc() for its buffers so that they can be
2697 Electric Fenced for debugging.
2698
2699 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2700 to read one or more bytes before the start of the subject string. Often this
2701 had no effect on PCRE's behaviour, but in some circumstances it could
2702 provoke a segmentation fault.
2703
2704 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2705 to try to read one or more bytes before the start of the subject string.
2706
2707 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2708 UTF-8 support could misbehave in various ways if the subject string
2709 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2710 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2711 back over UTF-8 characters.)
2712
2713
2714 Version 4.2 14-Apr-03
2715 ---------------------
2716
2717 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2718
2719 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2720 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2721 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2722 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2723 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2724 and BUILD_EXEEXT
2725 Note: automatic setting of the BUILD variables is not yet working
2726 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2727 compile-time but not at link-time
2728 [LINK]: use for linking executables only
2729 make different versions for Windows and non-Windows
2730 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2731 libraries
2732 [LINK_FOR_BUILD]: new variable
2733 [OBJEXT]: use throughout
2734 [EXEEXT]: use throughout
2735 <winshared>: new target
2736 <wininstall>: new target
2737 <dftables.o>: use native compiler
2738 <dftables>: use native linker
2739 <install>: handle Windows platform correctly
2740 <clean>: ditto
2741 <check>: ditto
2742 copy DLL to top builddir before testing
2743
2744 As part of these changes, -no-undefined was removed again. This was reported
2745 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2746 in any case.
2747
2748 3. Some tidies to get rid of compiler warnings:
2749
2750 . In the match_data structure, match_limit was an unsigned long int, whereas
2751 match_call_count was an int. I've made them both unsigned long ints.
2752
2753 . In pcretest the fact that a const uschar * doesn't automatically cast to
2754 a void * provoked a warning.
2755
2756 . Turning on some more compiler warnings threw up some "shadow" variables
2757 and a few more missing casts.
2758
2759 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2760 option, a class that contained a single character with a value between 128
2761 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2762
2763 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2764 option, a class that contained several characters, but with at least one
2765 whose value was between 128 and 255 caused PCRE to crash.
2766
2767
2768 Version 4.1 12-Mar-03
2769 ---------------------
2770
2771 1. Compiling with gcc -pedantic found a couple of places where casts were
2772 needed, and a string in dftables.c that was longer than standard compilers are
2773 required to support.
2774
2775 2. Compiling with Sun's compiler found a few more places where the code could
2776 be tidied up in order to avoid warnings.
2777
2778 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2779 first of these names is deprecated in the latest Autoconf in favour of the name
2780 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2781 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2782 analogy I have changed it to CFLAGS_FOR_BUILD.
2783
2784 4. Added -no-undefined to the linking command in the Makefile, because this is
2785 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2786 linking step for the pcreposix library.
2787
2788 5. PCRE was failing to diagnose the case of two named groups with the same
2789 name.
2790
2791 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2792 literal character that is needed in the subject for a match, and scans along to
2793 ensure that it is present before embarking on the full matching process. This
2794 saves time in cases of nested unlimited repeats that are never going to match.
2795 Problem: the scan can take a lot of time if the subject is very long (e.g.
2796 megabytes), thus penalizing straightforward matches. It is now done only if the
2797 amount of subject to be scanned is less than 1000 bytes.
2798
2799 7. A lesser problem with the same optimization is that it was recording the
2800 first character of an anchored pattern as "needed", thus provoking a search
2801 right along the subject, even when the first match of the pattern was going to
2802 fail. The "needed" character is now not set for anchored patterns, unless it
2803 follows something in the pattern that is of non-fixed length. Thus, it still
2804 fulfils its original purpose of finding quick non-matches in cases of nested
2805 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2806
2807
2808 Version 4.0 17-Feb-03
2809 ---------------------
2810
2811 1. If a comment in an extended regex that started immediately after a meta-item
2812 extended to the end of string, PCRE compiled incorrect data. This could lead to
2813 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2814
2815 2. Moved to autoconf 2.53 and libtool 1.4.2.
2816
2817 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2818 the special perltest8 script is no longer needed - all the tests can be run
2819 from a single perltest script.
2820
2821 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2822 by \s. It has now been removed in PCRE. This means it isn't recognized as
2823 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2824 class [:space:] *does* include VT, thereby creating a mess.
2825
2826 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2827 space and tab.
2828
2829 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2830 its new features into the main test script, reducing the number of scripts.
2831
2832 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2833 were backward compatible, and made the (?i) apply to the whole pattern, as if
2834 /i were given. Now it behaves more logically, and applies the option setting
2835 only to what follows. PCRE has been changed to follow suit. However, if it
2836 finds options settings right at the start of the pattern, it extracts them into
2837 the global options, as before. Thus, they show up in the info data.
2838
2839 8. Added support for the \Q...\E escape sequence. Characters in between are
2840 treated as literals. This is slightly different from Perl in that $ and @ are
2841 also handled as literals inside the quotes. In Perl, they will cause variable
2842 interpolation. Note the following examples:
2843
2844 Pattern PCRE matches Perl matches
2845
2846 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2847 \Qabc\$xyz\E abc\$xyz abc\$xyz
2848 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2849
2850 For compatibility with Perl, \Q...\E sequences are recognized inside character
2851 classes as well as outside them.
2852
2853 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2854 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2855 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2856 signed/unsigned warnings.
2857
2858 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2859 option for pcretest, so I've replaced it by a simple function that does just
2860 that job.
2861
2862 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2863 "pcregrep -".
2864
2865 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2866 Java package. This provides some syntactic sugar for simple cases of what my
2867 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2868 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2869 item, you can use this simplified notation. Note that only makes sense with
2870 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2871 greediness, whatever the setting of the PCRE_UNGREEDY option.
2872
2873 13. A change of greediness default within a pattern was not taking effect at
2874 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2875 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2876 was abstracted outside.
2877
2878 14. PCRE now supports the \G assertion. It is true when the current matching
2879 position is at the start point of the match. This differs from \A when the
2880 starting offset is non-zero. Used with the /g option of pcretest (or similar
2881 code), it works in the same way as it does for Perl's /g option. If all
2882 alternatives of a regex begin with \G, the expression is anchored to the start
2883 match position, and the "anchored" flag is set in the compiled expression.
2884
2885 15. Some bugs concerning the handling of certain option changes within patterns
2886 have been fixed. These applied to options other than (?ims). For example,
2887 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2888 been the other way round. Some of this was related to change 7 above.
2889
2890 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2891 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2892 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2893 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2894
2895 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2896 mode. Unlike ".", it always matches newline, whatever the setting of
2897 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2898 assertions. Perl allows it, but it doesn't (in general) work because it can't
2899 calculate the length of the lookbehind. At least, that's the case for Perl
2900 5.8.0 - I've been told they are going to document that it doesn't work in
2901 future.
2902
2903 18. Added an error diagnosis for escapes that PCRE does not support: these are
2904 \L, \l, \N, \P, \p, \U, \u, and \X.
2905
2906 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2907 reading past the end of the pattern in cases such as /[abcd/.
2908
2909 20. PCRE was getting more memory than necessary for patterns with classes that
2910 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2911
2912 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2913 compiling PCRE for use with Virtual Pascal.
2914
2915 22. Small fix to the Makefile to make it work properly if the build is done
2916 outside the source tree.
2917
2918 23. Added a new extension: a condition to go with recursion. If a conditional
2919 subpattern starts with (?(R) the "true" branch is used if recursion has
2920 happened, whereas the "false" branch is used only at the top level.
2921
2922 24. When there was a very long string of literal characters (over 255 bytes
2923 without UTF support, over 250 bytes with UTF support), the computation of how
2924 much memory was required could be incorrect, leading to segfaults or other
2925 strange effects.
2926
2927 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2928 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2929 there was a subsequent back reference to those brackets. This meant that, for
2930 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2931 possible to check for precisely this case. All we can do is abandon the
2932 optimization if .* occurs inside capturing brackets when there are any back
2933 references whatsoever. (See below for a better fix that came later.)
2934
2935 26. The handling of the optimization for finding the first character of a
2936 non-anchored pattern, and for finding a character that is required later in the
2937 match were failing in some cases. This didn't break the matching; it just
2938 failed to optimize when it could. The way this is done has been re-implemented.
2939
2940 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2941
2942 28. Added a new feature that provides some of the functionality that Perl
2943 provides with (?{...}). The facility is termed a "callout". The way it is done
2944 in PCRE is for the caller to provide an optional function, by setting
2945 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2946 global variable. By default it is unset, which disables all calling out. To get
2947 the function called, the regex must include (?C) at appropriate points. This
2948 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2949 This provides a means of identifying different callout points. When PCRE
2950 reaches such a point in the regex, if pcre_callout has been set, the external
2951 function is called. It is provided with data in a structure called
2952 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2953 matching continues; if it returns a non-zero value, the match at the current
2954 point fails. However, backtracking will occur if possible. [This was changed
2955 later and other features added - see item 49 below.]
2956
2957 29. pcretest is upgraded to test the callout functionality. It provides a
2958 callout function that displays information. By default, it shows the start of
2959 the match and the current position in the text. There are some new data escapes
2960 to vary what happens:
2961
2962 \C+ in addition, show current contents of captured substrings
2963 \C- do not supply a callout function
2964 \C!n return 1 when callout number n is reached
2965 \C!n!m return 1 when callout number n is reached for the mth time
2966
2967 30. If pcregrep was called with the -l option and just a single file name, it
2968 output "<stdin>" if a match was found, instead of the file name.
2969
2970 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2971 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2972 pcre_exec(). This saves a malloc/free per call. The default value of
2973 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2974 when configuring.
2975
2976 32. The default maximum size of a compiled pattern is 64K. There have been a
2977 few cases of people hitting this limit. The code now uses macros to handle the
2978 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2979 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2980 configuring. Tests 2 and 5 work only with 2-byte links because they output
2981 debugging information about compiled patterns.
2982
2983 33. Internal code re-arrangements:
2984
2985 (a) Moved the debugging function for printing out a compiled regex into
2986 its own source file (printint.c) and used #include to pull it into
2987 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2988 separate copies.
2989
2990 (b) Defined the list of op-code names for debugging as a macro in
2991 internal.h so that it is next to the definition of the opcodes.
2992
2993 (c) Defined a table of op-code lengths for simpler skipping along compiled
2994 code. This is again a macro in internal.h so that it is next to the
2995 definition of the opcodes.
2996
2997 34. Added support for recursive calls to individual subpatterns, along the
2998 lines of Robin Houston's patch (but implemented somewhat differently).
2999
3000 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3001 allow it to read and process whole directories in Win32. This code was
3002 contributed by Lionel Fourquaux; it has not been tested by me.
3003
3004 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3005 used to name a group. Names consist of alphanumerics and underscores, and must
3006 be unique. Back references use the syntax (?P=name) and recursive calls use
3007 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3008 numbers. The function pcre_fullinfo() can be used after compilation to extract
3009 a name/number map. There are three relevant calls:
3010
3011 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3012 PCRE_INFO_NAMECOUNT yields the number of entries
3013 PCRE_INFO_NAMETABLE yields a pointer to the map.
3014
3015 The map is a vector of fixed-size entries. The size of each entry depends on
3016 the length of the longest name used. The first two bytes of each entry are the
3017 group number, most significant byte first. There follows the corresponding
3018 name, zero terminated. The names are in alphabetical order.
3019
3020 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3021 case instead of 255. Making it the same both with and without UTF-8 support
3022 means that the same test output works with both.
3023
3024 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3025 calling malloc() with a zero argument.
3026
3027 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3028 optimization. I've improved things by keeping a bitmap of backreferences with
3029 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3030 fact referenced, the optimization can be applied. It is unlikely that a
3031 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3032 the match to follow \n) will appear inside brackets with a number greater than
3033 31, but if it does, any back reference > 31 suppresses the optimization.
3034
3035 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3036 of disabling numbered capturing parentheses. Any opening parenthesis that is
3037 not followed by ? behaves as if it were followed by ?: but named parentheses
3038 can still be used for capturing (and they will acquire numbers in the usual
3039 way).
3040
3041 41. Redesigned the return codes from the match() function into yes/no/error so
3042 that errors can be passed back from deep inside the nested calls. A malloc
3043 failure while inside a recursive subpattern call now causes the
3044 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3045
3046 42. It is now possible to set a limit on the number of times the match()
3047 function is called in a call to pcre_exec(). This facility makes it possible to
3048 limit the amount of recursion and backtracking, though not in a directly
3049 obvious way, because the match() function is used in a number of different
3050 circumstances. The count starts from zero for each position in the subject
3051 string (for non-anchored patterns). The default limit is, for compatibility, a
3052 large number, namely 10 000 000. You can change this in two ways:
3053
3054 (a) When configuring PCRE before making, you can use --with-match-limit=n
3055 to set a default value for the compiled library.
3056
3057 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3058 a different value is set. See 45 below.
3059
3060 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3061
3062 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3063 of things that can be changed at compile time. The first argument specifies
3064 what is wanted and the second points to where the information is to be placed.
3065 The current list of available information is:
3066
3067 PCRE_CONFIG_UTF8
3068
3069 The output is an integer that is set to one if UTF-8 support is available;
3070 otherwise it is set to zero.
3071
3072 PCRE_CONFIG_NEWLINE
3073
3074 The output is an integer that it set to the value of the code that is used for
3075 newline. It is either LF (10) or CR (13).
3076
3077 PCRE_CONFIG_LINK_SIZE
3078
3079 The output is an integer that contains the number of bytes used for internal
3080 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3081
3082 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3083
3084 The output is an integer that contains the threshold above which the POSIX
3085 interface uses malloc() for output vectors. See item 31 above.
3086
3087 PCRE_CONFIG_MATCH_LIMIT
3088
3089 The output is an unsigned integer that contains the default limit of the number
3090 of match() calls in a pcre_exec() execution. See 42 above.
3091
3092 44. pcretest has been upgraded by the addition of the -C option. This causes it
3093 to extract all the available output from the new pcre_config() function, and to
3094 output it. The program then exits immediately.
3095
3096 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3097 order to support additional features. One way would have been to define
3098 pcre_exec2() (for example) with extra arguments, but this would not have been
3099 extensible, and would also have required all calls to the original function to
3100 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3101 is used for passing in "extra" data from pcre_study().
3102
3103 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3104 contains the following fields:
3105
3106 flags a bitmap indicating which of the following fields are set
3107 study_data opaque data from pcre_study()
3108 match_limit a way of specifying a limit on match() calls for a specific
3109 call to pcre_exec()
3110 callout_data data for callouts (see 49 below)
3111
3112 The flag bits are also defined in pcre.h, and are
3113
3114 PCRE_EXTRA_STUDY_DATA
3115 PCRE_EXTRA_MATCH_LIMIT
3116 PCRE_EXTRA_CALLOUT_DATA
3117
3118 The pcre_study() function now returns one of these new pcre_extra blocks, with
3119 the actual study data pointed to by the study_data field, and the
3120 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3121 before. That is, this change is entirely upwards-compatible and requires no
3122 change to existing code.
3123
3124 If you want to pass in additional data to pcre_exec(), you can either place it
3125 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3126 block.
3127
3128 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3129 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3130 times with different match limits, until it finds the minimum value needed for
3131 pcre_exec() to complete. The value is then output. This can be instructive; for
3132 most simple matches the number is quite small, but for pathological cases it
3133 gets very large very quickly.
3134
3135 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3136 returns the size of the data block pointed to by the study_data field in a
3137 pcre_extra block, that is, the value that was passed as the argument to
3138 pcre_malloc() when PCRE was getting memory in which to place the information
3139 created by pcre_study(). The fourth argument should point to a size_t variable.
3140 pcretest has been extended so that this information is shown after a successful
3141 pcre_study() call when information about the compiled regex is being displayed.
3142
3143 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3144 because what follows is always an absolute path. (Later: it turns out that this
3145 is more than cosmetic for MinGW, because it doesn't like empty path
3146 components.)
3147
3148 49. Some changes have been made to the callout feature (see 28 above):
3149
3150 (i) A callout function now has three choices for what it returns:
3151
3152 0 => success, carry on matching
3153 > 0 => failure at this point, but backtrack if possible
3154 < 0 => serious error, return this value from pcre_exec()
3155
3156 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3157 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3158 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3159 use by callout functions. It will never be used by PCRE itself.
3160
3161 (ii) The pcre_extra structure (see 45 above) has a void * field called
3162 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3163 pcre_callout_block structure has a field of the same name. The contents of
3164 the field passed in the pcre_extra structure are passed to the callout
3165 function in the corresponding field in the callout block. This makes it
3166 easier to use the same callout-containing regex from multiple threads. For
3167 testing, the pcretest program has a new data escape
3168
3169 \C*n pass the number n (may be negative) as callout_data
3170
3171 If the callout function in pcretest receives a non-zero value as
3172 callout_data, it returns that value.
3173
3174 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3175 there were some redundant $(CFLAGS) in commands that are now specified as
3176 $(LINK), which already includes $(CFLAGS).
3177
3178 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3179 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3180 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3181 one-byte characters throughout. Note that case-insensitive matching applies
3182 only to characters whose values are less than 256. PCRE doesn't support the
3183 notion of cases for higher-valued characters.
3184
3185 (i) A character class whose characters are all within 0-255 is handled as
3186 a bit map, and the map is inverted for negative classes. Previously, a
3187 character > 255 always failed to match such a class; however it should
3188 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3189
3190 (ii) A negated character class with a single character < 255 is coded as
3191 "not this character" (OP_NOT). This wasn't working properly when the test
3192 character was multibyte, either singly or repeated.
3193
3194 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3195 mode, for example: \x{100}{2,3}.
3196
3197 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3198 singly or repeated) now correctly test multibyte characters. However,
3199 PCRE doesn't recognize any characters with values greater than 255 as
3200 digits, spaces, or word characters. Such characters always match \D, \S,
3201 and \W, and never match \d, \s, or \w.
3202
3203 (v) Classes may now contain characters and character ranges with values
3204 greater than 255. For example: [ab\x{100}-\x{400}].
3205
3206 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3207 PCRE in UTF-8 mode.
3208
3209 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3210 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3211 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3212 value.)
3213
3214 53. The single man page has become too large. I have therefore split it up into
3215 a number of separate man pages. These also give rise to individual HTML pages;
3216 these are now put in a separate directory, and there is an index.html page that
3217 lists them all. Some hyperlinking between the pages has been installed.
3218
3219 54. Added convenience functions for handling named capturing parentheses.
3220
3221 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3222 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3223 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3224 are faulted.
3225
3226 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3227 calling configure. These values are used when compiling the dftables.c program
3228 which is run to generate the source of the default character tables. They
3229 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3230 you will need to set these values.
3231
3232 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3233
3234
3235 Version 3.9 02-Jan-02
3236 ---------------------
3237
3238 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3239
3240 2. If --disable-static was given, the building process failed when trying to
3241 build pcretest and pcregrep. (For some reason it was using libtool to compile
3242 them, which is not right, as they aren't part of the library.)
3243
3244
3245 Version 3.8 18-Dec-01
3246 ---------------------
3247
3248 1. The experimental UTF-8 code was completely screwed up. It was packing the
3249 bytes in the wrong order. How dumb can you get?
3250
3251
3252 Version 3.7 29-Oct-01
3253 ---------------------
3254
3255 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3256 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3257 this didn't happen under Solaris 8, where I normally test things.
3258
3259 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3260 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3261 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3262 this fix an hour or so after the initial 3.7 release.)
3263
3264
3265 Version 3.6 23-Oct-01
3266 ---------------------
3267
3268 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3269 offsets passed as NULL with zero offset count.
3270
3271 2. The config.guess and config.sub files had not been updated when I moved to
3272 the latest autoconf.
3273
3274
3275 Version 3.5 15-Aug-01
3276 ---------------------
3277
3278 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3279 had been forgotten.
3280
3281 2. By using declared but undefined structures, we can avoid using "void"
3282 definitions in pcre.h while keeping the internal definitions of the structures
3283 private.
3284
3285 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3286 user point of view, this means that both static and shared libraries are built
3287 by default, but this can be individually controlled. More of the work of
3288 handling this static/shared cases is now inside libtool instead of PCRE's make
3289 file.
3290
3291 4. The pcretest utility is now installed along with pcregrep because it is
3292 useful for users (to test regexs) and by doing this, it automatically gets
3293 relinked by libtool. The documentation has been turned into a man page, so
3294 there are now .1, .txt, and .html versions in /doc.
3295
3296 5. Upgrades to pcregrep:
3297 (i) Added long-form option names like gnu grep.
3298 (ii) Added --help to list all options with an explanatory phrase.
3299 (iii) Added -r, --recursive to recurse into sub-directories.
3300 (iv) Added -f, --file to read patterns from a file.
3301
3302 6. pcre_exec() was referring to its "code" argument before testing that
3303 argument for NULL (and giving an error if it was NULL).
3304
3305 7. Upgraded Makefile.in to allow for compiling in a different directory from
3306 the source directory.
3307
3308 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3309 options bits, the pointer it was passed was to an int instead of to an unsigned
3310 long int. This mattered only on 64-bit systems.
3311
3312 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3313 generated) instead of pcre.in, which it its source. Also made the same change
3314 in several of the .c files.
3315
3316 10. A new release of gcc defines printf() as a macro, which broke pcretest
3317 because it had an ifdef in the middle of a string argument for printf(). Fixed
3318 by using separate calls to printf().
3319
3320 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3321 script, to force use of CR or LF instead of \n in the source. On non-Unix
3322 systems, the value can be set in config.h.
3323
3324 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3325 absolute limit. Changed the text of the error message to make this clear, and
3326 likewise updated the man page.
3327
3328 13. The limit of 99 on the number of capturing subpatterns has been removed.
3329 The new limit is 65535, which I hope will not be a "real" limit.
3330
3331
3332 Version 3.4 22-Aug-00
3333 ---------------------
3334
3335 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3336
3337 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3338
3339
3340 Version 3.3 01-Aug-00
3341 ---------------------
3342
3343 1. If an octal character was given, but the value was greater than \377, it
3344 was not getting masked to the least significant bits, as documented. This could
3345 lead to crashes in some systems.
3346
3347 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3348 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3349
3350 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3351 These just pass their arguments on to (pcre_free)(), but they are provided
3352 because some uses of PCRE bind it to non-C systems that can call its functions,
3353 but cannot call free() or pcre_free() directly.
3354
3355 4. Add "make test" as a synonym for "make check". Corrected some comments in
3356 the Makefile.
3357
3358 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3359 Makefile.
3360
3361 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3362 command called pgrep for grepping around the active processes.
3363
3364 7. Added the beginnings of support for UTF-8 character strings.
3365
3366 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3367 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3368 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3369 out for the ar command.)
3370
3371
3372 Version 3.2 12-May-00
3373 ---------------------
3374
3375 This is purely a bug fixing release.
3376
3377 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3378 of ZA. This was just one example of several cases that could provoke this bug,
3379 which was introduced by change 9 of version 2.00. The code for breaking
3380 infinite loops after an iteration that matches an empty string was't working
3381 correctly.
3382
3383 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3384 when matched against abbab (for example). After matching an empty string, it
3385 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3386 caused it to match further down the string than it should.
3387
3388 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3389 was there because it doesn't seem to be needed, and it causes trouble on some
3390 systems, as it is not a Standard C header. It has been removed.
3391
3392 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3393 were reported on the Macintosh. The changes were from
3394
3395 while ((c = *(++ptr)) != 0 && c != '\n');
3396 to
3397 while ((c = *(++ptr)) != 0 && c != '\n') ;
3398
3399 Totally extraordinary, but if that's what it takes...
3400
3401 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3402 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3403 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3404 assumes the way PCRE uses memmove() (always moving upwards).
3405
3406 6. PCRE is being used in one environment where strchr() is not available. There
3407 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3408 faster code anyway.
3409
3410
3411 Version 3.1 09-Feb-00
3412 ---------------------
3413
3414 The only change in this release is the fixing of some bugs in Makefile.in for
3415 the "install" target:
3416
3417 (1) It was failing to install pcreposix.h.
3418
3419 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3420
3421
3422 Version 3.0 01-Feb-00
3423 ---------------------
3424
3425 1. Add support for the /+ modifier to perltest (to output $` like it does in
3426 pcretest).
3427
3428 2. Add support for the /g modifier to perltest.
3429
3430 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3431 matches null strings.
3432
3433 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3434 pattern. Perl treats empty patterns specially - it reuses the most recent
3435 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3436 effect.
3437
3438 5. The POSIX interface was broken in that it was just handing over the POSIX
3439 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3440 required a bigger vector, with some working space on the end. This means that
3441 the POSIX wrapper now has to get and free some memory, and copy the results.
3442
3443 6. Added some simple autoconf support, placing the test data and the
3444 documentation in separate directories, re-organizing some of the
3445 information files, and making it build pcre-config (a GNU standard). Also added
3446 libtool support for building PCRE as a shared library, which is now the
3447 default.
3448
3449 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3450 09 are not valid octal constants. Single digits will be used for minor values
3451 less than 10.
3452
3453 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3454 existing programs that set these in the POSIX interface can use PCRE without
3455 modification.
3456
3457 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3458 return all that pcre_info() returns, plus additional data. The pcre_info()
3459 function is retained for compatibility, but is considered to be obsolete.
3460
3461 10. Added experimental recursion feature (?R) to handle one common case that
3462 Perl 5.6 will be able to do with (?p{...}).
3463
3464 11. Added support for POSIX character classes like [:alpha:], which Perl is
3465 adopting.
3466
3467
3468 Version 2.08 31-Aug-99
3469 ----------------------
3470
3471 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3472 trying to match at the startoffset position, but instead was moving forward to
3473 the next newline as if a previous match had failed.
3474
3475 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3476 and could get into a loop if a null string was matched other than at the start
3477 of the subject.
3478
3479 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3480 be distinguished at compile time, and for completeness also added PCRE_DATE.
3481
3482 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3483 in GnuWin32 environments.
3484
3485
3486 Version 2.07 29-Jul-99
3487 ----------------------
3488
3489 1. The documentation is now supplied in plain text form and HTML as well as in
3490 the form of man page sources.
3491
3492 2. C++ compilers don't like assigning (void *) values to other pointer types.
3493 In particular this affects malloc(). Although there is no problem in Standard
3494 C, I've put in casts to keep C++ compilers happy.
3495
3496 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3497 should be (const char *).
3498
3499 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3500 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3501 However, I haven't made this a standard facility. The documentation doesn't
3502 mention it, and the Makefile doesn't support it.
3503
3504 5. The Makefile now contains an "install" target, with editable destinations at
3505 the top of the file. The pcretest program is not installed.
3506
3507 6. pgrep -V now gives the PCRE version number and date.
3508
3509 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3510 causing the entire string to be ignored, instead of just the last character.
3511
3512 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3513 non-matching string, it can take a very, very long time, even for strings of
3514 quite modest length, because of the nested recursion. PCRE now does better in
3515 some of these cases. It does this by remembering the last required literal
3516 character in the pattern, and pre-searching the subject to ensure it is present
3517 before running the real match. In other words, it applies a heuristic to detect
3518 some types of certain failure quickly, and in the above example, if presented
3519 with a string that has no trailing " it gives "no match" very quickly.
3520
3521 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3522 other alternatives are tried instead.
3523
3524
3525 Version 2.06 09-Jun-99
3526 ----------------------
3527
3528 1. Change pcretest's output for amount of store used to show just the code
3529 space, because the remainder (the data block) varies in size between 32-bit and
3530 64-bit systems.
3531
3532 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3533 start matching at. This allows lookbehinds to work when searching for multiple
3534 occurrences in a string.
3535
3536 3. Added additional options to pcretest for testing multiple occurrences:
3537
3538 /+ outputs the rest of the string that follows a match
3539 /g loops for multiple occurrences, using the new startoffset argument
3540 /G loops for multiple occurrences by passing an incremented pointer
3541
3542 4. PCRE wasn't doing the "first character" optimization for patterns starting
3543 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3544 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3545 the letter 'x'. On long subject strings, this gives a significant speed-up.
3546
3547
3548 Version 2.05 21-Apr-99
3549 ----------------------
3550
3551 1. Changed the type of magic_number from int to long int so that it works
3552 properly on 16-bit systems.
3553
3554 2. Fixed a bug which caused patterns starting with .* not to work correctly
3555 when the subject string contained newline characters. PCRE was assuming
3556 anchoring for such patterns in all cases, which is not correct because .* will
3557 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3558 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3559 must be retried after every newline in the subject.
3560
3561
3562 Version 2.04 18-Feb-99
3563 ----------------------
3564
3565 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3566 computation of the store needed to hold the pattern was incorrect (too large).
3567 If such patterns were nested a few deep, this could multiply and become a real
3568 problem.
3569
3570 2. Added /M option to pcretest to show the memory requirement of a specific
3571 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3572
3573 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3574 compiled in such a way that the backtracking after subsequent failure was
3575 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3576 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3577
3578
3579 Version 2.03 02-Feb-99
3580 ----------------------
3581
3582 1. Fixed typo and small mistake in man page.
3583
3584 2. Added 4th condition (GPL supersedes if conflict) and created separate
3585 LICENCE file containing the conditions.
3586
3587 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3588 Perl, that is the internal \ allows the delimiter to be included in the
3589 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3590 the final delimiter, add \ to the end of the pattern (to test the error).
3591
3592 4. Added the convenience functions for extracting substrings after a successful
3593 match. Updated pcretest to make it able to test these functions.
3594
3595
3596 Version 2.02 14-Jan-99
3597 ----------------------
3598
3599 1. Initialized the working variables associated with each extraction so that
3600 their saving and restoring doesn't refer to uninitialized store.
3601
3602 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3603 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3604 fix the problem.
3605
3606 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3607 calls, and wasn't printing the correct value for compiling calls. Increased the
3608 default value of LOOPREPEAT, and the number of significant figures in the
3609 times.
3610
3611 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3612
3613 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3614 a building problem on Windows NT with a FAT file system.
3615
3616
3617 Version 2.01 21-Oct-98
3618 ----------------------
3619
3620 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3621 to character tables built by pcre_maketables() in the current locale. If NULL
3622 is passed, the default tables are used.
3623
3624
3625 Version 2.00 24-Sep-98
3626 ----------------------
3627
3628 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3629 it any more.
3630
3631 2. Allow quantification of (?>) groups, and make it work correctly.
3632
3633 3. The first character computation wasn't working for (?>) groups.
3634
3635 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3636 end of the subject) and add 5.005's \z, which really does match only at the
3637 very end of the subject.
3638
3639 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3640
3641 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3642 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3643 localized options. All options to pcre_study() were also removed.
3644
3645 7. Add other new features from 5.005:
3646
3647 $(?<= positive lookbehind
3648 $(?<! negative lookbehind
3649 (?imsx-imsx) added the unsetting capability
3650 such a setting is global if at outer level; local otherwise
3651 (?imsx-imsx:) non-capturing groups with option setting
3652 (?(cond)re|re) conditional pattern matching
3653
3654 A backreference to itself in a repeated group matches the previous
3655 captured string.
3656
3657 8. General tidying up of studying (both automatic and via "study")
3658 consequential on the addition of new assertions.
3659
3660 9. As in 5.005, unlimited repeated groups that could match an empty substring
3661 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3662 runtime if any iteration does actually match an empty substring.
3663
3664 10. Include the RunTest script in the distribution.
3665
3666 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3667 discrepancies, some of which were old and were also with respect to 5.004. They
3668 have now been fixed.
3669
3670
3671 Version 1.09 28-Apr-98
3672 ----------------------
3673
3674 1. A negated single character class followed by a quantifier with a minimum
3675 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3676 program crashes, or just wrong answers. This did not apply to negated classes
3677 containing more than one character, or to minima other than one.
3678
3679
3680 Version 1.08 27-Mar-98
3681 ----------------------
3682
3683 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3684
3685 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3686 latter must appear before anything that relies on it in the pattern.
3687
3688
3689 Version 1.07 16-Feb-98
3690 ----------------------
3691
3692 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3693 repeat of a potentially empty string).
3694
3695
3696 Version 1.06 23-Jan-98
3697 ----------------------
3698
3699 1. Added Markus Oberhumer's little patches for C++.
3700
3701 2. Literal strings longer than 255 characters were broken.
3702
3703
3704 Version 1.05 23-Dec-97
3705 ----------------------
3706
3707 1. Negated character classes containing more than one character were failing if
3708 PCRE_CASELESS was set at run time.
3709
3710
3711 Version 1.04 19-Dec-97
3712 ----------------------
3713
3714 1. Corrected the man page, where some "const" qualifiers had been omitted.
3715
3716 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3717 input syntax.
3718
3719 3. Fixed memory leak which occurred when a regex with back references was
3720 matched with an offsets vector that wasn't big enough. The temporary memory
3721 that is used in this case wasn't being freed if the match failed.
3722
3723 4. Tidied pcretest to ensure it frees memory that it gets.
3724
3725 5. Temporary memory was being obtained in the case where the passed offsets
3726 vector was exactly big enough.
3727
3728 6. Corrected definition of offsetof() from change 5 below.
3729
3730 7. I had screwed up change 6 below and broken the rules for the use of
3731 setjmp(). Now fixed.
3732
3733
3734 Version 1.03 18-Dec-97
3735 ----------------------
3736
3737 1. A erroneous regex with a missing opening parenthesis was correctly
3738 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3739 on some systems.
3740
3741 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3742 it was reported that one broken compiler failed on the former because "code" is
3743 also an independent variable.
3744
3745 3. The erroneous regex a[]b caused an array overrun reference.
3746
3747 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3748 fail on data ending with that character. (It was going on too far, and checking
3749 the next character, typically a binary zero.) This was specific to the
3750 optimized code for single-character negative classes.
3751
3752 5. Added a contributed patch from the TIN world which does the following:
3753
3754 + Add an undef for memmove, in case the the system defines a macro for it.
3755
3756 + Add a definition of offsetof(), in case there isn't one. (I don't know
3757 the reason behind this - offsetof() is part of the ANSI standard - but
3758 it does no harm).
3759
3760 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3761 most of the places where whitespace preceded '#'. I have given up and
3762 allowed the remaining 2 cases to be at the margin.
3763
3764 + Rename some variables in pcre to eliminate shadowing. This seems very
3765 pedantic, but does no harm, of course.
3766
3767 6. Moved the call to setjmp() into its own function, to get rid of warnings
3768 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3769
3770 7. Constructs such as \d{8,} were compiling into the equivalent of
3771 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3772 outcome, but in this particular case used more store than had been allocated,
3773 which caused the bug to be discovered because it threw up an internal error.
3774
3775 8. The debugging code in both pcre and pcretest for outputting the compiled
3776 form of a regex was going wrong in the case of back references followed by
3777 curly-bracketed repeats.
3778
3779
3780 Version 1.02 12-Dec-97
3781 ----------------------
3782
3783 1. Typos in pcre.3 and comments in the source fixed.
3784
3785 2. Applied a contributed patch to get rid of places where it used to remove
3786 'const' from variables, and fixed some signed/unsigned and uninitialized
3787 variable warnings.
3788
3789 3. Added the "runtest" target to Makefile.
3790
3791 4. Set default compiler flag to -O2 rather than just -O.
3792
3793
3794 Version 1.01 19-Nov-97
3795 ----------------------
3796
3797 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3798 like /([ab]*)*/, that is, for classes with more than one character in them.
3799
3800 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3801 as /((?>a*))*/ (a PCRE_EXTRA facility).
3802
3803
3804 Version 1.00 18-Nov-97
3805 ----------------------
3806
3807 1. Added compile-time macros to support systems such as SunOS4 which don't have
3808 memmove() or strerror() but have other things that can be used instead.
3809
3810 2. Arranged that "make clean" removes the executables.
3811
3812
3813 Version 0.99 27-Oct-97
3814 ----------------------
3815
3816 1. Fixed bug in code for optimizing classes with only one character. It was
3817 initializing a 32-byte map regardless, which could cause it to run off the end
3818 of the memory it had got.
3819
3820 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3821
3822
3823 Version 0.98 22-Oct-97
3824 ----------------------
3825
3826 1. Fixed bug in code for handling temporary memory usage when there are more
3827 back references than supplied space in the ovector. This could cause segfaults.
3828
3829
3830 Version 0.97 21-Oct-97
3831 ----------------------
3832
3833 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3834
3835 2. Optimized negated single characters not to use a bit map.
3836
3837 3. Brought error texts together as macro definitions; clarified some of them;
3838 fixed one that was wrong - it said "range out of order" when it meant "invalid
3839 escape sequence".
3840
3841 4. Changed some char * arguments to const char *.
3842
3843 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3844
3845 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3846 pcretest.
3847
3848
3849 Version 0.96 16-Oct-97
3850 ----------------------
3851
3852 1. Added a simple "pgrep" utility to the distribution.
3853
3854 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3855 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3856 where "ddd" means "one or more decimal digits".
3857
3858 3. Fixed serious bug. If a pattern had a back reference, but the call to
3859 pcre_exec() didn't supply a large enough ovector to record the related
3860 identifying subpattern, the match always failed. PCRE now remembers the number
3861 of the largest back reference, and gets some temporary memory in which to save
3862 the offsets during matching if necessary, in order to ensure that
3863 backreferences always work.
3864
3865 4. Increased the compatibility with Perl in a number of ways:
3866
3867 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3868 to request this handling. The option can be set at compile or exec time.
3869
3870 (b) $ matches before a terminating newline by default; an option
3871 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3872 mode). The option can be set at compile or exec time.
3873
3874 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3875 the same as Perl's. If the decimal number it represents is less than 10
3876 or there aren't that many previous left capturing parentheses, an octal
3877 escape is read. Inside a character class, it's always an octal escape,
3878 even if it is a single digit.
3879
3880 (d) An escaped but undefined alphabetic character is taken as a literal,
3881 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3882 escapes.
3883
3884 (e) {0} is now permitted. (The previous item is removed from the compiled
3885 pattern).
3886
3887 5. Changed all the names of code files so that the basic parts are no longer
3888 than 10 characters, and abolished the teeny "globals.c" file.
3889
3890 6. Changed the handling of character classes; they are now done with a 32-byte
3891 bit map always.
3892
3893 7. Added the -d and /D options to pcretest to make it possible to look at the
3894 internals of compilation without having to recompile pcre.
3895
3896
3897 Version 0.95 23-Sep-97
3898 ----------------------
3899
3900 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3901 \x20 at the start of a run of normal characters. These were being treated as
3902 real characters, instead of the source characters being re-checked.
3903
3904
3905 Version 0.94 18-Sep-97
3906 ----------------------
3907
3908 1. The functions are now thread-safe, with the caveat that the global variables
3909 containing pointers to malloc() and free() or alternative functions are the
3910 same for all threads.
3911
3912 2. Get pcre_study() to generate a bitmap of initial characters for non-
3913 anchored patterns when this is possible, and use it if passed to pcre_exec().
3914
3915
3916 Version 0.93 15-Sep-97
3917 ----------------------
3918
3919 1. /(b)|(:+)/ was computing an incorrect first character.
3920
3921 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3922 but not actually doing anything yet.
3923
3924 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3925 as Perl does (e.g. [-az] or [az-]).
3926
3927 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3928 all possible positions.
3929
3930 5. Split up into different modules to avoid including unneeded functions in a
3931 compiled binary. However, compile and exec are still in one module. The "study"
3932 function is split off.
3933
3934 6. The character tables are now in a separate module whose source is generated
3935 by an auxiliary program - but can then be edited by hand if required. There are
3936 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3937 toupper() in the code.
3938
3939 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3940 make them global. Abolish the function for setting them, as the caller can now
3941 set them directly.
3942
3943
3944 Version 0.92 11-Sep-97
3945 ----------------------
3946
3947 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3948 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3949
3950 2. Caseless matching was not working in character classes if the characters in
3951 the pattern were in upper case.
3952
3953 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3954
3955 4. Make PCRE_ANCHORED public and accept as a compile option.
3956
3957 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3958 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3959 pass them.
3960
3961 6. Give an error if bad option bits passed at compile or run time.
3962
3963 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3964 pcretest to cause it to pass that flag.
3965
3966 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3967 options, and the first character, if set.
3968
3969 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3970
3971
3972 Version 0.91 10-Sep-97
3973 ----------------------
3974
3975 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3976 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3977
3978 2. PCRE was looping on encountering an indefinitely repeated back reference to
3979 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3980 Perl does - treats the match as successful.
3981
3982 ****

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