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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 571 - (show annotations)
Tue Nov 16 17:51:37 2010 UTC (4 years, 7 months ago) by ph10
File size: 195143 byte(s)
Error occurred while calculating annotation data.
Rename --{in,ex}clude_dir with hyphen instead of underscore in pcregrep, but 
leave the old versions as undocumented synonyms. (In GNU grep, hyphens are 
used.)
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 17. Nobody had reported that the --include_dir option, which was added in
100 release 7.7 should have been called --include-dir (hyphen, not underscore)
101 for compatibility with GNU grep. I have changed it to --include-dir, but
102 left --include_dir as an undocumented synonym, and the same for
103 --exclude-dir, though that is not available in GNU grep, at least as of
104 release 2.5.4.
105
106
107 Version 8.10 25-Jun-2010
108 ------------------------
109
110 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
111 THEN.
112
113 2. (*ACCEPT) was not working when inside an atomic group.
114
115 3. Inside a character class, \B is treated as a literal by default, but
116 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
117 causes the error). The code is unchanged, but I tidied the documentation.
118
119 4. Inside a character class, PCRE always treated \R and \X as literals,
120 whereas Perl faults them if its -w option is set. I have changed PCRE so
121 that it faults them when PCRE_EXTRA is set.
122
123 5. Added support for \N, which always matches any character other than
124 newline. (It is the same as "." when PCRE_DOTALL is not set.)
125
126 6. When compiling pcregrep with newer versions of gcc which may have
127 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
128 declared with attribute warn_unused_result" were given. Just casting the
129 result to (void) does not stop the warnings; a more elaborate fudge is
130 needed. I've used a macro to implement this.
131
132 7. Minor change to pcretest.c to avoid a compiler warning.
133
134 8. Added four artifical Unicode properties to help with an option to make
135 \s etc use properties (see next item). The new properties are: Xan
136 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
137
138 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
139 use Unicode properties. (*UCP) at the start of a pattern can be used to set
140 this option. Modified pcretest to add /W to test this facility. Added
141 REG_UCP to make it available via the POSIX interface.
142
143 10. Added --line-buffered to pcregrep.
144
145 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
146 studied, and the match started with a letter with a code point greater than
147 127 whose first byte was different to the first byte of the other case of
148 the letter, the other case of this starting letter was not recognized
149 (#976).
150
151 12. If a pattern that was studied started with a repeated Unicode property
152 test, for example, \p{Nd}+, there was the theoretical possibility of
153 setting up an incorrect bitmap of starting bytes, but fortunately it could
154 not have actually happened in practice until change 8 above was made (it
155 added property types that matched character-matching opcodes).
156
157 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
158 possible starting bytes for non-anchored patterns.
159
160 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
161 \R, and also a number of cases that involve Unicode properties, both
162 explicit and implicit when PCRE_UCP is set.
163
164 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
165 input, it could crash or give wrong results if characters with values
166 greater than 0xc0 were present in the subject string. (Detail: it assumed
167 UTF-8 input when processing these items.)
168
169 16. Added a lot of (int) casts to avoid compiler warnings in systems where
170 size_t is 64-bit (#991).
171
172 17. Added a check for running out of memory when PCRE is compiled with
173 --disable-stack-for-recursion (#990).
174
175 18. If the last data line in a file for pcretest does not have a newline on
176 the end, a newline was missing in the output.
177
178 19. The default pcre_chartables.c file recognizes only ASCII characters (values
179 less than 128) in its various bitmaps. However, there is a facility for
180 generating tables according to the current locale when PCRE is compiled. It
181 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
182 characters, are recognized by isspace() and therefore were getting set in
183 these tables, and indeed these tables seem to approximate to ISO 8859. This
184 caused a problem in UTF-8 mode when pcre_study() was used to create a list
185 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
186 which of course cannot start UTF-8 characters. I have changed the code so
187 that only real ASCII characters (less than 128) and the correct starting
188 bytes for UTF-8 encodings are set for characters greater than 127 when in
189 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
190 altogether.)
191
192 20. Added the /T option to pcretest so as to be able to run tests with non-
193 standard character tables, thus making it possible to include the tests
194 used for 19 above in the standard set of tests.
195
196 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
197 reference to a subpattern the other side of a comment that contains an
198 opening parenthesis caused either an internal compiling error, or a
199 reference to the wrong subpattern.
200
201
202 Version 8.02 19-Mar-2010
203 ------------------------
204
205 1. The Unicode data tables have been updated to Unicode 5.2.0.
206
207 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
208 configured.
209
210 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
211 original author of that file, following a query about its status.
212
213 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
214 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
215
216 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
217 quantifier applied to a forward-referencing subroutine call, could compile
218 incorrect code or give the error "internal error: previously-checked
219 referenced subpattern not found".
220
221 6. Both MS Visual Studio and Symbian OS have problems with initializing
222 variables to point to external functions. For these systems, therefore,
223 pcre_malloc etc. are now initialized to local functions that call the
224 relevant global functions.
225
226 7. There were two entries missing in the vectors called coptable and poptable
227 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
228 I've fixed the data, and added a kludgy way of testing at compile time that
229 the lengths are correct (equal to the number of opcodes).
230
231 8. Following on from 7, I added a similar kludge to check the length of the
232 eint vector in pcreposix.c.
233
234 9. Error texts for pcre_compile() are held as one long string to avoid too
235 much relocation at load time. To find a text, the string is searched,
236 counting zeros. There was no check for running off the end of the string,
237 which could happen if a new error number was added without updating the
238 string.
239
240 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
241
242 11. \K was not working if it appeared in an atomic group or in a group that
243 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
244 \K is "not well defined" if used in an assertion. PCRE now accepts it if
245 the assertion is positive, but not if it is negative.
246
247 12. Change 11 fortuitously reduced the size of the stack frame used in the
248 "match()" function of pcre_exec.c by one pointer. Forthcoming
249 implementation of support for (*MARK) will need an extra pointer on the
250 stack; I have reserved it now, so that the stack frame size does not
251 decrease.
252
253 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
254 item in branch that calls a recursion is a subroutine call - as in the
255 second branch in the above example - was incorrectly given the compile-
256 time error "recursive call could loop indefinitely" because pcre_compile()
257 was not correctly checking the subroutine for matching a non-empty string.
258
259 14. The checks for overrunning compiling workspace could trigger after an
260 overrun had occurred. This is a "should never occur" error, but it can be
261 triggered by pathological patterns such as hundreds of nested parentheses.
262 The checks now trigger 100 bytes before the end of the workspace.
263
264 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
265
266
267 Version 8.01 19-Jan-2010
268 ------------------------
269
270 1. If a pattern contained a conditional subpattern with only one branch (in
271 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
272 computed the wrong minimum data length (which is of course zero for such
273 subpatterns). This could cause incorrect "no match" results.
274
275 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
276 the pattern is reset in the first branch, pcre_compile() failed with
277 "internal error: code overflow at offset...". This happened only when
278 the reset was to the original external option setting. (An optimization
279 abstracts leading options settings into an external setting, which was the
280 cause of this.)
281
282 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
283 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
284 assertion pattern did not match (meaning that the assertion was true), it
285 was incorrectly treated as false if the SKIP had been reached during the
286 matching. This also applied to assertions used as conditions.
287
288 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
289 assertion subpattern, including such a pattern used as a condition,
290 unpredictable results occurred, instead of the error return
291 PCRE_ERROR_DFA_UITEM.
292
293 5. The C++ GlobalReplace function was not working like Perl for the special
294 situation when an empty string is matched. It now does the fancy magic
295 stuff that is necessary.
296
297 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
298 removed. (These were left over from very, very early versions of PCRE.)
299
300 7. Some cosmetic changes to the code to make life easier when compiling it
301 as part of something else:
302
303 (a) Change DEBUG to PCRE_DEBUG.
304
305 (b) In pcre_compile(), rename the member of the "branch_chain" structure
306 called "current" as "current_branch", to prevent a collision with the
307 Linux macro when compiled as a kernel module.
308
309 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
310 prevent a collision with the Linux macro when compiled as a kernel
311 module.
312
313 8. In pcre_compile() there are some checks for integer overflows that used to
314 cast potentially large values to (double). This has been changed to that
315 when building, a check for int64_t is made, and if it is found, it is used
316 instead, thus avoiding the use of floating point arithmetic. (There is no
317 other use of FP in PCRE.) If int64_t is not found, the fallback is to
318 double.
319
320 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
321 2005 (difference between two addresses compared to an unsigned value).
322
323 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
324 custom one, because of the following reported problem in Windows:
325
326 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
327 under Win32.
328 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
329 therefore missing the function definition.
330 - The compiler thus generates a "C" signature for the test function.
331 - The linker fails to find the "C" function.
332 - PCRE fails to configure if asked to do so against libbz2.
333
334 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
335 messages were output:
336
337 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
338 rerunning libtoolize, to keep the correct libtool macros in-tree.
339 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
340
341 I have done both of these things.
342
343 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
344 most of the time, it *can* run out if it is given a pattern that contains a
345 runaway infinite recursion. I updated the discussion in the pcrestack man
346 page.
347
348 13. Now that we have gone to the x.xx style of version numbers, the minor
349 version may start with zero. Using 08 or 09 is a bad idea because users
350 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
351 interpreted as invalid octal numbers. I've updated the previous comment in
352 configure.ac, and also added a check that gives an error if 08 or 09 are
353 used.
354
355 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
356 causing partial matching to fail when the end of the subject matched \W
357 in a UTF-8 pattern where \W was quantified with a minimum of 3.
358
359 15. There were some discrepancies between the declarations in pcre_internal.h
360 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
361 their definitions. The declarations used "const uschar *" and the
362 definitions used USPTR. Even though USPTR is normally defined as "const
363 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
364 reported that: "This difference in casting confuses some C++ compilers, for
365 example, SunCC recognizes above declarations as different functions and
366 generates broken code for hbpcre." I have changed the declarations to use
367 USPTR.
368
369 16. GNU libtool is named differently on some systems. The autogen.sh script now
370 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
371 (FreeBSD).
372
373 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
374 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
375 comment: "Figure out how to create a longlong from a string: strtoll and
376 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
377 instance, but it only takes 2 args instead of 3!"
378
379 18. A subtle bug concerned with back references has been fixed by a change of
380 specification, with a corresponding code fix. A pattern such as
381 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
382 refers, was giving matches when it shouldn't. For example, xa=xaaa would
383 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
384 same bug. Such groups have to be quantified to be useful, or contained
385 inside another quantified group. (If there's no repetition, the reference
386 can never match.) The problem arises because, having left the group and
387 moved on to the rest of the pattern, a later failure that backtracks into
388 the group uses the captured value from the final iteration of the group
389 rather than the correct earlier one. I have fixed this in PCRE by forcing
390 any group that contains a reference to itself to be an atomic group; that
391 is, there cannot be any backtracking into it once it has completed. This is
392 similar to recursive and subroutine calls.
393
394
395 Version 8.00 19-Oct-09
396 ----------------------
397
398 1. The table for translating pcre_compile() error codes into POSIX error codes
399 was out-of-date, and there was no check on the pcre_compile() error code
400 being within the table. This could lead to an OK return being given in
401 error.
402
403 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
404 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
405 in a Windows environment.
406
407 3. The pcregrep --count option prints the count for each file even when it is
408 zero, as does GNU grep. However, pcregrep was also printing all files when
409 --files-with-matches was added. Now, when both options are given, it prints
410 counts only for those files that have at least one match. (GNU grep just
411 prints the file name in this circumstance, but including the count seems
412 more useful - otherwise, why use --count?) Also ensured that the
413 combination -clh just lists non-zero counts, with no names.
414
415 4. The long form of the pcregrep -F option was incorrectly implemented as
416 --fixed_strings instead of --fixed-strings. This is an incompatible change,
417 but it seems right to fix it, and I didn't think it was worth preserving
418 the old behaviour.
419
420 5. The command line items --regex=pattern and --regexp=pattern were not
421 recognized by pcregrep, which required --regex pattern or --regexp pattern
422 (with a space rather than an '='). The man page documented the '=' forms,
423 which are compatible with GNU grep; these now work.
424
425 6. No libpcreposix.pc file was created for pkg-config; there was just
426 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
427
428 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
429 when UCP support is not needed, by modifying the Python script that
430 generates it from Unicode data files. This should not matter if the module
431 is correctly used as a library, but I received one complaint about 50K of
432 unwanted data. My guess is that the person linked everything into his
433 program rather than using a library. Anyway, it does no harm.
434
435 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
436 was a minimum greater than 1 for a wide character in a possessive
437 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
438 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
439 character. Chaos in the form of incorrect output or a compiling loop could
440 result.
441
442 9. The restrictions on what a pattern can contain when partial matching is
443 requested for pcre_exec() have been removed. All patterns can now be
444 partially matched by this function. In addition, if there are at least two
445 slots in the offset vector, the offset of the earliest inspected character
446 for the match and the offset of the end of the subject are set in them when
447 PCRE_ERROR_PARTIAL is returned.
448
449 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
450 synonymous with PCRE_PARTIAL, for backwards compatibility, and
451 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
452 and may be more useful for multi-segment matching.
453
454 11. Partial matching with pcre_exec() is now more intuitive. A partial match
455 used to be given if ever the end of the subject was reached; now it is
456 given only if matching could not proceed because another character was
457 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
458 string "Z", which now yields "no match" instead of "partial match". In the
459 case of pcre_dfa_exec(), "no match" is given if every matching path for the
460 final character ended with (*FAIL).
461
462 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
463 if the pattern had a "must contain" character that was already found in the
464 earlier partial match, unless partial matching was again requested. For
465 example, with the pattern /dog.(body)?/, the "must contain" character is
466 "g". If the first part-match was for the string "dog", restarting with
467 "sbody" failed. This bug has been fixed.
468
469 13. The string returned by pcre_dfa_exec() after a partial match has been
470 changed so that it starts at the first inspected character rather than the
471 first character of the match. This makes a difference only if the pattern
472 starts with a lookbehind assertion or \b or \B (\K is not supported by
473 pcre_dfa_exec()). It's an incompatible change, but it makes the two
474 matching functions compatible, and I think it's the right thing to do.
475
476 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
477 so that the demonstration program is easily available in environments where
478 PCRE has not been installed from source.
479
480 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
481 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
482 library.
483
484 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
485 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
486 is not the first non-POSIX option to be added. Clearly some people find
487 these options useful.
488
489 17. If a caller to the POSIX matching function regexec() passes a non-zero
490 value for nmatch with a NULL value for pmatch, the value of
491 nmatch is forced to zero.
492
493 18. RunGrepTest did not have a test for the availability of the -u option of
494 the diff command, as RunTest does. It now checks in the same way as
495 RunTest, and also checks for the -b option.
496
497 19. If an odd number of negated classes containing just a single character
498 interposed, within parentheses, between a forward reference to a named
499 subpattern and the definition of the subpattern, compilation crashed with
500 an internal error, complaining that it could not find the referenced
501 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
502 [The bug was that it was starting one character too far in when skipping
503 over the character class, thus treating the ] as data rather than
504 terminating the class. This meant it could skip too much.]
505
506 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
507 /g option in pcretest when the pattern contains \K, which makes it possible
508 to have an empty string match not at the start, even when the pattern is
509 anchored. Updated pcretest and pcredemo to use this option.
510
511 21. If the maximum number of capturing subpatterns in a recursion was greater
512 than the maximum at the outer level, the higher number was returned, but
513 with unset values at the outer level. The correct (outer level) value is
514 now given.
515
516 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
517 PCRE did not set those parentheses (unlike Perl). I have now found a way to
518 make it do so. The string so far is captured, making this feature
519 compatible with Perl.
520
521 23. The tests have been re-organized, adding tests 11 and 12, to make it
522 possible to check the Perl 5.10 features against Perl 5.10.
523
524 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
525 pattern matches a fixed length string. PCRE did not allow this; now it
526 does. Neither allows recursion.
527
528 25. I finally figured out how to implement a request to provide the minimum
529 length of subject string that was needed in order to match a given pattern.
530 (It was back references and recursion that I had previously got hung up
531 on.) This code has now been added to pcre_study(); it finds a lower bound
532 to the length of subject needed. It is not necessarily the greatest lower
533 bound, but using it to avoid searching strings that are too short does give
534 some useful speed-ups. The value is available to calling programs via
535 pcre_fullinfo().
536
537 26. While implementing 25, I discovered to my embarrassment that pcretest had
538 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
539 study optimizations had never been tested with that matching function.
540 Oops. What is worse, even when it was passed study data, there was a bug in
541 pcre_dfa_exec() that meant it never actually used it. Double oops. There
542 were also very few tests of studied patterns with pcre_dfa_exec().
543
544 27. If (?| is used to create subpatterns with duplicate numbers, they are now
545 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
546 on the other side of the coin, they are no longer allowed to have different
547 names, because these cannot be distinguished in PCRE, and this has caused
548 confusion. (This is a difference from Perl.)
549
550 28. When duplicate subpattern names are present (necessarily with different
551 numbers, as required by 27 above), and a test is made by name in a
552 conditional pattern, either for a subpattern having been matched, or for
553 recursion in such a pattern, all the associated numbered subpatterns are
554 tested, and the overall condition is true if the condition is true for any
555 one of them. This is the way Perl works, and is also more like the way
556 testing by number works.
557
558
559 Version 7.9 11-Apr-09
560 ---------------------
561
562 1. When building with support for bzlib/zlib (pcregrep) and/or readline
563 (pcretest), all targets were linked against these libraries. This included
564 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
565 libraries. This caused unwanted dependencies to be created. This problem
566 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
567 pcretest is linked with readline.
568
569 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
570 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
571 moved outside it again, because FALSE and TRUE are already defined in AIX,
572 but BOOL is not.
573
574 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
575 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
576
577 4. The pcregrep documentation said spaces were inserted as well as colons (or
578 hyphens) following file names and line numbers when outputting matching
579 lines. This is not true; no spaces are inserted. I have also clarified the
580 wording for the --colour (or --color) option.
581
582 5. In pcregrep, when --colour was used with -o, the list of matching strings
583 was not coloured; this is different to GNU grep, so I have changed it to be
584 the same.
585
586 6. When --colo(u)r was used in pcregrep, only the first matching substring in
587 each matching line was coloured. Now it goes on to look for further matches
588 of any of the test patterns, which is the same behaviour as GNU grep.
589
590 7. A pattern that could match an empty string could cause pcregrep to loop; it
591 doesn't make sense to accept an empty string match in pcregrep, so I have
592 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
593 seems to be how GNU grep behaves.
594
595 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
596 start or after a newline", because the conditional assertion was not being
597 correctly handled. The rule now is that both the assertion and what follows
598 in the first alternative must satisfy the test.
599
600 9. If auto-callout was enabled in a pattern with a conditional group whose
601 condition was an assertion, PCRE could crash during matching, both with
602 pcre_exec() and pcre_dfa_exec().
603
604 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
605 used for matching.
606
607 11. Unicode property support in character classes was not working for
608 characters (bytes) greater than 127 when not in UTF-8 mode.
609
610 12. Added the -M command line option to pcretest.
611
612 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
613
614 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
615
616 16. Added comments and documentation about mis-use of no_arg in the C++
617 wrapper.
618
619 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
620 from Martin Jerabek that uses macro names for all relevant character and
621 string constants.
622
623 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
624 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
625 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
626 these, but not everybody uses configure.
627
628 19. A conditional group that had only one branch was not being correctly
629 recognized as an item that could match an empty string. This meant that an
630 enclosing group might also not be so recognized, causing infinite looping
631 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
632 with the subject "ab", where knowledge that the repeated group can match
633 nothing is needed in order to break the loop.
634
635 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
636 exec(), but without supplying a callout function, matching went wrong.
637
638 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
639 leak if the size of the offset vector was greater than 30. When the vector
640 is smaller, the saved offsets during recursion go onto a local stack
641 vector, but for larger vectors malloc() is used. It was failing to free
642 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
643 error, in fact).
644
645 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
646 heapframe that is used only when UTF-8 support is enabled. This caused no
647 problem, but was untidy.
648
649 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
650 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
651 included within another project.
652
653 24. Steven Van Ingelgem's patches to add more options to the CMake support,
654 slightly modified by me:
655
656 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
657 not building pcregrep.
658
659 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
660 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
661
662 25. Forward references, both numeric and by name, in patterns that made use of
663 duplicate group numbers, could behave incorrectly or give incorrect errors,
664 because when scanning forward to find the reference group, PCRE was not
665 taking into account the duplicate group numbers. A pattern such as
666 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
667
668 26. Changed a few more instances of "const unsigned char *" to USPTR, making
669 the feature of a custom pointer more persuasive (as requested by a user).
670
671 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
672 pcretest.c, inside #ifndefs, because it seems they are sometimes already
673 pre-defined.
674
675 28. Added support for (*UTF8) at the start of a pattern.
676
677 29. Arrange for flags added by the "release type" setting in CMake to be shown
678 in the configuration summary.
679
680
681 Version 7.8 05-Sep-08
682 ---------------------
683
684 1. Replaced UCP searching code with optimized version as implemented for Ad
685 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
686 stage table and inline lookup instead of a function, giving speed ups of 2
687 to 5 times on some simple patterns that I tested. Permission was given to
688 distribute the MultiStage2.py script that generates the tables (it's not in
689 the tarball, but is in the Subversion repository).
690
691 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
692 scripts.
693
694 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
695 a group with a zero qualifier. The result of the study could be incorrect,
696 or the function might crash, depending on the pattern.
697
698 4. Caseless matching was not working for non-ASCII characters in back
699 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
700 It now works when Unicode Property Support is available.
701
702 5. In pcretest, an escape such as \x{de} in the data was always generating
703 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
704 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
705 truncation.
706
707 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
708
709 7. Added two (int) casts to pcregrep when printing the difference of two
710 pointers, in case they are 64-bit values.
711
712 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
713 test 2 if it fails.
714
715 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
716 and a #define of that name to empty if it is not externally set. This is to
717 allow users of MSVC to set it if necessary.
718
719 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
720 the convenience functions in the pcre_get.c source file.
721
722 11. An option change at the start of a pattern that had top-level alternatives
723 could cause overwriting and/or a crash. This command provoked a crash in
724 some environments:
725
726 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
727
728 This potential security problem was recorded as CVE-2008-2371.
729
730 12. For a pattern where the match had to start at the beginning or immediately
731 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
732 pcre_dfa_exec() could read past the end of the passed subject if there was
733 no match. To help with detecting such bugs (e.g. with valgrind), I modified
734 pcretest so that it places the subject at the end of its malloc-ed buffer.
735
736 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
737 exec() might read past the end of the data buffer in UTF-8 mode.
738
739 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
740 the data contained the byte 0x85 as part of a UTF-8 character within its
741 first line. This applied both to normal and DFA matching.
742
743 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
744 /^[^d]*?$/8 failed to match "abc".
745
746 16. Added a missing copyright notice to pcrecpp_internal.h.
747
748 17. Make it more clear in the documentation that values returned from
749 pcre_exec() in ovector are byte offsets, not character counts.
750
751 18. Tidied a few places to stop certain compilers from issuing warnings.
752
753 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
754 supplied by Stefan Weber. I made a further small update for 7.8 because
755 there is a change of source arrangements: the pcre_searchfuncs.c module is
756 replaced by pcre_ucd.c.
757
758
759 Version 7.7 07-May-08
760 ---------------------
761
762 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
763 a string to a long long, pretend we don't even have a long long." This is
764 done by checking for the strtoq, strtoll, and _strtoi64 functions.
765
766 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
767 pre-7.6 versions, which defined a global no_arg variable instead of putting
768 it in the RE class. (See also #8 below.)
769
770 3. Remove a line of dead code, identified by coverity and reported by Nuno
771 Lopes.
772
773 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
774
775 (1) The include/exclude patterns were being applied to the whole pathnames
776 of files, instead of just to the final components.
777
778 (2) If there was more than one level of directory, the subdirectories were
779 skipped unless they satisfied the include/exclude conditions. This is
780 inconsistent with GNU grep (and could even be seen as contrary to the
781 pcregrep specification - which I improved to make it absolutely clear).
782 The action now is always to scan all levels of directory, and just
783 apply the include/exclude patterns to regular files.
784
785 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
786 --exclude_dir in the tests to avoid scanning .svn directories.
787
788 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
789 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
790 doesn't support NULs in patterns.
791
792 7. Added some missing "const"s to declarations of static tables in
793 pcre_compile.c and pcre_dfa_exec.c.
794
795 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
796 caused by fix #2 above. (Subsequently also a second patch to fix the
797 first patch. And a third patch - this was a messy problem.)
798
799 9. Applied Craig's patch to remove the use of push_back().
800
801 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
802 matching function regexec().
803
804 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
805 which, however, unlike Perl's \g{...}, are subroutine calls, not back
806 references. PCRE supports relative numbers with this syntax (I don't think
807 Oniguruma does).
808
809 12. Previously, a group with a zero repeat such as (...){0} was completely
810 omitted from the compiled regex. However, this means that if the group
811 was called as a subroutine from elsewhere in the pattern, things went wrong
812 (an internal error was given). Such groups are now left in the compiled
813 pattern, with a new opcode that causes them to be skipped at execution
814 time.
815
816 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
817 to the way PCRE behaves:
818
819 (a) A lone ] character is dis-allowed (Perl treats it as data).
820
821 (b) A back reference to an unmatched subpattern matches an empty string
822 (Perl fails the current match path).
823
824 (c) A data ] in a character class must be notated as \] because if the
825 first data character in a class is ], it defines an empty class. (In
826 Perl it is not possible to have an empty class.) The empty class []
827 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
828 The negative empty class [^] matches any one character, independently
829 of the DOTALL setting.
830
831 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
832 non-existent subpattern following a character class starting with ']' and
833 containing () gave an internal compiling error instead of "reference to
834 non-existent subpattern". Fortunately, when the pattern did exist, the
835 compiled code was correct. (When scanning forwards to check for the
836 existencd of the subpattern, it was treating the data ']' as terminating
837 the class, so got the count wrong. When actually compiling, the reference
838 was subsequently set up correctly.)
839
840 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
841 it was being rejected as not supported by pcre_dfa_exec(), even though
842 other assertions are supported. I have made pcre_dfa_exec() support
843 (*FAIL).
844
845 16. The implementation of 13c above involved the invention of a new opcode,
846 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
847 cannot be changed at match time, I realized I could make a small
848 improvement to matching performance by compiling OP_ALLANY instead of
849 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
850 on the OP_ANY path.
851
852 17. Compiling pcretest on Windows with readline support failed without the
853 following two fixes: (1) Make the unistd.h include conditional on
854 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
855
856 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
857 ncurses library to be included for pcretest when ReadLine support is
858 requested, but also to allow for it to be overridden. This patch came from
859 Daniel Bergström.
860
861 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
862 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
863 any errors with the current Unicode tables. Thanks to Peter Kankowski for
864 spotting this.
865
866
867 Version 7.6 28-Jan-08
868 ---------------------
869
870 1. A character class containing a very large number of characters with
871 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
872 overflow.
873
874 2. Patch to cut out the "long long" test in pcrecpp_unittest when
875 HAVE_LONG_LONG is not defined.
876
877 3. Applied Christian Ehrlicher's patch to update the CMake build files to
878 bring them up to date and include new features. This patch includes:
879
880 - Fixed PH's badly added libz and libbz2 support.
881 - Fixed a problem with static linking.
882 - Added pcredemo. [But later removed - see 7 below.]
883 - Fixed dftables problem and added an option.
884 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
885 HAVE_LONG_LONG.
886 - Added readline support for pcretest.
887 - Added an listing of the option settings after cmake has run.
888
889 4. A user submitted a patch to Makefile that makes it easy to create
890 "pcre.dll" under mingw when using Configure/Make. I added stuff to
891 Makefile.am that cause it to include this special target, without
892 affecting anything else. Note that the same mingw target plus all
893 the other distribution libraries and programs are now supported
894 when configuring with CMake (see 6 below) instead of with
895 Configure/Make.
896
897 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
898 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
899 exported in the Windows port". It has not yet been confirmed that the patch
900 solves the problem, but it does no harm.
901
902 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
903 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
904 with CMake, and also correct the comment about stack recursion.
905
906 7. Remove the automatic building of pcredemo from the ./configure system and
907 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
908 of a program that users should build themselves after PCRE is installed, so
909 building it automatically is not really right. What is more, it gave
910 trouble in some build environments.
911
912 8. Further tidies to CMakeLists.txt from Sheri and Christian.
913
914
915 Version 7.5 10-Jan-08
916 ---------------------
917
918 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
919 values in parens when parsing an RE using the C++ wrapper."
920
921 2. Negative specials like \S did not work in character classes in UTF-8 mode.
922 Characters greater than 255 were excluded from the class instead of being
923 included.
924
925 3. The same bug as (2) above applied to negated POSIX classes such as
926 [:^space:].
927
928 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
929 defined or documented. It seems to have been a typo for PCRE_STATIC, so
930 I have changed it.
931
932 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
933 first named subpattern) and a construct such as (?&a) would reference the
934 first named subpattern whose name started with "a" (in other words, the
935 length check was missing). Both these problems are fixed. "Subpattern name
936 expected" is now given for (?&) (a zero-length name), and this patch also
937 makes it give the same error for \k'' (previously it complained that that
938 was a reference to a non-existent subpattern).
939
940 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
941 this is right because (?- can be followed by option settings as well as by
942 digits. I have, however, made the messages clearer.
943
944 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
945 than the number used in the conditional) now cause a compile-time error.
946 This is actually not compatible with Perl, which accepts such patterns, but
947 treats the conditional as always being FALSE (as PCRE used to), but it
948 seems to me that giving a diagnostic is better.
949
950 8. Change "alphameric" to the more common word "alphanumeric" in comments
951 and messages.
952
953 9. Fix two occurrences of "backslash" in comments that should have been
954 "backspace".
955
956 10. Remove two redundant lines of code that can never be obeyed (their function
957 was moved elsewhere).
958
959 11. The program that makes PCRE's Unicode character property table had a bug
960 which caused it to generate incorrect table entries for sequences of
961 characters that have the same character type, but are in different scripts.
962 It amalgamated them into a single range, with the script of the first of
963 them. In other words, some characters were in the wrong script. There were
964 thirteen such cases, affecting characters in the following ranges:
965
966 U+002b0 - U+002c1
967 U+0060c - U+0060d
968 U+0061e - U+00612
969 U+0064b - U+0065e
970 U+0074d - U+0076d
971 U+01800 - U+01805
972 U+01d00 - U+01d77
973 U+01d9b - U+01dbf
974 U+0200b - U+0200f
975 U+030fc - U+030fe
976 U+03260 - U+0327f
977 U+0fb46 - U+0fbb1
978 U+10450 - U+1049d
979
980 12. The -o option (show only the matching part of a line) for pcregrep was not
981 compatible with GNU grep in that, if there was more than one match in a
982 line, it showed only the first of them. It now behaves in the same way as
983 GNU grep.
984
985 13. If the -o and -v options were combined for pcregrep, it printed a blank
986 line for every non-matching line. GNU grep prints nothing, and pcregrep now
987 does the same. The return code can be used to tell if there were any
988 non-matching lines.
989
990 14. Added --file-offsets and --line-offsets to pcregrep.
991
992 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
993 infinitely looping recursion. The bug was that positive lookaheads were not
994 being skipped when checking for a possible empty match (negative lookaheads
995 and both kinds of lookbehind were skipped).
996
997 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
998 inclusion of <windows.h> to before rather than after the definition of
999 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1000
1001 17. Specifying a possessive quantifier with a specific limit for a Unicode
1002 character property caused pcre_compile() to compile bad code, which led at
1003 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1004 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1005 caused the error; without that there was no problem.
1006
1007 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1008
1009 19. Added --enable-pcretest-libreadline.
1010
1011 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1012 RE::GlobalReplace(). As a result, the number of replacements returned was
1013 double what it should be. I removed one of the increments, but Craig sent a
1014 later patch that removed the other one (the right fix) and added unit tests
1015 that check the return values (which was not done before).
1016
1017 21. Several CMake things:
1018
1019 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1020 the names libpcre and libpcreposix, not just pcre and pcreposix.
1021
1022 (2) The above change means that pcretest and pcregrep are now correctly
1023 linked with the newly-built libraries, not previously installed ones.
1024
1025 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1026
1027 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1028 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1029 UTF-8 newline character). The key issue is that the pattern starts .*;
1030 this means that the match must be either at the beginning, or after a
1031 newline. The bug was in the code for advancing after a failed match and
1032 checking that the new position followed a newline. It was not taking
1033 account of UTF-8 characters correctly.
1034
1035 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1036 character classes. PCRE was not treating the sequence [:...:] as a
1037 character class unless the ... were all letters. Perl, however, seems to
1038 allow any characters between [: and :], though of course it rejects as
1039 unknown any "names" that contain non-letters, because all the known class
1040 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1041 for example, whereas PCRE did not - it did not recognize a POSIX character
1042 class. This seemed a bit dangerous, so the code has been changed to be
1043 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1044 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1045 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1046 Perl does, and where it didn't before.
1047
1048 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1049 Windows environments %n is disabled by default.
1050
1051
1052 Version 7.4 21-Sep-07
1053 ---------------------
1054
1055 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1056 means that a class such as [\s] counted as "explicit reference to CR or
1057 LF". That isn't really right - the whole point of the change was to try to
1058 help when there was an actual mention of one of the two characters. So now
1059 the change happens only if \r or \n (or a literal CR or LF) character is
1060 encountered.
1061
1062 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1063 of both had grown to the point where there were only 3 bits left.
1064 Fortunately, there was spare space in the data structure, and so I have
1065 moved the internal flags into a new 16-bit field to free up more option
1066 bits.
1067
1068 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1069 but did not set the internal JCHANGED flag - either of these is enough to
1070 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1071 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1072 start sets both bits.
1073
1074 4. Added options (at build time, compile time, exec time) to change \R from
1075 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1076
1077 5. doc/pcresyntax.html was missing from the distribution.
1078
1079 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1080 compatibility, even though it is no longer used.
1081
1082 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1083 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1084 windows.h file is present (where different names are used). [This was
1085 reversed later after testing - see 16 below.]
1086
1087 8. Changed all #include <config.h> to #include "config.h". There were also
1088 some further <pcre.h> cases that I changed to "pcre.h".
1089
1090 9. When pcregrep was used with the --colour option, it missed the line ending
1091 sequence off the lines that it output.
1092
1093 10. It was pointed out to me that arrays of string pointers cause lots of
1094 relocations when a shared library is dynamically loaded. A technique of
1095 using a single long string with a table of offsets can drastically reduce
1096 these. I have refactored PCRE in four places to do this. The result is
1097 dramatic:
1098
1099 Originally: 290
1100 After changing UCP table: 187
1101 After changing error message table: 43
1102 After changing table of "verbs" 36
1103 After changing table of Posix names 22
1104
1105 Thanks to the folks working on Gregex for glib for this insight.
1106
1107 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1108 unicode-properties was also set.
1109
1110 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1111
1112 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1113 checked only for CRLF.
1114
1115 14. Added casts to pcretest.c to avoid compiler warnings.
1116
1117 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1118
1119 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1120 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1121 entirely. This removes changes made in 7 above.
1122
1123 17. The CMake files have been updated, and there is now more information about
1124 building with CMake in the NON-UNIX-USE document.
1125
1126
1127 Version 7.3 28-Aug-07
1128 ---------------------
1129
1130 1. In the rejigging of the build system that eventually resulted in 7.1, the
1131 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1132 brackets there is not right, since it causes compilers to look for an
1133 installed pcre.h, not the version that is in the source that is being
1134 compiled (which of course may be different). I have changed it back to:
1135
1136 #include "pcre.h"
1137
1138 I have a vague recollection that the change was concerned with compiling in
1139 different directories, but in the new build system, that is taken care of
1140 by the VPATH setting the Makefile.
1141
1142 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1143 when the subject happened to end in the byte 0x85 (e.g. if the last
1144 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1145 characters but of course it shouldn't be taken as a newline when it is part
1146 of another character. The bug was that, for an unlimited repeat of . in
1147 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1148 characters when looking for a newline.
1149
1150 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1151
1152 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1153 in debug output.
1154
1155 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1156 long printing in the pcrecpp unittest when running under MinGW.
1157
1158 6. ESC_K was left out of the EBCDIC table.
1159
1160 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1161 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1162 limit also applies to "virtual nesting" when a pattern is recursive, and in
1163 this case 1000 isn't so big. I have been able to remove this limit at the
1164 expense of backing off one optimization in certain circumstances. Normally,
1165 when pcre_exec() would call its internal match() function recursively and
1166 immediately return the result unconditionally, it uses a "tail recursion"
1167 feature to save stack. However, when a subpattern that can match an empty
1168 string has an unlimited repetition quantifier, it no longer makes this
1169 optimization. That gives it a stack frame in which to save the data for
1170 checking that an empty string has been matched. Previously this was taken
1171 from the 1000-entry workspace that had been reserved. So now there is no
1172 explicit limit, but more stack is used.
1173
1174 8. Applied Daniel's patches to solve problems with the import/export magic
1175 syntax that is required for Windows, and which was going wrong for the
1176 pcreposix and pcrecpp parts of the library. These were overlooked when this
1177 problem was solved for the main library.
1178
1179 9. There were some crude static tests to avoid integer overflow when computing
1180 the size of patterns that contain repeated groups with explicit upper
1181 limits. As the maximum quantifier is 65535, the maximum group length was
1182 set at 30,000 so that the product of these two numbers did not overflow a
1183 32-bit integer. However, it turns out that people want to use groups that
1184 are longer than 30,000 bytes (though not repeat them that many times).
1185 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1186 made it possible to implement the integer overflow checks in a much more
1187 dynamic way, which I have now done. The artificial limitation on group
1188 length has been removed - we now have only the limit on the total length of
1189 the compiled pattern, which depends on the LINK_SIZE setting.
1190
1191 10. Fixed a bug in the documentation for get/copy named substring when
1192 duplicate names are permitted. If none of the named substrings are set, the
1193 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1194 empty string.
1195
1196 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1197 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1198 because the ] is interpreted as the first data character and the
1199 terminating ] is not found. PCRE has been made compatible with Perl in this
1200 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1201 cause memory overwriting.
1202
1203 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1204 string has been matched (to stop an infinite loop). It was not recognizing
1205 a conditional subpattern that could match an empty string if that
1206 subpattern was within another subpattern. For example, it looped when
1207 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1208 condition was not nested. This bug has been fixed.
1209
1210 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1211 past the start of the subject in the presence of bytes with the top bit
1212 set, for example "\x8aBCD".
1213
1214 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1215 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1216
1217 14. Optimized (?!) to (*FAIL).
1218
1219 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1220 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1221 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1222 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1223 does: it's just the validity check that is more restrictive.
1224
1225 16. Inserted checks for integer overflows during escape sequence (backslash)
1226 processing, and also fixed erroneous offset values for syntax errors during
1227 backslash processing.
1228
1229 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1230 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1231
1232 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1233 caused an overrun.
1234
1235 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1236 something other than just ASCII characters) inside a group that had an
1237 unlimited repeat caused a loop at compile time (while checking to see
1238 whether the group could match an empty string).
1239
1240 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1241 [\P{Any}] did so. (Error in the code for printing property names.)
1242
1243 21. An orphan \E inside a character class could cause a crash.
1244
1245 22. A repeated capturing bracket such as (A)? could cause a wild memory
1246 reference during compilation.
1247
1248 23. There are several functions in pcre_compile() that scan along a compiled
1249 expression for various reasons (e.g. to see if it's fixed length for look
1250 behind). There were bugs in these functions when a repeated \p or \P was
1251 present in the pattern. These operators have additional parameters compared
1252 with \d, etc, and these were not being taken into account when moving along
1253 the compiled data. Specifically:
1254
1255 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1256 length.
1257
1258 (b) An item such as \pL+ within a repeated group could cause crashes or
1259 loops.
1260
1261 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1262 "reference to non-existent subpattern" error.
1263
1264 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1265
1266 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1267 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1268
1269 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1270
1271 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1272 character were causing crashes (broken optimization).
1273
1274 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1275 \p or \P) caused a compile-time loop.
1276
1277 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1278 break. For example, the unstudied pattern [\r\n]A does not match the string
1279 "\r\nA" because change 7.0/46 below moves the current point on by two
1280 characters after failing to match at the start. However, the pattern \nA
1281 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1282 the same is true. There doesn't seem any very clean way out of this, but
1283 what I have chosen to do makes the common cases work: PCRE now takes note
1284 of whether there can be an explicit match for \r or \n anywhere in the
1285 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1286 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1287 pattern has explicit CR or LF references.
1288
1289 29. Added (*CR) etc for changing newline setting at start of pattern.
1290
1291
1292 Version 7.2 19-Jun-07
1293 ---------------------
1294
1295 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1296 which is apparently normally available under Windows.
1297
1298 2. Re-jig the pcregrep tests with different newline settings in an attempt
1299 to make them independent of the local environment's newline setting.
1300
1301 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1302
1303 4. Some of the "internals" tests were previously cut out when the link size
1304 was not 2, because the output contained actual offsets. The recent new
1305 "Z" feature of pcretest means that these can be cut out, making the tests
1306 usable with all link sizes.
1307
1308 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1309 stack recursion. This gives a massive performance boost under BSD, but just
1310 a small improvement under Linux. However, it saves one field in the frame
1311 in all cases.
1312
1313 6. Added more features from the forthcoming Perl 5.10:
1314
1315 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1316 recursion call. It refers to the nth most recently opened parentheses.
1317
1318 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1319 to be opened parentheses.
1320
1321 (c) Conditions that refer to capturing parentheses can be specified
1322 relatively, for example, (?(-2)... or (?(+3)...
1323
1324 (d) \K resets the start of the current match so that everything before
1325 is not part of it.
1326
1327 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1328
1329 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1330 reference syntax.
1331
1332 (g) (?| introduces a group in which the numbering of parentheses in each
1333 alternative starts with the same number.
1334
1335 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1336
1337 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1338 PCRE_INFO_JCHANGED.
1339
1340 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1341 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1342 for detecting groups that can match an empty string.
1343
1344 9. A pattern with a very large number of alternatives (more than several
1345 hundred) was running out of internal workspace during the pre-compile
1346 phase, where pcre_compile() figures out how much memory will be needed. A
1347 bit of new cunning has reduced the workspace needed for groups with
1348 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1349 workspace instead of running out of the 4096 that are available.
1350
1351 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1352
1353 11. Applied patch from Google to remove an optimization that didn't quite work.
1354 The report of the bug said:
1355
1356 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1357 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1358 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1359
1360 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1361 it matched the wrong number of bytes.
1362
1363
1364 Version 7.1 24-Apr-07
1365 ---------------------
1366
1367 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1368 that is more "standard", making use of automake and other Autotools. There
1369 is some re-arrangement of the files and adjustment of comments consequent
1370 on this.
1371
1372 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1373 for recursive directory scanning broke on some systems because the files
1374 are not scanned in any specific order and on different systems the order
1375 was different. A call to "sort" has been inserted into RunGrepTest for the
1376 approprate test as a short-term fix. In the longer term there may be an
1377 alternative.
1378
1379 3. I had an email from Eric Raymond about problems translating some of PCRE's
1380 man pages to HTML (despite the fact that I distribute HTML pages, some
1381 people do their own conversions for various reasons). The problems
1382 concerned the use of low-level troff macros .br and .in. I have therefore
1383 removed all such uses from the man pages (some were redundant, some could
1384 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1385 HTML has been updated to handle .nf/.fi and to complain if it encounters
1386 .br or .in.
1387
1388 4. Updated comments in configure.ac that get placed in config.h.in and also
1389 arranged for config.h to be included in the distribution, with the name
1390 config.h.generic, for the benefit of those who have to compile without
1391 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1392
1393 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1394 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1395 makevp.bat for the current PCRE, using the additional files
1396 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1397
1398 6. A Windows user reported a minor discrepancy with test 2, which turned out
1399 to be caused by a trailing space on an input line that had got lost in his
1400 copy. The trailing space was an accident, so I've just removed it.
1401
1402 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1403 that is needed.
1404
1405 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1406 as "const" (a) because they are and (b) because it helps the PHP
1407 maintainers who have recently made a script to detect big data structures
1408 in the php code that should be moved to the .rodata section. I remembered
1409 to update Builducptable as well, so it won't revert if ucptable.h is ever
1410 re-created.
1411
1412 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1413 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1414 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1415 support is not required. This saves 1.5-2K of code, which is important in
1416 some applications.
1417
1418 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1419 so as not to refer to the tables, even though these functions will never be
1420 called when UTF-8 support is disabled. Otherwise there are problems with a
1421 shared library.
1422
1423 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1424
1425 (a) It was defining its arguments as char * instead of void *.
1426
1427 (b) It was assuming that all moves were upwards in memory; this was true
1428 a long time ago when I wrote it, but is no longer the case.
1429
1430 The emulated memove() is provided for those environments that have neither
1431 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1432 is clearly not the case, as these two bugs were recently reported.
1433
1434 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1435 and Detrail to create the HTML documentation, the .txt form of the man
1436 pages, and it removes trailing spaces from listed files. It also creates
1437 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1438 case, it wraps all the #defines with #ifndefs. This script should be run
1439 before "make dist".
1440
1441 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1442 with Unicode property support.
1443
1444 (a) For a maximizing quantifier, if the two different cases of the
1445 character were of different lengths in their UTF-8 codings (there are
1446 some cases like this - I found 11), and the matching function had to
1447 back up over a mixture of the two cases, it incorrectly assumed they
1448 were both the same length.
1449
1450 (b) When PCRE was configured to use the heap rather than the stack for
1451 recursion during matching, it was not correctly preserving the data for
1452 the other case of a UTF-8 character when checking ahead for a match
1453 while processing a minimizing repeat. If the check also involved
1454 matching a wide character, but failed, corruption could cause an
1455 erroneous result when trying to check for a repeat of the original
1456 character.
1457
1458 13. Some tidying changes to the testing mechanism:
1459
1460 (a) The RunTest script now detects the internal link size and whether there
1461 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1462 values substituted by "configure". (The RunGrepTest script already did
1463 this for UTF-8.) The configure.ac script no longer substitutes the
1464 relevant variables.
1465
1466 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1467 with length and offset values. This means that the output is different
1468 for different internal link sizes. Test 2 is skipped for link sizes
1469 other than 2 because of this, bypassing the problem. Unfortunately,
1470 there was also a test in test 3 (the locale tests) that used /B and
1471 failed for link sizes other than 2. Rather than cut the whole test out,
1472 I have added a new /Z option to pcretest that replaces the length and
1473 offset values with spaces. This is now used to make test 3 independent
1474 of link size. (Test 2 will be tidied up later.)
1475
1476 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1477 segmentation fault instead of returning the appropriate error message.
1478
1479 15. In multiline mode when the newline sequence was set to "any", the pattern
1480 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1481 This doesn't seem right; it now treats the CRLF combination as the line
1482 ending, and so does not match in that case. It's only a pattern such as ^$
1483 that would hit this one: something like ^ABC$ would have failed after \r
1484 and then tried again after \r\n.
1485
1486 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1487 in an attempt to make files that differ only in their line terminators
1488 compare equal. This works on Linux.
1489
1490 17. Under certain error circumstances pcregrep might try to free random memory
1491 as it exited. This is now fixed, thanks to valgrind.
1492
1493 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1494 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1495 was because its rules for how to advance for /g after matching an empty
1496 string at the end of a line did not allow for this case. They now check for
1497 it specially.
1498
1499 20. pcretest is supposed to handle patterns and data of any length, by
1500 extending its buffers when necessary. It was getting this wrong when the
1501 buffer for a data line had to be extended.
1502
1503 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1504 CRLF as a newline sequence.
1505
1506 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1507 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1508 I have nevertheless tidied it up.
1509
1510 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1511
1512 24. Added a man page for pcre-config.
1513
1514
1515 Version 7.0 19-Dec-06
1516 ---------------------
1517
1518 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1519 moving to gcc 4.1.1.
1520
1521 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1522 sys/time.h, which is documented as needed for this function. It doesn't
1523 seem to matter on Linux, but it showed up on some releases of OS X.
1524
1525 3. It seems that there are systems where bytes whose values are greater than
1526 127 match isprint() in the "C" locale. The "C" locale should be the
1527 default when a C program starts up. In most systems, only ASCII printing
1528 characters match isprint(). This difference caused the output from pcretest
1529 to vary, making some of the tests fail. I have changed pcretest so that:
1530
1531 (a) When it is outputting text in the compiled version of a pattern, bytes
1532 other than 32-126 are always shown as hex escapes.
1533
1534 (b) When it is outputting text that is a matched part of a subject string,
1535 it does the same, unless a different locale has been set for the match
1536 (using the /L modifier). In this case, it uses isprint() to decide.
1537
1538 4. Fixed a major bug that caused incorrect computation of the amount of memory
1539 required for a compiled pattern when options that changed within the
1540 pattern affected the logic of the preliminary scan that determines the
1541 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1542 that the computed length was too small. The symptoms of this bug were
1543 either the PCRE error "internal error: code overflow" from pcre_compile(),
1544 or a glibc crash with a message such as "pcretest: free(): invalid next
1545 size (fast)". Examples of patterns that provoked this bug (shown in
1546 pcretest format) are:
1547
1548 /(?-x: )/x
1549 /(?x)(?-x: \s*#\s*)/
1550 /((?i)[\x{c0}])/8
1551 /(?i:[\x{c0}])/8
1552
1553 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1554 is now done differently.
1555
1556 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1557 wrapper classes; (b) implement a new function in the C++ scanner that is
1558 more efficient than the old way of doing things because it avoids levels of
1559 recursion in the regex matching; (c) add a paragraph to the documentation
1560 for the FullMatch() function.
1561
1562 6. The escape sequence \n was being treated as whatever was defined as
1563 "newline". Not only was this contrary to the documentation, which states
1564 that \n is character 10 (hex 0A), but it also went horribly wrong when
1565 "newline" was defined as CRLF. This has been fixed.
1566
1567 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1568 was being set to -1 for the "end of line" case (supposedly a value that no
1569 character can have). Though this value is never used (the check for end of
1570 line is "zero bytes in current character"), it caused compiler complaints.
1571 I've changed it to 0xffffffff.
1572
1573 8. In pcre_version.c, the version string was being built by a sequence of
1574 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1575 string (as it is for production releases) called a macro with an empty
1576 argument. The C standard says the result of this is undefined. The gcc
1577 compiler treats it as an empty string (which was what was wanted) but it is
1578 reported that Visual C gives an error. The source has been hacked around to
1579 avoid this problem.
1580
1581 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1582 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1583 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1584 of them did).
1585
1586 10. Originally, pcretest opened its input and output without "b"; then I was
1587 told that "b" was needed in some environments, so it was added for release
1588 5.0 to both the input and output. (It makes no difference on Unix-like
1589 systems.) Later I was told that it is wrong for the input on Windows. I've
1590 now abstracted the modes into two macros, to make it easier to fiddle with
1591 them, and removed "b" from the input mode under Windows.
1592
1593 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1594
1595 12. Added -help and --help to pcretest as an official way of being reminded
1596 of the options.
1597
1598 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1599 and pcrecpp.cc because they annoy compilers at high warning levels.
1600
1601 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1602
1603 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1604 scripts are not C programs :-) and which was not noticed because it works
1605 on Linux.
1606
1607 16. pcretest is supposed to handle any length of pattern and data line (as one
1608 line or as a continued sequence of lines) by extending its input buffer if
1609 necessary. This feature was broken for very long pattern lines, leading to
1610 a string of junk being passed to pcre_compile() if the pattern was longer
1611 than about 50K.
1612
1613 17. I have done a major re-factoring of the way pcre_compile() computes the
1614 amount of memory needed for a compiled pattern. Previously, there was code
1615 that made a preliminary scan of the pattern in order to do this. That was
1616 OK when PCRE was new, but as the facilities have expanded, it has become
1617 harder and harder to keep it in step with the real compile phase, and there
1618 have been a number of bugs (see for example, 4 above). I have now found a
1619 cunning way of running the real compile function in a "fake" mode that
1620 enables it to compute how much memory it would need, while actually only
1621 ever using a few hundred bytes of working memory and without too many
1622 tests of the mode. This should make future maintenance and development
1623 easier. A side effect of this work is that the limit of 200 on the nesting
1624 depth of parentheses has been removed (though this was never a serious
1625 limitation, I suspect). However, there is a downside: pcre_compile() now
1626 runs more slowly than before (30% or more, depending on the pattern). I
1627 hope this isn't a big issue. There is no effect on runtime performance.
1628
1629 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1630 newline (only possible for the last line of a file) and it was a
1631 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1632
1633 19. Added additional timing features to pcretest. (1) The -tm option now times
1634 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1635 separate command line item, by a number that specifies the number of
1636 repeats to use when timing. The default is 50000; this gives better
1637 precision, but takes uncomfortably long for very large patterns.
1638
1639 20. Extended pcre_study() to be more clever in cases where a branch of a
1640 subpattern has no definite first character. For example, (a*|b*)[cd] would
1641 previously give no result from pcre_study(). Now it recognizes that the
1642 first character must be a, b, c, or d.
1643
1644 21. There was an incorrect error "recursive call could loop indefinitely" if
1645 a subpattern (or the entire pattern) that was being tested for matching an
1646 empty string contained only one non-empty item after a nested subpattern.
1647 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1648 incorrectly, because the \d was being skipped in the check.
1649
1650 22. The pcretest program now has a new pattern option /B and a command line
1651 option -b, which is equivalent to adding /B to every pattern. This causes
1652 it to show the compiled bytecode, without the additional information that
1653 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1654 is the same as /B/I).
1655
1656 23. A new optimization is now able automatically to treat some sequences such
1657 as a*b as a*+b. More specifically, if something simple (such as a character
1658 or a simple class like \d) has an unlimited quantifier, and is followed by
1659 something that cannot possibly match the quantified thing, the quantifier
1660 is automatically "possessified".
1661
1662 24. A recursive reference to a subpattern whose number was greater than 39
1663 went wrong under certain circumstances in UTF-8 mode. This bug could also
1664 have affected the operation of pcre_study().
1665
1666 25. Realized that a little bit of performance could be had by replacing
1667 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1668
1669 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1670
1671 27. Possessive quantifiers such as a++ were previously implemented by turning
1672 them into atomic groups such as ($>a+). Now they have their own opcodes,
1673 which improves performance. This includes the automatically created ones
1674 from 23 above.
1675
1676 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1677 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1678 the first matched character to be a colon. This applied both to named and
1679 numbered groups.
1680
1681 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1682
1683 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1684 building PCRE on Windows easier, so I have included it in the distribution.
1685
1686 31. There is now a check in pcretest against a ridiculously large number being
1687 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1688 loop, the loop is abandoned.
1689
1690 32. Forward references to subpatterns in conditions such as (?(2)...) where
1691 subpattern 2 is defined later cause pcre_compile() to search forwards in
1692 the pattern for the relevant set of parentheses. This search went wrong
1693 when there were unescaped parentheses in a character class, parentheses
1694 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1695
1696 33. "Subroutine" calls and backreferences were previously restricted to
1697 referencing subpatterns earlier in the regex. This restriction has now
1698 been removed.
1699
1700 34. Added a number of extra features that are going to be in Perl 5.10. On the
1701 whole, these are just syntactic alternatives for features that PCRE had
1702 previously implemented using the Python syntax or my own invention. The
1703 other formats are all retained for compatibility.
1704
1705 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1706 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1707 also .NET compatible.
1708
1709 (b) A recursion or subroutine call to a named group can now be defined as
1710 (?&name) as well as (?P>name).
1711
1712 (c) A backreference to a named group can now be defined as \k<name> or
1713 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1714 5.10, are also .NET compatible.
1715
1716 (d) A conditional reference to a named group can now use the syntax
1717 (?(<name>) or (?('name') as well as (?(name).
1718
1719 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1720 groups (named and numbered) that are never evaluated inline, but can be
1721 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1722 is always false. There may be only one alternative in such a group.
1723
1724 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1725 as the simple (?(R). The condition is true only if the most recent
1726 recursion is that of the given number or name. It does not search out
1727 through the entire recursion stack.
1728
1729 (g) The escape \gN or \g{N} has been added, where N is a positive or
1730 negative number, specifying an absolute or relative reference.
1731
1732 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1733 some "unreachable code" warnings.
1734
1735 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1736 things, this adds five new scripts.
1737
1738 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1739 There were also incompatibilities regarding the handling of \Q..\E inside
1740 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1741 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1742
1743 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1744 matches an empty string, and forcibly breaks the loop. There were bugs in
1745 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1746 against aaaa the result was just "a" rather than "aaaa", for example. Two
1747 separate and independent bugs (that affected different cases) have been
1748 fixed.
1749
1750 39. Refactored the code to abolish the use of different opcodes for small
1751 capturing bracket numbers. This is a tidy that I avoided doing when I
1752 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1753 The new approach is not only tidier, it makes it possible to reduce the
1754 memory needed to fix the previous bug (38).
1755
1756 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1757 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1758 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1759 mode.
1760
1761 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1762 report.
1763
1764 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1765 copy construction and assignment in the C++ wrapper.
1766
1767 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1768 couple of bugs that could have given wrong results in the "--newline=crlf"
1769 case.
1770
1771 44. Added a number of casts and did some reorganization of signed/unsigned int
1772 variables following suggestions from Dair Grant. Also renamed the variable
1773 "this" as "item" because it is a C++ keyword.
1774
1775 45. Arranged for dftables to add
1776
1777 #include "pcre_internal.h"
1778
1779 to pcre_chartables.c because without it, gcc 4.x may remove the array
1780 definition from the final binary if PCRE is built into a static library and
1781 dead code stripping is activated.
1782
1783 46. For an unanchored pattern, if a match attempt fails at the start of a
1784 newline sequence, and the newline setting is CRLF or ANY, and the next two
1785 characters are CRLF, advance by two characters instead of one.
1786
1787
1788 Version 6.7 04-Jul-06
1789 ---------------------
1790
1791 1. In order to handle tests when input lines are enormously long, pcretest has
1792 been re-factored so that it automatically extends its buffers when
1793 necessary. The code is crude, but this _is_ just a test program. The
1794 default size has been increased from 32K to 50K.
1795
1796 2. The code in pcre_study() was using the value of the re argument before
1797 testing it for NULL. (Of course, in any sensible call of the function, it
1798 won't be NULL.)
1799
1800 3. The memmove() emulation function in pcre_internal.h, which is used on
1801 systems that lack both memmove() and bcopy() - that is, hardly ever -
1802 was missing a "static" storage class specifier.
1803
1804 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1805 containing an extended class (one that cannot be represented by a bitmap
1806 because it contains high-valued characters or Unicode property items, e.g.
1807 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1808 pattern, but PCRE should not loop if you do not (it no longer does).
1809 [Detail: two cases were found: (a) a repeated subpattern containing an
1810 extended class; (b) a recursive reference to a subpattern that followed a
1811 previous extended class. It wasn't skipping over the extended class
1812 correctly when UTF-8 mode was not set.]
1813
1814 5. A negated single-character class was not being recognized as fixed-length
1815 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1816 compile error "lookbehind assertion is not fixed length".
1817
1818 6. The RunPerlTest auxiliary script was showing an unexpected difference
1819 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1820 write a Perl script that can interpret lines of an input file either as
1821 byte characters or as UTF-8, which is what "perltest" was being required to
1822 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1823 can't do is switch easily at run time between having the "use utf8;" pragma
1824 or not. In the end, I fudged it by using the RunPerlTest script to insert
1825 "use utf8;" explicitly for the UTF-8 tests.
1826
1827 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1828 the end of the subject string, contrary to the documentation and to what
1829 Perl does. This was true of both matching functions. Now it matches only at
1830 the start of the subject and immediately after *internal* newlines.
1831
1832 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1833 a pointer to an int instead of a pointer to an unsigned long int. This
1834 caused problems on 64-bit systems.
1835
1836 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1837 instance of the 'standard' template library not being so standard".
1838
1839 10. There was no check on the number of named subpatterns nor the maximum
1840 length of a subpattern name. The product of these values is used to compute
1841 the size of the memory block for a compiled pattern. By supplying a very
1842 long subpattern name and a large number of named subpatterns, the size
1843 computation could be caused to overflow. This is now prevented by limiting
1844 the length of names to 32 characters, and the number of named subpatterns
1845 to 10,000.
1846
1847 11. Subpatterns that are repeated with specific counts have to be replicated in
1848 the compiled pattern. The size of memory for this was computed from the
1849 length of the subpattern and the repeat count. The latter is limited to
1850 65535, but there was no limit on the former, meaning that integer overflow
1851 could in principle occur. The compiled length of a repeated subpattern is
1852 now limited to 30,000 bytes in order to prevent this.
1853
1854 12. Added the optional facility to have named substrings with the same name.
1855
1856 13. Added the ability to use a named substring as a condition, using the
1857 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1858 are numbers (not recommended). Forward references are permitted.
1859
1860 14. Added forward references in named backreferences (if you see what I mean).
1861
1862 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1863 pattern could run off the end of the subject. For example, the pattern
1864 "(?s)(.{1,5})"8 did this with the subject "ab".
1865
1866 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1867 PCRE_CASELESS was set when matching characters that were quantified with ?
1868 or *.
1869
1870 17. A character class other than a single negated character that had a minimum
1871 but no maximum quantifier - for example [ab]{6,} - was not handled
1872 correctly by pce_dfa_exec(). It would match only one character.
1873
1874 18. A valid (though odd) pattern that looked like a POSIX character
1875 class but used an invalid character after [ (for example [[,abc,]]) caused
1876 pcre_compile() to give the error "Failed: internal error: code overflow" or
1877 in some cases to crash with a glibc free() error. This could even happen if
1878 the pattern terminated after [[ but there just happened to be a sequence of
1879 letters, a binary zero, and a closing ] in the memory that followed.
1880
1881 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1882 over the years. Originally (before any Unicode support), just the bottom 8
1883 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1884 output from "man perlunicode" includes this:
1885
1886 The regular expression compiler produces polymorphic opcodes. That
1887 is, the pattern adapts to the data and automatically switches to
1888 the Unicode character scheme when presented with Unicode data--or
1889 instead uses a traditional byte scheme when presented with byte
1890 data.
1891
1892 Sadly, a wide octal escape does not cause a switch, and in a string with
1893 no other multibyte characters, these octal escapes are treated as before.
1894 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1895 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1896 Unicode string.
1897
1898 I have not perpetrated such confusion in PCRE. Up till now, it took just
1899 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1900 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1901 translate to the appropriate multibyte character.
1902
1903 29. Applied some refactoring to reduce the number of warnings from Microsoft
1904 and Borland compilers. This has included removing the fudge introduced
1905 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1906 a warning about an unused variable.
1907
1908 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1909 characters since release 4.0, because Perl (from release 5.004) does not.
1910 [Or at least, is documented not to: some releases seem to be in conflict
1911 with the documentation.] However, when a pattern was studied with
1912 pcre_study() and all its branches started with \s, PCRE still included VT
1913 as a possible starting character. Of course, this did no harm; it just
1914 caused an unnecessary match attempt.
1915
1916 22. Removed a now-redundant internal flag bit that recorded the fact that case
1917 dependency changed within the pattern. This was once needed for "required
1918 byte" processing, but is no longer used. This recovers a now-scarce options
1919 bit. Also moved the least significant internal flag bit to the most-
1920 significant bit of the word, which was not previously used (hangover from
1921 the days when it was an int rather than a uint) to free up another bit for
1922 the future.
1923
1924 23. Added support for CRLF line endings as well as CR and LF. As well as the
1925 default being selectable at build time, it can now be changed at runtime
1926 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1927 specify that it is scanning data with non-default line endings.
1928
1929 24. Changed the definition of CXXLINK to make it agree with the definition of
1930 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1931
1932 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1933 recursions. This makes a big different to stack usage for some patterns.
1934
1935 26. If a subpattern containing a named recursion or subroutine reference such
1936 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1937 the space required for the compiled pattern went wrong and gave too small a
1938 value. Depending on the environment, this could lead to "Failed: internal
1939 error: code overflow at offset 49" or "glibc detected double free or
1940 corruption" errors.
1941
1942 27. Applied patches from Google (a) to support the new newline modes and (b) to
1943 advance over multibyte UTF-8 characters in GlobalReplace.
1944
1945 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1946 difference for some implementation of PCRE in some Windows version.
1947
1948 29. Added some extra testing facilities to pcretest:
1949
1950 \q<number> in a data line sets the "match limit" value
1951 \Q<number> in a data line sets the "match recursion limt" value
1952 -S <number> sets the stack size, where <number> is in megabytes
1953
1954 The -S option isn't available for Windows.
1955
1956
1957 Version 6.6 06-Feb-06
1958 ---------------------
1959
1960 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1961 in pcreposix.h. I have copied the definition from pcre.h.
1962
1963 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1964 because pcre.h is no longer a built file.
1965
1966 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1967 not normally included in the compiled code.
1968
1969
1970 Version 6.5 01-Feb-06
1971 ---------------------
1972
1973 1. When using the partial match feature with pcre_dfa_exec(), it was not
1974 anchoring the second and subsequent partial matches at the new starting
1975 point. This could lead to incorrect results. For example, with the pattern
1976 /1234/, partially matching against "123" and then "a4" gave a match.
1977
1978 2. Changes to pcregrep:
1979
1980 (a) All non-match returns from pcre_exec() were being treated as failures
1981 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1982 error message is output. Some extra information is given for the
1983 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1984 probably the only errors that are likely to be caused by users (by
1985 specifying a regex that has nested indefinite repeats, for instance).
1986 If there are more than 20 of these errors, pcregrep is abandoned.
1987
1988 (b) A binary zero was treated as data while matching, but terminated the
1989 output line if it was written out. This has been fixed: binary zeroes
1990 are now no different to any other data bytes.
1991
1992 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1993 used to set a locale for matching. The --locale=xxxx long option has
1994 been added (no short equivalent) to specify a locale explicitly on the
1995 pcregrep command, overriding the environment variables.
1996
1997 (d) When -B was used with -n, some line numbers in the output were one less
1998 than they should have been.
1999
2000 (e) Added the -o (--only-matching) option.
2001
2002 (f) If -A or -C was used with -c (count only), some lines of context were
2003 accidentally printed for the final match.
2004
2005 (g) Added the -H (--with-filename) option.
2006
2007 (h) The combination of options -rh failed to suppress file names for files
2008 that were found from directory arguments.
2009
2010 (i) Added the -D (--devices) and -d (--directories) options.
2011
2012 (j) Added the -F (--fixed-strings) option.
2013
2014 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2015
2016 (l) Added the --colo(u)r option.
2017
2018 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2019 is not present by default.
2020
2021 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2022 items such as (?R) or (?1), when the recursion could match a number of
2023 alternatives. If it matched one of the alternatives, but subsequently,
2024 outside the recursion, there was a failure, the code tried to back up into
2025 the recursion. However, because of the way PCRE is implemented, this is not
2026 possible, and the result was an incorrect result from the match.
2027
2028 In order to prevent this happening, the specification of recursion has
2029 been changed so that all such subpatterns are automatically treated as
2030 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2031
2032 4. I had overlooked the fact that, in some locales, there are characters for
2033 which isalpha() is true but neither isupper() nor islower() are true. In
2034 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2035 and ordfeminine) are like this. This affected the treatment of \w and \W
2036 when they appeared in character classes, but not when they appeared outside
2037 a character class. The bit map for "word" characters is now created
2038 separately from the results of isalnum() instead of just taking it from the
2039 upper, lower, and digit maps. (Plus the underscore character, of course.)
2040
2041 5. The above bug also affected the handling of POSIX character classes such as
2042 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2043 permanent tables. Instead, the bit maps for such a class were previously
2044 created as the appropriate unions of the upper, lower, and digit bitmaps.
2045 Now they are created by subtraction from the [[:word:]] class, which has
2046 its own bitmap.
2047
2048 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2049 It is created by subtracting the vertical space characters (\x09, \x0a,
2050 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2051 subtraction was done in the overall bitmap for a character class, meaning
2052 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2053 be recognized. This bug has been fixed.
2054
2055 7. Patches from the folks at Google:
2056
2057 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2058 real life, but is still worth protecting against".
2059
2060 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2061 regular expressions".
2062
2063 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2064 have it.
2065
2066 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2067 "configure" and the latter not, in order to fix a problem somebody had
2068 with compiling the Arg class on HP-UX.
2069
2070 (e) Improve the error-handling of the C++ wrapper a little bit.
2071
2072 (f) New tests for checking recursion limiting.
2073
2074 8. The pcre_memmove() function, which is used only if the environment does not
2075 have a standard memmove() function (and is therefore rarely compiled),
2076 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2077 returning a result (though PCRE never actually uses the result).
2078
2079 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2080 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2081 returned instead of calling malloc() with an overflowing number that would
2082 most likely cause subsequent chaos.
2083
2084 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2085
2086 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2087 with this option is matched, the nmatch and pmatch options of regexec() are
2088 ignored.
2089
2090 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2091 provided in case anyone wants to the the POSIX interface with UTF-8
2092 strings.
2093
2094 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2095 C++ linking (needed for some HP-UX environments).
2096
2097 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2098 (unused parameter) and in the pcre_printint() function (omitted "default"
2099 switch label when the default is to do nothing).
2100
2101 15. Added some code to make it possible, when PCRE is compiled as a C++
2102 library, to replace subject pointers for pcre_exec() with a smart pointer
2103 class, thus making it possible to process discontinuous strings.
2104
2105 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2106 much the same function. They were added by different people who were trying
2107 to make PCRE easy to compile on non-Unix systems. It has been suggested
2108 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2109 for compiling on Windows systems. I have therefore replaced it with
2110 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2111 defaults to "extern" for C or "extern C" for C++, which works fine on
2112 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2113 SCOPE with something explicit in config.h. In addition:
2114
2115 (a) pcreposix.h still had just "extern" instead of either of these macros;
2116 I have replaced it with PCRE_DATA_SCOPE.
2117
2118 (b) Functions such as _pcre_xclass(), which are internal to the library,
2119 but external in the C sense, all had PCRE_EXPORT in their definitions.
2120 This is apparently wrong for the Windows case, so I have removed it.
2121 (It makes no difference on Unix-like systems.)
2122
2123 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2124 of recursive calls to match(). This is different to MATCH_LIMIT because
2125 that limits the total number of calls to match(), not all of which increase
2126 the depth of recursion. Limiting the recursion depth limits the amount of
2127 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2128 when PCRE is compiled, and changed at run time. A patch from Google adds
2129 this functionality to the C++ interface.
2130
2131 18. Changes to the handling of Unicode character properties:
2132
2133 (a) Updated the table to Unicode 4.1.0.
2134
2135 (b) Recognize characters that are not in the table as "Cn" (undefined).
2136
2137 (c) I revised the way the table is implemented to a much improved format
2138 which includes recognition of ranges. It now supports the ranges that
2139 are defined in UnicodeData.txt, and it also amalgamates other
2140 characters into ranges. This has reduced the number of entries in the
2141 table from around 16,000 to around 3,000, thus reducing its size
2142 considerably. I realized I did not need to use a tree structure after
2143 all - a binary chop search is just as efficient. Having reduced the
2144 number of entries, I extended their size from 6 bytes to 8 bytes to
2145 allow for more data.
2146
2147 (d) Added support for Unicode script names via properties such as \p{Han}.
2148
2149 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2150 matching that character.
2151
2152 20. When matching a repeated Unicode property with a minimum greater than zero,
2153 (for example \pL{2,}), PCRE could look past the end of the subject if it
2154 reached it while seeking the minimum number of characters. This could
2155 happen only if some of the characters were more than one byte long, because
2156 there is a check for at least the minimum number of bytes.
2157
2158 21. Refactored the implementation of \p and \P so as to be more general, to
2159 allow for more different types of property in future. This has changed the
2160 compiled form incompatibly. Anybody with saved compiled patterns that use
2161 \p or \P will have to recompile them.
2162
2163 22. Added "Any" and "L&" to the supported property types.
2164
2165 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2166 but give a compile time error if the value is greater than 0xff.
2167
2168 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2169 accidentally not being installed or uninstalled.
2170
2171 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2172 made were to insert the current release number. This seemed silly, because
2173 it made things harder for people building PCRE on systems that don't run
2174 "configure". I have turned pcre.h into a distributed file, no longer built
2175 by "configure", with the version identification directly included. There is
2176 no longer a pcre.h.in file.
2177
2178 However, this change necessitated a change to the pcre-config script as
2179 well. It is built from pcre-config.in, and one of the substitutions was the
2180 release number. I have updated configure.ac so that ./configure now finds
2181 the release number by grepping pcre.h.
2182
2183 26. Added the ability to run the tests under valgrind.
2184
2185
2186 Version 6.4 05-Sep-05
2187 ---------------------
2188
2189 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2190 "--" to be printed when multiple files were scanned, even when none of the
2191 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2192 consider it to be a bug, and have restored the previous behaviour.
2193
2194 2. A couple of code tidies to get rid of compiler warnings.
2195
2196 3. The pcretest program used to cheat by referring to symbols in the library
2197 whose names begin with _pcre_. These are internal symbols that are not
2198 really supposed to be visible externally, and in some environments it is
2199 possible to suppress them. The cheating is now confined to including
2200 certain files from the library's source, which is a bit cleaner.
2201
2202 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2203 file's purpose clearer.
2204
2205 5. Reorganized pcre_ucp_findchar().
2206
2207
2208 Version 6.3 15-Aug-05
2209 ---------------------
2210
2211 1. The file libpcre.pc.in did not have general read permission in the tarball.
2212
2213 2. There were some problems when building without C++ support:
2214
2215 (a) If C++ support was not built, "make install" and "make test" still
2216 tried to test it.
2217
2218 (b) There were problems when the value of CXX was explicitly set. Some
2219 changes have been made to try to fix these, and ...
2220
2221 (c) --disable-cpp can now be used to explicitly disable C++ support.
2222
2223 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2224 backslash in a target when C++ was disabled. This confuses some
2225 versions of "make", apparently. Using an intermediate variable solves
2226 this. (Same for CPP_LOBJ.)
2227
2228 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2229 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2230 necessary on certain architectures.
2231
2232 4. Added a setting of -export-symbols-regex to the link command to remove
2233 those symbols that are exported in the C sense, but actually are local
2234 within the library, and not documented. Their names all begin with
2235 "_pcre_". This is not a perfect job, because (a) we have to except some
2236 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2237 available (and never for static libraries). I have made a note to try to
2238 find a way round (a) in the future.
2239
2240
2241 Version 6.2 01-Aug-05
2242 ---------------------
2243
2244 1. There was no test for integer overflow of quantifier values. A construction
2245 such as {1111111111111111} would give undefined results. What is worse, if
2246 a minimum quantifier for a parenthesized subpattern overflowed and became
2247 negative, the calculation of the memory size went wrong. This could have
2248 led to memory overwriting.
2249
2250 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2251
2252 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2253 operating environments where this matters.
2254
2255 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2256 PCRE options from within the C++ wrapper.
2257
2258 5. Named capturing subpatterns were not being correctly counted when a pattern
2259 was compiled. This caused two problems: (a) If there were more than 100
2260 such subpatterns, the calculation of the memory needed for the whole
2261 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2262 back references of the form \12, where the number was greater than 9, were
2263 not recognized as back references, even though there were sufficient
2264 previous subpatterns.
2265
2266 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2267 versions of gcc, e.g. 2.95.4.
2268
2269
2270 Version 6.1 21-Jun-05
2271 ---------------------
2272
2273 1. There was one reference to the variable "posix" in pcretest.c that was not
2274 surrounded by "#if !defined NOPOSIX".
2275
2276 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2277 the cross-check on the old pcre_info() function, for the benefit of the
2278 cut-down version of PCRE that is currently imported into Exim.
2279
2280 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2281 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2282 patterns that start (?i) but I don't think that matters. The use of (?i) is
2283 just an example; this all applies to the other options as well.
2284
2285 4. Since libtool seems to echo the compile commands it is issuing, the output
2286 from "make" can be reduced a bit by putting "@" in front of each libtool
2287 compile command.
2288
2289 5. Patch from the folks at Google for configure.in to be a bit more thorough
2290 in checking for a suitable C++ installation before trying to compile the
2291 C++ stuff. This should fix a reported problem when a compiler was present,
2292 but no suitable headers.
2293
2294 6. The man pages all had just "PCRE" as their title. I have changed them to
2295 be the relevant file name. I have also arranged that these names are
2296 retained in the file doc/pcre.txt, which is a concatenation in text format
2297 of all the man pages except the little individual ones for each function.
2298
2299 7. The NON-UNIX-USE file had not been updated for the different set of source
2300 files that come with release 6. I also added a few comments about the C++
2301 wrapper.
2302
2303
2304 Version 6.0 07-Jun-05
2305 ---------------------
2306
2307 1. Some minor internal re-organization to help with my DFA experiments.
2308
2309 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2310 didn't matter for the library itself when fully configured, but did matter
2311 when compiling without UCP support, or within Exim, where the ucp files are
2312 not imported.
2313
2314 3. Refactoring of the library code to split up the various functions into
2315 different source modules. The addition of the new DFA matching code (see
2316 below) to a single monolithic source would have made it really too
2317 unwieldy, quite apart from causing all the code to be include in a
2318 statically linked application, when only some functions are used. This is
2319 relevant even without the DFA addition now that patterns can be compiled in
2320 one application and matched in another.
2321
2322 The downside of splitting up is that there have to be some external
2323 functions and data tables that are used internally in different modules of
2324 the library but which are not part of the API. These have all had their
2325 names changed to start with "_pcre_" so that they are unlikely to clash
2326 with other external names.
2327
2328 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2329 a different (DFA) algorithm. Although it is slower than the original
2330 function, it does have some advantages for certain types of matching
2331 problem.
2332
2333 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2334 including restarting after a partial match.
2335
2336 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2337 defined when compiling for Windows was sent to me. I have put it into the
2338 code, though I have no means of testing or verifying it.
2339
2340 7. Added the pcre_refcount() auxiliary function.
2341
2342 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2343 match before or at the first newline in the subject string. In pcretest,
2344 the /f option on a pattern can be used to set this.
2345
2346 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2347 would behave wrongly. This has been present in PCRE since release 4.0.
2348
2349 10. A number of changes to the pcregrep command:
2350
2351 (a) Refactored how -x works; insert ^(...)$ instead of setting
2352 PCRE_ANCHORED and checking the length, in preparation for adding
2353 something similar for -w.
2354
2355 (b) Added the -w (match as a word) option.
2356
2357 (c) Refactored the way lines are read and buffered so as to have more
2358 than one at a time available.
2359
2360 (d) Implemented a pcregrep test script.
2361
2362 (e) Added the -M (multiline match) option. This allows patterns to match
2363 over several lines of the subject. The buffering ensures that at least
2364 8K, or the rest of the document (whichever is the shorter) is available
2365 for matching (and similarly the previous 8K for lookbehind assertions).
2366
2367 (f) Changed the --help output so that it now says
2368
2369 -w, --word-regex(p)
2370
2371 instead of two lines, one with "regex" and the other with "regexp"
2372 because that confused at least one person since the short forms are the
2373 same. (This required a bit of code, as the output is generated
2374 automatically from a table. It wasn't just a text change.)
2375
2376 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2377 option but starts with a hyphen. Could be a pattern or a path name
2378 starting with a hyphen, for instance.
2379
2380 (h) "-" can be given as a file name to represent stdin.
2381
2382 (i) When file names are being printed, "(standard input)" is used for
2383 the standard input, for compatibility with GNU grep. Previously
2384 "<stdin>" was used.
2385
2386 (j) The option --label=xxx can be used to supply a name to be used for
2387 stdin when file names are being printed. There is no short form.
2388
2389 (k) Re-factored the options decoding logic because we are going to add
2390 two more options that take data. Such options can now be given in four
2391 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2392
2393 (l) Added the -A, -B, and -C options for requesting that lines of context
2394 around matches be printed.
2395
2396 (m) Added the -L option to print the names of files that do not contain
2397 any matching lines, that is, the complement of -l.
2398
2399 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2400 continue to scan other files.
2401
2402 (o) The -s option was incorrectly implemented. For compatibility with other
2403 greps, it now suppresses the error message for a non-existent or non-
2404 accessible file (but not the return code). There is a new option called
2405 -q that suppresses the output of matching lines, which was what -s was
2406 previously doing.
2407
2408 (p) Added --include and --exclude options to specify files for inclusion
2409 and exclusion when recursing.
2410
2411 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2412 Hopefully, it now does.
2413
2414 12. Missing cast in pcre_study().
2415
2416 13. Added an "uninstall" target to the makefile.
2417
2418 14. Replaced "extern" in the function prototypes in Makefile.in with
2419 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2420 world, but is set differently for Windows.
2421
2422 15. Added a second compiling function called pcre_compile2(). The only
2423 difference is that it has an extra argument, which is a pointer to an
2424 integer error code. When there is a compile-time failure, this is set
2425 non-zero, in addition to the error test pointer being set to point to an
2426 error message. The new argument may be NULL if no error number is required
2427 (but then you may as well call pcre_compile(), which is now just a
2428 wrapper). This facility is provided because some applications need a
2429 numeric error indication, but it has also enabled me to tidy up the way
2430 compile-time errors are handled in the POSIX wrapper.
2431
2432 16. Added VPATH=.libs to the makefile; this should help when building with one
2433 prefix path and installing with another. (Or so I'm told by someone who
2434 knows more about this stuff than I do.)
2435
2436 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2437 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2438 match everything, including newlines. This is not POSIX-compatible, but
2439 somebody wanted the feature. From pcretest it can be activated by using
2440 both the P and the s flags.
2441
2442 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2443
2444 19. libpcre.pc was being incorrectly installed as executable.
2445
2446 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2447 it now also looks for '\r' so that it will work unmodified on Windows.
2448
2449 21. Added Google's contributed C++ wrapper to the distribution.
2450
2451 22. Added some untidy missing memory free() calls in pcretest, to keep
2452 Electric Fence happy when testing.
2453
2454
2455
2456 Version 5.0 13-Sep-04
2457 ---------------------
2458
2459 1. Internal change: literal characters are no longer packed up into items
2460 containing multiple characters in a single byte-string. Each character
2461 is now matched using a separate opcode. However, there may be more than one
2462 byte in the character in UTF-8 mode.
2463
2464 2. The pcre_callout_block structure has two new fields: pattern_position and
2465 next_item_length. These contain the offset in the pattern to the next match
2466 item, and its length, respectively.
2467
2468 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2469 insertion of callouts before each pattern item. Added the /C option to
2470 pcretest to make use of this.
2471
2472 4. On the advice of a Windows user, the lines
2473
2474 #if defined(_WIN32) || defined(WIN32)
2475 _setmode( _fileno( stdout ), 0x8000 );
2476 #endif /* defined(_WIN32) || defined(WIN32) */
2477
2478 have been added to the source of pcretest. This apparently does useful
2479 magic in relation to line terminators.
2480
2481 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2482 for the benefit of those environments where the "b" makes a difference.
2483
2484 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2485 to know about it. I have put a hack into configure.in that adds in code
2486 to set GCC=yes if CC=icc. This seems to end up at a point in the
2487 generated configure script that is early enough to affect the setting of
2488 compiler options, which is what is needed, but I have no means of testing
2489 whether it really works. (The user who reported this had patched the
2490 generated configure script, which of course I cannot do.)
2491
2492 LATER: After change 22 below (new libtool files), the configure script
2493 seems to know about icc (and also ecc). Therefore, I have commented out
2494 this hack in configure.in.
2495
2496 7. Added support for pkg-config (2 patches were sent in).
2497
2498 8. Negated POSIX character classes that used a combination of internal tables
2499 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2500 [[:^ascii]]. Typically, they would match almost any characters. The other
2501 POSIX classes were not broken in this way.
2502
2503 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2504 to find the match, as PCRE was deluded into thinking that the match had to
2505 start at the start point or following a newline. The same bug applied to
2506 patterns with negative forward assertions or any backward assertions
2507 preceding ".*" at the start, unless the pattern required a fixed first
2508 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2509
2510 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2511 starting at the last subject character, bytes beyond the end of the subject
2512 string were read.
2513
2514 11. Renamed the variable "class" as "classbits" to make life easier for C++
2515 users. (Previously there was a macro definition, but it apparently wasn't
2516 enough.)
2517
2518 12. Added the new field "tables" to the extra data so that tables can be passed
2519 in at exec time, or the internal tables can be re-selected. This allows
2520 a compiled regex to be saved and re-used at a later time by a different
2521 program that might have everything at different addresses.
2522
2523 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2524 -R library as well as a -L library.
2525
2526 14. The debugging options of pcretest (-d on the command line or D on a
2527 pattern) showed incorrect output for anything following an extended class
2528 that contained multibyte characters and which was followed by a quantifier.
2529
2530 15. Added optional support for general category Unicode character properties
2531 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2532 support. It adds about 90K to the size of the library. The meanings of the
2533 inbuilt class escapes such as \d and \s have NOT been changed.
2534
2535 16. Updated pcredemo.c to include calls to free() to release the memory for the
2536 compiled pattern.
2537
2538 17. The generated file chartables.c was being created in the source directory
2539 instead of in the building directory. This caused the build to fail if the
2540 source directory was different from the building directory, and was
2541 read-only.
2542
2543 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2544 file. No doubt somebody will tell me if they don't make sense... Also added
2545 Dan Mooney's comments about building on OpenVMS.
2546
2547 19. Added support for partial matching via the PCRE_PARTIAL option for
2548 pcre_exec() and the \P data escape in pcretest.
2549
2550 20. Extended pcretest with 3 new pattern features:
2551
2552 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2553 write the compiled pattern to the file whose name is "rest-of-line".
2554 This is a straight binary dump of the data, with the saved pointer to
2555 the character tables forced to be NULL. The study data, if any, is
2556 written too. After writing, pcretest reads a new pattern.
2557
2558 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2559 compiled pattern from the given file. There must not be any
2560 occurrences of "<" in the file name (pretty unlikely); if there are,
2561 pcretest will instead treat the initial "<" as a pattern delimiter.
2562 After reading in the pattern, pcretest goes on to read data lines as
2563 usual.
2564
2565 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2566 and 16-bit fields in a compiled pattern, to simulate a pattern that
2567 was compiled on a host of opposite endianness.
2568
2569 21. The pcre-exec() function can now cope with patterns that were compiled on
2570 hosts of opposite endianness, with this restriction:
2571
2572 As for any compiled expression that is saved and used later, the tables
2573 pointer field cannot be preserved; the extra_data field in the arguments
2574 to pcre_exec() should be used to pass in a tables address if a value
2575 other than the default internal tables were used at compile time.
2576
2577 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2578 now diagnosed as an error. Previously, most of the time, a negative number
2579 would have been treated as zero, but if in addition "ovector" was passed as
2580 NULL, a crash could occur.
2581
2582 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2583 new versions from the libtool 1.5 distribution (the last one is a copy of
2584 a file called libtool.m4). This seems to have fixed the need to patch
2585 "configure" to support Darwin 1.3 (which I used to do). However, I still
2586 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2587 workstation).
2588
2589 24. Changed the PCRE licence to be the more standard "BSD" licence.
2590
2591
2592 Version 4.5 01-Dec-03
2593 ---------------------
2594
2595 1. There has been some re-arrangement of the code for the match() function so
2596 that it can be compiled in a version that does not call itself recursively.
2597 Instead, it keeps those local variables that need separate instances for
2598 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2599 needs to "recurse". Keeping track of where control must go is done by means
2600 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2601 hide most of the details from the main code, and operates only if
2602 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2603 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2604 operating.
2605
2606 To make it easier for callers to provide specially tailored get/free
2607 functions for this usage, two new functions, pcre_stack_malloc, and
2608 pcre_stack_free, are used. They are always called in strict stacking order,
2609 and the size of block requested is always the same.
2610
2611 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2612 PCRE has been compiled to use the stack or the heap for recursion. The
2613 -C option of pcretest uses this to show which version is compiled.
2614
2615 A new data escape \S, is added to pcretest; it causes the amounts of store
2616 obtained and freed by both kinds of malloc/free at match time to be added
2617 to the output.
2618
2619 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2620 what's available on my current Linux desktop machine.
2621
2622 3. When matching a UTF-8 string, the test for a valid string at the start has
2623 been extended. If start_offset is not zero, PCRE now checks that it points
2624 to a byte that is the start of a UTF-8 character. If not, it returns
2625 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2626 this is necessary because there may be backward assertions in the pattern.
2627 When matching the same subject several times, it may save resources to use
2628 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2629
2630 4. The code for checking the validity of UTF-8 strings has been tightened so
2631 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2632 containing "overlong sequences".
2633
2634 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2635 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2636 should have been "&", but it just so happened that all the cases this let
2637 through by mistake were picked up later in the function.
2638
2639 6. I had used a variable called "isblank" - this is a C99 function, causing
2640 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2641
2642 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2643 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2644 so the version is shown; (c) stop "make test" echoing "./RunTest".
2645
2646 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2647
2648 9. The prototype for memmove() for systems that don't have it was using
2649 size_t, but the inclusion of the header that defines size_t was later. I've
2650 moved the #includes for the C headers earlier to avoid this.
2651
2652 10. Added some adjustments to the code to make it easier to compiler on certain
2653 special systems:
2654
2655 (a) Some "const" qualifiers were missing.
2656 (b) Added the macro EXPORT before all exported functions; by default this
2657 is defined to be empty.
2658 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2659 that it reads its output file name as an argument instead of writing
2660 to the standard output and assuming this can be redirected.
2661
2662 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2663 class containing characters with values greater than 255, PCRE compilation
2664 went into a loop.
2665
2666 12. A recursive reference to a subpattern that was within another subpattern
2667 that had a minimum quantifier of zero caused PCRE to crash. For example,
2668 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2669 recursion. If the recursively-called subpattern itself had a zero repeat,
2670 that was OK.
2671
2672 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2673 buffer into which it was copied (for escape processing) was still set at
2674 1024, so long lines caused crashes.
2675
2676 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2677 "internal error: code overflow...". This applied to any character class
2678 that was followed by a possessive quantifier.
2679
2680 15. Modified the Makefile to add libpcre.la as a prerequisite for
2681 libpcreposix.la because I was told this is needed for a parallel build to
2682 work.
2683
2684 16. If a pattern that contained .* following optional items at the start was
2685 studied, the wrong optimizing data was generated, leading to matching
2686 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2687 matching string must start with a or b or c. The correct conclusion for
2688 this pattern is that a match can start with any character.
2689
2690
2691 Version 4.4 13-Aug-03
2692 ---------------------
2693
2694 1. In UTF-8 mode, a character class containing characters with values between
2695 127 and 255 was not handled correctly if the compiled pattern was studied.
2696 In fixing this, I have also improved the studying algorithm for such
2697 classes (slightly).
2698
2699 2. Three internal functions had redundant arguments passed to them. Removal
2700 might give a very teeny performance improvement.
2701
2702 3. Documentation bug: the value of the capture_top field in a callout is *one
2703 more than* the number of the hightest numbered captured substring.
2704
2705 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2706 in incorrectly linking with a previously installed version. They now link
2707 explicitly with libpcre.la.
2708
2709 5. configure.in no longer needs to recognize Cygwin specially.
2710
2711 6. A problem in pcre.in for Windows platforms is fixed.
2712
2713 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2714 pcretest, it used to include the size of the study block as part of its
2715 output. Unfortunately, the structure contains a field that has a different
2716 size on different hardware architectures. This meant that the tests that
2717 showed this size failed. As the block is currently always of a fixed size,
2718 this information isn't actually particularly useful in pcretest output, so
2719 I have just removed it.
2720
2721 8. Three pre-processor statements accidentally did not start in column 1.
2722 Sadly, there are *still* compilers around that complain, even though
2723 standard C has not required this for well over a decade. Sigh.
2724
2725 9. In pcretest, the code for checking callouts passed small integers in the
2726 callout_data field, which is a void * field. However, some picky compilers
2727 complained about the casts involved for this on 64-bit systems. Now
2728 pcretest passes the address of the small integer instead, which should get
2729 rid of the warnings.
2730
2731 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2732 both compile and run time, and gives an error if an invalid UTF-8 sequence
2733 is found. There is a option for disabling this check in cases where the
2734 string is known to be correct and/or the maximum performance is wanted.
2735
2736 11. In response to a bug report, I changed one line in Makefile.in from
2737
2738 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2739 to
2740 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2741
2742 to look similar to other lines, but I have no way of telling whether this
2743 is the right thing to do, as I do not use Windows. No doubt I'll get told
2744 if it's wrong...
2745
2746
2747 Version 4.3 21-May-03
2748 ---------------------
2749
2750 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2751 Makefile.
2752
2753 2. Some refactoring to improve the quality of the code:
2754
2755 (i) The utf8_table... variables are now declared "const".
2756
2757 (ii) The code for \cx, which used the "case flipping" table to upper case
2758 lower case letters, now just substracts 32. This is ASCII-specific,
2759 but the whole concept of \cx is ASCII-specific, so it seems
2760 reasonable.
2761
2762 (iii) PCRE was using its character types table to recognize decimal and
2763 hexadecimal digits in the pattern. This is silly, because it handles
2764 only 0-9, a-f, and A-F, but the character types table is locale-
2765 specific, which means strange things might happen. A private
2766 table is now used for this - though it costs 256 bytes, a table is
2767 much faster than multiple explicit tests. Of course, the standard
2768 character types table is still used for matching digits in subject
2769 strings against \d.
2770
2771 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2772 ending in _t are). So I've renamed it as ESC_tee.
2773
2774 3. The first argument for regexec() in the POSIX wrapper should have been
2775 defined as "const".
2776
2777 4. Changed pcretest to use malloc() for its buffers so that they can be
2778 Electric Fenced for debugging.
2779
2780 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2781 to read one or more bytes before the start of the subject string. Often this
2782 had no effect on PCRE's behaviour, but in some circumstances it could
2783 provoke a segmentation fault.
2784
2785 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2786 to try to read one or more bytes before the start of the subject string.
2787
2788 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2789 UTF-8 support could misbehave in various ways if the subject string
2790 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2791 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2792 back over UTF-8 characters.)
2793
2794
2795 Version 4.2 14-Apr-03
2796 ---------------------
2797
2798 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2799
2800 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2801 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2802 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2803 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2804 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2805 and BUILD_EXEEXT
2806 Note: automatic setting of the BUILD variables is not yet working
2807 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2808 compile-time but not at link-time
2809 [LINK]: use for linking executables only
2810 make different versions for Windows and non-Windows
2811 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2812 libraries
2813 [LINK_FOR_BUILD]: new variable
2814 [OBJEXT]: use throughout
2815 [EXEEXT]: use throughout
2816 <winshared>: new target
2817 <wininstall>: new target
2818 <dftables.o>: use native compiler
2819 <dftables>: use native linker
2820 <install>: handle Windows platform correctly
2821 <clean>: ditto
2822 <check>: ditto
2823 copy DLL to top builddir before testing
2824
2825 As part of these changes, -no-undefined was removed again. This was reported
2826 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2827 in any case.
2828
2829 3. Some tidies to get rid of compiler warnings:
2830
2831 . In the match_data structure, match_limit was an unsigned long int, whereas
2832 match_call_count was an int. I've made them both unsigned long ints.
2833
2834 . In pcretest the fact that a const uschar * doesn't automatically cast to
2835 a void * provoked a warning.
2836
2837 . Turning on some more compiler warnings threw up some "shadow" variables
2838 and a few more missing casts.
2839
2840 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2841 option, a class that contained a single character with a value between 128
2842 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2843
2844 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2845 option, a class that contained several characters, but with at least one
2846 whose value was between 128 and 255 caused PCRE to crash.
2847
2848
2849 Version 4.1 12-Mar-03
2850 ---------------------
2851
2852 1. Compiling with gcc -pedantic found a couple of places where casts were
2853 needed, and a string in dftables.c that was longer than standard compilers are
2854 required to support.
2855
2856 2. Compiling with Sun's compiler found a few more places where the code could
2857 be tidied up in order to avoid warnings.
2858
2859 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2860 first of these names is deprecated in the latest Autoconf in favour of the name
2861 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2862 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2863 analogy I have changed it to CFLAGS_FOR_BUILD.
2864
2865 4. Added -no-undefined to the linking command in the Makefile, because this is
2866 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2867 linking step for the pcreposix library.
2868
2869 5. PCRE was failing to diagnose the case of two named groups with the same
2870 name.
2871
2872 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2873 literal character that is needed in the subject for a match, and scans along to
2874 ensure that it is present before embarking on the full matching process. This
2875 saves time in cases of nested unlimited repeats that are never going to match.
2876 Problem: the scan can take a lot of time if the subject is very long (e.g.
2877 megabytes), thus penalizing straightforward matches. It is now done only if the
2878 amount of subject to be scanned is less than 1000 bytes.
2879
2880 7. A lesser problem with the same optimization is that it was recording the
2881 first character of an anchored pattern as "needed", thus provoking a search
2882 right along the subject, even when the first match of the pattern was going to
2883 fail. The "needed" character is now not set for anchored patterns, unless it
2884 follows something in the pattern that is of non-fixed length. Thus, it still
2885 fulfils its original purpose of finding quick non-matches in cases of nested
2886 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2887
2888
2889 Version 4.0 17-Feb-03
2890 ---------------------
2891
2892 1. If a comment in an extended regex that started immediately after a meta-item
2893 extended to the end of string, PCRE compiled incorrect data. This could lead to
2894 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2895
2896 2. Moved to autoconf 2.53 and libtool 1.4.2.
2897
2898 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2899 the special perltest8 script is no longer needed - all the tests can be run
2900 from a single perltest script.
2901
2902 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2903 by \s. It has now been removed in PCRE. This means it isn't recognized as
2904 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2905 class [:space:] *does* include VT, thereby creating a mess.
2906
2907 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2908 space and tab.
2909
2910 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2911 its new features into the main test script, reducing the number of scripts.
2912
2913 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2914 were backward compatible, and made the (?i) apply to the whole pattern, as if
2915 /i were given. Now it behaves more logically, and applies the option setting
2916 only to what follows. PCRE has been changed to follow suit. However, if it
2917 finds options settings right at the start of the pattern, it extracts them into
2918 the global options, as before. Thus, they show up in the info data.
2919
2920 8. Added support for the \Q...\E escape sequence. Characters in between are
2921 treated as literals. This is slightly different from Perl in that $ and @ are
2922 also handled as literals inside the quotes. In Perl, they will cause variable
2923 interpolation. Note the following examples:
2924
2925 Pattern PCRE matches Perl matches
2926
2927 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2928 \Qabc\$xyz\E abc\$xyz abc\$xyz
2929 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2930
2931 For compatibility with Perl, \Q...\E sequences are recognized inside character
2932 classes as well as outside them.
2933
2934 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2935 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2936 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2937 signed/unsigned warnings.
2938
2939 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2940 option for pcretest, so I've replaced it by a simple function that does just
2941 that job.
2942
2943 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2944 "pcregrep -".
2945
2946 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2947 Java package. This provides some syntactic sugar for simple cases of what my
2948 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2949 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2950 item, you can use this simplified notation. Note that only makes sense with
2951 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2952 greediness, whatever the setting of the PCRE_UNGREEDY option.
2953
2954 13. A change of greediness default within a pattern was not taking effect at
2955 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2956 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2957 was abstracted outside.
2958
2959 14. PCRE now supports the \G assertion. It is true when the current matching
2960 position is at the start point of the match. This differs from \A when the
2961 starting offset is non-zero. Used with the /g option of pcretest (or similar
2962 code), it works in the same way as it does for Perl's /g option. If all
2963 alternatives of a regex begin with \G, the expression is anchored to the start
2964 match position, and the "anchored" flag is set in the compiled expression.
2965
2966 15. Some bugs concerning the handling of certain option changes within patterns
2967 have been fixed. These applied to options other than (?ims). For example,
2968 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2969 been the other way round. Some of this was related to change 7 above.
2970
2971 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2972 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2973 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2974 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2975
2976 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2977 mode. Unlike ".", it always matches newline, whatever the setting of
2978 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2979 assertions. Perl allows it, but it doesn't (in general) work because it can't
2980 calculate the length of the lookbehind. At least, that's the case for Perl
2981 5.8.0 - I've been told they are going to document that it doesn't work in
2982 future.
2983
2984 18. Added an error diagnosis for escapes that PCRE does not support: these are
2985 \L, \l, \N, \P, \p, \U, \u, and \X.
2986
2987 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2988 reading past the end of the pattern in cases such as /[abcd/.
2989
2990 20. PCRE was getting more memory than necessary for patterns with classes that
2991 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2992
2993 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2994 compiling PCRE for use with Virtual Pascal.
2995
2996 22. Small fix to the Makefile to make it work properly if the build is done
2997 outside the source tree.
2998
2999 23. Added a new extension: a condition to go with recursion. If a conditional
3000 subpattern starts with (?(R) the "true" branch is used if recursion has
3001 happened, whereas the "false" branch is used only at the top level.
3002
3003 24. When there was a very long string of literal characters (over 255 bytes
3004 without UTF support, over 250 bytes with UTF support), the computation of how
3005 much memory was required could be incorrect, leading to segfaults or other
3006 strange effects.
3007
3008 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3009 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3010 there was a subsequent back reference to those brackets. This meant that, for
3011 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3012 possible to check for precisely this case. All we can do is abandon the
3013 optimization if .* occurs inside capturing brackets when there are any back
3014 references whatsoever. (See below for a better fix that came later.)
3015
3016 26. The handling of the optimization for finding the first character of a
3017 non-anchored pattern, and for finding a character that is required later in the
3018 match were failing in some cases. This didn't break the matching; it just
3019 failed to optimize when it could. The way this is done has been re-implemented.
3020
3021 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3022
3023 28. Added a new feature that provides some of the functionality that Perl
3024 provides with (?{...}). The facility is termed a "callout". The way it is done
3025 in PCRE is for the caller to provide an optional function, by setting
3026 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3027 global variable. By default it is unset, which disables all calling out. To get
3028 the function called, the regex must include (?C) at appropriate points. This
3029 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3030 This provides a means of identifying different callout points. When PCRE
3031 reaches such a point in the regex, if pcre_callout has been set, the external
3032 function is called. It is provided with data in a structure called
3033 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3034 matching continues; if it returns a non-zero value, the match at the current
3035 point fails. However, backtracking will occur if possible. [This was changed
3036 later and other features added - see item 49 below.]
3037
3038 29. pcretest is upgraded to test the callout functionality. It provides a
3039 callout function that displays information. By default, it shows the start of
3040 the match and the current position in the text. There are some new data escapes
3041 to vary what happens:
3042
3043 \C+ in addition, show current contents of captured substrings
3044 \C- do not supply a callout function
3045 \C!n return 1 when callout number n is reached
3046 \C!n!m return 1 when callout number n is reached for the mth time
3047
3048 30. If pcregrep was called with the -l option and just a single file name, it
3049 output "<stdin>" if a match was found, instead of the file name.
3050
3051 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3052 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3053 pcre_exec(). This saves a malloc/free per call. The default value of
3054 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3055 when configuring.
3056
3057 32. The default maximum size of a compiled pattern is 64K. There have been a
3058 few cases of people hitting this limit. The code now uses macros to handle the
3059 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3060 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3061 configuring. Tests 2 and 5 work only with 2-byte links because they output
3062 debugging information about compiled patterns.
3063
3064 33. Internal code re-arrangements:
3065
3066 (a) Moved the debugging function for printing out a compiled regex into
3067 its own source file (printint.c) and used #include to pull it into
3068 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3069 separate copies.
3070
3071 (b) Defined the list of op-code names for debugging as a macro in
3072 internal.h so that it is next to the definition of the opcodes.
3073
3074 (c) Defined a table of op-code lengths for simpler skipping along compiled
3075 code. This is again a macro in internal.h so that it is next to the
3076 definition of the opcodes.
3077
3078 34. Added support for recursive calls to individual subpatterns, along the
3079 lines of Robin Houston's patch (but implemented somewhat differently).
3080
3081 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3082 allow it to read and process whole directories in Win32. This code was
3083 contributed by Lionel Fourquaux; it has not been tested by me.
3084
3085 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3086 used to name a group. Names consist of alphanumerics and underscores, and must
3087 be unique. Back references use the syntax (?P=name) and recursive calls use
3088 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3089 numbers. The function pcre_fullinfo() can be used after compilation to extract
3090 a name/number map. There are three relevant calls:
3091
3092 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3093 PCRE_INFO_NAMECOUNT yields the number of entries
3094 PCRE_INFO_NAMETABLE yields a pointer to the map.
3095
3096 The map is a vector of fixed-size entries. The size of each entry depends on
3097 the length of the longest name used. The first two bytes of each entry are the
3098 group number, most significant byte first. There follows the corresponding
3099 name, zero terminated. The names are in alphabetical order.
3100
3101 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3102 case instead of 255. Making it the same both with and without UTF-8 support
3103 means that the same test output works with both.
3104
3105 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3106 calling malloc() with a zero argument.
3107
3108 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3109 optimization. I've improved things by keeping a bitmap of backreferences with
3110 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3111 fact referenced, the optimization can be applied. It is unlikely that a
3112 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3113 the match to follow \n) will appear inside brackets with a number greater than
3114 31, but if it does, any back reference > 31 suppresses the optimization.
3115
3116 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3117 of disabling numbered capturing parentheses. Any opening parenthesis that is
3118 not followed by ? behaves as if it were followed by ?: but named parentheses
3119 can still be used for capturing (and they will acquire numbers in the usual
3120 way).
3121
3122 41. Redesigned the return codes from the match() function into yes/no/error so
3123 that errors can be passed back from deep inside the nested calls. A malloc
3124 failure while inside a recursive subpattern call now causes the
3125 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3126
3127 42. It is now possible to set a limit on the number of times the match()
3128 function is called in a call to pcre_exec(). This facility makes it possible to
3129 limit the amount of recursion and backtracking, though not in a directly
3130 obvious way, because the match() function is used in a number of different
3131 circumstances. The count starts from zero for each position in the subject
3132 string (for non-anchored patterns). The default limit is, for compatibility, a
3133 large number, namely 10 000 000. You can change this in two ways:
3134
3135 (a) When configuring PCRE before making, you can use --with-match-limit=n
3136 to set a default value for the compiled library.
3137
3138 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3139 a different value is set. See 45 below.
3140
3141 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3142
3143 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3144 of things that can be changed at compile time. The first argument specifies
3145 what is wanted and the second points to where the information is to be placed.
3146 The current list of available information is:
3147
3148 PCRE_CONFIG_UTF8
3149
3150 The output is an integer that is set to one if UTF-8 support is available;
3151 otherwise it is set to zero.
3152
3153 PCRE_CONFIG_NEWLINE
3154
3155 The output is an integer that it set to the value of the code that is used for
3156 newline. It is either LF (10) or CR (13).
3157
3158 PCRE_CONFIG_LINK_SIZE
3159
3160 The output is an integer that contains the number of bytes used for internal
3161 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3162
3163 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3164
3165 The output is an integer that contains the threshold above which the POSIX
3166 interface uses malloc() for output vectors. See item 31 above.
3167
3168 PCRE_CONFIG_MATCH_LIMIT
3169
3170 The output is an unsigned integer that contains the default limit of the number
3171 of match() calls in a pcre_exec() execution. See 42 above.
3172
3173 44. pcretest has been upgraded by the addition of the -C option. This causes it
3174 to extract all the available output from the new pcre_config() function, and to
3175 output it. The program then exits immediately.
3176
3177 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3178 order to support additional features. One way would have been to define
3179 pcre_exec2() (for example) with extra arguments, but this would not have been
3180 extensible, and would also have required all calls to the original function to
3181 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3182 is used for passing in "extra" data from pcre_study().
3183
3184 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3185 contains the following fields:
3186
3187 flags a bitmap indicating which of the following fields are set
3188 study_data opaque data from pcre_study()
3189 match_limit a way of specifying a limit on match() calls for a specific
3190 call to pcre_exec()
3191 callout_data data for callouts (see 49 below)
3192
3193 The flag bits are also defined in pcre.h, and are
3194
3195 PCRE_EXTRA_STUDY_DATA
3196 PCRE_EXTRA_MATCH_LIMIT
3197 PCRE_EXTRA_CALLOUT_DATA
3198
3199 The pcre_study() function now returns one of these new pcre_extra blocks, with
3200 the actual study data pointed to by the study_data field, and the
3201 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3202 before. That is, this change is entirely upwards-compatible and requires no
3203 change to existing code.
3204
3205 If you want to pass in additional data to pcre_exec(), you can either place it
3206 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3207 block.
3208
3209 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3210 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3211 times with different match limits, until it finds the minimum value needed for
3212 pcre_exec() to complete. The value is then output. This can be instructive; for
3213 most simple matches the number is quite small, but for pathological cases it
3214 gets very large very quickly.
3215
3216 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3217 returns the size of the data block pointed to by the study_data field in a
3218 pcre_extra block, that is, the value that was passed as the argument to
3219 pcre_malloc() when PCRE was getting memory in which to place the information
3220 created by pcre_study(). The fourth argument should point to a size_t variable.
3221 pcretest has been extended so that this information is shown after a successful
3222 pcre_study() call when information about the compiled regex is being displayed.
3223
3224 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3225 because what follows is always an absolute path. (Later: it turns out that this
3226 is more than cosmetic for MinGW, because it doesn't like empty path
3227 components.)
3228
3229 49. Some changes have been made to the callout feature (see 28 above):
3230
3231 (i) A callout function now has three choices for what it returns:
3232
3233 0 => success, carry on matching
3234 > 0 => failure at this point, but backtrack if possible
3235 < 0 => serious error, return this value from pcre_exec()
3236
3237 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3238 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3239 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3240 use by callout functions. It will never be used by PCRE itself.
3241
3242 (ii) The pcre_extra structure (see 45 above) has a void * field called
3243 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3244 pcre_callout_block structure has a field of the same name. The contents of
3245 the field passed in the pcre_extra structure are passed to the callout
3246 function in the corresponding field in the callout block. This makes it
3247 easier to use the same callout-containing regex from multiple threads. For
3248 testing, the pcretest program has a new data escape
3249
3250 \C*n pass the number n (may be negative) as callout_data
3251
3252 If the callout function in pcretest receives a non-zero value as
3253 callout_data, it returns that value.
3254
3255 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3256 there were some redundant $(CFLAGS) in commands that are now specified as
3257 $(LINK), which already includes $(CFLAGS).
3258
3259 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3260 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3261 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3262 one-byte characters throughout. Note that case-insensitive matching applies
3263 only to characters whose values are less than 256. PCRE doesn't support the
3264 notion of cases for higher-valued characters.
3265
3266 (i) A character class whose characters are all within 0-255 is handled as
3267 a bit map, and the map is inverted for negative classes. Previously, a
3268 character > 255 always failed to match such a class; however it should
3269 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3270
3271 (ii) A negated character class with a single character < 255 is coded as
3272 "not this character" (OP_NOT). This wasn't working properly when the test
3273 character was multibyte, either singly or repeated.
3274
3275 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3276 mode, for example: \x{100}{2,3}.
3277
3278 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3279 singly or repeated) now correctly test multibyte characters. However,
3280 PCRE doesn't recognize any characters with values greater than 255 as
3281 digits, spaces, or word characters. Such characters always match \D, \S,
3282 and \W, and never match \d, \s, or \w.
3283
3284 (v) Classes may now contain characters and character ranges with values
3285 greater than 255. For example: [ab\x{100}-\x{400}].
3286
3287 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3288 PCRE in UTF-8 mode.
3289
3290 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3291 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3292 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3293 value.)
3294
3295 53. The single man page has become too large. I have therefore split it up into
3296 a number of separate man pages. These also give rise to individual HTML pages;
3297 these are now put in a separate directory, and there is an index.html page that
3298 lists them all. Some hyperlinking between the pages has been installed.
3299
3300 54. Added convenience functions for handling named capturing parentheses.
3301
3302 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3303 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3304 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3305 are faulted.
3306
3307 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3308 calling configure. These values are used when compiling the dftables.c program
3309 which is run to generate the source of the default character tables. They
3310 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3311 you will need to set these values.
3312
3313 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3314
3315
3316 Version 3.9 02-Jan-02
3317 ---------------------
3318
3319 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3320
3321 2. If --disable-static was given, the building process failed when trying to
3322 build pcretest and pcregrep. (For some reason it was using libtool to compile
3323 them, which is not right, as they aren't part of the library.)
3324
3325
3326 Version 3.8 18-Dec-01
3327 ---------------------
3328
3329 1. The experimental UTF-8 code was completely screwed up. It was packing the
3330 bytes in the wrong order. How dumb can you get?
3331
3332
3333 Version 3.7 29-Oct-01
3334 ---------------------
3335
3336 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3337 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3338 this didn't happen under Solaris 8, where I normally test things.
3339
3340 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3341 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3342 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3343 this fix an hour or so after the initial 3.7 release.)
3344
3345
3346 Version 3.6 23-Oct-01
3347 ---------------------
3348
3349 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3350 offsets passed as NULL with zero offset count.
3351
3352 2. The config.guess and config.sub files had not been updated when I moved to
3353 the latest autoconf.
3354
3355
3356 Version 3.5 15-Aug-01
3357 ---------------------
3358
3359 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3360 had been forgotten.
3361
3362 2. By using declared but undefined structures, we can avoid using "void"
3363 definitions in pcre.h while keeping the internal definitions of the structures
3364 private.
3365
3366 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3367 user point of view, this means that both static and shared libraries are built
3368 by default, but this can be individually controlled. More of the work of
3369 handling this static/shared cases is now inside libtool instead of PCRE's make
3370 file.
3371
3372 4. The pcretest utility is now installed along with pcregrep because it is
3373 useful for users (to test regexs) and by doing this, it automatically gets
3374 relinked by libtool. The documentation has been turned into a man page, so
3375 there are now .1, .txt, and .html versions in /doc.
3376
3377 5. Upgrades to pcregrep:
3378 (i) Added long-form option names like gnu grep.
3379 (ii) Added --help to list all options with an explanatory phrase.
3380 (iii) Added -r, --recursive to recurse into sub-directories.
3381 (iv) Added -f, --file to read patterns from a file.
3382
3383 6. pcre_exec() was referring to its "code" argument before testing that
3384 argument for NULL (and giving an error if it was NULL).
3385
3386 7. Upgraded Makefile.in to allow for compiling in a different directory from
3387 the source directory.
3388
3389 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3390 options bits, the pointer it was passed was to an int instead of to an unsigned
3391 long int. This mattered only on 64-bit systems.
3392
3393 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3394 generated) instead of pcre.in, which it its source. Also made the same change
3395 in several of the .c files.
3396
3397 10. A new release of gcc defines printf() as a macro, which broke pcretest
3398 because it had an ifdef in the middle of a string argument for printf(). Fixed
3399 by using separate calls to printf().
3400
3401 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3402 script, to force use of CR or LF instead of \n in the source. On non-Unix
3403 systems, the value can be set in config.h.
3404
3405 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3406 absolute limit. Changed the text of the error message to make this clear, and
3407 likewise updated the man page.
3408
3409 13. The limit of 99 on the number of capturing subpatterns has been removed.
3410 The new limit is 65535, which I hope will not be a "real" limit.
3411
3412
3413 Version 3.4 22-Aug-00
3414 ---------------------
3415
3416 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3417
3418 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3419
3420
3421 Version 3.3 01-Aug-00
3422 ---------------------
3423
3424 1. If an octal character was given, but the value was greater than \377, it
3425 was not getting masked to the least significant bits, as documented. This could
3426 lead to crashes in some systems.
3427
3428 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3429 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3430
3431 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3432 These just pass their arguments on to (pcre_free)(), but they are provided
3433 because some uses of PCRE bind it to non-C systems that can call its functions,
3434 but cannot call free() or pcre_free() directly.
3435
3436 4. Add "make test" as a synonym for "make check". Corrected some comments in
3437 the Makefile.
3438
3439 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3440 Makefile.
3441
3442 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3443 command called pgrep for grepping around the active processes.
3444
3445 7. Added the beginnings of support for UTF-8 character strings.
3446
3447 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3448 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3449 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3450 out for the ar command.)
3451
3452
3453 Version 3.2 12-May-00
3454 ---------------------
3455
3456 This is purely a bug fixing release.
3457
3458 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3459 of ZA. This was just one example of several cases that could provoke this bug,
3460 which was introduced by change 9 of version 2.00. The code for breaking
3461 infinite loops after an iteration that matches an empty string was't working
3462 correctly.
3463
3464 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3465 when matched against abbab (for example). After matching an empty string, it
3466 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3467 caused it to match further down the string than it should.
3468
3469 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3470 was there because it doesn't seem to be needed, and it causes trouble on some
3471 systems, as it is not a Standard C header. It has been removed.
3472
3473 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3474 were reported on the Macintosh. The changes were from
3475
3476 while ((c = *(++ptr)) != 0 && c != '\n');
3477 to
3478 while ((c = *(++ptr)) != 0 && c != '\n') ;
3479
3480 Totally extraordinary, but if that's what it takes...
3481
3482 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3483 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3484 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3485 assumes the way PCRE uses memmove() (always moving upwards).
3486
3487 6. PCRE is being used in one environment where strchr() is not available. There
3488 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3489 faster code anyway.
3490
3491
3492 Version 3.1 09-Feb-00
3493 ---------------------
3494
3495 The only change in this release is the fixing of some bugs in Makefile.in for
3496 the "install" target:
3497
3498 (1) It was failing to install pcreposix.h.
3499
3500 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3501
3502
3503 Version 3.0 01-Feb-00
3504 ---------------------
3505
3506 1. Add support for the /+ modifier to perltest (to output $` like it does in
3507 pcretest).
3508
3509 2. Add support for the /g modifier to perltest.
3510
3511 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3512 matches null strings.
3513
3514 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3515 pattern. Perl treats empty patterns specially - it reuses the most recent
3516 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3517 effect.
3518
3519 5. The POSIX interface was broken in that it was just handing over the POSIX
3520 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3521 required a bigger vector, with some working space on the end. This means that
3522 the POSIX wrapper now has to get and free some memory, and copy the results.
3523
3524 6. Added some simple autoconf support, placing the test data and the
3525 documentation in separate directories, re-organizing some of the
3526 information files, and making it build pcre-config (a GNU standard). Also added
3527 libtool support for building PCRE as a shared library, which is now the
3528 default.
3529
3530 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3531 09 are not valid octal constants. Single digits will be used for minor values
3532 less than 10.
3533
3534 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3535 existing programs that set these in the POSIX interface can use PCRE without
3536 modification.
3537
3538 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3539 return all that pcre_info() returns, plus additional data. The pcre_info()
3540 function is retained for compatibility, but is considered to be obsolete.
3541
3542 10. Added experimental recursion feature (?R) to handle one common case that
3543 Perl 5.6 will be able to do with (?p{...}).
3544
3545 11. Added support for POSIX character classes like [:alpha:], which Perl is
3546 adopting.
3547
3548
3549 Version 2.08 31-Aug-99
3550 ----------------------
3551
3552 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3553 trying to match at the startoffset position, but instead was moving forward to
3554 the next newline as if a previous match had failed.
3555
3556 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3557 and could get into a loop if a null string was matched other than at the start
3558 of the subject.
3559
3560 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3561 be distinguished at compile time, and for completeness also added PCRE_DATE.
3562
3563 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3564 in GnuWin32 environments.
3565
3566
3567 Version 2.07 29-Jul-99
3568 ----------------------
3569
3570 1. The documentation is now supplied in plain text form and HTML as well as in
3571 the form of man page sources.
3572
3573 2. C++ compilers don't like assigning (void *) values to other pointer types.
3574 In particular this affects malloc(). Although there is no problem in Standard
3575 C, I've put in casts to keep C++ compilers happy.
3576
3577 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3578 should be (const char *).
3579
3580 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3581 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3582 However, I haven't made this a standard facility. The documentation doesn't
3583 mention it, and the Makefile doesn't support it.
3584
3585 5. The Makefile now contains an "install" target, with editable destinations at
3586 the top of the file. The pcretest program is not installed.
3587
3588 6. pgrep -V now gives the PCRE version number and date.
3589
3590 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3591 causing the entire string to be ignored, instead of just the last character.
3592
3593 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3594 non-matching string, it can take a very, very long time, even for strings of
3595 quite modest length, because of the nested recursion. PCRE now does better in
3596 some of these cases. It does this by remembering the last required literal
3597 character in the pattern, and pre-searching the subject to ensure it is present
3598 before running the real match. In other words, it applies a heuristic to detect
3599 some types of certain failure quickly, and in the above example, if presented
3600 with a string that has no trailing " it gives "no match" very quickly.
3601
3602 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3603 other alternatives are tried instead.
3604
3605
3606 Version 2.06 09-Jun-99
3607 ----------------------
3608
3609 1. Change pcretest's output for amount of store used to show just the code
3610 space, because the remainder (the data block) varies in size between 32-bit and
3611 64-bit systems.
3612
3613 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3614 start matching at. This allows lookbehinds to work when searching for multiple
3615 occurrences in a string.
3616
3617 3. Added additional options to pcretest for testing multiple occurrences:
3618
3619 /+ outputs the rest of the string that follows a match
3620 /g loops for multiple occurrences, using the new startoffset argument
3621 /G loops for multiple occurrences by passing an incremented pointer
3622
3623 4. PCRE wasn't doing the "first character" optimization for patterns starting
3624 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3625 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3626 the letter 'x'. On long subject strings, this gives a significant speed-up.
3627
3628
3629 Version 2.05 21-Apr-99
3630 ----------------------
3631
3632 1. Changed the type of magic_number from int to long int so that it works
3633 properly on 16-bit systems.
3634
3635 2. Fixed a bug which caused patterns starting with .* not to work correctly
3636 when the subject string contained newline characters. PCRE was assuming
3637 anchoring for such patterns in all cases, which is not correct because .* will
3638 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3639 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3640 must be retried after every newline in the subject.
3641
3642
3643 Version 2.04 18-Feb-99
3644 ----------------------
3645
3646 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3647 computation of the store needed to hold the pattern was incorrect (too large).
3648 If such patterns were nested a few deep, this could multiply and become a real
3649 problem.
3650
3651 2. Added /M option to pcretest to show the memory requirement of a specific
3652 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3653
3654 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3655 compiled in such a way that the backtracking after subsequent failure was
3656 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3657 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3658
3659
3660 Version 2.03 02-Feb-99
3661 ----------------------
3662
3663 1. Fixed typo and small mistake in man page.
3664
3665 2. Added 4th condition (GPL supersedes if conflict) and created separate
3666 LICENCE file containing the conditions.
3667
3668 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3669 Perl, that is the internal \ allows the delimiter to be included in the
3670 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3671 the final delimiter, add \ to the end of the pattern (to test the error).
3672
3673 4. Added the convenience functions for extracting substrings after a successful
3674 match. Updated pcretest to make it able to test these functions.
3675
3676
3677 Version 2.02 14-Jan-99
3678 ----------------------
3679
3680 1. Initialized the working variables associated with each extraction so that
3681 their saving and restoring doesn't refer to uninitialized store.
3682
3683 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3684 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3685 fix the problem.
3686
3687 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3688 calls, and wasn't printing the correct value for compiling calls. Increased the
3689 default value of LOOPREPEAT, and the number of significant figures in the
3690 times.
3691
3692 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3693
3694 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3695 a building problem on Windows NT with a FAT file system.
3696
3697
3698 Version 2.01 21-Oct-98
3699 ----------------------
3700
3701 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3702 to character tables built by pcre_maketables() in the current locale. If NULL
3703 is passed, the default tables are used.
3704
3705
3706 Version 2.00 24-Sep-98
3707 ----------------------
3708
3709 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3710 it any more.
3711
3712 2. Allow quantification of (?>) groups, and make it work correctly.
3713
3714 3. The first character computation wasn't working for (?>) groups.
3715
3716 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3717 end of the subject) and add 5.005's \z, which really does match only at the
3718 very end of the subject.
3719
3720 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3721
3722 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3723 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3724 localized options. All options to pcre_study() were also removed.
3725
3726 7. Add other new features from 5.005:
3727
3728 $(?<= positive lookbehind
3729 $(?<! negative lookbehind
3730 (?imsx-imsx) added the unsetting capability
3731 such a setting is global if at outer level; local otherwise
3732 (?imsx-imsx:) non-capturing groups with option setting
3733 (?(cond)re|re) conditional pattern matching
3734
3735 A backreference to itself in a repeated group matches the previous
3736 captured string.
3737
3738 8. General tidying up of studying (both automatic and via "study")
3739 consequential on the addition of new assertions.
3740
3741 9. As in 5.005, unlimited repeated groups that could match an empty substring
3742 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3743 runtime if any iteration does actually match an empty substring.
3744
3745 10. Include the RunTest script in the distribution.
3746
3747 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3748 discrepancies, some of which were old and were also with respect to 5.004. They
3749 have now been fixed.
3750
3751
3752 Version 1.09 28-Apr-98
3753 ----------------------
3754
3755 1. A negated single character class followed by a quantifier with a minimum
3756 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3757 program crashes, or just wrong answers. This did not apply to negated classes
3758 containing more than one character, or to minima other than one.
3759
3760
3761 Version 1.08 27-Mar-98
3762 ----------------------
3763
3764 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3765
3766 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3767 latter must appear before anything that relies on it in the pattern.
3768
3769
3770 Version 1.07 16-Feb-98
3771 ----------------------
3772
3773 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3774 repeat of a potentially empty string).
3775
3776
3777 Version 1.06 23-Jan-98
3778 ----------------------
3779
3780 1. Added Markus Oberhumer's little patches for C++.
3781
3782 2. Literal strings longer than 255 characters were broken.
3783
3784
3785 Version 1.05 23-Dec-97
3786 ----------------------
3787
3788 1. Negated character classes containing more than one character were failing if
3789 PCRE_CASELESS was set at run time.
3790
3791
3792 Version 1.04 19-Dec-97
3793 ----------------------
3794
3795 1. Corrected the man page, where some "const" qualifiers had been omitted.
3796
3797 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3798 input syntax.
3799
3800 3. Fixed memory leak which occurred when a regex with back references was
3801 matched with an offsets vector that wasn't big enough. The temporary memory
3802 that is used in this case wasn't being freed if the match failed.
3803
3804 4. Tidied pcretest to ensure it frees memory that it gets.
3805
3806 5. Temporary memory was being obtained in the case where the passed offsets
3807 vector was exactly big enough.
3808
3809 6. Corrected definition of offsetof() from change 5 below.
3810
3811 7. I had screwed up change 6 below and broken the rules for the use of
3812 setjmp(). Now fixed.
3813
3814
3815 Version 1.03 18-Dec-97
3816 ----------------------
3817
3818 1. A erroneous regex with a missing opening parenthesis was correctly
3819 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3820 on some systems.
3821
3822 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3823 it was reported that one broken compiler failed on the former because "code" is
3824 also an independent variable.
3825
3826 3. The erroneous regex a[]b caused an array overrun reference.
3827
3828 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3829 fail on data ending with that character. (It was going on too far, and checking
3830 the next character, typically a binary zero.) This was specific to the
3831 optimized code for single-character negative classes.
3832
3833 5. Added a contributed patch from the TIN world which does the following:
3834
3835 + Add an undef for memmove, in case the the system defines a macro for it.
3836
3837 + Add a definition of offsetof(), in case there isn't one. (I don't know
3838 the reason behind this - offsetof() is part of the ANSI standard - but
3839 it does no harm).
3840
3841 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3842 most of the places where whitespace preceded '#'. I have given up and
3843 allowed the remaining 2 cases to be at the margin.
3844
3845 + Rename some variables in pcre to eliminate shadowing. This seems very
3846 pedantic, but does no harm, of course.
3847
3848 6. Moved the call to setjmp() into its own function, to get rid of warnings
3849 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3850
3851 7. Constructs such as \d{8,} were compiling into the equivalent of
3852 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3853 outcome, but in this particular case used more store than had been allocated,
3854 which caused the bug to be discovered because it threw up an internal error.
3855
3856 8. The debugging code in both pcre and pcretest for outputting the compiled
3857 form of a regex was going wrong in the case of back references followed by
3858 curly-bracketed repeats.
3859
3860
3861 Version 1.02 12-Dec-97
3862 ----------------------
3863
3864 1. Typos in pcre.3 and comments in the source fixed.
3865
3866 2. Applied a contributed patch to get rid of places where it used to remove
3867 'const' from variables, and fixed some signed/unsigned and uninitialized
3868 variable warnings.
3869
3870 3. Added the "runtest" target to Makefile.
3871
3872 4. Set default compiler flag to -O2 rather than just -O.
3873
3874
3875 Version 1.01 19-Nov-97
3876 ----------------------
3877
3878 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3879 like /([ab]*)*/, that is, for classes with more than one character in them.
3880
3881 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3882 as /((?>a*))*/ (a PCRE_EXTRA facility).
3883
3884
3885 Version 1.00 18-Nov-97
3886 ----------------------
3887
3888 1. Added compile-time macros to support systems such as SunOS4 which don't have
3889 memmove() or strerror() but have other things that can be used instead.
3890
3891 2. Arranged that "make clean" removes the executables.
3892
3893
3894 Version 0.99 27-Oct-97
3895 ----------------------
3896
3897 1. Fixed bug in code for optimizing classes with only one character. It was
3898 initializing a 32-byte map regardless, which could cause it to run off the end
3899 of the memory it had got.
3900
3901 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3902
3903
3904 Version 0.98 22-Oct-97
3905 ----------------------
3906
3907 1. Fixed bug in code for handling temporary memory usage when there are more
3908 back references than supplied space in the ovector. This could cause segfaults.
3909
3910
3911 Version 0.97 21-Oct-97
3912 ----------------------
3913
3914 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3915
3916 2. Optimized negated single characters not to use a bit map.
3917
3918 3. Brought error texts together as macro definitions; clarified some of them;
3919 fixed one that was wrong - it said "range out of order" when it meant "invalid
3920 escape sequence".
3921
3922 4. Changed some char * arguments to const char *.
3923
3924 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3925
3926 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3927 pcretest.
3928
3929
3930 Version 0.96 16-Oct-97
3931 ----------------------
3932
3933 1. Added a simple "pgrep" utility to the distribution.
3934
3935 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3936 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3937 where "ddd" means "one or more decimal digits".
3938
3939 3. Fixed serious bug. If a pattern had a back reference, but the call to
3940 pcre_exec() didn't supply a large enough ovector to record the related
3941 identifying subpattern, the match always failed. PCRE now remembers the number
3942 of the largest back reference, and gets some temporary memory in which to save
3943 the offsets during matching if necessary, in order to ensure that
3944 backreferences always work.
3945
3946 4. Increased the compatibility with Perl in a number of ways:
3947
3948 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3949 to request this handling. The option can be set at compile or exec time.
3950
3951 (b) $ matches before a terminating newline by default; an option
3952 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3953 mode). The option can be set at compile or exec time.
3954
3955 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3956 the same as Perl's. If the decimal number it represents is less than 10
3957 or there aren't that many previous left capturing parentheses, an octal
3958 escape is read. Inside a character class, it's always an octal escape,
3959 even if it is a single digit.
3960
3961 (d) An escaped but undefined alphabetic character is taken as a literal,
3962 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3963 escapes.
3964
3965 (e) {0} is now permitted. (The previous item is removed from the compiled
3966 pattern).
3967
3968 5. Changed all the names of code files so that the basic parts are no longer
3969 than 10 characters, and abolished the teeny "globals.c" file.
3970
3971 6. Changed the handling of character classes; they are now done with a 32-byte
3972 bit map always.
3973
3974 7. Added the -d and /D options to pcretest to make it possible to look at the
3975 internals of compilation without having to recompile pcre.
3976
3977
3978 Version 0.95 23-Sep-97
3979 ----------------------
3980
3981 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3982 \x20 at the start of a run of normal characters. These were being treated as
3983 real characters, instead of the source characters being re-checked.
3984
3985
3986 Version 0.94 18-Sep-97
3987 ----------------------
3988
3989 1. The functions are now thread-safe, with the caveat that the global variables
3990 containing pointers to malloc() and free() or alternative functions are the
3991 same for all threads.
3992
3993 2. Get pcre_study() to generate a bitmap of initial characters for non-
3994 anchored patterns when this is possible, and use it if passed to pcre_exec().
3995
3996
3997 Version 0.93 15-Sep-97
3998 ----------------------
3999
4000 1. /(b)|(:+)/ was computing an incorrect first character.
4001
4002 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4003 but not actually doing anything yet.
4004
4005 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4006 as Perl does (e.g. [-az] or [az-]).
4007
4008 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4009 all possible positions.
4010
4011 5. Split up into different modules to avoid including unneeded functions in a
4012 compiled binary. However, compile and exec are still in one module. The "study"
4013 function is split off.
4014
4015 6. The character tables are now in a separate module whose source is generated
4016 by an auxiliary program - but can then be edited by hand if required. There are
4017 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4018 toupper() in the code.
4019
4020 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4021 make them global. Abolish the function for setting them, as the caller can now
4022 set them directly.
4023
4024
4025 Version 0.92 11-Sep-97
4026 ----------------------
4027
4028 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4029 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4030
4031 2. Caseless matching was not working in character classes if the characters in
4032 the pattern were in upper case.
4033
4034 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4035
4036 4. Make PCRE_ANCHORED public and accept as a compile option.
4037
4038 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4039 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4040 pass them.
4041
4042 6. Give an error if bad option bits passed at compile or run time.
4043
4044 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4045 pcretest to cause it to pass that flag.
4046
4047 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4048 options, and the first character, if set.
4049
4050 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4051
4052
4053 Version 0.91 10-Sep-97
4054 ----------------------
4055
4056 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4057 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4058
4059 2. PCRE was looping on encountering an indefinitely repeated back reference to
4060 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4061 Perl does - treats the match as successful.
4062
4063 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5