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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 580 - (show annotations) (download)
Fri Nov 26 11:16:43 2010 UTC (3 years, 11 months ago) by ph10
File size: 196599 byte(s)
Patches to avoid build problems in some Borland environments.

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

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