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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 585 - (show annotations) (download)
Wed Jan 12 17:21:11 2011 UTC (3 years, 7 months ago) by ph10
File size: 197324 byte(s)
Fix pcregrep bug when --colour was used with -v.

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

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