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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 567 - (show annotations) (download)
Sat Nov 6 17:10:00 2010 UTC (3 years, 10 months ago) by ph10
File size: 194606 byte(s)
Test for ridiculous values of starting offsets; tidy UTF-8 code.

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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12