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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 618 - (show annotations)
Sat Jul 16 17:24:16 2011 UTC (8 years, 2 months ago) by ph10
File size: 205755 byte(s)
Re-do atomic group processing to fix backtrack capture bugs. Recursion is also 
re-worked.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.13 30-Apr-2011
5 ------------------------
6
7 1. The Unicode data tables have been updated to Unicode 6.0.0.
8
9 2. Two minor typos in pcre_internal.h have been fixed.
10
11 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
12 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
13 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
14
15 4. There were a number of related bugs in the code for matching backrefences
16 caselessly in UTF-8 mode when codes for the characters concerned were
17 different numbers of bytes. For example, U+023A and U+2C65 are an upper
18 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
19 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
20 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
21 2-byte code at the end of the subject (it thought there wasn't enough data
22 left).
23
24 5. Comprehensive information about what went wrong is now returned by
25 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
26 as the output vector has at least 2 elements. The offset of the start of
27 the failing character and a reason code are placed in the vector.
28
29 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
30 now returned is for the first byte of the failing character, instead of the
31 last byte inspected. This is an incompatible change, but I hope it is small
32 enough not to be a problem. It makes the returned offset consistent with
33 pcre_exec() and pcre_dfa_exec().
34
35 7. pcretest now gives a text phrase as well as the error number when
36 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
37 failure, the offset and reason code are output.
38
39 8. When \R was used with a maximizing quantifier it failed to skip backwards
40 over a \r\n pair if the subsequent match failed. Instead, it just skipped
41 back over a single character (\n). This seems wrong (because it treated the
42 two characters as a single entity when going forwards), conflicts with the
43 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
44 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
45 has been changed.
46
47 9. Some internal refactoring has changed the processing so that the handling
48 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
49 time (the PCRE_DOTALL option was changed this way some time ago: version
50 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
51 which was always a bit of a fudge. It also means that there is one less
52 argument for the match() function, which reduces its stack requirements
53 slightly. This change also fixes an incompatibility with Perl: the pattern
54 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
55
56 10. More internal refactoring has drastically reduced the number of recursive
57 calls to match() for possessively repeated groups such as (abc)++ when
58 using pcre_exec().
59
60 11. While implementing 10, a number of bugs in the handling of groups were
61 discovered and fixed:
62
63 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
64 (a|)*(?1) gave a compile-time internal error.
65 ((a|)+)+ did not notice that the outer group could match an empty string.
66 (^a|^)+ was not marked as anchored.
67 (.*a|.*)+ was not marked as matching at start or after a newline.
68
69 12. Yet more internal refactoring has removed another argument from the match()
70 function. Special calls to this function are now indicated by setting a
71 value in a variable in the "match data" data block.
72
73 13. Be more explicit in pcre_study() instead of relying on "default" for
74 opcodes that mean there is no starting character; this means that when new
75 ones are added and accidentally left out of pcre_study(), testing should
76 pick them up.
77
78 14. The -s option of pcretest has been documented for ages as being an old
79 synonym of -m (show memory usage). I have changed it to mean "force study
80 for every regex", that is, assume /S for every regex. This is similar to -i
81 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
82 using it. It makes it easier to run collections of tests with and without
83 study enabled, and thereby test pcre_study() more easily. All the standard
84 tests are now run with and without -s (but some patterns can be marked as
85 "never study" - see 20 below).
86
87 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
88 restoration of the capturing data to the outer values was not happening
89 correctly.
90
91 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
92 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
93 pattern had matched an empty string, and so incorrectly returned a no
94 match.
95
96 17. There was optimizing code for the last branch of non-capturing parentheses,
97 and also for the obeyed branch of a conditional subexpression, which used
98 tail recursion to cut down on stack usage. Unfortunately, not that there is
99 the possibility of (*THEN) occurring in these branches, tail recursion is
100 no longer possible because the return has to be checked for (*THEN). These
101 two optimizations have therefore been removed.
102
103 18. If a pattern containing \R was studied, it was assumed that \R always
104 matched two bytes, thus causing the minimum subject length to be
105 incorrectly computed because \R can also match just one byte.
106
107 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
108 was incorrectly computed.
109
110 20. If /S is present twice on a test pattern in pcretest input, it *disables*
111 studying, thereby overriding the use of -s on the command line. This is
112 necessary for one or two tests to keep the output identical in both cases.
113
114 21. When (*ACCEPT) was used in an assertion that matched an empty string and
115 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
116
117 22. When an atomic group that contained a capturing parenthesis was
118 successfully matched, but the branch in which it appeared failed, the
119 capturing was not being forgotten if a higher numbered group was later
120 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
121 group 1 to "a", when in fact it should be unset. This applied to multi-
122 branched capturing and non-capturing groups, repeated or not, and also to
123 positive assertions (capturing in negative assertions is not well defined
124 in PCRE) and also to nested atomic groups.
125
126 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
127 subject after a captured substring (to make it easier to tell which of a
128 number of identical substrings has been captured).
129
130 24. The way atomic groups are processed by pcre_exec() has been changed so that
131 if they are repeated, backtracking one repetition now resets captured
132 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
133 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
134 "aaa". Previously, it would have been "a". As part of this code
135 refactoring, the way recursive calls are handled has also been changed.
136
137
138 Version 8.12 15-Jan-2011
139 ------------------------
140
141 1. Fixed some typos in the markup of the man pages, and wrote a script that
142 checks for such things as part of the documentation building process.
143
144 2. On a big-endian 64-bit system, pcregrep did not correctly process the
145 --match-limit and --recursion-limit options (added for 8.11). In
146 particular, this made one of the standard tests fail. (The integer value
147 went into the wrong half of a long int.)
148
149 3. If the --colour option was given to pcregrep with -v (invert match), it
150 did strange things, either producing crazy output, or crashing. It should,
151 of course, ignore a request for colour when reporting lines that do not
152 match.
153
154 4. Another pcregrep bug caused similar problems if --colour was specified with
155 -M (multiline) and the pattern match finished with a line ending.
156
157 5. In pcregrep, when a pattern that ended with a literal newline sequence was
158 matched in multiline mode, the following line was shown as part of the
159 match. This seems wrong, so I have changed it.
160
161 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
162 the check for further matches in the same line (so they could be coloured)
163 to overrun the end of the current line. If another match was found, it was
164 incorrectly shown (and then shown again when found in the next line).
165
166 7. If pcregrep was compiled under Windows, there was a reference to the
167 function pcregrep_exit() before it was defined. I am assuming this was
168 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
169 reported by a user. I've moved the definition above the reference.
170
171
172 Version 8.11 10-Dec-2010
173 ------------------------
174
175 1. (*THEN) was not working properly if there were untried alternatives prior
176 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
177 backtracked to try for "b" instead of moving to the next alternative branch
178 at the same level (in this case, to look for "c"). The Perl documentation
179 is clear that when (*THEN) is backtracked onto, it goes to the "next
180 alternative in the innermost enclosing group".
181
182 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
183 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
184 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
185 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
186 (*THEN).
187
188 3. If \s appeared in a character class, it removed the VT character from
189 the class, even if it had been included by some previous item, for example
190 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
191 of \s, but is part of the POSIX "space" class.)
192
193 4. A partial match never returns an empty string (because you can always
194 match an empty string at the end of the subject); however the checking for
195 an empty string was starting at the "start of match" point. This has been
196 changed to the "earliest inspected character" point, because the returned
197 data for a partial match starts at this character. This means that, for
198 example, /(?<=abc)def/ gives a partial match for the subject "abc"
199 (previously it gave "no match").
200
201 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
202 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
203 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
204 has an implication that the given string is incomplete (because a partial
205 match is preferred over a full match). For this reason, these items now
206 give a partial match in this situation. [Aside: previously, the one case
207 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
208 match rather than a full match, which was wrong by the old rules, but is
209 now correct.]
210
211 6. There was a bug in the handling of #-introduced comments, recognized when
212 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
213 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
214 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
215 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
216 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
217 places in pcre_compile().
218
219 7. Related to (6) above, when pcre_compile() was skipping #-introduced
220 comments when looking ahead for named forward references to subpatterns,
221 the only newline sequence it recognized was NL. It now handles newlines
222 according to the set newline convention.
223
224 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
225 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
226 cater for a lack of strerror(). These oversights have been fixed.
227
228 9. Added --match-limit and --recursion-limit to pcregrep.
229
230 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
231
232 11. When the -o option was used, pcregrep was setting a return code of 1, even
233 when matches were found, and --line-buffered was not being honoured.
234
235 12. Added an optional parentheses number to the -o and --only-matching options
236 of pcregrep.
237
238 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
239 can match an empty string. The code to do it in pcretest and pcredemo
240 needed fixing:
241
242 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
243 only one byte after an empty string match just before CRLF (this case
244 just got forgotten; "any" and "anycrlf" were OK).
245
246 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
247 mode when an empty string match preceded an ASCII character followed by
248 a non-ASCII character. (The code for advancing by one character rather
249 than one byte was nonsense.)
250
251 (c) The pcredemo.c sample program did not have any code at all to handle
252 the cases when CRLF is a valid newline sequence.
253
254 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
255 as a starting offset was within the subject string. There is now a new
256 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
257 negative or greater than the length of the string. In order to test this,
258 pcretest is extended to allow the setting of negative starting offsets.
259
260 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
261 starting offset points to the beginning of a UTF-8 character was
262 unnecessarily clumsy. I tidied it up.
263
264 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
265 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
266
267 17. Nobody had reported that the --include_dir option, which was added in
268 release 7.7 should have been called --include-dir (hyphen, not underscore)
269 for compatibility with GNU grep. I have changed it to --include-dir, but
270 left --include_dir as an undocumented synonym, and the same for
271 --exclude-dir, though that is not available in GNU grep, at least as of
272 release 2.5.4.
273
274 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
275 characters from a string of bytes) have been redefined so as not to use
276 loops, in order to improve performance in some environments. At the same
277 time, I abstracted some of the common code into auxiliary macros to save
278 repetition (this should not affect the compiled code).
279
280 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
281 compile-time error is now given if \c is not followed by an ASCII
282 character, that is, a byte less than 128. (In EBCDIC mode, the code is
283 different, and any byte value is allowed.)
284
285 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
286 START_OPTIMIZE option, which is now allowed at compile time - but just
287 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
288 to pcregrep and other applications that have no direct access to PCRE
289 options. The new /Y option in pcretest sets this option when calling
290 pcre_compile().
291
292 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
293 back references. Groups containing recursive back references were forced to
294 be atomic by that change, but in the case of named groups, the amount of
295 memory required was incorrectly computed, leading to "Failed: internal
296 error: code overflow". This has been fixed.
297
298 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
299 pcretest.c, to avoid build problems in some Borland environments.
300
301
302 Version 8.10 25-Jun-2010
303 ------------------------
304
305 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
306 THEN.
307
308 2. (*ACCEPT) was not working when inside an atomic group.
309
310 3. Inside a character class, \B is treated as a literal by default, but
311 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
312 causes the error). The code is unchanged, but I tidied the documentation.
313
314 4. Inside a character class, PCRE always treated \R and \X as literals,
315 whereas Perl faults them if its -w option is set. I have changed PCRE so
316 that it faults them when PCRE_EXTRA is set.
317
318 5. Added support for \N, which always matches any character other than
319 newline. (It is the same as "." when PCRE_DOTALL is not set.)
320
321 6. When compiling pcregrep with newer versions of gcc which may have
322 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
323 declared with attribute warn_unused_result" were given. Just casting the
324 result to (void) does not stop the warnings; a more elaborate fudge is
325 needed. I've used a macro to implement this.
326
327 7. Minor change to pcretest.c to avoid a compiler warning.
328
329 8. Added four artifical Unicode properties to help with an option to make
330 \s etc use properties (see next item). The new properties are: Xan
331 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
332
333 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
334 use Unicode properties. (*UCP) at the start of a pattern can be used to set
335 this option. Modified pcretest to add /W to test this facility. Added
336 REG_UCP to make it available via the POSIX interface.
337
338 10. Added --line-buffered to pcregrep.
339
340 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
341 studied, and the match started with a letter with a code point greater than
342 127 whose first byte was different to the first byte of the other case of
343 the letter, the other case of this starting letter was not recognized
344 (#976).
345
346 12. If a pattern that was studied started with a repeated Unicode property
347 test, for example, \p{Nd}+, there was the theoretical possibility of
348 setting up an incorrect bitmap of starting bytes, but fortunately it could
349 not have actually happened in practice until change 8 above was made (it
350 added property types that matched character-matching opcodes).
351
352 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
353 possible starting bytes for non-anchored patterns.
354
355 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
356 \R, and also a number of cases that involve Unicode properties, both
357 explicit and implicit when PCRE_UCP is set.
358
359 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
360 input, it could crash or give wrong results if characters with values
361 greater than 0xc0 were present in the subject string. (Detail: it assumed
362 UTF-8 input when processing these items.)
363
364 16. Added a lot of (int) casts to avoid compiler warnings in systems where
365 size_t is 64-bit (#991).
366
367 17. Added a check for running out of memory when PCRE is compiled with
368 --disable-stack-for-recursion (#990).
369
370 18. If the last data line in a file for pcretest does not have a newline on
371 the end, a newline was missing in the output.
372
373 19. The default pcre_chartables.c file recognizes only ASCII characters (values
374 less than 128) in its various bitmaps. However, there is a facility for
375 generating tables according to the current locale when PCRE is compiled. It
376 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
377 characters, are recognized by isspace() and therefore were getting set in
378 these tables, and indeed these tables seem to approximate to ISO 8859. This
379 caused a problem in UTF-8 mode when pcre_study() was used to create a list
380 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
381 which of course cannot start UTF-8 characters. I have changed the code so
382 that only real ASCII characters (less than 128) and the correct starting
383 bytes for UTF-8 encodings are set for characters greater than 127 when in
384 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
385 altogether.)
386
387 20. Added the /T option to pcretest so as to be able to run tests with non-
388 standard character tables, thus making it possible to include the tests
389 used for 19 above in the standard set of tests.
390
391 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
392 reference to a subpattern the other side of a comment that contains an
393 opening parenthesis caused either an internal compiling error, or a
394 reference to the wrong subpattern.
395
396
397 Version 8.02 19-Mar-2010
398 ------------------------
399
400 1. The Unicode data tables have been updated to Unicode 5.2.0.
401
402 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
403 configured.
404
405 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
406 original author of that file, following a query about its status.
407
408 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
409 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
410
411 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
412 quantifier applied to a forward-referencing subroutine call, could compile
413 incorrect code or give the error "internal error: previously-checked
414 referenced subpattern not found".
415
416 6. Both MS Visual Studio and Symbian OS have problems with initializing
417 variables to point to external functions. For these systems, therefore,
418 pcre_malloc etc. are now initialized to local functions that call the
419 relevant global functions.
420
421 7. There were two entries missing in the vectors called coptable and poptable
422 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
423 I've fixed the data, and added a kludgy way of testing at compile time that
424 the lengths are correct (equal to the number of opcodes).
425
426 8. Following on from 7, I added a similar kludge to check the length of the
427 eint vector in pcreposix.c.
428
429 9. Error texts for pcre_compile() are held as one long string to avoid too
430 much relocation at load time. To find a text, the string is searched,
431 counting zeros. There was no check for running off the end of the string,
432 which could happen if a new error number was added without updating the
433 string.
434
435 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
436
437 11. \K was not working if it appeared in an atomic group or in a group that
438 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
439 \K is "not well defined" if used in an assertion. PCRE now accepts it if
440 the assertion is positive, but not if it is negative.
441
442 12. Change 11 fortuitously reduced the size of the stack frame used in the
443 "match()" function of pcre_exec.c by one pointer. Forthcoming
444 implementation of support for (*MARK) will need an extra pointer on the
445 stack; I have reserved it now, so that the stack frame size does not
446 decrease.
447
448 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
449 item in branch that calls a recursion is a subroutine call - as in the
450 second branch in the above example - was incorrectly given the compile-
451 time error "recursive call could loop indefinitely" because pcre_compile()
452 was not correctly checking the subroutine for matching a non-empty string.
453
454 14. The checks for overrunning compiling workspace could trigger after an
455 overrun had occurred. This is a "should never occur" error, but it can be
456 triggered by pathological patterns such as hundreds of nested parentheses.
457 The checks now trigger 100 bytes before the end of the workspace.
458
459 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
460
461
462 Version 8.01 19-Jan-2010
463 ------------------------
464
465 1. If a pattern contained a conditional subpattern with only one branch (in
466 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
467 computed the wrong minimum data length (which is of course zero for such
468 subpatterns). This could cause incorrect "no match" results.
469
470 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
471 the pattern is reset in the first branch, pcre_compile() failed with
472 "internal error: code overflow at offset...". This happened only when
473 the reset was to the original external option setting. (An optimization
474 abstracts leading options settings into an external setting, which was the
475 cause of this.)
476
477 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
478 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
479 assertion pattern did not match (meaning that the assertion was true), it
480 was incorrectly treated as false if the SKIP had been reached during the
481 matching. This also applied to assertions used as conditions.
482
483 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
484 assertion subpattern, including such a pattern used as a condition,
485 unpredictable results occurred, instead of the error return
486 PCRE_ERROR_DFA_UITEM.
487
488 5. The C++ GlobalReplace function was not working like Perl for the special
489 situation when an empty string is matched. It now does the fancy magic
490 stuff that is necessary.
491
492 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
493 removed. (These were left over from very, very early versions of PCRE.)
494
495 7. Some cosmetic changes to the code to make life easier when compiling it
496 as part of something else:
497
498 (a) Change DEBUG to PCRE_DEBUG.
499
500 (b) In pcre_compile(), rename the member of the "branch_chain" structure
501 called "current" as "current_branch", to prevent a collision with the
502 Linux macro when compiled as a kernel module.
503
504 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
505 prevent a collision with the Linux macro when compiled as a kernel
506 module.
507
508 8. In pcre_compile() there are some checks for integer overflows that used to
509 cast potentially large values to (double). This has been changed to that
510 when building, a check for int64_t is made, and if it is found, it is used
511 instead, thus avoiding the use of floating point arithmetic. (There is no
512 other use of FP in PCRE.) If int64_t is not found, the fallback is to
513 double.
514
515 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
516 2005 (difference between two addresses compared to an unsigned value).
517
518 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
519 custom one, because of the following reported problem in Windows:
520
521 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
522 under Win32.
523 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
524 therefore missing the function definition.
525 - The compiler thus generates a "C" signature for the test function.
526 - The linker fails to find the "C" function.
527 - PCRE fails to configure if asked to do so against libbz2.
528
529 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
530 messages were output:
531
532 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
533 rerunning libtoolize, to keep the correct libtool macros in-tree.
534 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
535
536 I have done both of these things.
537
538 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
539 most of the time, it *can* run out if it is given a pattern that contains a
540 runaway infinite recursion. I updated the discussion in the pcrestack man
541 page.
542
543 13. Now that we have gone to the x.xx style of version numbers, the minor
544 version may start with zero. Using 08 or 09 is a bad idea because users
545 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
546 interpreted as invalid octal numbers. I've updated the previous comment in
547 configure.ac, and also added a check that gives an error if 08 or 09 are
548 used.
549
550 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
551 causing partial matching to fail when the end of the subject matched \W
552 in a UTF-8 pattern where \W was quantified with a minimum of 3.
553
554 15. There were some discrepancies between the declarations in pcre_internal.h
555 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
556 their definitions. The declarations used "const uschar *" and the
557 definitions used USPTR. Even though USPTR is normally defined as "const
558 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
559 reported that: "This difference in casting confuses some C++ compilers, for
560 example, SunCC recognizes above declarations as different functions and
561 generates broken code for hbpcre." I have changed the declarations to use
562 USPTR.
563
564 16. GNU libtool is named differently on some systems. The autogen.sh script now
565 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
566 (FreeBSD).
567
568 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
569 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
570 comment: "Figure out how to create a longlong from a string: strtoll and
571 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
572 instance, but it only takes 2 args instead of 3!"
573
574 18. A subtle bug concerned with back references has been fixed by a change of
575 specification, with a corresponding code fix. A pattern such as
576 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
577 refers, was giving matches when it shouldn't. For example, xa=xaaa would
578 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
579 same bug. Such groups have to be quantified to be useful, or contained
580 inside another quantified group. (If there's no repetition, the reference
581 can never match.) The problem arises because, having left the group and
582 moved on to the rest of the pattern, a later failure that backtracks into
583 the group uses the captured value from the final iteration of the group
584 rather than the correct earlier one. I have fixed this in PCRE by forcing
585 any group that contains a reference to itself to be an atomic group; that
586 is, there cannot be any backtracking into it once it has completed. This is
587 similar to recursive and subroutine calls.
588
589
590 Version 8.00 19-Oct-09
591 ----------------------
592
593 1. The table for translating pcre_compile() error codes into POSIX error codes
594 was out-of-date, and there was no check on the pcre_compile() error code
595 being within the table. This could lead to an OK return being given in
596 error.
597
598 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
599 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
600 in a Windows environment.
601
602 3. The pcregrep --count option prints the count for each file even when it is
603 zero, as does GNU grep. However, pcregrep was also printing all files when
604 --files-with-matches was added. Now, when both options are given, it prints
605 counts only for those files that have at least one match. (GNU grep just
606 prints the file name in this circumstance, but including the count seems
607 more useful - otherwise, why use --count?) Also ensured that the
608 combination -clh just lists non-zero counts, with no names.
609
610 4. The long form of the pcregrep -F option was incorrectly implemented as
611 --fixed_strings instead of --fixed-strings. This is an incompatible change,
612 but it seems right to fix it, and I didn't think it was worth preserving
613 the old behaviour.
614
615 5. The command line items --regex=pattern and --regexp=pattern were not
616 recognized by pcregrep, which required --regex pattern or --regexp pattern
617 (with a space rather than an '='). The man page documented the '=' forms,
618 which are compatible with GNU grep; these now work.
619
620 6. No libpcreposix.pc file was created for pkg-config; there was just
621 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
622
623 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
624 when UCP support is not needed, by modifying the Python script that
625 generates it from Unicode data files. This should not matter if the module
626 is correctly used as a library, but I received one complaint about 50K of
627 unwanted data. My guess is that the person linked everything into his
628 program rather than using a library. Anyway, it does no harm.
629
630 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
631 was a minimum greater than 1 for a wide character in a possessive
632 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
633 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
634 character. Chaos in the form of incorrect output or a compiling loop could
635 result.
636
637 9. The restrictions on what a pattern can contain when partial matching is
638 requested for pcre_exec() have been removed. All patterns can now be
639 partially matched by this function. In addition, if there are at least two
640 slots in the offset vector, the offset of the earliest inspected character
641 for the match and the offset of the end of the subject are set in them when
642 PCRE_ERROR_PARTIAL is returned.
643
644 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
645 synonymous with PCRE_PARTIAL, for backwards compatibility, and
646 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
647 and may be more useful for multi-segment matching.
648
649 11. Partial matching with pcre_exec() is now more intuitive. A partial match
650 used to be given if ever the end of the subject was reached; now it is
651 given only if matching could not proceed because another character was
652 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
653 string "Z", which now yields "no match" instead of "partial match". In the
654 case of pcre_dfa_exec(), "no match" is given if every matching path for the
655 final character ended with (*FAIL).
656
657 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
658 if the pattern had a "must contain" character that was already found in the
659 earlier partial match, unless partial matching was again requested. For
660 example, with the pattern /dog.(body)?/, the "must contain" character is
661 "g". If the first part-match was for the string "dog", restarting with
662 "sbody" failed. This bug has been fixed.
663
664 13. The string returned by pcre_dfa_exec() after a partial match has been
665 changed so that it starts at the first inspected character rather than the
666 first character of the match. This makes a difference only if the pattern
667 starts with a lookbehind assertion or \b or \B (\K is not supported by
668 pcre_dfa_exec()). It's an incompatible change, but it makes the two
669 matching functions compatible, and I think it's the right thing to do.
670
671 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
672 so that the demonstration program is easily available in environments where
673 PCRE has not been installed from source.
674
675 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
676 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
677 library.
678
679 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
680 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
681 is not the first non-POSIX option to be added. Clearly some people find
682 these options useful.
683
684 17. If a caller to the POSIX matching function regexec() passes a non-zero
685 value for nmatch with a NULL value for pmatch, the value of
686 nmatch is forced to zero.
687
688 18. RunGrepTest did not have a test for the availability of the -u option of
689 the diff command, as RunTest does. It now checks in the same way as
690 RunTest, and also checks for the -b option.
691
692 19. If an odd number of negated classes containing just a single character
693 interposed, within parentheses, between a forward reference to a named
694 subpattern and the definition of the subpattern, compilation crashed with
695 an internal error, complaining that it could not find the referenced
696 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
697 [The bug was that it was starting one character too far in when skipping
698 over the character class, thus treating the ] as data rather than
699 terminating the class. This meant it could skip too much.]
700
701 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
702 /g option in pcretest when the pattern contains \K, which makes it possible
703 to have an empty string match not at the start, even when the pattern is
704 anchored. Updated pcretest and pcredemo to use this option.
705
706 21. If the maximum number of capturing subpatterns in a recursion was greater
707 than the maximum at the outer level, the higher number was returned, but
708 with unset values at the outer level. The correct (outer level) value is
709 now given.
710
711 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
712 PCRE did not set those parentheses (unlike Perl). I have now found a way to
713 make it do so. The string so far is captured, making this feature
714 compatible with Perl.
715
716 23. The tests have been re-organized, adding tests 11 and 12, to make it
717 possible to check the Perl 5.10 features against Perl 5.10.
718
719 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
720 pattern matches a fixed length string. PCRE did not allow this; now it
721 does. Neither allows recursion.
722
723 25. I finally figured out how to implement a request to provide the minimum
724 length of subject string that was needed in order to match a given pattern.
725 (It was back references and recursion that I had previously got hung up
726 on.) This code has now been added to pcre_study(); it finds a lower bound
727 to the length of subject needed. It is not necessarily the greatest lower
728 bound, but using it to avoid searching strings that are too short does give
729 some useful speed-ups. The value is available to calling programs via
730 pcre_fullinfo().
731
732 26. While implementing 25, I discovered to my embarrassment that pcretest had
733 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
734 study optimizations had never been tested with that matching function.
735 Oops. What is worse, even when it was passed study data, there was a bug in
736 pcre_dfa_exec() that meant it never actually used it. Double oops. There
737 were also very few tests of studied patterns with pcre_dfa_exec().
738
739 27. If (?| is used to create subpatterns with duplicate numbers, they are now
740 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
741 on the other side of the coin, they are no longer allowed to have different
742 names, because these cannot be distinguished in PCRE, and this has caused
743 confusion. (This is a difference from Perl.)
744
745 28. When duplicate subpattern names are present (necessarily with different
746 numbers, as required by 27 above), and a test is made by name in a
747 conditional pattern, either for a subpattern having been matched, or for
748 recursion in such a pattern, all the associated numbered subpatterns are
749 tested, and the overall condition is true if the condition is true for any
750 one of them. This is the way Perl works, and is also more like the way
751 testing by number works.
752
753
754 Version 7.9 11-Apr-09
755 ---------------------
756
757 1. When building with support for bzlib/zlib (pcregrep) and/or readline
758 (pcretest), all targets were linked against these libraries. This included
759 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
760 libraries. This caused unwanted dependencies to be created. This problem
761 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
762 pcretest is linked with readline.
763
764 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
765 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
766 moved outside it again, because FALSE and TRUE are already defined in AIX,
767 but BOOL is not.
768
769 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
770 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
771
772 4. The pcregrep documentation said spaces were inserted as well as colons (or
773 hyphens) following file names and line numbers when outputting matching
774 lines. This is not true; no spaces are inserted. I have also clarified the
775 wording for the --colour (or --color) option.
776
777 5. In pcregrep, when --colour was used with -o, the list of matching strings
778 was not coloured; this is different to GNU grep, so I have changed it to be
779 the same.
780
781 6. When --colo(u)r was used in pcregrep, only the first matching substring in
782 each matching line was coloured. Now it goes on to look for further matches
783 of any of the test patterns, which is the same behaviour as GNU grep.
784
785 7. A pattern that could match an empty string could cause pcregrep to loop; it
786 doesn't make sense to accept an empty string match in pcregrep, so I have
787 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
788 seems to be how GNU grep behaves.
789
790 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
791 start or after a newline", because the conditional assertion was not being
792 correctly handled. The rule now is that both the assertion and what follows
793 in the first alternative must satisfy the test.
794
795 9. If auto-callout was enabled in a pattern with a conditional group whose
796 condition was an assertion, PCRE could crash during matching, both with
797 pcre_exec() and pcre_dfa_exec().
798
799 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
800 used for matching.
801
802 11. Unicode property support in character classes was not working for
803 characters (bytes) greater than 127 when not in UTF-8 mode.
804
805 12. Added the -M command line option to pcretest.
806
807 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
808
809 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
810
811 16. Added comments and documentation about mis-use of no_arg in the C++
812 wrapper.
813
814 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
815 from Martin Jerabek that uses macro names for all relevant character and
816 string constants.
817
818 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
819 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
820 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
821 these, but not everybody uses configure.
822
823 19. A conditional group that had only one branch was not being correctly
824 recognized as an item that could match an empty string. This meant that an
825 enclosing group might also not be so recognized, causing infinite looping
826 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
827 with the subject "ab", where knowledge that the repeated group can match
828 nothing is needed in order to break the loop.
829
830 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
831 exec(), but without supplying a callout function, matching went wrong.
832
833 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
834 leak if the size of the offset vector was greater than 30. When the vector
835 is smaller, the saved offsets during recursion go onto a local stack
836 vector, but for larger vectors malloc() is used. It was failing to free
837 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
838 error, in fact).
839
840 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
841 heapframe that is used only when UTF-8 support is enabled. This caused no
842 problem, but was untidy.
843
844 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
845 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
846 included within another project.
847
848 24. Steven Van Ingelgem's patches to add more options to the CMake support,
849 slightly modified by me:
850
851 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
852 not building pcregrep.
853
854 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
855 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
856
857 25. Forward references, both numeric and by name, in patterns that made use of
858 duplicate group numbers, could behave incorrectly or give incorrect errors,
859 because when scanning forward to find the reference group, PCRE was not
860 taking into account the duplicate group numbers. A pattern such as
861 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
862
863 26. Changed a few more instances of "const unsigned char *" to USPTR, making
864 the feature of a custom pointer more persuasive (as requested by a user).
865
866 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
867 pcretest.c, inside #ifndefs, because it seems they are sometimes already
868 pre-defined.
869
870 28. Added support for (*UTF8) at the start of a pattern.
871
872 29. Arrange for flags added by the "release type" setting in CMake to be shown
873 in the configuration summary.
874
875
876 Version 7.8 05-Sep-08
877 ---------------------
878
879 1. Replaced UCP searching code with optimized version as implemented for Ad
880 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
881 stage table and inline lookup instead of a function, giving speed ups of 2
882 to 5 times on some simple patterns that I tested. Permission was given to
883 distribute the MultiStage2.py script that generates the tables (it's not in
884 the tarball, but is in the Subversion repository).
885
886 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
887 scripts.
888
889 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
890 a group with a zero qualifier. The result of the study could be incorrect,
891 or the function might crash, depending on the pattern.
892
893 4. Caseless matching was not working for non-ASCII characters in back
894 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
895 It now works when Unicode Property Support is available.
896
897 5. In pcretest, an escape such as \x{de} in the data was always generating
898 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
899 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
900 truncation.
901
902 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
903
904 7. Added two (int) casts to pcregrep when printing the difference of two
905 pointers, in case they are 64-bit values.
906
907 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
908 test 2 if it fails.
909
910 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
911 and a #define of that name to empty if it is not externally set. This is to
912 allow users of MSVC to set it if necessary.
913
914 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
915 the convenience functions in the pcre_get.c source file.
916
917 11. An option change at the start of a pattern that had top-level alternatives
918 could cause overwriting and/or a crash. This command provoked a crash in
919 some environments:
920
921 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
922
923 This potential security problem was recorded as CVE-2008-2371.
924
925 12. For a pattern where the match had to start at the beginning or immediately
926 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
927 pcre_dfa_exec() could read past the end of the passed subject if there was
928 no match. To help with detecting such bugs (e.g. with valgrind), I modified
929 pcretest so that it places the subject at the end of its malloc-ed buffer.
930
931 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
932 exec() might read past the end of the data buffer in UTF-8 mode.
933
934 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
935 the data contained the byte 0x85 as part of a UTF-8 character within its
936 first line. This applied both to normal and DFA matching.
937
938 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
939 /^[^d]*?$/8 failed to match "abc".
940
941 16. Added a missing copyright notice to pcrecpp_internal.h.
942
943 17. Make it more clear in the documentation that values returned from
944 pcre_exec() in ovector are byte offsets, not character counts.
945
946 18. Tidied a few places to stop certain compilers from issuing warnings.
947
948 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
949 supplied by Stefan Weber. I made a further small update for 7.8 because
950 there is a change of source arrangements: the pcre_searchfuncs.c module is
951 replaced by pcre_ucd.c.
952
953
954 Version 7.7 07-May-08
955 ---------------------
956
957 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
958 a string to a long long, pretend we don't even have a long long." This is
959 done by checking for the strtoq, strtoll, and _strtoi64 functions.
960
961 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
962 pre-7.6 versions, which defined a global no_arg variable instead of putting
963 it in the RE class. (See also #8 below.)
964
965 3. Remove a line of dead code, identified by coverity and reported by Nuno
966 Lopes.
967
968 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
969
970 (1) The include/exclude patterns were being applied to the whole pathnames
971 of files, instead of just to the final components.
972
973 (2) If there was more than one level of directory, the subdirectories were
974 skipped unless they satisfied the include/exclude conditions. This is
975 inconsistent with GNU grep (and could even be seen as contrary to the
976 pcregrep specification - which I improved to make it absolutely clear).
977 The action now is always to scan all levels of directory, and just
978 apply the include/exclude patterns to regular files.
979
980 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
981 --exclude_dir in the tests to avoid scanning .svn directories.
982
983 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
984 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
985 doesn't support NULs in patterns.
986
987 7. Added some missing "const"s to declarations of static tables in
988 pcre_compile.c and pcre_dfa_exec.c.
989
990 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
991 caused by fix #2 above. (Subsequently also a second patch to fix the
992 first patch. And a third patch - this was a messy problem.)
993
994 9. Applied Craig's patch to remove the use of push_back().
995
996 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
997 matching function regexec().
998
999 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1000 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1001 references. PCRE supports relative numbers with this syntax (I don't think
1002 Oniguruma does).
1003
1004 12. Previously, a group with a zero repeat such as (...){0} was completely
1005 omitted from the compiled regex. However, this means that if the group
1006 was called as a subroutine from elsewhere in the pattern, things went wrong
1007 (an internal error was given). Such groups are now left in the compiled
1008 pattern, with a new opcode that causes them to be skipped at execution
1009 time.
1010
1011 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1012 to the way PCRE behaves:
1013
1014 (a) A lone ] character is dis-allowed (Perl treats it as data).
1015
1016 (b) A back reference to an unmatched subpattern matches an empty string
1017 (Perl fails the current match path).
1018
1019 (c) A data ] in a character class must be notated as \] because if the
1020 first data character in a class is ], it defines an empty class. (In
1021 Perl it is not possible to have an empty class.) The empty class []
1022 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1023 The negative empty class [^] matches any one character, independently
1024 of the DOTALL setting.
1025
1026 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1027 non-existent subpattern following a character class starting with ']' and
1028 containing () gave an internal compiling error instead of "reference to
1029 non-existent subpattern". Fortunately, when the pattern did exist, the
1030 compiled code was correct. (When scanning forwards to check for the
1031 existencd of the subpattern, it was treating the data ']' as terminating
1032 the class, so got the count wrong. When actually compiling, the reference
1033 was subsequently set up correctly.)
1034
1035 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1036 it was being rejected as not supported by pcre_dfa_exec(), even though
1037 other assertions are supported. I have made pcre_dfa_exec() support
1038 (*FAIL).
1039
1040 16. The implementation of 13c above involved the invention of a new opcode,
1041 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1042 cannot be changed at match time, I realized I could make a small
1043 improvement to matching performance by compiling OP_ALLANY instead of
1044 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1045 on the OP_ANY path.
1046
1047 17. Compiling pcretest on Windows with readline support failed without the
1048 following two fixes: (1) Make the unistd.h include conditional on
1049 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1050
1051 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1052 ncurses library to be included for pcretest when ReadLine support is
1053 requested, but also to allow for it to be overridden. This patch came from
1054 Daniel Bergström.
1055
1056 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1057 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1058 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1059 spotting this.
1060
1061
1062 Version 7.6 28-Jan-08
1063 ---------------------
1064
1065 1. A character class containing a very large number of characters with
1066 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1067 overflow.
1068
1069 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1070 HAVE_LONG_LONG is not defined.
1071
1072 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1073 bring them up to date and include new features. This patch includes:
1074
1075 - Fixed PH's badly added libz and libbz2 support.
1076 - Fixed a problem with static linking.
1077 - Added pcredemo. [But later removed - see 7 below.]
1078 - Fixed dftables problem and added an option.
1079 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1080 HAVE_LONG_LONG.
1081 - Added readline support for pcretest.
1082 - Added an listing of the option settings after cmake has run.
1083
1084 4. A user submitted a patch to Makefile that makes it easy to create
1085 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1086 Makefile.am that cause it to include this special target, without
1087 affecting anything else. Note that the same mingw target plus all
1088 the other distribution libraries and programs are now supported
1089 when configuring with CMake (see 6 below) instead of with
1090 Configure/Make.
1091
1092 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1093 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1094 exported in the Windows port". It has not yet been confirmed that the patch
1095 solves the problem, but it does no harm.
1096
1097 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1098 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1099 with CMake, and also correct the comment about stack recursion.
1100
1101 7. Remove the automatic building of pcredemo from the ./configure system and
1102 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1103 of a program that users should build themselves after PCRE is installed, so
1104 building it automatically is not really right. What is more, it gave
1105 trouble in some build environments.
1106
1107 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1108
1109
1110 Version 7.5 10-Jan-08
1111 ---------------------
1112
1113 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1114 values in parens when parsing an RE using the C++ wrapper."
1115
1116 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1117 Characters greater than 255 were excluded from the class instead of being
1118 included.
1119
1120 3. The same bug as (2) above applied to negated POSIX classes such as
1121 [:^space:].
1122
1123 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1124 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1125 I have changed it.
1126
1127 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1128 first named subpattern) and a construct such as (?&a) would reference the
1129 first named subpattern whose name started with "a" (in other words, the
1130 length check was missing). Both these problems are fixed. "Subpattern name
1131 expected" is now given for (?&) (a zero-length name), and this patch also
1132 makes it give the same error for \k'' (previously it complained that that
1133 was a reference to a non-existent subpattern).
1134
1135 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1136 this is right because (?- can be followed by option settings as well as by
1137 digits. I have, however, made the messages clearer.
1138
1139 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1140 than the number used in the conditional) now cause a compile-time error.
1141 This is actually not compatible with Perl, which accepts such patterns, but
1142 treats the conditional as always being FALSE (as PCRE used to), but it
1143 seems to me that giving a diagnostic is better.
1144
1145 8. Change "alphameric" to the more common word "alphanumeric" in comments
1146 and messages.
1147
1148 9. Fix two occurrences of "backslash" in comments that should have been
1149 "backspace".
1150
1151 10. Remove two redundant lines of code that can never be obeyed (their function
1152 was moved elsewhere).
1153
1154 11. The program that makes PCRE's Unicode character property table had a bug
1155 which caused it to generate incorrect table entries for sequences of
1156 characters that have the same character type, but are in different scripts.
1157 It amalgamated them into a single range, with the script of the first of
1158 them. In other words, some characters were in the wrong script. There were
1159 thirteen such cases, affecting characters in the following ranges:
1160
1161 U+002b0 - U+002c1
1162 U+0060c - U+0060d
1163 U+0061e - U+00612
1164 U+0064b - U+0065e
1165 U+0074d - U+0076d
1166 U+01800 - U+01805
1167 U+01d00 - U+01d77
1168 U+01d9b - U+01dbf
1169 U+0200b - U+0200f
1170 U+030fc - U+030fe
1171 U+03260 - U+0327f
1172 U+0fb46 - U+0fbb1
1173 U+10450 - U+1049d
1174
1175 12. The -o option (show only the matching part of a line) for pcregrep was not
1176 compatible with GNU grep in that, if there was more than one match in a
1177 line, it showed only the first of them. It now behaves in the same way as
1178 GNU grep.
1179
1180 13. If the -o and -v options were combined for pcregrep, it printed a blank
1181 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1182 does the same. The return code can be used to tell if there were any
1183 non-matching lines.
1184
1185 14. Added --file-offsets and --line-offsets to pcregrep.
1186
1187 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1188 infinitely looping recursion. The bug was that positive lookaheads were not
1189 being skipped when checking for a possible empty match (negative lookaheads
1190 and both kinds of lookbehind were skipped).
1191
1192 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1193 inclusion of <windows.h> to before rather than after the definition of
1194 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1195
1196 17. Specifying a possessive quantifier with a specific limit for a Unicode
1197 character property caused pcre_compile() to compile bad code, which led at
1198 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1199 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1200 caused the error; without that there was no problem.
1201
1202 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1203
1204 19. Added --enable-pcretest-libreadline.
1205
1206 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1207 RE::GlobalReplace(). As a result, the number of replacements returned was
1208 double what it should be. I removed one of the increments, but Craig sent a
1209 later patch that removed the other one (the right fix) and added unit tests
1210 that check the return values (which was not done before).
1211
1212 21. Several CMake things:
1213
1214 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1215 the names libpcre and libpcreposix, not just pcre and pcreposix.
1216
1217 (2) The above change means that pcretest and pcregrep are now correctly
1218 linked with the newly-built libraries, not previously installed ones.
1219
1220 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1221
1222 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1223 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1224 UTF-8 newline character). The key issue is that the pattern starts .*;
1225 this means that the match must be either at the beginning, or after a
1226 newline. The bug was in the code for advancing after a failed match and
1227 checking that the new position followed a newline. It was not taking
1228 account of UTF-8 characters correctly.
1229
1230 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1231 character classes. PCRE was not treating the sequence [:...:] as a
1232 character class unless the ... were all letters. Perl, however, seems to
1233 allow any characters between [: and :], though of course it rejects as
1234 unknown any "names" that contain non-letters, because all the known class
1235 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1236 for example, whereas PCRE did not - it did not recognize a POSIX character
1237 class. This seemed a bit dangerous, so the code has been changed to be
1238 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1239 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1240 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1241 Perl does, and where it didn't before.
1242
1243 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1244 Windows environments %n is disabled by default.
1245
1246
1247 Version 7.4 21-Sep-07
1248 ---------------------
1249
1250 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1251 means that a class such as [\s] counted as "explicit reference to CR or
1252 LF". That isn't really right - the whole point of the change was to try to
1253 help when there was an actual mention of one of the two characters. So now
1254 the change happens only if \r or \n (or a literal CR or LF) character is
1255 encountered.
1256
1257 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1258 of both had grown to the point where there were only 3 bits left.
1259 Fortunately, there was spare space in the data structure, and so I have
1260 moved the internal flags into a new 16-bit field to free up more option
1261 bits.
1262
1263 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1264 but did not set the internal JCHANGED flag - either of these is enough to
1265 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1266 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1267 start sets both bits.
1268
1269 4. Added options (at build time, compile time, exec time) to change \R from
1270 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1271
1272 5. doc/pcresyntax.html was missing from the distribution.
1273
1274 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1275 compatibility, even though it is no longer used.
1276
1277 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1278 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1279 windows.h file is present (where different names are used). [This was
1280 reversed later after testing - see 16 below.]
1281
1282 8. Changed all #include <config.h> to #include "config.h". There were also
1283 some further <pcre.h> cases that I changed to "pcre.h".
1284
1285 9. When pcregrep was used with the --colour option, it missed the line ending
1286 sequence off the lines that it output.
1287
1288 10. It was pointed out to me that arrays of string pointers cause lots of
1289 relocations when a shared library is dynamically loaded. A technique of
1290 using a single long string with a table of offsets can drastically reduce
1291 these. I have refactored PCRE in four places to do this. The result is
1292 dramatic:
1293
1294 Originally: 290
1295 After changing UCP table: 187
1296 After changing error message table: 43
1297 After changing table of "verbs" 36
1298 After changing table of Posix names 22
1299
1300 Thanks to the folks working on Gregex for glib for this insight.
1301
1302 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1303 unicode-properties was also set.
1304
1305 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1306
1307 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1308 checked only for CRLF.
1309
1310 14. Added casts to pcretest.c to avoid compiler warnings.
1311
1312 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1313
1314 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1315 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1316 entirely. This removes changes made in 7 above.
1317
1318 17. The CMake files have been updated, and there is now more information about
1319 building with CMake in the NON-UNIX-USE document.
1320
1321
1322 Version 7.3 28-Aug-07
1323 ---------------------
1324
1325 1. In the rejigging of the build system that eventually resulted in 7.1, the
1326 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1327 brackets there is not right, since it causes compilers to look for an
1328 installed pcre.h, not the version that is in the source that is being
1329 compiled (which of course may be different). I have changed it back to:
1330
1331 #include "pcre.h"
1332
1333 I have a vague recollection that the change was concerned with compiling in
1334 different directories, but in the new build system, that is taken care of
1335 by the VPATH setting the Makefile.
1336
1337 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1338 when the subject happened to end in the byte 0x85 (e.g. if the last
1339 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1340 characters but of course it shouldn't be taken as a newline when it is part
1341 of another character. The bug was that, for an unlimited repeat of . in
1342 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1343 characters when looking for a newline.
1344
1345 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1346
1347 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1348 in debug output.
1349
1350 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1351 long printing in the pcrecpp unittest when running under MinGW.
1352
1353 6. ESC_K was left out of the EBCDIC table.
1354
1355 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1356 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1357 limit also applies to "virtual nesting" when a pattern is recursive, and in
1358 this case 1000 isn't so big. I have been able to remove this limit at the
1359 expense of backing off one optimization in certain circumstances. Normally,
1360 when pcre_exec() would call its internal match() function recursively and
1361 immediately return the result unconditionally, it uses a "tail recursion"
1362 feature to save stack. However, when a subpattern that can match an empty
1363 string has an unlimited repetition quantifier, it no longer makes this
1364 optimization. That gives it a stack frame in which to save the data for
1365 checking that an empty string has been matched. Previously this was taken
1366 from the 1000-entry workspace that had been reserved. So now there is no
1367 explicit limit, but more stack is used.
1368
1369 8. Applied Daniel's patches to solve problems with the import/export magic
1370 syntax that is required for Windows, and which was going wrong for the
1371 pcreposix and pcrecpp parts of the library. These were overlooked when this
1372 problem was solved for the main library.
1373
1374 9. There were some crude static tests to avoid integer overflow when computing
1375 the size of patterns that contain repeated groups with explicit upper
1376 limits. As the maximum quantifier is 65535, the maximum group length was
1377 set at 30,000 so that the product of these two numbers did not overflow a
1378 32-bit integer. However, it turns out that people want to use groups that
1379 are longer than 30,000 bytes (though not repeat them that many times).
1380 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1381 made it possible to implement the integer overflow checks in a much more
1382 dynamic way, which I have now done. The artificial limitation on group
1383 length has been removed - we now have only the limit on the total length of
1384 the compiled pattern, which depends on the LINK_SIZE setting.
1385
1386 10. Fixed a bug in the documentation for get/copy named substring when
1387 duplicate names are permitted. If none of the named substrings are set, the
1388 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1389 empty string.
1390
1391 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1392 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1393 because the ] is interpreted as the first data character and the
1394 terminating ] is not found. PCRE has been made compatible with Perl in this
1395 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1396 cause memory overwriting.
1397
1398 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1399 string has been matched (to stop an infinite loop). It was not recognizing
1400 a conditional subpattern that could match an empty string if that
1401 subpattern was within another subpattern. For example, it looped when
1402 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1403 condition was not nested. This bug has been fixed.
1404
1405 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1406 past the start of the subject in the presence of bytes with the top bit
1407 set, for example "\x8aBCD".
1408
1409 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1410 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1411
1412 14. Optimized (?!) to (*FAIL).
1413
1414 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1415 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1416 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1417 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1418 does: it's just the validity check that is more restrictive.
1419
1420 16. Inserted checks for integer overflows during escape sequence (backslash)
1421 processing, and also fixed erroneous offset values for syntax errors during
1422 backslash processing.
1423
1424 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1425 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1426
1427 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1428 caused an overrun.
1429
1430 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1431 something other than just ASCII characters) inside a group that had an
1432 unlimited repeat caused a loop at compile time (while checking to see
1433 whether the group could match an empty string).
1434
1435 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1436 [\P{Any}] did so. (Error in the code for printing property names.)
1437
1438 21. An orphan \E inside a character class could cause a crash.
1439
1440 22. A repeated capturing bracket such as (A)? could cause a wild memory
1441 reference during compilation.
1442
1443 23. There are several functions in pcre_compile() that scan along a compiled
1444 expression for various reasons (e.g. to see if it's fixed length for look
1445 behind). There were bugs in these functions when a repeated \p or \P was
1446 present in the pattern. These operators have additional parameters compared
1447 with \d, etc, and these were not being taken into account when moving along
1448 the compiled data. Specifically:
1449
1450 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1451 length.
1452
1453 (b) An item such as \pL+ within a repeated group could cause crashes or
1454 loops.
1455
1456 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1457 "reference to non-existent subpattern" error.
1458
1459 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1460
1461 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1462 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1463
1464 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1465
1466 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1467 character were causing crashes (broken optimization).
1468
1469 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1470 \p or \P) caused a compile-time loop.
1471
1472 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1473 break. For example, the unstudied pattern [\r\n]A does not match the string
1474 "\r\nA" because change 7.0/46 below moves the current point on by two
1475 characters after failing to match at the start. However, the pattern \nA
1476 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1477 the same is true. There doesn't seem any very clean way out of this, but
1478 what I have chosen to do makes the common cases work: PCRE now takes note
1479 of whether there can be an explicit match for \r or \n anywhere in the
1480 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1481 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1482 pattern has explicit CR or LF references.
1483
1484 29. Added (*CR) etc for changing newline setting at start of pattern.
1485
1486
1487 Version 7.2 19-Jun-07
1488 ---------------------
1489
1490 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1491 which is apparently normally available under Windows.
1492
1493 2. Re-jig the pcregrep tests with different newline settings in an attempt
1494 to make them independent of the local environment's newline setting.
1495
1496 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1497
1498 4. Some of the "internals" tests were previously cut out when the link size
1499 was not 2, because the output contained actual offsets. The recent new
1500 "Z" feature of pcretest means that these can be cut out, making the tests
1501 usable with all link sizes.
1502
1503 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1504 stack recursion. This gives a massive performance boost under BSD, but just
1505 a small improvement under Linux. However, it saves one field in the frame
1506 in all cases.
1507
1508 6. Added more features from the forthcoming Perl 5.10:
1509
1510 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1511 recursion call. It refers to the nth most recently opened parentheses.
1512
1513 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1514 to be opened parentheses.
1515
1516 (c) Conditions that refer to capturing parentheses can be specified
1517 relatively, for example, (?(-2)... or (?(+3)...
1518
1519 (d) \K resets the start of the current match so that everything before
1520 is not part of it.
1521
1522 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1523
1524 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1525 reference syntax.
1526
1527 (g) (?| introduces a group in which the numbering of parentheses in each
1528 alternative starts with the same number.
1529
1530 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1531
1532 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1533 PCRE_INFO_JCHANGED.
1534
1535 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1536 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1537 for detecting groups that can match an empty string.
1538
1539 9. A pattern with a very large number of alternatives (more than several
1540 hundred) was running out of internal workspace during the pre-compile
1541 phase, where pcre_compile() figures out how much memory will be needed. A
1542 bit of new cunning has reduced the workspace needed for groups with
1543 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1544 workspace instead of running out of the 4096 that are available.
1545
1546 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1547
1548 11. Applied patch from Google to remove an optimization that didn't quite work.
1549 The report of the bug said:
1550
1551 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1552 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1553 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1554
1555 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1556 it matched the wrong number of bytes.
1557
1558
1559 Version 7.1 24-Apr-07
1560 ---------------------
1561
1562 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1563 that is more "standard", making use of automake and other Autotools. There
1564 is some re-arrangement of the files and adjustment of comments consequent
1565 on this.
1566
1567 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1568 for recursive directory scanning broke on some systems because the files
1569 are not scanned in any specific order and on different systems the order
1570 was different. A call to "sort" has been inserted into RunGrepTest for the
1571 approprate test as a short-term fix. In the longer term there may be an
1572 alternative.
1573
1574 3. I had an email from Eric Raymond about problems translating some of PCRE's
1575 man pages to HTML (despite the fact that I distribute HTML pages, some
1576 people do their own conversions for various reasons). The problems
1577 concerned the use of low-level troff macros .br and .in. I have therefore
1578 removed all such uses from the man pages (some were redundant, some could
1579 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1580 HTML has been updated to handle .nf/.fi and to complain if it encounters
1581 .br or .in.
1582
1583 4. Updated comments in configure.ac that get placed in config.h.in and also
1584 arranged for config.h to be included in the distribution, with the name
1585 config.h.generic, for the benefit of those who have to compile without
1586 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1587
1588 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1589 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1590 makevp.bat for the current PCRE, using the additional files
1591 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1592
1593 6. A Windows user reported a minor discrepancy with test 2, which turned out
1594 to be caused by a trailing space on an input line that had got lost in his
1595 copy. The trailing space was an accident, so I've just removed it.
1596
1597 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1598 that is needed.
1599
1600 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1601 as "const" (a) because they are and (b) because it helps the PHP
1602 maintainers who have recently made a script to detect big data structures
1603 in the php code that should be moved to the .rodata section. I remembered
1604 to update Builducptable as well, so it won't revert if ucptable.h is ever
1605 re-created.
1606
1607 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1608 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1609 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1610 support is not required. This saves 1.5-2K of code, which is important in
1611 some applications.
1612
1613 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1614 so as not to refer to the tables, even though these functions will never be
1615 called when UTF-8 support is disabled. Otherwise there are problems with a
1616 shared library.
1617
1618 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1619
1620 (a) It was defining its arguments as char * instead of void *.
1621
1622 (b) It was assuming that all moves were upwards in memory; this was true
1623 a long time ago when I wrote it, but is no longer the case.
1624
1625 The emulated memove() is provided for those environments that have neither
1626 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1627 is clearly not the case, as these two bugs were recently reported.
1628
1629 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1630 and Detrail to create the HTML documentation, the .txt form of the man
1631 pages, and it removes trailing spaces from listed files. It also creates
1632 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1633 case, it wraps all the #defines with #ifndefs. This script should be run
1634 before "make dist".
1635
1636 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1637 with Unicode property support.
1638
1639 (a) For a maximizing quantifier, if the two different cases of the
1640 character were of different lengths in their UTF-8 codings (there are
1641 some cases like this - I found 11), and the matching function had to
1642 back up over a mixture of the two cases, it incorrectly assumed they
1643 were both the same length.
1644
1645 (b) When PCRE was configured to use the heap rather than the stack for
1646 recursion during matching, it was not correctly preserving the data for
1647 the other case of a UTF-8 character when checking ahead for a match
1648 while processing a minimizing repeat. If the check also involved
1649 matching a wide character, but failed, corruption could cause an
1650 erroneous result when trying to check for a repeat of the original
1651 character.
1652
1653 13. Some tidying changes to the testing mechanism:
1654
1655 (a) The RunTest script now detects the internal link size and whether there
1656 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1657 values substituted by "configure". (The RunGrepTest script already did
1658 this for UTF-8.) The configure.ac script no longer substitutes the
1659 relevant variables.
1660
1661 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1662 with length and offset values. This means that the output is different
1663 for different internal link sizes. Test 2 is skipped for link sizes
1664 other than 2 because of this, bypassing the problem. Unfortunately,
1665 there was also a test in test 3 (the locale tests) that used /B and
1666 failed for link sizes other than 2. Rather than cut the whole test out,
1667 I have added a new /Z option to pcretest that replaces the length and
1668 offset values with spaces. This is now used to make test 3 independent
1669 of link size. (Test 2 will be tidied up later.)
1670
1671 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1672 segmentation fault instead of returning the appropriate error message.
1673
1674 15. In multiline mode when the newline sequence was set to "any", the pattern
1675 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1676 This doesn't seem right; it now treats the CRLF combination as the line
1677 ending, and so does not match in that case. It's only a pattern such as ^$
1678 that would hit this one: something like ^ABC$ would have failed after \r
1679 and then tried again after \r\n.
1680
1681 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1682 in an attempt to make files that differ only in their line terminators
1683 compare equal. This works on Linux.
1684
1685 17. Under certain error circumstances pcregrep might try to free random memory
1686 as it exited. This is now fixed, thanks to valgrind.
1687
1688 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1689 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1690 was because its rules for how to advance for /g after matching an empty
1691 string at the end of a line did not allow for this case. They now check for
1692 it specially.
1693
1694 20. pcretest is supposed to handle patterns and data of any length, by
1695 extending its buffers when necessary. It was getting this wrong when the
1696 buffer for a data line had to be extended.
1697
1698 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1699 CRLF as a newline sequence.
1700
1701 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1702 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1703 I have nevertheless tidied it up.
1704
1705 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1706
1707 24. Added a man page for pcre-config.
1708
1709
1710 Version 7.0 19-Dec-06
1711 ---------------------
1712
1713 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1714 moving to gcc 4.1.1.
1715
1716 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1717 sys/time.h, which is documented as needed for this function. It doesn't
1718 seem to matter on Linux, but it showed up on some releases of OS X.
1719
1720 3. It seems that there are systems where bytes whose values are greater than
1721 127 match isprint() in the "C" locale. The "C" locale should be the
1722 default when a C program starts up. In most systems, only ASCII printing
1723 characters match isprint(). This difference caused the output from pcretest
1724 to vary, making some of the tests fail. I have changed pcretest so that:
1725
1726 (a) When it is outputting text in the compiled version of a pattern, bytes
1727 other than 32-126 are always shown as hex escapes.
1728
1729 (b) When it is outputting text that is a matched part of a subject string,
1730 it does the same, unless a different locale has been set for the match
1731 (using the /L modifier). In this case, it uses isprint() to decide.
1732
1733 4. Fixed a major bug that caused incorrect computation of the amount of memory
1734 required for a compiled pattern when options that changed within the
1735 pattern affected the logic of the preliminary scan that determines the
1736 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1737 that the computed length was too small. The symptoms of this bug were
1738 either the PCRE error "internal error: code overflow" from pcre_compile(),
1739 or a glibc crash with a message such as "pcretest: free(): invalid next
1740 size (fast)". Examples of patterns that provoked this bug (shown in
1741 pcretest format) are:
1742
1743 /(?-x: )/x
1744 /(?x)(?-x: \s*#\s*)/
1745 /((?i)[\x{c0}])/8
1746 /(?i:[\x{c0}])/8
1747
1748 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1749 is now done differently.
1750
1751 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1752 wrapper classes; (b) implement a new function in the C++ scanner that is
1753 more efficient than the old way of doing things because it avoids levels of
1754 recursion in the regex matching; (c) add a paragraph to the documentation
1755 for the FullMatch() function.
1756
1757 6. The escape sequence \n was being treated as whatever was defined as
1758 "newline". Not only was this contrary to the documentation, which states
1759 that \n is character 10 (hex 0A), but it also went horribly wrong when
1760 "newline" was defined as CRLF. This has been fixed.
1761
1762 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1763 was being set to -1 for the "end of line" case (supposedly a value that no
1764 character can have). Though this value is never used (the check for end of
1765 line is "zero bytes in current character"), it caused compiler complaints.
1766 I've changed it to 0xffffffff.
1767
1768 8. In pcre_version.c, the version string was being built by a sequence of
1769 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1770 string (as it is for production releases) called a macro with an empty
1771 argument. The C standard says the result of this is undefined. The gcc
1772 compiler treats it as an empty string (which was what was wanted) but it is
1773 reported that Visual C gives an error. The source has been hacked around to
1774 avoid this problem.
1775
1776 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1777 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1778 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1779 of them did).
1780
1781 10. Originally, pcretest opened its input and output without "b"; then I was
1782 told that "b" was needed in some environments, so it was added for release
1783 5.0 to both the input and output. (It makes no difference on Unix-like
1784 systems.) Later I was told that it is wrong for the input on Windows. I've
1785 now abstracted the modes into two macros, to make it easier to fiddle with
1786 them, and removed "b" from the input mode under Windows.
1787
1788 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1789
1790 12. Added -help and --help to pcretest as an official way of being reminded
1791 of the options.
1792
1793 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1794 and pcrecpp.cc because they annoy compilers at high warning levels.
1795
1796 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1797
1798 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1799 scripts are not C programs :-) and which was not noticed because it works
1800 on Linux.
1801
1802 16. pcretest is supposed to handle any length of pattern and data line (as one
1803 line or as a continued sequence of lines) by extending its input buffer if
1804 necessary. This feature was broken for very long pattern lines, leading to
1805 a string of junk being passed to pcre_compile() if the pattern was longer
1806 than about 50K.
1807
1808 17. I have done a major re-factoring of the way pcre_compile() computes the
1809 amount of memory needed for a compiled pattern. Previously, there was code
1810 that made a preliminary scan of the pattern in order to do this. That was
1811 OK when PCRE was new, but as the facilities have expanded, it has become
1812 harder and harder to keep it in step with the real compile phase, and there
1813 have been a number of bugs (see for example, 4 above). I have now found a
1814 cunning way of running the real compile function in a "fake" mode that
1815 enables it to compute how much memory it would need, while actually only
1816 ever using a few hundred bytes of working memory and without too many
1817 tests of the mode. This should make future maintenance and development
1818 easier. A side effect of this work is that the limit of 200 on the nesting
1819 depth of parentheses has been removed (though this was never a serious
1820 limitation, I suspect). However, there is a downside: pcre_compile() now
1821 runs more slowly than before (30% or more, depending on the pattern). I
1822 hope this isn't a big issue. There is no effect on runtime performance.
1823
1824 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1825 newline (only possible for the last line of a file) and it was a
1826 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1827
1828 19. Added additional timing features to pcretest. (1) The -tm option now times
1829 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1830 separate command line item, by a number that specifies the number of
1831 repeats to use when timing. The default is 50000; this gives better
1832 precision, but takes uncomfortably long for very large patterns.
1833
1834 20. Extended pcre_study() to be more clever in cases where a branch of a
1835 subpattern has no definite first character. For example, (a*|b*)[cd] would
1836 previously give no result from pcre_study(). Now it recognizes that the
1837 first character must be a, b, c, or d.
1838
1839 21. There was an incorrect error "recursive call could loop indefinitely" if
1840 a subpattern (or the entire pattern) that was being tested for matching an
1841 empty string contained only one non-empty item after a nested subpattern.
1842 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1843 incorrectly, because the \d was being skipped in the check.
1844
1845 22. The pcretest program now has a new pattern option /B and a command line
1846 option -b, which is equivalent to adding /B to every pattern. This causes
1847 it to show the compiled bytecode, without the additional information that
1848 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1849 is the same as /B/I).
1850
1851 23. A new optimization is now able automatically to treat some sequences such
1852 as a*b as a*+b. More specifically, if something simple (such as a character
1853 or a simple class like \d) has an unlimited quantifier, and is followed by
1854 something that cannot possibly match the quantified thing, the quantifier
1855 is automatically "possessified".
1856
1857 24. A recursive reference to a subpattern whose number was greater than 39
1858 went wrong under certain circumstances in UTF-8 mode. This bug could also
1859 have affected the operation of pcre_study().
1860
1861 25. Realized that a little bit of performance could be had by replacing
1862 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1863
1864 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1865
1866 27. Possessive quantifiers such as a++ were previously implemented by turning
1867 them into atomic groups such as ($>a+). Now they have their own opcodes,
1868 which improves performance. This includes the automatically created ones
1869 from 23 above.
1870
1871 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1872 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1873 the first matched character to be a colon. This applied both to named and
1874 numbered groups.
1875
1876 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1877
1878 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1879 building PCRE on Windows easier, so I have included it in the distribution.
1880
1881 31. There is now a check in pcretest against a ridiculously large number being
1882 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1883 loop, the loop is abandoned.
1884
1885 32. Forward references to subpatterns in conditions such as (?(2)...) where
1886 subpattern 2 is defined later cause pcre_compile() to search forwards in
1887 the pattern for the relevant set of parentheses. This search went wrong
1888 when there were unescaped parentheses in a character class, parentheses
1889 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1890
1891 33. "Subroutine" calls and backreferences were previously restricted to
1892 referencing subpatterns earlier in the regex. This restriction has now
1893 been removed.
1894
1895 34. Added a number of extra features that are going to be in Perl 5.10. On the
1896 whole, these are just syntactic alternatives for features that PCRE had
1897 previously implemented using the Python syntax or my own invention. The
1898 other formats are all retained for compatibility.
1899
1900 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1901 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1902 also .NET compatible.
1903
1904 (b) A recursion or subroutine call to a named group can now be defined as
1905 (?&name) as well as (?P>name).
1906
1907 (c) A backreference to a named group can now be defined as \k<name> or
1908 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1909 5.10, are also .NET compatible.
1910
1911 (d) A conditional reference to a named group can now use the syntax
1912 (?(<name>) or (?('name') as well as (?(name).
1913
1914 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1915 groups (named and numbered) that are never evaluated inline, but can be
1916 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1917 is always false. There may be only one alternative in such a group.
1918
1919 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1920 as the simple (?(R). The condition is true only if the most recent
1921 recursion is that of the given number or name. It does not search out
1922 through the entire recursion stack.
1923
1924 (g) The escape \gN or \g{N} has been added, where N is a positive or
1925 negative number, specifying an absolute or relative reference.
1926
1927 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1928 some "unreachable code" warnings.
1929
1930 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1931 things, this adds five new scripts.
1932
1933 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1934 There were also incompatibilities regarding the handling of \Q..\E inside
1935 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1936 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1937
1938 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1939 matches an empty string, and forcibly breaks the loop. There were bugs in
1940 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1941 against aaaa the result was just "a" rather than "aaaa", for example. Two
1942 separate and independent bugs (that affected different cases) have been
1943 fixed.
1944
1945 39. Refactored the code to abolish the use of different opcodes for small
1946 capturing bracket numbers. This is a tidy that I avoided doing when I
1947 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1948 The new approach is not only tidier, it makes it possible to reduce the
1949 memory needed to fix the previous bug (38).
1950
1951 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1952 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1953 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1954 mode.
1955
1956 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1957 report.
1958
1959 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1960 copy construction and assignment in the C++ wrapper.
1961
1962 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1963 couple of bugs that could have given wrong results in the "--newline=crlf"
1964 case.
1965
1966 44. Added a number of casts and did some reorganization of signed/unsigned int
1967 variables following suggestions from Dair Grant. Also renamed the variable
1968 "this" as "item" because it is a C++ keyword.
1969
1970 45. Arranged for dftables to add
1971
1972 #include "pcre_internal.h"
1973
1974 to pcre_chartables.c because without it, gcc 4.x may remove the array
1975 definition from the final binary if PCRE is built into a static library and
1976 dead code stripping is activated.
1977
1978 46. For an unanchored pattern, if a match attempt fails at the start of a
1979 newline sequence, and the newline setting is CRLF or ANY, and the next two
1980 characters are CRLF, advance by two characters instead of one.
1981
1982
1983 Version 6.7 04-Jul-06
1984 ---------------------
1985
1986 1. In order to handle tests when input lines are enormously long, pcretest has
1987 been re-factored so that it automatically extends its buffers when
1988 necessary. The code is crude, but this _is_ just a test program. The
1989 default size has been increased from 32K to 50K.
1990
1991 2. The code in pcre_study() was using the value of the re argument before
1992 testing it for NULL. (Of course, in any sensible call of the function, it
1993 won't be NULL.)
1994
1995 3. The memmove() emulation function in pcre_internal.h, which is used on
1996 systems that lack both memmove() and bcopy() - that is, hardly ever -
1997 was missing a "static" storage class specifier.
1998
1999 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2000 containing an extended class (one that cannot be represented by a bitmap
2001 because it contains high-valued characters or Unicode property items, e.g.
2002 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2003 pattern, but PCRE should not loop if you do not (it no longer does).
2004 [Detail: two cases were found: (a) a repeated subpattern containing an
2005 extended class; (b) a recursive reference to a subpattern that followed a
2006 previous extended class. It wasn't skipping over the extended class
2007 correctly when UTF-8 mode was not set.]
2008
2009 5. A negated single-character class was not being recognized as fixed-length
2010 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2011 compile error "lookbehind assertion is not fixed length".
2012
2013 6. The RunPerlTest auxiliary script was showing an unexpected difference
2014 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2015 write a Perl script that can interpret lines of an input file either as
2016 byte characters or as UTF-8, which is what "perltest" was being required to
2017 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2018 can't do is switch easily at run time between having the "use utf8;" pragma
2019 or not. In the end, I fudged it by using the RunPerlTest script to insert
2020 "use utf8;" explicitly for the UTF-8 tests.
2021
2022 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2023 the end of the subject string, contrary to the documentation and to what
2024 Perl does. This was true of both matching functions. Now it matches only at
2025 the start of the subject and immediately after *internal* newlines.
2026
2027 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2028 a pointer to an int instead of a pointer to an unsigned long int. This
2029 caused problems on 64-bit systems.
2030
2031 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2032 instance of the 'standard' template library not being so standard".
2033
2034 10. There was no check on the number of named subpatterns nor the maximum
2035 length of a subpattern name. The product of these values is used to compute
2036 the size of the memory block for a compiled pattern. By supplying a very
2037 long subpattern name and a large number of named subpatterns, the size
2038 computation could be caused to overflow. This is now prevented by limiting
2039 the length of names to 32 characters, and the number of named subpatterns
2040 to 10,000.
2041
2042 11. Subpatterns that are repeated with specific counts have to be replicated in
2043 the compiled pattern. The size of memory for this was computed from the
2044 length of the subpattern and the repeat count. The latter is limited to
2045 65535, but there was no limit on the former, meaning that integer overflow
2046 could in principle occur. The compiled length of a repeated subpattern is
2047 now limited to 30,000 bytes in order to prevent this.
2048
2049 12. Added the optional facility to have named substrings with the same name.
2050
2051 13. Added the ability to use a named substring as a condition, using the
2052 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2053 are numbers (not recommended). Forward references are permitted.
2054
2055 14. Added forward references in named backreferences (if you see what I mean).
2056
2057 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2058 pattern could run off the end of the subject. For example, the pattern
2059 "(?s)(.{1,5})"8 did this with the subject "ab".
2060
2061 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2062 PCRE_CASELESS was set when matching characters that were quantified with ?
2063 or *.
2064
2065 17. A character class other than a single negated character that had a minimum
2066 but no maximum quantifier - for example [ab]{6,} - was not handled
2067 correctly by pce_dfa_exec(). It would match only one character.
2068
2069 18. A valid (though odd) pattern that looked like a POSIX character
2070 class but used an invalid character after [ (for example [[,abc,]]) caused
2071 pcre_compile() to give the error "Failed: internal error: code overflow" or
2072 in some cases to crash with a glibc free() error. This could even happen if
2073 the pattern terminated after [[ but there just happened to be a sequence of
2074 letters, a binary zero, and a closing ] in the memory that followed.
2075
2076 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2077 over the years. Originally (before any Unicode support), just the bottom 8
2078 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2079 output from "man perlunicode" includes this:
2080
2081 The regular expression compiler produces polymorphic opcodes. That
2082 is, the pattern adapts to the data and automatically switches to
2083 the Unicode character scheme when presented with Unicode data--or
2084 instead uses a traditional byte scheme when presented with byte
2085 data.
2086
2087 Sadly, a wide octal escape does not cause a switch, and in a string with
2088 no other multibyte characters, these octal escapes are treated as before.
2089 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2090 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2091 Unicode string.
2092
2093 I have not perpetrated such confusion in PCRE. Up till now, it took just
2094 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2095 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2096 translate to the appropriate multibyte character.
2097
2098 29. Applied some refactoring to reduce the number of warnings from Microsoft
2099 and Borland compilers. This has included removing the fudge introduced
2100 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2101 a warning about an unused variable.
2102
2103 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2104 characters since release 4.0, because Perl (from release 5.004) does not.
2105 [Or at least, is documented not to: some releases seem to be in conflict
2106 with the documentation.] However, when a pattern was studied with
2107 pcre_study() and all its branches started with \s, PCRE still included VT
2108 as a possible starting character. Of course, this did no harm; it just
2109 caused an unnecessary match attempt.
2110
2111 22. Removed a now-redundant internal flag bit that recorded the fact that case
2112 dependency changed within the pattern. This was once needed for "required
2113 byte" processing, but is no longer used. This recovers a now-scarce options
2114 bit. Also moved the least significant internal flag bit to the most-
2115 significant bit of the word, which was not previously used (hangover from
2116 the days when it was an int rather than a uint) to free up another bit for
2117 the future.
2118
2119 23. Added support for CRLF line endings as well as CR and LF. As well as the
2120 default being selectable at build time, it can now be changed at runtime
2121 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2122 specify that it is scanning data with non-default line endings.
2123
2124 24. Changed the definition of CXXLINK to make it agree with the definition of
2125 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2126
2127 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2128 recursions. This makes a big different to stack usage for some patterns.
2129
2130 26. If a subpattern containing a named recursion or subroutine reference such
2131 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2132 the space required for the compiled pattern went wrong and gave too small a
2133 value. Depending on the environment, this could lead to "Failed: internal
2134 error: code overflow at offset 49" or "glibc detected double free or
2135 corruption" errors.
2136
2137 27. Applied patches from Google (a) to support the new newline modes and (b) to
2138 advance over multibyte UTF-8 characters in GlobalReplace.
2139
2140 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2141 difference for some implementation of PCRE in some Windows version.
2142
2143 29. Added some extra testing facilities to pcretest:
2144
2145 \q<number> in a data line sets the "match limit" value
2146 \Q<number> in a data line sets the "match recursion limt" value
2147 -S <number> sets the stack size, where <number> is in megabytes
2148
2149 The -S option isn't available for Windows.
2150
2151
2152 Version 6.6 06-Feb-06
2153 ---------------------
2154
2155 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2156 in pcreposix.h. I have copied the definition from pcre.h.
2157
2158 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2159 because pcre.h is no longer a built file.
2160
2161 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2162 not normally included in the compiled code.
2163
2164
2165 Version 6.5 01-Feb-06
2166 ---------------------
2167
2168 1. When using the partial match feature with pcre_dfa_exec(), it was not
2169 anchoring the second and subsequent partial matches at the new starting
2170 point. This could lead to incorrect results. For example, with the pattern
2171 /1234/, partially matching against "123" and then "a4" gave a match.
2172
2173 2. Changes to pcregrep:
2174
2175 (a) All non-match returns from pcre_exec() were being treated as failures
2176 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2177 error message is output. Some extra information is given for the
2178 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2179 probably the only errors that are likely to be caused by users (by
2180 specifying a regex that has nested indefinite repeats, for instance).
2181 If there are more than 20 of these errors, pcregrep is abandoned.
2182
2183 (b) A binary zero was treated as data while matching, but terminated the
2184 output line if it was written out. This has been fixed: binary zeroes
2185 are now no different to any other data bytes.
2186
2187 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2188 used to set a locale for matching. The --locale=xxxx long option has
2189 been added (no short equivalent) to specify a locale explicitly on the
2190 pcregrep command, overriding the environment variables.
2191
2192 (d) When -B was used with -n, some line numbers in the output were one less
2193 than they should have been.
2194
2195 (e) Added the -o (--only-matching) option.
2196
2197 (f) If -A or -C was used with -c (count only), some lines of context were
2198 accidentally printed for the final match.
2199
2200 (g) Added the -H (--with-filename) option.
2201
2202 (h) The combination of options -rh failed to suppress file names for files
2203 that were found from directory arguments.
2204
2205 (i) Added the -D (--devices) and -d (--directories) options.
2206
2207 (j) Added the -F (--fixed-strings) option.
2208
2209 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2210
2211 (l) Added the --colo(u)r option.
2212
2213 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2214 is not present by default.
2215
2216 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2217 items such as (?R) or (?1), when the recursion could match a number of
2218 alternatives. If it matched one of the alternatives, but subsequently,
2219 outside the recursion, there was a failure, the code tried to back up into
2220 the recursion. However, because of the way PCRE is implemented, this is not
2221 possible, and the result was an incorrect result from the match.
2222
2223 In order to prevent this happening, the specification of recursion has
2224 been changed so that all such subpatterns are automatically treated as
2225 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2226
2227 4. I had overlooked the fact that, in some locales, there are characters for
2228 which isalpha() is true but neither isupper() nor islower() are true. In
2229 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2230 and ordfeminine) are like this. This affected the treatment of \w and \W
2231 when they appeared in character classes, but not when they appeared outside
2232 a character class. The bit map for "word" characters is now created
2233 separately from the results of isalnum() instead of just taking it from the
2234 upper, lower, and digit maps. (Plus the underscore character, of course.)
2235
2236 5. The above bug also affected the handling of POSIX character classes such as
2237 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2238 permanent tables. Instead, the bit maps for such a class were previously
2239 created as the appropriate unions of the upper, lower, and digit bitmaps.
2240 Now they are created by subtraction from the [[:word:]] class, which has
2241 its own bitmap.
2242
2243 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2244 It is created by subtracting the vertical space characters (\x09, \x0a,
2245 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2246 subtraction was done in the overall bitmap for a character class, meaning
2247 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2248 be recognized. This bug has been fixed.
2249
2250 7. Patches from the folks at Google:
2251
2252 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2253 real life, but is still worth protecting against".
2254
2255 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2256 regular expressions".
2257
2258 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2259 have it.
2260
2261 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2262 "configure" and the latter not, in order to fix a problem somebody had
2263 with compiling the Arg class on HP-UX.
2264
2265 (e) Improve the error-handling of the C++ wrapper a little bit.
2266
2267 (f) New tests for checking recursion limiting.
2268
2269 8. The pcre_memmove() function, which is used only if the environment does not
2270 have a standard memmove() function (and is therefore rarely compiled),
2271 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2272 returning a result (though PCRE never actually uses the result).
2273
2274 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2275 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2276 returned instead of calling malloc() with an overflowing number that would
2277 most likely cause subsequent chaos.
2278
2279 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2280
2281 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2282 with this option is matched, the nmatch and pmatch options of regexec() are
2283 ignored.
2284
2285 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2286 provided in case anyone wants to the the POSIX interface with UTF-8
2287 strings.
2288
2289 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2290 C++ linking (needed for some HP-UX environments).
2291
2292 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2293 (unused parameter) and in the pcre_printint() function (omitted "default"
2294 switch label when the default is to do nothing).
2295
2296 15. Added some code to make it possible, when PCRE is compiled as a C++
2297 library, to replace subject pointers for pcre_exec() with a smart pointer
2298 class, thus making it possible to process discontinuous strings.
2299
2300 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2301 much the same function. They were added by different people who were trying
2302 to make PCRE easy to compile on non-Unix systems. It has been suggested
2303 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2304 for compiling on Windows systems. I have therefore replaced it with
2305 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2306 defaults to "extern" for C or "extern C" for C++, which works fine on
2307 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2308 SCOPE with something explicit in config.h. In addition:
2309
2310 (a) pcreposix.h still had just "extern" instead of either of these macros;
2311 I have replaced it with PCRE_DATA_SCOPE.
2312
2313 (b) Functions such as _pcre_xclass(), which are internal to the library,
2314 but external in the C sense, all had PCRE_EXPORT in their definitions.
2315 This is apparently wrong for the Windows case, so I have removed it.
2316 (It makes no difference on Unix-like systems.)
2317
2318 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2319 of recursive calls to match(). This is different to MATCH_LIMIT because
2320 that limits the total number of calls to match(), not all of which increase
2321 the depth of recursion. Limiting the recursion depth limits the amount of
2322 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2323 when PCRE is compiled, and changed at run time. A patch from Google adds
2324 this functionality to the C++ interface.
2325
2326 18. Changes to the handling of Unicode character properties:
2327
2328 (a) Updated the table to Unicode 4.1.0.
2329
2330 (b) Recognize characters that are not in the table as "Cn" (undefined).
2331
2332 (c) I revised the way the table is implemented to a much improved format
2333 which includes recognition of ranges. It now supports the ranges that
2334 are defined in UnicodeData.txt, and it also amalgamates other
2335 characters into ranges. This has reduced the number of entries in the
2336 table from around 16,000 to around 3,000, thus reducing its size
2337 considerably. I realized I did not need to use a tree structure after
2338 all - a binary chop search is just as efficient. Having reduced the
2339 number of entries, I extended their size from 6 bytes to 8 bytes to
2340 allow for more data.
2341
2342 (d) Added support for Unicode script names via properties such as \p{Han}.
2343
2344 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2345 matching that character.
2346
2347 20. When matching a repeated Unicode property with a minimum greater than zero,
2348 (for example \pL{2,}), PCRE could look past the end of the subject if it
2349 reached it while seeking the minimum number of characters. This could
2350 happen only if some of the characters were more than one byte long, because
2351 there is a check for at least the minimum number of bytes.
2352
2353 21. Refactored the implementation of \p and \P so as to be more general, to
2354 allow for more different types of property in future. This has changed the
2355 compiled form incompatibly. Anybody with saved compiled patterns that use
2356 \p or \P will have to recompile them.
2357
2358 22. Added "Any" and "L&" to the supported property types.
2359
2360 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2361 but give a compile time error if the value is greater than 0xff.
2362
2363 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2364 accidentally not being installed or uninstalled.
2365
2366 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2367 made were to insert the current release number. This seemed silly, because
2368 it made things harder for people building PCRE on systems that don't run
2369 "configure". I have turned pcre.h into a distributed file, no longer built
2370 by "configure", with the version identification directly included. There is
2371 no longer a pcre.h.in file.
2372
2373 However, this change necessitated a change to the pcre-config script as
2374 well. It is built from pcre-config.in, and one of the substitutions was the
2375 release number. I have updated configure.ac so that ./configure now finds
2376 the release number by grepping pcre.h.
2377
2378 26. Added the ability to run the tests under valgrind.
2379
2380
2381 Version 6.4 05-Sep-05
2382 ---------------------
2383
2384 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2385 "--" to be printed when multiple files were scanned, even when none of the
2386 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2387 consider it to be a bug, and have restored the previous behaviour.
2388
2389 2. A couple of code tidies to get rid of compiler warnings.
2390
2391 3. The pcretest program used to cheat by referring to symbols in the library
2392 whose names begin with _pcre_. These are internal symbols that are not
2393 really supposed to be visible externally, and in some environments it is
2394 possible to suppress them. The cheating is now confined to including
2395 certain files from the library's source, which is a bit cleaner.
2396
2397 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2398 file's purpose clearer.
2399
2400 5. Reorganized pcre_ucp_findchar().
2401
2402
2403 Version 6.3 15-Aug-05
2404 ---------------------
2405
2406 1. The file libpcre.pc.in did not have general read permission in the tarball.
2407
2408 2. There were some problems when building without C++ support:
2409
2410 (a) If C++ support was not built, "make install" and "make test" still
2411 tried to test it.
2412
2413 (b) There were problems when the value of CXX was explicitly set. Some
2414 changes have been made to try to fix these, and ...
2415
2416 (c) --disable-cpp can now be used to explicitly disable C++ support.
2417
2418 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2419 backslash in a target when C++ was disabled. This confuses some
2420 versions of "make", apparently. Using an intermediate variable solves
2421 this. (Same for CPP_LOBJ.)
2422
2423 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2424 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2425 necessary on certain architectures.
2426
2427 4. Added a setting of -export-symbols-regex to the link command to remove
2428 those symbols that are exported in the C sense, but actually are local
2429 within the library, and not documented. Their names all begin with
2430 "_pcre_". This is not a perfect job, because (a) we have to except some
2431 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2432 available (and never for static libraries). I have made a note to try to
2433 find a way round (a) in the future.
2434
2435
2436 Version 6.2 01-Aug-05
2437 ---------------------
2438
2439 1. There was no test for integer overflow of quantifier values. A construction
2440 such as {1111111111111111} would give undefined results. What is worse, if
2441 a minimum quantifier for a parenthesized subpattern overflowed and became
2442 negative, the calculation of the memory size went wrong. This could have
2443 led to memory overwriting.
2444
2445 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2446
2447 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2448 operating environments where this matters.
2449
2450 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2451 PCRE options from within the C++ wrapper.
2452
2453 5. Named capturing subpatterns were not being correctly counted when a pattern
2454 was compiled. This caused two problems: (a) If there were more than 100
2455 such subpatterns, the calculation of the memory needed for the whole
2456 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2457 back references of the form \12, where the number was greater than 9, were
2458 not recognized as back references, even though there were sufficient
2459 previous subpatterns.
2460
2461 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2462 versions of gcc, e.g. 2.95.4.
2463
2464
2465 Version 6.1 21-Jun-05
2466 ---------------------
2467
2468 1. There was one reference to the variable "posix" in pcretest.c that was not
2469 surrounded by "#if !defined NOPOSIX".
2470
2471 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2472 the cross-check on the old pcre_info() function, for the benefit of the
2473 cut-down version of PCRE that is currently imported into Exim.
2474
2475 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2476 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2477 patterns that start (?i) but I don't think that matters. The use of (?i) is
2478 just an example; this all applies to the other options as well.
2479
2480 4. Since libtool seems to echo the compile commands it is issuing, the output
2481 from "make" can be reduced a bit by putting "@" in front of each libtool
2482 compile command.
2483
2484 5. Patch from the folks at Google for configure.in to be a bit more thorough
2485 in checking for a suitable C++ installation before trying to compile the
2486 C++ stuff. This should fix a reported problem when a compiler was present,
2487 but no suitable headers.
2488
2489 6. The man pages all had just "PCRE" as their title. I have changed them to
2490 be the relevant file name. I have also arranged that these names are
2491 retained in the file doc/pcre.txt, which is a concatenation in text format
2492 of all the man pages except the little individual ones for each function.
2493
2494 7. The NON-UNIX-USE file had not been updated for the different set of source
2495 files that come with release 6. I also added a few comments about the C++
2496 wrapper.
2497
2498
2499 Version 6.0 07-Jun-05
2500 ---------------------
2501
2502 1. Some minor internal re-organization to help with my DFA experiments.
2503
2504 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2505 didn't matter for the library itself when fully configured, but did matter
2506 when compiling without UCP support, or within Exim, where the ucp files are
2507 not imported.
2508
2509 3. Refactoring of the library code to split up the various functions into
2510 different source modules. The addition of the new DFA matching code (see
2511 below) to a single monolithic source would have made it really too
2512 unwieldy, quite apart from causing all the code to be include in a
2513 statically linked application, when only some functions are used. This is
2514 relevant even without the DFA addition now that patterns can be compiled in
2515 one application and matched in another.
2516
2517 The downside of splitting up is that there have to be some external
2518 functions and data tables that are used internally in different modules of
2519 the library but which are not part of the API. These have all had their
2520 names changed to start with "_pcre_" so that they are unlikely to clash
2521 with other external names.
2522
2523 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2524 a different (DFA) algorithm. Although it is slower than the original
2525 function, it does have some advantages for certain types of matching
2526 problem.
2527
2528 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2529 including restarting after a partial match.
2530
2531 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2532 defined when compiling for Windows was sent to me. I have put it into the
2533 code, though I have no means of testing or verifying it.
2534
2535 7. Added the pcre_refcount() auxiliary function.
2536
2537 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2538 match before or at the first newline in the subject string. In pcretest,
2539 the /f option on a pattern can be used to set this.
2540
2541 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2542 would behave wrongly. This has been present in PCRE since release 4.0.
2543
2544 10. A number of changes to the pcregrep command:
2545
2546 (a) Refactored how -x works; insert ^(...)$ instead of setting
2547 PCRE_ANCHORED and checking the length, in preparation for adding
2548 something similar for -w.
2549
2550 (b) Added the -w (match as a word) option.
2551
2552 (c) Refactored the way lines are read and buffered so as to have more
2553 than one at a time available.
2554
2555 (d) Implemented a pcregrep test script.
2556
2557 (e) Added the -M (multiline match) option. This allows patterns to match
2558 over several lines of the subject. The buffering ensures that at least
2559 8K, or the rest of the document (whichever is the shorter) is available
2560 for matching (and similarly the previous 8K for lookbehind assertions).
2561
2562 (f) Changed the --help output so that it now says
2563
2564 -w, --word-regex(p)
2565
2566 instead of two lines, one with "regex" and the other with "regexp"
2567 because that confused at least one person since the short forms are the
2568 same. (This required a bit of code, as the output is generated
2569 automatically from a table. It wasn't just a text change.)
2570
2571 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2572 option but starts with a hyphen. Could be a pattern or a path name
2573 starting with a hyphen, for instance.
2574
2575 (h) "-" can be given as a file name to represent stdin.
2576
2577 (i) When file names are being printed, "(standard input)" is used for
2578 the standard input, for compatibility with GNU grep. Previously
2579 "<stdin>" was used.
2580
2581 (j) The option --label=xxx can be used to supply a name to be used for
2582 stdin when file names are being printed. There is no short form.
2583
2584 (k) Re-factored the options decoding logic because we are going to add
2585 two more options that take data. Such options can now be given in four
2586 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2587
2588 (l) Added the -A, -B, and -C options for requesting that lines of context
2589 around matches be printed.
2590
2591 (m) Added the -L option to print the names of files that do not contain
2592 any matching lines, that is, the complement of -l.
2593
2594 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2595 continue to scan other files.
2596
2597 (o) The -s option was incorrectly implemented. For compatibility with other
2598 greps, it now suppresses the error message for a non-existent or non-
2599 accessible file (but not the return code). There is a new option called
2600 -q that suppresses the output of matching lines, which was what -s was
2601 previously doing.
2602
2603 (p) Added --include and --exclude options to specify files for inclusion
2604 and exclusion when recursing.
2605
2606 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2607 Hopefully, it now does.
2608
2609 12. Missing cast in pcre_study().
2610
2611 13. Added an "uninstall" target to the makefile.
2612
2613 14. Replaced "extern" in the function prototypes in Makefile.in with
2614 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2615 world, but is set differently for Windows.
2616
2617 15. Added a second compiling function called pcre_compile2(). The only
2618 difference is that it has an extra argument, which is a pointer to an
2619 integer error code. When there is a compile-time failure, this is set
2620 non-zero, in addition to the error test pointer being set to point to an
2621 error message. The new argument may be NULL if no error number is required
2622 (but then you may as well call pcre_compile(), which is now just a
2623 wrapper). This facility is provided because some applications need a
2624 numeric error indication, but it has also enabled me to tidy up the way
2625 compile-time errors are handled in the POSIX wrapper.
2626
2627 16. Added VPATH=.libs to the makefile; this should help when building with one
2628 prefix path and installing with another. (Or so I'm told by someone who
2629 knows more about this stuff than I do.)
2630
2631 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2632 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2633 match everything, including newlines. This is not POSIX-compatible, but
2634 somebody wanted the feature. From pcretest it can be activated by using
2635 both the P and the s flags.
2636
2637 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2638
2639 19. libpcre.pc was being incorrectly installed as executable.
2640
2641 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2642 it now also looks for '\r' so that it will work unmodified on Windows.
2643
2644 21. Added Google's contributed C++ wrapper to the distribution.
2645
2646 22. Added some untidy missing memory free() calls in pcretest, to keep
2647 Electric Fence happy when testing.
2648
2649
2650
2651 Version 5.0 13-Sep-04
2652 ---------------------
2653
2654 1. Internal change: literal characters are no longer packed up into items
2655 containing multiple characters in a single byte-string. Each character
2656 is now matched using a separate opcode. However, there may be more than one
2657 byte in the character in UTF-8 mode.
2658
2659 2. The pcre_callout_block structure has two new fields: pattern_position and
2660 next_item_length. These contain the offset in the pattern to the next match
2661 item, and its length, respectively.
2662
2663 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2664 insertion of callouts before each pattern item. Added the /C option to
2665 pcretest to make use of this.
2666
2667 4. On the advice of a Windows user, the lines
2668
2669 #if defined(_WIN32) || defined(WIN32)
2670 _setmode( _fileno( stdout ), 0x8000 );
2671 #endif /* defined(_WIN32) || defined(WIN32) */
2672
2673 have been added to the source of pcretest. This apparently does useful
2674 magic in relation to line terminators.
2675
2676 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2677 for the benefit of those environments where the "b" makes a difference.
2678
2679 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2680 to know about it. I have put a hack into configure.in that adds in code
2681 to set GCC=yes if CC=icc. This seems to end up at a point in the
2682 generated configure script that is early enough to affect the setting of
2683 compiler options, which is what is needed, but I have no means of testing
2684 whether it really works. (The user who reported this had patched the
2685 generated configure script, which of course I cannot do.)
2686
2687 LATER: After change 22 below (new libtool files), the configure script
2688 seems to know about icc (and also ecc). Therefore, I have commented out
2689 this hack in configure.in.
2690
2691 7. Added support for pkg-config (2 patches were sent in).
2692
2693 8. Negated POSIX character classes that used a combination of internal tables
2694 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2695 [[:^ascii]]. Typically, they would match almost any characters. The other
2696 POSIX classes were not broken in this way.
2697
2698 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2699 to find the match, as PCRE was deluded into thinking that the match had to
2700 start at the start point or following a newline. The same bug applied to
2701 patterns with negative forward assertions or any backward assertions
2702 preceding ".*" at the start, unless the pattern required a fixed first
2703 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2704
2705 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2706 starting at the last subject character, bytes beyond the end of the subject
2707 string were read.
2708
2709 11. Renamed the variable "class" as "classbits" to make life easier for C++
2710 users. (Previously there was a macro definition, but it apparently wasn't
2711 enough.)
2712
2713 12. Added the new field "tables" to the extra data so that tables can be passed
2714 in at exec time, or the internal tables can be re-selected. This allows
2715 a compiled regex to be saved and re-used at a later time by a different
2716 program that might have everything at different addresses.
2717
2718 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2719 -R library as well as a -L library.
2720
2721 14. The debugging options of pcretest (-d on the command line or D on a
2722 pattern) showed incorrect output for anything following an extended class
2723 that contained multibyte characters and which was followed by a quantifier.
2724
2725 15. Added optional support for general category Unicode character properties
2726 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2727 support. It adds about 90K to the size of the library. The meanings of the
2728 inbuilt class escapes such as \d and \s have NOT been changed.
2729
2730 16. Updated pcredemo.c to include calls to free() to release the memory for the
2731 compiled pattern.
2732
2733 17. The generated file chartables.c was being created in the source directory
2734 instead of in the building directory. This caused the build to fail if the
2735 source directory was different from the building directory, and was
2736 read-only.
2737
2738 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2739 file. No doubt somebody will tell me if they don't make sense... Also added
2740 Dan Mooney's comments about building on OpenVMS.
2741
2742 19. Added support for partial matching via the PCRE_PARTIAL option for
2743 pcre_exec() and the \P data escape in pcretest.
2744
2745 20. Extended pcretest with 3 new pattern features:
2746
2747 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2748 write the compiled pattern to the file whose name is "rest-of-line".
2749 This is a straight binary dump of the data, with the saved pointer to
2750 the character tables forced to be NULL. The study data, if any, is
2751 written too. After writing, pcretest reads a new pattern.
2752
2753 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2754 compiled pattern from the given file. There must not be any
2755 occurrences of "<" in the file name (pretty unlikely); if there are,
2756 pcretest will instead treat the initial "<" as a pattern delimiter.
2757 After reading in the pattern, pcretest goes on to read data lines as
2758 usual.
2759
2760 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2761 and 16-bit fields in a compiled pattern, to simulate a pattern that
2762 was compiled on a host of opposite endianness.
2763
2764 21. The pcre-exec() function can now cope with patterns that were compiled on
2765 hosts of opposite endianness, with this restriction:
2766
2767 As for any compiled expression that is saved and used later, the tables
2768 pointer field cannot be preserved; the extra_data field in the arguments
2769 to pcre_exec() should be used to pass in a tables address if a value
2770 other than the default internal tables were used at compile time.
2771
2772 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2773 now diagnosed as an error. Previously, most of the time, a negative number
2774 would have been treated as zero, but if in addition "ovector" was passed as
2775 NULL, a crash could occur.
2776
2777 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2778 new versions from the libtool 1.5 distribution (the last one is a copy of
2779 a file called libtool.m4). This seems to have fixed the need to patch
2780 "configure" to support Darwin 1.3 (which I used to do). However, I still
2781 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2782 workstation).
2783
2784 24. Changed the PCRE licence to be the more standard "BSD" licence.
2785
2786
2787 Version 4.5 01-Dec-03
2788 ---------------------
2789
2790 1. There has been some re-arrangement of the code for the match() function so
2791 that it can be compiled in a version that does not call itself recursively.
2792 Instead, it keeps those local variables that need separate instances for
2793 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2794 needs to "recurse". Keeping track of where control must go is done by means
2795 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2796 hide most of the details from the main code, and operates only if
2797 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2798 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2799 operating.
2800
2801 To make it easier for callers to provide specially tailored get/free
2802 functions for this usage, two new functions, pcre_stack_malloc, and
2803 pcre_stack_free, are used. They are always called in strict stacking order,
2804 and the size of block requested is always the same.
2805
2806 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2807 PCRE has been compiled to use the stack or the heap for recursion. The
2808 -C option of pcretest uses this to show which version is compiled.
2809
2810 A new data escape \S, is added to pcretest; it causes the amounts of store
2811 obtained and freed by both kinds of malloc/free at match time to be added
2812 to the output.
2813
2814 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2815 what's available on my current Linux desktop machine.
2816
2817 3. When matching a UTF-8 string, the test for a valid string at the start has
2818 been extended. If start_offset is not zero, PCRE now checks that it points
2819 to a byte that is the start of a UTF-8 character. If not, it returns
2820 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2821 this is necessary because there may be backward assertions in the pattern.
2822 When matching the same subject several times, it may save resources to use
2823 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2824
2825 4. The code for checking the validity of UTF-8 strings has been tightened so
2826 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2827 containing "overlong sequences".
2828
2829 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2830 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2831 should have been "&", but it just so happened that all the cases this let
2832 through by mistake were picked up later in the function.
2833
2834 6. I had used a variable called "isblank" - this is a C99 function, causing
2835 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2836
2837 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2838 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2839 so the version is shown; (c) stop "make test" echoing "./RunTest".
2840
2841 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2842
2843 9. The prototype for memmove() for systems that don't have it was using
2844 size_t, but the inclusion of the header that defines size_t was later. I've
2845 moved the #includes for the C headers earlier to avoid this.
2846
2847 10. Added some adjustments to the code to make it easier to compiler on certain
2848 special systems:
2849
2850 (a) Some "const" qualifiers were missing.
2851 (b) Added the macro EXPORT before all exported functions; by default this
2852 is defined to be empty.
2853 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2854 that it reads its output file name as an argument instead of writing
2855 to the standard output and assuming this can be redirected.
2856
2857 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2858 class containing characters with values greater than 255, PCRE compilation
2859 went into a loop.
2860
2861 12. A recursive reference to a subpattern that was within another subpattern
2862 that had a minimum quantifier of zero caused PCRE to crash. For example,
2863 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2864 recursion. If the recursively-called subpattern itself had a zero repeat,
2865 that was OK.
2866
2867 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2868 buffer into which it was copied (for escape processing) was still set at
2869 1024, so long lines caused crashes.
2870
2871 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2872 "internal error: code overflow...". This applied to any character class
2873 that was followed by a possessive quantifier.
2874
2875 15. Modified the Makefile to add libpcre.la as a prerequisite for
2876 libpcreposix.la because I was told this is needed for a parallel build to
2877 work.
2878
2879 16. If a pattern that contained .* following optional items at the start was
2880 studied, the wrong optimizing data was generated, leading to matching
2881 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2882 matching string must start with a or b or c. The correct conclusion for
2883 this pattern is that a match can start with any character.
2884
2885
2886 Version 4.4 13-Aug-03
2887 ---------------------
2888
2889 1. In UTF-8 mode, a character class containing characters with values between
2890 127 and 255 was not handled correctly if the compiled pattern was studied.
2891 In fixing this, I have also improved the studying algorithm for such
2892 classes (slightly).
2893
2894 2. Three internal functions had redundant arguments passed to them. Removal
2895 might give a very teeny performance improvement.
2896
2897 3. Documentation bug: the value of the capture_top field in a callout is *one
2898 more than* the number of the hightest numbered captured substring.
2899
2900 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2901 in incorrectly linking with a previously installed version. They now link
2902 explicitly with libpcre.la.
2903
2904 5. configure.in no longer needs to recognize Cygwin specially.
2905
2906 6. A problem in pcre.in for Windows platforms is fixed.
2907
2908 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2909 pcretest, it used to include the size of the study block as part of its
2910 output. Unfortunately, the structure contains a field that has a different
2911 size on different hardware architectures. This meant that the tests that
2912 showed this size failed. As the block is currently always of a fixed size,
2913 this information isn't actually particularly useful in pcretest output, so
2914 I have just removed it.
2915
2916 8. Three pre-processor statements accidentally did not start in column 1.
2917 Sadly, there are *still* compilers around that complain, even though
2918 standard C has not required this for well over a decade. Sigh.
2919
2920 9. In pcretest, the code for checking callouts passed small integers in the
2921 callout_data field, which is a void * field. However, some picky compilers
2922 complained about the casts involved for this on 64-bit systems. Now
2923 pcretest passes the address of the small integer instead, which should get
2924 rid of the warnings.
2925
2926 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2927 both compile and run time, and gives an error if an invalid UTF-8 sequence
2928 is found. There is a option for disabling this check in cases where the
2929 string is known to be correct and/or the maximum performance is wanted.
2930
2931 11. In response to a bug report, I changed one line in Makefile.in from
2932
2933 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2934 to
2935 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2936
2937 to look similar to other lines, but I have no way of telling whether this
2938 is the right thing to do, as I do not use Windows. No doubt I'll get told
2939 if it's wrong...
2940
2941
2942 Version 4.3 21-May-03
2943 ---------------------
2944
2945 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2946 Makefile.
2947
2948 2. Some refactoring to improve the quality of the code:
2949
2950 (i) The utf8_table... variables are now declared "const".
2951
2952 (ii) The code for \cx, which used the "case flipping" table to upper case
2953 lower case letters, now just substracts 32. This is ASCII-specific,
2954 but the whole concept of \cx is ASCII-specific, so it seems
2955 reasonable.
2956
2957 (iii) PCRE was using its character types table to recognize decimal and
2958 hexadecimal digits in the pattern. This is silly, because it handles
2959 only 0-9, a-f, and A-F, but the character types table is locale-
2960 specific, which means strange things might happen. A private
2961 table is now used for this - though it costs 256 bytes, a table is
2962 much faster than multiple explicit tests. Of course, the standard
2963 character types table is still used for matching digits in subject
2964 strings against \d.
2965
2966 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2967 ending in _t are). So I've renamed it as ESC_tee.
2968
2969 3. The first argument for regexec() in the POSIX wrapper should have been
2970 defined as "const".
2971
2972 4. Changed pcretest to use malloc() for its buffers so that they can be
2973 Electric Fenced for debugging.
2974
2975 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2976 to read one or more bytes before the start of the subject string. Often this
2977 had no effect on PCRE's behaviour, but in some circumstances it could
2978 provoke a segmentation fault.
2979
2980 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2981 to try to read one or more bytes before the start of the subject string.
2982
2983 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2984 UTF-8 support could misbehave in various ways if the subject string
2985 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2986 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2987 back over UTF-8 characters.)
2988
2989
2990 Version 4.2 14-Apr-03
2991 ---------------------
2992
2993 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2994
2995 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2996 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2997 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2998 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2999 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3000 and BUILD_EXEEXT
3001 Note: automatic setting of the BUILD variables is not yet working
3002 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3003 compile-time but not at link-time
3004 [LINK]: use for linking executables only
3005 make different versions for Windows and non-Windows
3006 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3007 libraries
3008 [LINK_FOR_BUILD]: new variable
3009 [OBJEXT]: use throughout
3010 [EXEEXT]: use throughout
3011 <winshared>: new target
3012 <wininstall>: new target
3013 <dftables.o>: use native compiler
3014 <dftables>: use native linker
3015 <install>: handle Windows platform correctly
3016 <clean>: ditto
3017 <check>: ditto
3018 copy DLL to top builddir before testing
3019
3020 As part of these changes, -no-undefined was removed again. This was reported
3021 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3022 in any case.
3023
3024 3. Some tidies to get rid of compiler warnings:
3025
3026 . In the match_data structure, match_limit was an unsigned long int, whereas
3027 match_call_count was an int. I've made them both unsigned long ints.
3028
3029 . In pcretest the fact that a const uschar * doesn't automatically cast to
3030 a void * provoked a warning.
3031
3032 . Turning on some more compiler warnings threw up some "shadow" variables
3033 and a few more missing casts.
3034
3035 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3036 option, a class that contained a single character with a value between 128
3037 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3038
3039 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3040 option, a class that contained several characters, but with at least one
3041 whose value was between 128 and 255 caused PCRE to crash.
3042
3043
3044 Version 4.1 12-Mar-03
3045 ---------------------
3046
3047 1. Compiling with gcc -pedantic found a couple of places where casts were
3048 needed, and a string in dftables.c that was longer than standard compilers are
3049 required to support.
3050
3051 2. Compiling with Sun's compiler found a few more places where the code could
3052 be tidied up in order to avoid warnings.
3053
3054 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3055 first of these names is deprecated in the latest Autoconf in favour of the name
3056 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3057 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3058 analogy I have changed it to CFLAGS_FOR_BUILD.
3059
3060 4. Added -no-undefined to the linking command in the Makefile, because this is
3061 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3062 linking step for the pcreposix library.
3063
3064 5. PCRE was failing to diagnose the case of two named groups with the same
3065 name.
3066
3067 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3068 literal character that is needed in the subject for a match, and scans along to
3069 ensure that it is present before embarking on the full matching process. This
3070 saves time in cases of nested unlimited repeats that are never going to match.
3071 Problem: the scan can take a lot of time if the subject is very long (e.g.
3072 megabytes), thus penalizing straightforward matches. It is now done only if the
3073 amount of subject to be scanned is less than 1000 bytes.
3074
3075 7. A lesser problem with the same optimization is that it was recording the
3076 first character of an anchored pattern as "needed", thus provoking a search
3077 right along the subject, even when the first match of the pattern was going to
3078 fail. The "needed" character is now not set for anchored patterns, unless it
3079 follows something in the pattern that is of non-fixed length. Thus, it still
3080 fulfils its original purpose of finding quick non-matches in cases of nested
3081 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3082
3083
3084 Version 4.0 17-Feb-03
3085 ---------------------
3086
3087 1. If a comment in an extended regex that started immediately after a meta-item
3088 extended to the end of string, PCRE compiled incorrect data. This could lead to
3089 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3090
3091 2. Moved to autoconf 2.53 and libtool 1.4.2.
3092
3093 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3094 the special perltest8 script is no longer needed - all the tests can be run
3095 from a single perltest script.
3096
3097 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3098 by \s. It has now been removed in PCRE. This means it isn't recognized as
3099 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3100 class [:space:] *does* include VT, thereby creating a mess.
3101
3102 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3103 space and tab.
3104
3105 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3106 its new features into the main test script, reducing the number of scripts.
3107
3108 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3109 were backward compatible, and made the (?i) apply to the whole pattern, as if
3110 /i were given. Now it behaves more logically, and applies the option setting
3111 only to what follows. PCRE has been changed to follow suit. However, if it
3112 finds options settings right at the start of the pattern, it extracts them into
3113 the global options, as before. Thus, they show up in the info data.
3114
3115 8. Added support for the \Q...\E escape sequence. Characters in between are
3116 treated as literals. This is slightly different from Perl in that $ and @ are
3117 also handled as literals inside the quotes. In Perl, they will cause variable
3118 interpolation. Note the following examples:
3119
3120 Pattern PCRE matches Perl matches
3121
3122 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3123 \Qabc\$xyz\E abc\$xyz abc\$xyz
3124 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3125
3126 For compatibility with Perl, \Q...\E sequences are recognized inside character
3127 classes as well as outside them.
3128
3129 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3130 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3131 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3132 signed/unsigned warnings.
3133
3134 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3135 option for pcretest, so I've replaced it by a simple function that does just
3136 that job.
3137
3138 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3139 "pcregrep -".
3140
3141 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3142 Java package. This provides some syntactic sugar for simple cases of what my
3143 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3144 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3145 item, you can use this simplified notation. Note that only makes sense with
3146 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3147 greediness, whatever the setting of the PCRE_UNGREEDY option.
3148
3149 13. A change of greediness default within a pattern was not taking effect at
3150 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3151 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3152 was abstracted outside.
3153
3154 14. PCRE now supports the \G assertion. It is true when the current matching
3155 position is at the start point of the match. This differs from \A when the
3156 starting offset is non-zero. Used with the /g option of pcretest (or similar
3157 code), it works in the same way as it does for Perl's /g option. If all
3158 alternatives of a regex begin with \G, the expression is anchored to the start
3159 match position, and the "anchored" flag is set in the compiled expression.
3160
3161 15. Some bugs concerning the handling of certain option changes within patterns
3162 have been fixed. These applied to options other than (?ims). For example,
3163 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3164 been the other way round. Some of this was related to change 7 above.
3165
3166 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3167 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3168 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3169 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3170
3171 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3172 mode. Unlike ".", it always matches newline, whatever the setting of
3173 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3174 assertions. Perl allows it, but it doesn't (in general) work because it can't
3175 calculate the length of the lookbehind. At least, that's the case for Perl
3176 5.8.0 - I've been told they are going to document that it doesn't work in
3177 future.
3178
3179 18. Added an error diagnosis for escapes that PCRE does not support: these are
3180 \L, \l, \N, \P, \p, \U, \u, and \X.
3181
3182 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3183 reading past the end of the pattern in cases such as /[abcd/.
3184
3185 20. PCRE was getting more memory than necessary for patterns with classes that
3186 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3187
3188 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3189 compiling PCRE for use with Virtual Pascal.
3190
3191 22. Small fix to the Makefile to make it work properly if the build is done
3192 outside the source tree.
3193
3194 23. Added a new extension: a condition to go with recursion. If a conditional
3195 subpattern starts with (?(R) the "true" branch is used if recursion has
3196 happened, whereas the "false" branch is used only at the top level.
3197
3198 24. When there was a very long string of literal characters (over 255 bytes
3199 without UTF support, over 250 bytes with UTF support), the computation of how
3200 much memory was required could be incorrect, leading to segfaults or other
3201 strange effects.
3202
3203 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3204 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3205 there was a subsequent back reference to those brackets. This meant that, for
3206 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3207 possible to check for precisely this case. All we can do is abandon the
3208 optimization if .* occurs inside capturing brackets when there are any back
3209 references whatsoever. (See below for a better fix that came later.)
3210
3211 26. The handling of the optimization for finding the first character of a
3212 non-anchored pattern, and for finding a character that is required later in the
3213 match were failing in some cases. This didn't break the matching; it just
3214 failed to optimize when it could. The way this is done has been re-implemented.
3215
3216 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3217
3218 28. Added a new feature that provides some of the functionality that Perl
3219 provides with (?{...}). The facility is termed a "callout". The way it is done
3220 in PCRE is for the caller to provide an optional function, by setting
3221 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3222 global variable. By default it is unset, which disables all calling out. To get
3223 the function called, the regex must include (?C) at appropriate points. This
3224 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3225 This provides a means of identifying different callout points. When PCRE
3226 reaches such a point in the regex, if pcre_callout has been set, the external
3227 function is called. It is provided with data in a structure called
3228 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3229 matching continues; if it returns a non-zero value, the match at the current
3230 point fails. However, backtracking will occur if possible. [This was changed
3231 later and other features added - see item 49 below.]
3232
3233 29. pcretest is upgraded to test the callout functionality. It provides a
3234 callout function that displays information. By default, it shows the start of
3235 the match and the current position in the text. There are some new data escapes
3236 to vary what happens:
3237
3238 \C+ in addition, show current contents of captured substrings
3239 \C- do not supply a callout function
3240 \C!n return 1 when callout number n is reached
3241 \C!n!m return 1 when callout number n is reached for the mth time
3242
3243 30. If pcregrep was called with the -l option and just a single file name, it
3244 output "<stdin>" if a match was found, instead of the file name.
3245
3246 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3247 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3248 pcre_exec(). This saves a malloc/free per call. The default value of
3249 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3250 when configuring.
3251
3252 32. The default maximum size of a compiled pattern is 64K. There have been a
3253 few cases of people hitting this limit. The code now uses macros to handle the
3254 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3255 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3256 configuring. Tests 2 and 5 work only with 2-byte links because they output
3257 debugging information about compiled patterns.
3258
3259 33. Internal code re-arrangements:
3260
3261 (a) Moved the debugging function for printing out a compiled regex into
3262 its own source file (printint.c) and used #include to pull it into
3263 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3264 separate copies.
3265
3266 (b) Defined the list of op-code names for debugging as a macro in
3267 internal.h so that it is next to the definition of the opcodes.
3268
3269 (c) Defined a table of op-code lengths for simpler skipping along compiled
3270 code. This is again a macro in internal.h so that it is next to the
3271 definition of the opcodes.
3272
3273 34. Added support for recursive calls to individual subpatterns, along the
3274 lines of Robin Houston's patch (but implemented somewhat differently).
3275
3276 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3277 allow it to read and process whole directories in Win32. This code was
3278 contributed by Lionel Fourquaux; it has not been tested by me.
3279
3280 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3281 used to name a group. Names consist of alphanumerics and underscores, and must
3282 be unique. Back references use the syntax (?P=name) and recursive calls use
3283 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3284 numbers. The function pcre_fullinfo() can be used after compilation to extract
3285 a name/number map. There are three relevant calls:
3286
3287 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3288 PCRE_INFO_NAMECOUNT yields the number of entries
3289 PCRE_INFO_NAMETABLE yields a pointer to the map.
3290
3291 The map is a vector of fixed-size entries. The size of each entry depends on
3292 the length of the longest name used. The first two bytes of each entry are the
3293 group number, most significant byte first. There follows the corresponding
3294 name, zero terminated. The names are in alphabetical order.
3295
3296 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3297 case instead of 255. Making it the same both with and without UTF-8 support
3298 means that the same test output works with both.
3299
3300 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3301 calling malloc() with a zero argument.
3302
3303 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3304 optimization. I've improved things by keeping a bitmap of backreferences with
3305 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3306 fact referenced, the optimization can be applied. It is unlikely that a
3307 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3308 the match to follow \n) will appear inside brackets with a number greater than
3309 31, but if it does, any back reference > 31 suppresses the optimization.
3310
3311 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3312 of disabling numbered capturing parentheses. Any opening parenthesis that is
3313 not followed by ? behaves as if it were followed by ?: but named parentheses
3314 can still be used for capturing (and they will acquire numbers in the usual
3315 way).
3316
3317 41. Redesigned the return codes from the match() function into yes/no/error so
3318 that errors can be passed back from deep inside the nested calls. A malloc
3319 failure while inside a recursive subpattern call now causes the
3320 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3321
3322 42. It is now possible to set a limit on the number of times the match()
3323 function is called in a call to pcre_exec(). This facility makes it possible to
3324 limit the amount of recursion and backtracking, though not in a directly
3325 obvious way, because the match() function is used in a number of different
3326 circumstances. The count starts from zero for each position in the subject
3327 string (for non-anchored patterns). The default limit is, for compatibility, a
3328 large number, namely 10 000 000. You can change this in two ways:
3329
3330 (a) When configuring PCRE before making, you can use --with-match-limit=n
3331 to set a default value for the compiled library.
3332
3333 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3334 a different value is set. See 45 below.
3335
3336 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3337
3338 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3339 of things that can be changed at compile time. The first argument specifies
3340 what is wanted and the second points to where the information is to be placed.
3341 The current list of available information is:
3342
3343 PCRE_CONFIG_UTF8
3344
3345 The output is an integer that is set to one if UTF-8 support is available;
3346 otherwise it is set to zero.
3347
3348 PCRE_CONFIG_NEWLINE
3349
3350 The output is an integer that it set to the value of the code that is used for
3351 newline. It is either LF (10) or CR (13).
3352
3353 PCRE_CONFIG_LINK_SIZE
3354
3355 The output is an integer that contains the number of bytes used for internal
3356 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3357
3358 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3359
3360 The output is an integer that contains the threshold above which the POSIX
3361 interface uses malloc() for output vectors. See item 31 above.
3362
3363 PCRE_CONFIG_MATCH_LIMIT
3364
3365 The output is an unsigned integer that contains the default limit of the number
3366 of match() calls in a pcre_exec() execution. See 42 above.
3367
3368 44. pcretest has been upgraded by the addition of the -C option. This causes it
3369 to extract all the available output from the new pcre_config() function, and to
3370 output it. The program then exits immediately.
3371
3372 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3373 order to support additional features. One way would have been to define
3374 pcre_exec2() (for example) with extra arguments, but this would not have been
3375 extensible, and would also have required all calls to the original function to
3376 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3377 is used for passing in "extra" data from pcre_study().
3378
3379 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3380 contains the following fields:
3381
3382 flags a bitmap indicating which of the following fields are set
3383 study_data opaque data from pcre_study()
3384 match_limit a way of specifying a limit on match() calls for a specific
3385 call to pcre_exec()
3386 callout_data data for callouts (see 49 below)
3387
3388 The flag bits are also defined in pcre.h, and are
3389
3390 PCRE_EXTRA_STUDY_DATA
3391 PCRE_EXTRA_MATCH_LIMIT
3392 PCRE_EXTRA_CALLOUT_DATA
3393
3394 The pcre_study() function now returns one of these new pcre_extra blocks, with
3395 the actual study data pointed to by the study_data field, and the
3396 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3397 before. That is, this change is entirely upwards-compatible and requires no
3398 change to existing code.
3399
3400 If you want to pass in additional data to pcre_exec(), you can either place it
3401 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3402 block.
3403
3404 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3405 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3406 times with different match limits, until it finds the minimum value needed for
3407 pcre_exec() to complete. The value is then output. This can be instructive; for
3408 most simple matches the number is quite small, but for pathological cases it
3409 gets very large very quickly.
3410
3411 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3412 returns the size of the data block pointed to by the study_data field in a
3413 pcre_extra block, that is, the value that was passed as the argument to
3414 pcre_malloc() when PCRE was getting memory in which to place the information
3415 created by pcre_study(). The fourth argument should point to a size_t variable.
3416 pcretest has been extended so that this information is shown after a successful
3417 pcre_study() call when information about the compiled regex is being displayed.
3418
3419 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3420 because what follows is always an absolute path. (Later: it turns out that this
3421 is more than cosmetic for MinGW, because it doesn't like empty path
3422 components.)
3423
3424 49. Some changes have been made to the callout feature (see 28 above):
3425
3426 (i) A callout function now has three choices for what it returns:
3427
3428 0 => success, carry on matching
3429 > 0 => failure at this point, but backtrack if possible
3430 < 0 => serious error, return this value from pcre_exec()
3431
3432 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3433 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3434 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3435 use by callout functions. It will never be used by PCRE itself.
3436
3437 (ii) The pcre_extra structure (see 45 above) has a void * field called
3438 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3439 pcre_callout_block structure has a field of the same name. The contents of
3440 the field passed in the pcre_extra structure are passed to the callout
3441 function in the corresponding field in the callout block. This makes it
3442 easier to use the same callout-containing regex from multiple threads. For
3443 testing, the pcretest program has a new data escape
3444
3445 \C*n pass the number n (may be negative) as callout_data
3446
3447 If the callout function in pcretest receives a non-zero value as
3448 callout_data, it returns that value.
3449
3450 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3451 there were some redundant $(CFLAGS) in commands that are now specified as
3452 $(LINK), which already includes $(CFLAGS).
3453
3454 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3455 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3456 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3457 one-byte characters throughout. Note that case-insensitive matching applies
3458 only to characters whose values are less than 256. PCRE doesn't support the
3459 notion of cases for higher-valued characters.
3460
3461 (i) A character class whose characters are all within 0-255 is handled as
3462 a bit map, and the map is inverted for negative classes. Previously, a
3463 character > 255 always failed to match such a class; however it should
3464 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3465
3466 (ii) A negated character class with a single character < 255 is coded as
3467 "not this character" (OP_NOT). This wasn't working properly when the test
3468 character was multibyte, either singly or repeated.
3469
3470 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3471 mode, for example: \x{100}{2,3}.
3472
3473 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3474 singly or repeated) now correctly test multibyte characters. However,
3475 PCRE doesn't recognize any characters with values greater than 255 as
3476 digits, spaces, or word characters. Such characters always match \D, \S,
3477 and \W, and never match \d, \s, or \w.
3478
3479 (v) Classes may now contain characters and character ranges with values
3480 greater than 255. For example: [ab\x{100}-\x{400}].
3481
3482 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3483 PCRE in UTF-8 mode.
3484
3485 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3486 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3487 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3488 value.)
3489
3490 53. The single man page has become too large. I have therefore split it up into
3491 a number of separate man pages. These also give rise to individual HTML pages;
3492 these are now put in a separate directory, and there is an index.html page that
3493 lists them all. Some hyperlinking between the pages has been installed.
3494
3495 54. Added convenience functions for handling named capturing parentheses.
3496
3497 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3498 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3499 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3500 are faulted.
3501
3502 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3503 calling configure. These values are used when compiling the dftables.c program
3504 which is run to generate the source of the default character tables. They
3505 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3506 you will need to set these values.
3507
3508 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3509
3510
3511 Version 3.9 02-Jan-02
3512 ---------------------
3513
3514 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3515
3516 2. If --disable-static was given, the building process failed when trying to
3517 build pcretest and pcregrep. (For some reason it was using libtool to compile
3518 them, which is not right, as they aren't part of the library.)
3519
3520
3521 Version 3.8 18-Dec-01
3522 ---------------------
3523
3524 1. The experimental UTF-8 code was completely screwed up. It was packing the
3525 bytes in the wrong order. How dumb can you get?
3526
3527
3528 Version 3.7 29-Oct-01
3529 ---------------------
3530
3531 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3532 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3533 this didn't happen under Solaris 8, where I normally test things.
3534
3535 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3536 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3537 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3538 this fix an hour or so after the initial 3.7 release.)
3539
3540
3541 Version 3.6 23-Oct-01
3542 ---------------------
3543
3544 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3545 offsets passed as NULL with zero offset count.
3546
3547 2. The config.guess and config.sub files had not been updated when I moved to
3548 the latest autoconf.
3549
3550
3551 Version 3.5 15-Aug-01
3552 ---------------------
3553
3554 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3555 had been forgotten.
3556
3557 2. By using declared but undefined structures, we can avoid using "void"
3558 definitions in pcre.h while keeping the internal definitions of the structures
3559 private.
3560
3561 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3562 user point of view, this means that both static and shared libraries are built
3563 by default, but this can be individually controlled. More of the work of
3564 handling this static/shared cases is now inside libtool instead of PCRE's make
3565 file.
3566
3567 4. The pcretest utility is now installed along with pcregrep because it is
3568 useful for users (to test regexs) and by doing this, it automatically gets
3569 relinked by libtool. The documentation has been turned into a man page, so
3570 there are now .1, .txt, and .html versions in /doc.
3571
3572 5. Upgrades to pcregrep:
3573 (i) Added long-form option names like gnu grep.
3574 (ii) Added --help to list all options with an explanatory phrase.
3575 (iii) Added -r, --recursive to recurse into sub-directories.
3576 (iv) Added -f, --file to read patterns from a file.
3577
3578 6. pcre_exec() was referring to its "code" argument before testing that
3579 argument for NULL (and giving an error if it was NULL).
3580
3581 7. Upgraded Makefile.in to allow for compiling in a different directory from
3582 the source directory.
3583
3584 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3585 options bits, the pointer it was passed was to an int instead of to an unsigned
3586 long int. This mattered only on 64-bit systems.
3587
3588 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3589 generated) instead of pcre.in, which it its source. Also made the same change
3590 in several of the .c files.
3591
3592 10. A new release of gcc defines printf() as a macro, which broke pcretest
3593 because it had an ifdef in the middle of a string argument for printf(). Fixed
3594 by using separate calls to printf().
3595
3596 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3597 script, to force use of CR or LF instead of \n in the source. On non-Unix
3598 systems, the value can be set in config.h.
3599
3600 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3601 absolute limit. Changed the text of the error message to make this clear, and
3602 likewise updated the man page.
3603
3604 13. The limit of 99 on the number of capturing subpatterns has been removed.
3605 The new limit is 65535, which I hope will not be a "real" limit.
3606
3607
3608 Version 3.4 22-Aug-00
3609 ---------------------
3610
3611 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3612
3613 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3614
3615
3616 Version 3.3 01-Aug-00
3617 ---------------------
3618
3619 1. If an octal character was given, but the value was greater than \377, it
3620 was not getting masked to the least significant bits, as documented. This could
3621 lead to crashes in some systems.
3622
3623 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3624 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3625
3626 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3627 These just pass their arguments on to (pcre_free)(), but they are provided
3628 because some uses of PCRE bind it to non-C systems that can call its functions,
3629 but cannot call free() or pcre_free() directly.
3630
3631 4. Add "make test" as a synonym for "make check". Corrected some comments in
3632 the Makefile.
3633
3634 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3635 Makefile.
3636
3637 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3638 command called pgrep for grepping around the active processes.
3639
3640 7. Added the beginnings of support for UTF-8 character strings.
3641
3642 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3643 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3644 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3645 out for the ar command.)
3646
3647
3648 Version 3.2 12-May-00
3649 ---------------------
3650
3651 This is purely a bug fixing release.
3652
3653 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3654 of ZA. This was just one example of several cases that could provoke this bug,
3655 which was introduced by change 9 of version 2.00. The code for breaking
3656 infinite loops after an iteration that matches an empty string was't working
3657 correctly.
3658
3659 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3660 when matched against abbab (for example). After matching an empty string, it
3661 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3662 caused it to match further down the string than it should.
3663
3664 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3665 was there because it doesn't seem to be needed, and it causes trouble on some
3666 systems, as it is not a Standard C header. It has been removed.
3667
3668 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3669 were reported on the Macintosh. The changes were from
3670
3671 while ((c = *(++ptr)) != 0 && c != '\n');
3672 to
3673 while ((c = *(++ptr)) != 0 && c != '\n') ;
3674
3675 Totally extraordinary, but if that's what it takes...
3676
3677 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3678 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3679 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3680 assumes the way PCRE uses memmove() (always moving upwards).
3681
3682 6. PCRE is being used in one environment where strchr() is not available. There
3683 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3684 faster code anyway.
3685
3686
3687 Version 3.1 09-Feb-00
3688 ---------------------
3689
3690 The only change in this release is the fixing of some bugs in Makefile.in for
3691 the "install" target:
3692
3693 (1) It was failing to install pcreposix.h.
3694
3695 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3696
3697
3698 Version 3.0 01-Feb-00
3699 ---------------------
3700
3701 1. Add support for the /+ modifier to perltest (to output $` like it does in
3702 pcretest).
3703
3704 2. Add support for the /g modifier to perltest.
3705
3706 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3707 matches null strings.
3708
3709 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3710 pattern. Perl treats empty patterns specially - it reuses the most recent
3711 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3712 effect.
3713
3714 5. The POSIX interface was broken in that it was just handing over the POSIX
3715 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3716 required a bigger vector, with some working space on the end. This means that
3717 the POSIX wrapper now has to get and free some memory, and copy the results.
3718
3719 6. Added some simple autoconf support, placing the test data and the
3720 documentation in separate directories, re-organizing some of the
3721 information files, and making it build pcre-config (a GNU standard). Also added
3722 libtool support for building PCRE as a shared library, which is now the
3723 default.
3724
3725 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3726 09 are not valid octal constants. Single digits will be used for minor values
3727 less than 10.
3728
3729 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3730 existing programs that set these in the POSIX interface can use PCRE without
3731 modification.
3732
3733 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3734 return all that pcre_info() returns, plus additional data. The pcre_info()
3735 function is retained for compatibility, but is considered to be obsolete.
3736
3737 10. Added experimental recursion feature (?R) to handle one common case that
3738 Perl 5.6 will be able to do with (?p{...}).
3739
3740 11. Added support for POSIX character classes like [:alpha:], which Perl is
3741 adopting.
3742
3743
3744 Version 2.08 31-Aug-99
3745 ----------------------
3746
3747 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3748 trying to match at the startoffset position, but instead was moving forward to
3749 the next newline as if a previous match had failed.
3750
3751 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3752 and could get into a loop if a null string was matched other than at the start
3753 of the subject.
3754
3755 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3756 be distinguished at compile time, and for completeness also added PCRE_DATE.
3757
3758 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3759 in GnuWin32 environments.
3760
3761
3762 Version 2.07 29-Jul-99
3763 ----------------------
3764
3765 1. The documentation is now supplied in plain text form and HTML as well as in
3766 the form of man page sources.
3767
3768 2. C++ compilers don't like assigning (void *) values to other pointer types.
3769 In particular this affects malloc(). Although there is no problem in Standard
3770 C, I've put in casts to keep C++ compilers happy.
3771
3772 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3773 should be (const char *).
3774
3775 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3776 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3777 However, I haven't made this a standard facility. The documentation doesn't
3778 mention it, and the Makefile doesn't support it.
3779
3780 5. The Makefile now contains an "install" target, with editable destinations at
3781 the top of the file. The pcretest program is not installed.
3782
3783 6. pgrep -V now gives the PCRE version number and date.
3784
3785 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3786 causing the entire string to be ignored, instead of just the last character.
3787
3788 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3789 non-matching string, it can take a very, very long time, even for strings of
3790 quite modest length, because of the nested recursion. PCRE now does better in
3791 some of these cases. It does this by remembering the last required literal
3792 character in the pattern, and pre-searching the subject to ensure it is present
3793 before running the real match. In other words, it applies a heuristic to detect
3794 some types of certain failure quickly, and in the above example, if presented
3795 with a string that has no trailing " it gives "no match" very quickly.
3796
3797 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3798 other alternatives are tried instead.
3799
3800
3801 Version 2.06 09-Jun-99
3802 ----------------------
3803
3804 1. Change pcretest's output for amount of store used to show just the code
3805 space, because the remainder (the data block) varies in size between 32-bit and
3806 64-bit systems.
3807
3808 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3809 start matching at. This allows lookbehinds to work when searching for multiple
3810 occurrences in a string.
3811
3812 3. Added additional options to pcretest for testing multiple occurrences:
3813
3814 /+ outputs the rest of the string that follows a match
3815 /g loops for multiple occurrences, using the new startoffset argument
3816 /G loops for multiple occurrences by passing an incremented pointer
3817
3818 4. PCRE wasn't doing the "first character" optimization for patterns starting
3819 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3820 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3821 the letter 'x'. On long subject strings, this gives a significant speed-up.
3822
3823
3824 Version 2.05 21-Apr-99
3825 ----------------------
3826
3827 1. Changed the type of magic_number from int to long int so that it works
3828 properly on 16-bit systems.
3829
3830 2. Fixed a bug which caused patterns starting with .* not to work correctly
3831 when the subject string contained newline characters. PCRE was assuming
3832 anchoring for such patterns in all cases, which is not correct because .* will
3833 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3834 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3835 must be retried after every newline in the subject.
3836
3837
3838 Version 2.04 18-Feb-99
3839 ----------------------
3840
3841 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3842 computation of the store needed to hold the pattern was incorrect (too large).
3843 If such patterns were nested a few deep, this could multiply and become a real
3844 problem.
3845
3846 2. Added /M option to pcretest to show the memory requirement of a specific
3847 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3848
3849 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3850 compiled in such a way that the backtracking after subsequent failure was
3851 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3852 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3853
3854
3855 Version 2.03 02-Feb-99
3856 ----------------------
3857
3858 1. Fixed typo and small mistake in man page.
3859
3860 2. Added 4th condition (GPL supersedes if conflict) and created separate
3861 LICENCE file containing the conditions.
3862
3863 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3864 Perl, that is the internal \ allows the delimiter to be included in the
3865 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3866 the final delimiter, add \ to the end of the pattern (to test the error).
3867
3868 4. Added the convenience functions for extracting substrings after a successful
3869 match. Updated pcretest to make it able to test these functions.
3870
3871
3872 Version 2.02 14-Jan-99
3873 ----------------------
3874
3875 1. Initialized the working variables associated with each extraction so that
3876 their saving and restoring doesn't refer to uninitialized store.
3877
3878 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3879 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3880 fix the problem.
3881
3882 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3883 calls, and wasn't printing the correct value for compiling calls. Increased the
3884 default value of LOOPREPEAT, and the number of significant figures in the
3885 times.
3886
3887 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3888
3889 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3890 a building problem on Windows NT with a FAT file system.
3891
3892
3893 Version 2.01 21-Oct-98
3894 ----------------------
3895
3896 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3897 to character tables built by pcre_maketables() in the current locale. If NULL
3898 is passed, the default tables are used.
3899
3900
3901 Version 2.00 24-Sep-98
3902 ----------------------
3903
3904 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3905 it any more.
3906
3907 2. Allow quantification of (?>) groups, and make it work correctly.
3908
3909 3. The first character computation wasn't working for (?>) groups.
3910
3911 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3912 end of the subject) and add 5.005's \z, which really does match only at the
3913 very end of the subject.
3914
3915 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3916
3917 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3918 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3919 localized options. All options to pcre_study() were also removed.
3920
3921 7. Add other new features from 5.005:
3922
3923 $(?<= positive lookbehind
3924 $(?<! negative lookbehind
3925 (?imsx-imsx) added the unsetting capability
3926 such a setting is global if at outer level; local otherwise
3927 (?imsx-imsx:) non-capturing groups with option setting
3928 (?(cond)re|re) conditional pattern matching
3929
3930 A backreference to itself in a repeated group matches the previous
3931 captured string.
3932
3933 8. General tidying up of studying (both automatic and via "study")
3934 consequential on the addition of new assertions.
3935
3936 9. As in 5.005, unlimited repeated groups that could match an empty substring
3937 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3938 runtime if any iteration does actually match an empty substring.
3939
3940 10. Include the RunTest script in the distribution.
3941
3942 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3943 discrepancies, some of which were old and were also with respect to 5.004. They
3944 have now been fixed.
3945
3946
3947 Version 1.09 28-Apr-98
3948 ----------------------
3949
3950 1. A negated single character class followed by a quantifier with a minimum
3951 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3952 program crashes, or just wrong answers. This did not apply to negated classes
3953 containing more than one character, or to minima other than one.
3954
3955
3956 Version 1.08 27-Mar-98
3957 ----------------------
3958
3959 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3960
3961 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3962 latter must appear before anything that relies on it in the pattern.
3963
3964
3965 Version 1.07 16-Feb-98
3966 ----------------------
3967
3968 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3969 repeat of a potentially empty string).
3970
3971
3972 Version 1.06 23-Jan-98
3973 ----------------------
3974
3975 1. Added Markus Oberhumer's little patches for C++.
3976
3977 2. Literal strings longer than 255 characters were broken.
3978
3979
3980 Version 1.05 23-Dec-97
3981 ----------------------
3982
3983 1. Negated character classes containing more than one character were failing if
3984 PCRE_CASELESS was set at run time.
3985
3986
3987 Version 1.04 19-Dec-97
3988 ----------------------
3989
3990 1. Corrected the man page, where some "const" qualifiers had been omitted.
3991
3992 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3993 input syntax.
3994
3995 3. Fixed memory leak which occurred when a regex with back references was
3996 matched with an offsets vector that wasn't big enough. The temporary memory
3997 that is used in this case wasn't being freed if the match failed.
3998
3999 4. Tidied pcretest to ensure it frees memory that it gets.
4000
4001 5. Temporary memory was being obtained in the case where the passed offsets
4002 vector was exactly big enough.
4003
4004 6. Corrected definition of offsetof() from change 5 below.
4005
4006 7. I had screwed up change 6 below and broken the rules for the use of
4007 setjmp(). Now fixed.
4008
4009
4010 Version 1.03 18-Dec-97
4011 ----------------------
4012
4013 1. A erroneous regex with a missing opening parenthesis was correctly
4014 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4015 on some systems.
4016
4017 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4018 it was reported that one broken compiler failed on the former because "code" is
4019 also an independent variable.
4020
4021 3. The erroneous regex a[]b caused an array overrun reference.
4022
4023 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4024 fail on data ending with that character. (It was going on too far, and checking
4025 the next character, typically a binary zero.) This was specific to the
4026 optimized code for single-character negative classes.
4027
4028 5. Added a contributed patch from the TIN world which does the following:
4029
4030 + Add an undef for memmove, in case the the system defines a macro for it.
4031
4032 + Add a definition of offsetof(), in case there isn't one. (I don't know
4033 the reason behind this - offsetof() is part of the ANSI standard - but
4034 it does no harm).
4035
4036 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4037 most of the places where whitespace preceded '#'. I have given up and
4038 allowed the remaining 2 cases to be at the margin.
4039
4040 + Rename some variables in pcre to eliminate shadowing. This seems very
4041 pedantic, but does no harm, of course.
4042
4043 6. Moved the call to setjmp() into its own function, to get rid of warnings
4044 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4045
4046 7. Constructs such as \d{8,} were compiling into the equivalent of
4047 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4048 outcome, but in this particular case used more store than had been allocated,
4049 which caused the bug to be discovered because it threw up an internal error.
4050
4051 8. The debugging code in both pcre and pcretest for outputting the compiled
4052 form of a regex was going wrong in the case of back references followed by
4053 curly-bracketed repeats.
4054
4055
4056 Version 1.02 12-Dec-97
4057 ----------------------
4058
4059 1. Typos in pcre.3 and comments in the source fixed.
4060
4061 2. Applied a contributed patch to get rid of places where it used to remove
4062 'const' from variables, and fixed some signed/unsigned and uninitialized
4063 variable warnings.
4064
4065 3. Added the "runtest" target to Makefile.
4066
4067 4. Set default compiler flag to -O2 rather than just -O.
4068
4069
4070 Version 1.01 19-Nov-97
4071 ----------------------
4072
4073 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4074 like /([ab]*)*/, that is, for classes with more than one character in them.
4075
4076 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4077 as /((?>a*))*/ (a PCRE_EXTRA facility).
4078
4079
4080 Version 1.00 18-Nov-97
4081 ----------------------
4082
4083 1. Added compile-time macros to support systems such as SunOS4 which don't have
4084 memmove() or strerror() but have other things that can be used instead.
4085
4086 2. Arranged that "make clean" removes the executables.
4087
4088
4089 Version 0.99 27-Oct-97
4090 ----------------------
4091
4092 1. Fixed bug in code for optimizing classes with only one character. It was
4093 initializing a 32-byte map regardless, which could cause it to run off the end
4094 of the memory it had got.
4095
4096 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4097
4098
4099 Version 0.98 22-Oct-97
4100 ----------------------
4101
4102 1. Fixed bug in code for handling temporary memory usage when there are more
4103 back references than supplied space in the ovector. This could cause segfaults.
4104
4105
4106 Version 0.97 21-Oct-97
4107 ----------------------
4108
4109 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4110
4111 2. Optimized negated single characters not to use a bit map.
4112
4113 3. Brought error texts together as macro definitions; clarified some of them;
4114 fixed one that was wrong - it said "range out of order" when it meant "invalid
4115 escape sequence".
4116
4117 4. Changed some char * arguments to const char *.
4118
4119 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4120
4121 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4122 pcretest.
4123
4124
4125 Version 0.96 16-Oct-97
4126 ----------------------
4127
4128 1. Added a simple "pgrep" utility to the distribution.
4129
4130 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4131 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4132 where "ddd" means "one or more decimal digits".
4133
4134 3. Fixed serious bug. If a pattern had a back reference, but the call to
4135 pcre_exec() didn't supply a large enough ovector to record the related
4136 identifying subpattern, the match always failed. PCRE now remembers the number
4137 of the largest back reference, and gets some temporary memory in which to save
4138 the offsets during matching if necessary, in order to ensure that
4139 backreferences always work.
4140
4141 4. Increased the compatibility with Perl in a number of ways:
4142
4143 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4144 to request this handling. The option can be set at compile or exec time.
4145
4146 (b) $ matches before a terminating newline by default; an option
4147 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4148 mode). The option can be set at compile or exec time.
4149
4150 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4151 the same as Perl's. If the decimal number it represents is less than 10
4152 or there aren't that many previous left capturing parentheses, an octal
4153 escape is read. Inside a character class, it's always an octal escape,
4154 even if it is a single digit.
4155
4156 (d) An escaped but undefined alphabetic character is taken as a literal,
4157 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4158 escapes.
4159
4160 (e) {0} is now permitted. (The previous item is removed from the compiled
4161 pattern).
4162
4163 5. Changed all the names of code files so that the basic parts are no longer
4164 than 10 characters, and abolished the teeny "globals.c" file.
4165
4166 6. Changed the handling of character classes; they are now done with a 32-byte
4167 bit map always.
4168
4169 7. Added the -d and /D options to pcretest to make it possible to look at the
4170 internals of compilation without having to recompile pcre.
4171
4172
4173 Version 0.95 23-Sep-97
4174 ----------------------
4175
4176 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4177 \x20 at the start of a run of normal characters. These were being treated as
4178 real characters, instead of the source characters being re-checked.
4179
4180
4181 Version 0.94 18-Sep-97
4182 ----------------------
4183
4184 1. The functions are now thread-safe, with the caveat that the global variables
4185 containing pointers to malloc() and free() or alternative functions are the
4186 same for all threads.
4187
4188 2. Get pcre_study() to generate a bitmap of initial characters for non-
4189 anchored patterns when this is possible, and use it if passed to pcre_exec().
4190
4191
4192 Version 0.93 15-Sep-97
4193 ----------------------
4194
4195 1. /(b)|(:+)/ was computing an incorrect first character.
4196
4197 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4198 but not actually doing anything yet.
4199
4200 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4201 as Perl does (e.g. [-az] or [az-]).
4202
4203 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4204 all possible positions.
4205
4206 5. Split up into different modules to avoid including unneeded functions in a
4207 compiled binary. However, compile and exec are still in one module. The "study"
4208 function is split off.
4209
4210 6. The character tables are now in a separate module whose source is generated
4211 by an auxiliary program - but can then be edited by hand if required. There are
4212 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4213 toupper() in the code.
4214
4215 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4216 make them global. Abolish the function for setting them, as the caller can now
4217 set them directly.
4218
4219
4220 Version 0.92 11-Sep-97
4221 ----------------------
4222
4223 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4224 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4225
4226 2. Caseless matching was not working in character classes if the characters in
4227 the pattern were in upper case.
4228
4229 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4230
4231 4. Make PCRE_ANCHORED public and accept as a compile option.
4232
4233 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4234 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4235 pass them.
4236
4237 6. Give an error if bad option bits passed at compile or run time.
4238
4239 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4240 pcretest to cause it to pass that flag.
4241
4242 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4243 options, and the first character, if set.
4244
4245 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4246
4247
4248 Version 0.91 10-Sep-97
4249 ----------------------
4250
4251 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4252 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4253
4254 2. PCRE was looping on encountering an indefinitely repeated back reference to
4255 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4256 Perl does - treats the match as successful.
4257
4258 ****

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

  ViewVC Help
Powered by ViewVC 1.1.5