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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 583 - (show annotations) (download)
Tue Jan 11 16:49:55 2011 UTC (3 years, 8 months ago) by ph10
File size: 196803 byte(s)
Add script to check for man page markup typos. Fix some typos.

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

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