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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 569 - (show annotations) (download)
Sun Nov 7 16:14:50 2010 UTC (3 years, 11 months ago) by ph10
File size: 194739 byte(s)
Add PCRE_ERROR_SHORTUTF8 to PCRE_PARTIAL_HARD processing.

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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12