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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 584 - (show annotations) (download)
Wed Jan 12 11:41:55 2011 UTC (3 years, 9 months ago) by ph10
File size: 197075 byte(s)
Fix pcregrep cast issue for 64-bit big-endian systems.

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

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