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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 592 - (show annotations)
Sat Apr 30 17:37:37 2011 UTC (4 years, 2 months ago) by ph10
File size: 198396 byte(s)
Error occurred while calculating annotation data.
Update tables to Unicode 6.0.0.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.13 30-Apr-2011
5 ------------------------
6
7 1. The Unicode data tables have been updated to Unicode 6.0.0.
8
9
10 Version 8.12 15-Jan-2011
11 ------------------------
12
13 1. Fixed some typos in the markup of the man pages, and wrote a script that
14 checks for such things as part of the documentation building process.
15
16 2. On a big-endian 64-bit system, pcregrep did not correctly process the
17 --match-limit and --recursion-limit options (added for 8.11). In
18 particular, this made one of the standard tests fail. (The integer value
19 went into the wrong half of a long int.)
20
21 3. If the --colour option was given to pcregrep with -v (invert match), it
22 did strange things, either producing crazy output, or crashing. It should,
23 of course, ignore a request for colour when reporting lines that do not
24 match.
25
26 4. Another pcregrep bug caused similar problems if --colour was specified with
27 -M (multiline) and the pattern match finished with a line ending.
28
29 5. In pcregrep, when a pattern that ended with a literal newline sequence was
30 matched in multiline mode, the following line was shown as part of the
31 match. This seems wrong, so I have changed it.
32
33 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
34 the check for further matches in the same line (so they could be coloured)
35 to overrun the end of the current line. If another match was found, it was
36 incorrectly shown (and then shown again when found in the next line).
37
38 7. If pcregrep was compiled under Windows, there was a reference to the
39 function pcregrep_exit() before it was defined. I am assuming this was
40 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
41 reported by a user. I've moved the definition above the reference.
42
43
44 Version 8.11 10-Dec-2010
45 ------------------------
46
47 1. (*THEN) was not working properly if there were untried alternatives prior
48 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
49 backtracked to try for "b" instead of moving to the next alternative branch
50 at the same level (in this case, to look for "c"). The Perl documentation
51 is clear that when (*THEN) is backtracked onto, it goes to the "next
52 alternative in the innermost enclosing group".
53
54 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
55 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
56 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
57 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
58 (*THEN).
59
60 3. If \s appeared in a character class, it removed the VT character from
61 the class, even if it had been included by some previous item, for example
62 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
63 of \s, but is part of the POSIX "space" class.)
64
65 4. A partial match never returns an empty string (because you can always
66 match an empty string at the end of the subject); however the checking for
67 an empty string was starting at the "start of match" point. This has been
68 changed to the "earliest inspected character" point, because the returned
69 data for a partial match starts at this character. This means that, for
70 example, /(?<=abc)def/ gives a partial match for the subject "abc"
71 (previously it gave "no match").
72
73 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
74 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
75 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
76 has an implication that the given string is incomplete (because a partial
77 match is preferred over a full match). For this reason, these items now
78 give a partial match in this situation. [Aside: previously, the one case
79 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
80 match rather than a full match, which was wrong by the old rules, but is
81 now correct.]
82
83 6. There was a bug in the handling of #-introduced comments, recognized when
84 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
85 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
86 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
87 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
88 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
89 places in pcre_compile().
90
91 7. Related to (6) above, when pcre_compile() was skipping #-introduced
92 comments when looking ahead for named forward references to subpatterns,
93 the only newline sequence it recognized was NL. It now handles newlines
94 according to the set newline convention.
95
96 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
97 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
98 cater for a lack of strerror(). These oversights have been fixed.
99
100 9. Added --match-limit and --recursion-limit to pcregrep.
101
102 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
103
104 11. When the -o option was used, pcregrep was setting a return code of 1, even
105 when matches were found, and --line-buffered was not being honoured.
106
107 12. Added an optional parentheses number to the -o and --only-matching options
108 of pcregrep.
109
110 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
111 can match an empty string. The code to do it in pcretest and pcredemo
112 needed fixing:
113
114 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
115 only one byte after an empty string match just before CRLF (this case
116 just got forgotten; "any" and "anycrlf" were OK).
117
118 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
119 mode when an empty string match preceded an ASCII character followed by
120 a non-ASCII character. (The code for advancing by one character rather
121 than one byte was nonsense.)
122
123 (c) The pcredemo.c sample program did not have any code at all to handle
124 the cases when CRLF is a valid newline sequence.
125
126 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
127 as a starting offset was within the subject string. There is now a new
128 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
129 negative or greater than the length of the string. In order to test this,
130 pcretest is extended to allow the setting of negative starting offsets.
131
132 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
133 starting offset points to the beginning of a UTF-8 character was
134 unnecessarily clumsy. I tidied it up.
135
136 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
137 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
138
139 17. Nobody had reported that the --include_dir option, which was added in
140 release 7.7 should have been called --include-dir (hyphen, not underscore)
141 for compatibility with GNU grep. I have changed it to --include-dir, but
142 left --include_dir as an undocumented synonym, and the same for
143 --exclude-dir, though that is not available in GNU grep, at least as of
144 release 2.5.4.
145
146 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
147 characters from a string of bytes) have been redefined so as not to use
148 loops, in order to improve performance in some environments. At the same
149 time, I abstracted some of the common code into auxiliary macros to save
150 repetition (this should not affect the compiled code).
151
152 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
153 compile-time error is now given if \c is not followed by an ASCII
154 character, that is, a byte less than 128. (In EBCDIC mode, the code is
155 different, and any byte value is allowed.)
156
157 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
158 START_OPTIMIZE option, which is now allowed at compile time - but just
159 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
160 to pcregrep and other applications that have no direct access to PCRE
161 options. The new /Y option in pcretest sets this option when calling
162 pcre_compile().
163
164 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
165 back references. Groups containing recursive back references were forced to
166 be atomic by that change, but in the case of named groups, the amount of
167 memory required was incorrectly computed, leading to "Failed: internal
168 error: code overflow". This has been fixed.
169
170 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
171 pcretest.c, to avoid build problems in some Borland environments.
172
173
174 Version 8.10 25-Jun-2010
175 ------------------------
176
177 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
178 THEN.
179
180 2. (*ACCEPT) was not working when inside an atomic group.
181
182 3. Inside a character class, \B is treated as a literal by default, but
183 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
184 causes the error). The code is unchanged, but I tidied the documentation.
185
186 4. Inside a character class, PCRE always treated \R and \X as literals,
187 whereas Perl faults them if its -w option is set. I have changed PCRE so
188 that it faults them when PCRE_EXTRA is set.
189
190 5. Added support for \N, which always matches any character other than
191 newline. (It is the same as "." when PCRE_DOTALL is not set.)
192
193 6. When compiling pcregrep with newer versions of gcc which may have
194 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
195 declared with attribute warn_unused_result" were given. Just casting the
196 result to (void) does not stop the warnings; a more elaborate fudge is
197 needed. I've used a macro to implement this.
198
199 7. Minor change to pcretest.c to avoid a compiler warning.
200
201 8. Added four artifical Unicode properties to help with an option to make
202 \s etc use properties (see next item). The new properties are: Xan
203 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
204
205 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
206 use Unicode properties. (*UCP) at the start of a pattern can be used to set
207 this option. Modified pcretest to add /W to test this facility. Added
208 REG_UCP to make it available via the POSIX interface.
209
210 10. Added --line-buffered to pcregrep.
211
212 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
213 studied, and the match started with a letter with a code point greater than
214 127 whose first byte was different to the first byte of the other case of
215 the letter, the other case of this starting letter was not recognized
216 (#976).
217
218 12. If a pattern that was studied started with a repeated Unicode property
219 test, for example, \p{Nd}+, there was the theoretical possibility of
220 setting up an incorrect bitmap of starting bytes, but fortunately it could
221 not have actually happened in practice until change 8 above was made (it
222 added property types that matched character-matching opcodes).
223
224 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
225 possible starting bytes for non-anchored patterns.
226
227 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
228 \R, and also a number of cases that involve Unicode properties, both
229 explicit and implicit when PCRE_UCP is set.
230
231 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
232 input, it could crash or give wrong results if characters with values
233 greater than 0xc0 were present in the subject string. (Detail: it assumed
234 UTF-8 input when processing these items.)
235
236 16. Added a lot of (int) casts to avoid compiler warnings in systems where
237 size_t is 64-bit (#991).
238
239 17. Added a check for running out of memory when PCRE is compiled with
240 --disable-stack-for-recursion (#990).
241
242 18. If the last data line in a file for pcretest does not have a newline on
243 the end, a newline was missing in the output.
244
245 19. The default pcre_chartables.c file recognizes only ASCII characters (values
246 less than 128) in its various bitmaps. However, there is a facility for
247 generating tables according to the current locale when PCRE is compiled. It
248 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
249 characters, are recognized by isspace() and therefore were getting set in
250 these tables, and indeed these tables seem to approximate to ISO 8859. This
251 caused a problem in UTF-8 mode when pcre_study() was used to create a list
252 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
253 which of course cannot start UTF-8 characters. I have changed the code so
254 that only real ASCII characters (less than 128) and the correct starting
255 bytes for UTF-8 encodings are set for characters greater than 127 when in
256 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
257 altogether.)
258
259 20. Added the /T option to pcretest so as to be able to run tests with non-
260 standard character tables, thus making it possible to include the tests
261 used for 19 above in the standard set of tests.
262
263 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
264 reference to a subpattern the other side of a comment that contains an
265 opening parenthesis caused either an internal compiling error, or a
266 reference to the wrong subpattern.
267
268
269 Version 8.02 19-Mar-2010
270 ------------------------
271
272 1. The Unicode data tables have been updated to Unicode 5.2.0.
273
274 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
275 configured.
276
277 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
278 original author of that file, following a query about its status.
279
280 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
281 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
282
283 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
284 quantifier applied to a forward-referencing subroutine call, could compile
285 incorrect code or give the error "internal error: previously-checked
286 referenced subpattern not found".
287
288 6. Both MS Visual Studio and Symbian OS have problems with initializing
289 variables to point to external functions. For these systems, therefore,
290 pcre_malloc etc. are now initialized to local functions that call the
291 relevant global functions.
292
293 7. There were two entries missing in the vectors called coptable and poptable
294 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
295 I've fixed the data, and added a kludgy way of testing at compile time that
296 the lengths are correct (equal to the number of opcodes).
297
298 8. Following on from 7, I added a similar kludge to check the length of the
299 eint vector in pcreposix.c.
300
301 9. Error texts for pcre_compile() are held as one long string to avoid too
302 much relocation at load time. To find a text, the string is searched,
303 counting zeros. There was no check for running off the end of the string,
304 which could happen if a new error number was added without updating the
305 string.
306
307 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
308
309 11. \K was not working if it appeared in an atomic group or in a group that
310 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
311 \K is "not well defined" if used in an assertion. PCRE now accepts it if
312 the assertion is positive, but not if it is negative.
313
314 12. Change 11 fortuitously reduced the size of the stack frame used in the
315 "match()" function of pcre_exec.c by one pointer. Forthcoming
316 implementation of support for (*MARK) will need an extra pointer on the
317 stack; I have reserved it now, so that the stack frame size does not
318 decrease.
319
320 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
321 item in branch that calls a recursion is a subroutine call - as in the
322 second branch in the above example - was incorrectly given the compile-
323 time error "recursive call could loop indefinitely" because pcre_compile()
324 was not correctly checking the subroutine for matching a non-empty string.
325
326 14. The checks for overrunning compiling workspace could trigger after an
327 overrun had occurred. This is a "should never occur" error, but it can be
328 triggered by pathological patterns such as hundreds of nested parentheses.
329 The checks now trigger 100 bytes before the end of the workspace.
330
331 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
332
333
334 Version 8.01 19-Jan-2010
335 ------------------------
336
337 1. If a pattern contained a conditional subpattern with only one branch (in
338 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
339 computed the wrong minimum data length (which is of course zero for such
340 subpatterns). This could cause incorrect "no match" results.
341
342 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
343 the pattern is reset in the first branch, pcre_compile() failed with
344 "internal error: code overflow at offset...". This happened only when
345 the reset was to the original external option setting. (An optimization
346 abstracts leading options settings into an external setting, which was the
347 cause of this.)
348
349 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
350 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
351 assertion pattern did not match (meaning that the assertion was true), it
352 was incorrectly treated as false if the SKIP had been reached during the
353 matching. This also applied to assertions used as conditions.
354
355 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
356 assertion subpattern, including such a pattern used as a condition,
357 unpredictable results occurred, instead of the error return
358 PCRE_ERROR_DFA_UITEM.
359
360 5. The C++ GlobalReplace function was not working like Perl for the special
361 situation when an empty string is matched. It now does the fancy magic
362 stuff that is necessary.
363
364 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
365 removed. (These were left over from very, very early versions of PCRE.)
366
367 7. Some cosmetic changes to the code to make life easier when compiling it
368 as part of something else:
369
370 (a) Change DEBUG to PCRE_DEBUG.
371
372 (b) In pcre_compile(), rename the member of the "branch_chain" structure
373 called "current" as "current_branch", to prevent a collision with the
374 Linux macro when compiled as a kernel module.
375
376 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
377 prevent a collision with the Linux macro when compiled as a kernel
378 module.
379
380 8. In pcre_compile() there are some checks for integer overflows that used to
381 cast potentially large values to (double). This has been changed to that
382 when building, a check for int64_t is made, and if it is found, it is used
383 instead, thus avoiding the use of floating point arithmetic. (There is no
384 other use of FP in PCRE.) If int64_t is not found, the fallback is to
385 double.
386
387 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
388 2005 (difference between two addresses compared to an unsigned value).
389
390 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
391 custom one, because of the following reported problem in Windows:
392
393 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
394 under Win32.
395 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
396 therefore missing the function definition.
397 - The compiler thus generates a "C" signature for the test function.
398 - The linker fails to find the "C" function.
399 - PCRE fails to configure if asked to do so against libbz2.
400
401 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
402 messages were output:
403
404 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
405 rerunning libtoolize, to keep the correct libtool macros in-tree.
406 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
407
408 I have done both of these things.
409
410 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
411 most of the time, it *can* run out if it is given a pattern that contains a
412 runaway infinite recursion. I updated the discussion in the pcrestack man
413 page.
414
415 13. Now that we have gone to the x.xx style of version numbers, the minor
416 version may start with zero. Using 08 or 09 is a bad idea because users
417 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
418 interpreted as invalid octal numbers. I've updated the previous comment in
419 configure.ac, and also added a check that gives an error if 08 or 09 are
420 used.
421
422 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
423 causing partial matching to fail when the end of the subject matched \W
424 in a UTF-8 pattern where \W was quantified with a minimum of 3.
425
426 15. There were some discrepancies between the declarations in pcre_internal.h
427 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
428 their definitions. The declarations used "const uschar *" and the
429 definitions used USPTR. Even though USPTR is normally defined as "const
430 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
431 reported that: "This difference in casting confuses some C++ compilers, for
432 example, SunCC recognizes above declarations as different functions and
433 generates broken code for hbpcre." I have changed the declarations to use
434 USPTR.
435
436 16. GNU libtool is named differently on some systems. The autogen.sh script now
437 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
438 (FreeBSD).
439
440 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
441 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
442 comment: "Figure out how to create a longlong from a string: strtoll and
443 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
444 instance, but it only takes 2 args instead of 3!"
445
446 18. A subtle bug concerned with back references has been fixed by a change of
447 specification, with a corresponding code fix. A pattern such as
448 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
449 refers, was giving matches when it shouldn't. For example, xa=xaaa would
450 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
451 same bug. Such groups have to be quantified to be useful, or contained
452 inside another quantified group. (If there's no repetition, the reference
453 can never match.) The problem arises because, having left the group and
454 moved on to the rest of the pattern, a later failure that backtracks into
455 the group uses the captured value from the final iteration of the group
456 rather than the correct earlier one. I have fixed this in PCRE by forcing
457 any group that contains a reference to itself to be an atomic group; that
458 is, there cannot be any backtracking into it once it has completed. This is
459 similar to recursive and subroutine calls.
460
461
462 Version 8.00 19-Oct-09
463 ----------------------
464
465 1. The table for translating pcre_compile() error codes into POSIX error codes
466 was out-of-date, and there was no check on the pcre_compile() error code
467 being within the table. This could lead to an OK return being given in
468 error.
469
470 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
471 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
472 in a Windows environment.
473
474 3. The pcregrep --count option prints the count for each file even when it is
475 zero, as does GNU grep. However, pcregrep was also printing all files when
476 --files-with-matches was added. Now, when both options are given, it prints
477 counts only for those files that have at least one match. (GNU grep just
478 prints the file name in this circumstance, but including the count seems
479 more useful - otherwise, why use --count?) Also ensured that the
480 combination -clh just lists non-zero counts, with no names.
481
482 4. The long form of the pcregrep -F option was incorrectly implemented as
483 --fixed_strings instead of --fixed-strings. This is an incompatible change,
484 but it seems right to fix it, and I didn't think it was worth preserving
485 the old behaviour.
486
487 5. The command line items --regex=pattern and --regexp=pattern were not
488 recognized by pcregrep, which required --regex pattern or --regexp pattern
489 (with a space rather than an '='). The man page documented the '=' forms,
490 which are compatible with GNU grep; these now work.
491
492 6. No libpcreposix.pc file was created for pkg-config; there was just
493 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
494
495 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
496 when UCP support is not needed, by modifying the Python script that
497 generates it from Unicode data files. This should not matter if the module
498 is correctly used as a library, but I received one complaint about 50K of
499 unwanted data. My guess is that the person linked everything into his
500 program rather than using a library. Anyway, it does no harm.
501
502 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
503 was a minimum greater than 1 for a wide character in a possessive
504 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
505 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
506 character. Chaos in the form of incorrect output or a compiling loop could
507 result.
508
509 9. The restrictions on what a pattern can contain when partial matching is
510 requested for pcre_exec() have been removed. All patterns can now be
511 partially matched by this function. In addition, if there are at least two
512 slots in the offset vector, the offset of the earliest inspected character
513 for the match and the offset of the end of the subject are set in them when
514 PCRE_ERROR_PARTIAL is returned.
515
516 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
517 synonymous with PCRE_PARTIAL, for backwards compatibility, and
518 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
519 and may be more useful for multi-segment matching.
520
521 11. Partial matching with pcre_exec() is now more intuitive. A partial match
522 used to be given if ever the end of the subject was reached; now it is
523 given only if matching could not proceed because another character was
524 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
525 string "Z", which now yields "no match" instead of "partial match". In the
526 case of pcre_dfa_exec(), "no match" is given if every matching path for the
527 final character ended with (*FAIL).
528
529 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
530 if the pattern had a "must contain" character that was already found in the
531 earlier partial match, unless partial matching was again requested. For
532 example, with the pattern /dog.(body)?/, the "must contain" character is
533 "g". If the first part-match was for the string "dog", restarting with
534 "sbody" failed. This bug has been fixed.
535
536 13. The string returned by pcre_dfa_exec() after a partial match has been
537 changed so that it starts at the first inspected character rather than the
538 first character of the match. This makes a difference only if the pattern
539 starts with a lookbehind assertion or \b or \B (\K is not supported by
540 pcre_dfa_exec()). It's an incompatible change, but it makes the two
541 matching functions compatible, and I think it's the right thing to do.
542
543 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
544 so that the demonstration program is easily available in environments where
545 PCRE has not been installed from source.
546
547 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
548 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
549 library.
550
551 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
552 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
553 is not the first non-POSIX option to be added. Clearly some people find
554 these options useful.
555
556 17. If a caller to the POSIX matching function regexec() passes a non-zero
557 value for nmatch with a NULL value for pmatch, the value of
558 nmatch is forced to zero.
559
560 18. RunGrepTest did not have a test for the availability of the -u option of
561 the diff command, as RunTest does. It now checks in the same way as
562 RunTest, and also checks for the -b option.
563
564 19. If an odd number of negated classes containing just a single character
565 interposed, within parentheses, between a forward reference to a named
566 subpattern and the definition of the subpattern, compilation crashed with
567 an internal error, complaining that it could not find the referenced
568 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
569 [The bug was that it was starting one character too far in when skipping
570 over the character class, thus treating the ] as data rather than
571 terminating the class. This meant it could skip too much.]
572
573 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
574 /g option in pcretest when the pattern contains \K, which makes it possible
575 to have an empty string match not at the start, even when the pattern is
576 anchored. Updated pcretest and pcredemo to use this option.
577
578 21. If the maximum number of capturing subpatterns in a recursion was greater
579 than the maximum at the outer level, the higher number was returned, but
580 with unset values at the outer level. The correct (outer level) value is
581 now given.
582
583 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
584 PCRE did not set those parentheses (unlike Perl). I have now found a way to
585 make it do so. The string so far is captured, making this feature
586 compatible with Perl.
587
588 23. The tests have been re-organized, adding tests 11 and 12, to make it
589 possible to check the Perl 5.10 features against Perl 5.10.
590
591 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
592 pattern matches a fixed length string. PCRE did not allow this; now it
593 does. Neither allows recursion.
594
595 25. I finally figured out how to implement a request to provide the minimum
596 length of subject string that was needed in order to match a given pattern.
597 (It was back references and recursion that I had previously got hung up
598 on.) This code has now been added to pcre_study(); it finds a lower bound
599 to the length of subject needed. It is not necessarily the greatest lower
600 bound, but using it to avoid searching strings that are too short does give
601 some useful speed-ups. The value is available to calling programs via
602 pcre_fullinfo().
603
604 26. While implementing 25, I discovered to my embarrassment that pcretest had
605 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
606 study optimizations had never been tested with that matching function.
607 Oops. What is worse, even when it was passed study data, there was a bug in
608 pcre_dfa_exec() that meant it never actually used it. Double oops. There
609 were also very few tests of studied patterns with pcre_dfa_exec().
610
611 27. If (?| is used to create subpatterns with duplicate numbers, they are now
612 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
613 on the other side of the coin, they are no longer allowed to have different
614 names, because these cannot be distinguished in PCRE, and this has caused
615 confusion. (This is a difference from Perl.)
616
617 28. When duplicate subpattern names are present (necessarily with different
618 numbers, as required by 27 above), and a test is made by name in a
619 conditional pattern, either for a subpattern having been matched, or for
620 recursion in such a pattern, all the associated numbered subpatterns are
621 tested, and the overall condition is true if the condition is true for any
622 one of them. This is the way Perl works, and is also more like the way
623 testing by number works.
624
625
626 Version 7.9 11-Apr-09
627 ---------------------
628
629 1. When building with support for bzlib/zlib (pcregrep) and/or readline
630 (pcretest), all targets were linked against these libraries. This included
631 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
632 libraries. This caused unwanted dependencies to be created. This problem
633 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
634 pcretest is linked with readline.
635
636 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
637 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
638 moved outside it again, because FALSE and TRUE are already defined in AIX,
639 but BOOL is not.
640
641 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
642 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
643
644 4. The pcregrep documentation said spaces were inserted as well as colons (or
645 hyphens) following file names and line numbers when outputting matching
646 lines. This is not true; no spaces are inserted. I have also clarified the
647 wording for the --colour (or --color) option.
648
649 5. In pcregrep, when --colour was used with -o, the list of matching strings
650 was not coloured; this is different to GNU grep, so I have changed it to be
651 the same.
652
653 6. When --colo(u)r was used in pcregrep, only the first matching substring in
654 each matching line was coloured. Now it goes on to look for further matches
655 of any of the test patterns, which is the same behaviour as GNU grep.
656
657 7. A pattern that could match an empty string could cause pcregrep to loop; it
658 doesn't make sense to accept an empty string match in pcregrep, so I have
659 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
660 seems to be how GNU grep behaves.
661
662 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
663 start or after a newline", because the conditional assertion was not being
664 correctly handled. The rule now is that both the assertion and what follows
665 in the first alternative must satisfy the test.
666
667 9. If auto-callout was enabled in a pattern with a conditional group whose
668 condition was an assertion, PCRE could crash during matching, both with
669 pcre_exec() and pcre_dfa_exec().
670
671 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
672 used for matching.
673
674 11. Unicode property support in character classes was not working for
675 characters (bytes) greater than 127 when not in UTF-8 mode.
676
677 12. Added the -M command line option to pcretest.
678
679 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
680
681 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
682
683 16. Added comments and documentation about mis-use of no_arg in the C++
684 wrapper.
685
686 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
687 from Martin Jerabek that uses macro names for all relevant character and
688 string constants.
689
690 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
691 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
692 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
693 these, but not everybody uses configure.
694
695 19. A conditional group that had only one branch was not being correctly
696 recognized as an item that could match an empty string. This meant that an
697 enclosing group might also not be so recognized, causing infinite looping
698 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
699 with the subject "ab", where knowledge that the repeated group can match
700 nothing is needed in order to break the loop.
701
702 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
703 exec(), but without supplying a callout function, matching went wrong.
704
705 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
706 leak if the size of the offset vector was greater than 30. When the vector
707 is smaller, the saved offsets during recursion go onto a local stack
708 vector, but for larger vectors malloc() is used. It was failing to free
709 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
710 error, in fact).
711
712 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
713 heapframe that is used only when UTF-8 support is enabled. This caused no
714 problem, but was untidy.
715
716 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
717 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
718 included within another project.
719
720 24. Steven Van Ingelgem's patches to add more options to the CMake support,
721 slightly modified by me:
722
723 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
724 not building pcregrep.
725
726 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
727 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
728
729 25. Forward references, both numeric and by name, in patterns that made use of
730 duplicate group numbers, could behave incorrectly or give incorrect errors,
731 because when scanning forward to find the reference group, PCRE was not
732 taking into account the duplicate group numbers. A pattern such as
733 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
734
735 26. Changed a few more instances of "const unsigned char *" to USPTR, making
736 the feature of a custom pointer more persuasive (as requested by a user).
737
738 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
739 pcretest.c, inside #ifndefs, because it seems they are sometimes already
740 pre-defined.
741
742 28. Added support for (*UTF8) at the start of a pattern.
743
744 29. Arrange for flags added by the "release type" setting in CMake to be shown
745 in the configuration summary.
746
747
748 Version 7.8 05-Sep-08
749 ---------------------
750
751 1. Replaced UCP searching code with optimized version as implemented for Ad
752 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
753 stage table and inline lookup instead of a function, giving speed ups of 2
754 to 5 times on some simple patterns that I tested. Permission was given to
755 distribute the MultiStage2.py script that generates the tables (it's not in
756 the tarball, but is in the Subversion repository).
757
758 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
759 scripts.
760
761 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
762 a group with a zero qualifier. The result of the study could be incorrect,
763 or the function might crash, depending on the pattern.
764
765 4. Caseless matching was not working for non-ASCII characters in back
766 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
767 It now works when Unicode Property Support is available.
768
769 5. In pcretest, an escape such as \x{de} in the data was always generating
770 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
771 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
772 truncation.
773
774 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
775
776 7. Added two (int) casts to pcregrep when printing the difference of two
777 pointers, in case they are 64-bit values.
778
779 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
780 test 2 if it fails.
781
782 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
783 and a #define of that name to empty if it is not externally set. This is to
784 allow users of MSVC to set it if necessary.
785
786 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
787 the convenience functions in the pcre_get.c source file.
788
789 11. An option change at the start of a pattern that had top-level alternatives
790 could cause overwriting and/or a crash. This command provoked a crash in
791 some environments:
792
793 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
794
795 This potential security problem was recorded as CVE-2008-2371.
796
797 12. For a pattern where the match had to start at the beginning or immediately
798 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
799 pcre_dfa_exec() could read past the end of the passed subject if there was
800 no match. To help with detecting such bugs (e.g. with valgrind), I modified
801 pcretest so that it places the subject at the end of its malloc-ed buffer.
802
803 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
804 exec() might read past the end of the data buffer in UTF-8 mode.
805
806 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
807 the data contained the byte 0x85 as part of a UTF-8 character within its
808 first line. This applied both to normal and DFA matching.
809
810 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
811 /^[^d]*?$/8 failed to match "abc".
812
813 16. Added a missing copyright notice to pcrecpp_internal.h.
814
815 17. Make it more clear in the documentation that values returned from
816 pcre_exec() in ovector are byte offsets, not character counts.
817
818 18. Tidied a few places to stop certain compilers from issuing warnings.
819
820 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
821 supplied by Stefan Weber. I made a further small update for 7.8 because
822 there is a change of source arrangements: the pcre_searchfuncs.c module is
823 replaced by pcre_ucd.c.
824
825
826 Version 7.7 07-May-08
827 ---------------------
828
829 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
830 a string to a long long, pretend we don't even have a long long." This is
831 done by checking for the strtoq, strtoll, and _strtoi64 functions.
832
833 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
834 pre-7.6 versions, which defined a global no_arg variable instead of putting
835 it in the RE class. (See also #8 below.)
836
837 3. Remove a line of dead code, identified by coverity and reported by Nuno
838 Lopes.
839
840 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
841
842 (1) The include/exclude patterns were being applied to the whole pathnames
843 of files, instead of just to the final components.
844
845 (2) If there was more than one level of directory, the subdirectories were
846 skipped unless they satisfied the include/exclude conditions. This is
847 inconsistent with GNU grep (and could even be seen as contrary to the
848 pcregrep specification - which I improved to make it absolutely clear).
849 The action now is always to scan all levels of directory, and just
850 apply the include/exclude patterns to regular files.
851
852 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
853 --exclude_dir in the tests to avoid scanning .svn directories.
854
855 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
856 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
857 doesn't support NULs in patterns.
858
859 7. Added some missing "const"s to declarations of static tables in
860 pcre_compile.c and pcre_dfa_exec.c.
861
862 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
863 caused by fix #2 above. (Subsequently also a second patch to fix the
864 first patch. And a third patch - this was a messy problem.)
865
866 9. Applied Craig's patch to remove the use of push_back().
867
868 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
869 matching function regexec().
870
871 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
872 which, however, unlike Perl's \g{...}, are subroutine calls, not back
873 references. PCRE supports relative numbers with this syntax (I don't think
874 Oniguruma does).
875
876 12. Previously, a group with a zero repeat such as (...){0} was completely
877 omitted from the compiled regex. However, this means that if the group
878 was called as a subroutine from elsewhere in the pattern, things went wrong
879 (an internal error was given). Such groups are now left in the compiled
880 pattern, with a new opcode that causes them to be skipped at execution
881 time.
882
883 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
884 to the way PCRE behaves:
885
886 (a) A lone ] character is dis-allowed (Perl treats it as data).
887
888 (b) A back reference to an unmatched subpattern matches an empty string
889 (Perl fails the current match path).
890
891 (c) A data ] in a character class must be notated as \] because if the
892 first data character in a class is ], it defines an empty class. (In
893 Perl it is not possible to have an empty class.) The empty class []
894 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
895 The negative empty class [^] matches any one character, independently
896 of the DOTALL setting.
897
898 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
899 non-existent subpattern following a character class starting with ']' and
900 containing () gave an internal compiling error instead of "reference to
901 non-existent subpattern". Fortunately, when the pattern did exist, the
902 compiled code was correct. (When scanning forwards to check for the
903 existencd of the subpattern, it was treating the data ']' as terminating
904 the class, so got the count wrong. When actually compiling, the reference
905 was subsequently set up correctly.)
906
907 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
908 it was being rejected as not supported by pcre_dfa_exec(), even though
909 other assertions are supported. I have made pcre_dfa_exec() support
910 (*FAIL).
911
912 16. The implementation of 13c above involved the invention of a new opcode,
913 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
914 cannot be changed at match time, I realized I could make a small
915 improvement to matching performance by compiling OP_ALLANY instead of
916 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
917 on the OP_ANY path.
918
919 17. Compiling pcretest on Windows with readline support failed without the
920 following two fixes: (1) Make the unistd.h include conditional on
921 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
922
923 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
924 ncurses library to be included for pcretest when ReadLine support is
925 requested, but also to allow for it to be overridden. This patch came from
926 Daniel Bergström.
927
928 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
929 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
930 any errors with the current Unicode tables. Thanks to Peter Kankowski for
931 spotting this.
932
933
934 Version 7.6 28-Jan-08
935 ---------------------
936
937 1. A character class containing a very large number of characters with
938 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
939 overflow.
940
941 2. Patch to cut out the "long long" test in pcrecpp_unittest when
942 HAVE_LONG_LONG is not defined.
943
944 3. Applied Christian Ehrlicher's patch to update the CMake build files to
945 bring them up to date and include new features. This patch includes:
946
947 - Fixed PH's badly added libz and libbz2 support.
948 - Fixed a problem with static linking.
949 - Added pcredemo. [But later removed - see 7 below.]
950 - Fixed dftables problem and added an option.
951 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
952 HAVE_LONG_LONG.
953 - Added readline support for pcretest.
954 - Added an listing of the option settings after cmake has run.
955
956 4. A user submitted a patch to Makefile that makes it easy to create
957 "pcre.dll" under mingw when using Configure/Make. I added stuff to
958 Makefile.am that cause it to include this special target, without
959 affecting anything else. Note that the same mingw target plus all
960 the other distribution libraries and programs are now supported
961 when configuring with CMake (see 6 below) instead of with
962 Configure/Make.
963
964 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
965 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
966 exported in the Windows port". It has not yet been confirmed that the patch
967 solves the problem, but it does no harm.
968
969 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
970 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
971 with CMake, and also correct the comment about stack recursion.
972
973 7. Remove the automatic building of pcredemo from the ./configure system and
974 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
975 of a program that users should build themselves after PCRE is installed, so
976 building it automatically is not really right. What is more, it gave
977 trouble in some build environments.
978
979 8. Further tidies to CMakeLists.txt from Sheri and Christian.
980
981
982 Version 7.5 10-Jan-08
983 ---------------------
984
985 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
986 values in parens when parsing an RE using the C++ wrapper."
987
988 2. Negative specials like \S did not work in character classes in UTF-8 mode.
989 Characters greater than 255 were excluded from the class instead of being
990 included.
991
992 3. The same bug as (2) above applied to negated POSIX classes such as
993 [:^space:].
994
995 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
996 defined or documented. It seems to have been a typo for PCRE_STATIC, so
997 I have changed it.
998
999 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1000 first named subpattern) and a construct such as (?&a) would reference the
1001 first named subpattern whose name started with "a" (in other words, the
1002 length check was missing). Both these problems are fixed. "Subpattern name
1003 expected" is now given for (?&) (a zero-length name), and this patch also
1004 makes it give the same error for \k'' (previously it complained that that
1005 was a reference to a non-existent subpattern).
1006
1007 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1008 this is right because (?- can be followed by option settings as well as by
1009 digits. I have, however, made the messages clearer.
1010
1011 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1012 than the number used in the conditional) now cause a compile-time error.
1013 This is actually not compatible with Perl, which accepts such patterns, but
1014 treats the conditional as always being FALSE (as PCRE used to), but it
1015 seems to me that giving a diagnostic is better.
1016
1017 8. Change "alphameric" to the more common word "alphanumeric" in comments
1018 and messages.
1019
1020 9. Fix two occurrences of "backslash" in comments that should have been
1021 "backspace".
1022
1023 10. Remove two redundant lines of code that can never be obeyed (their function
1024 was moved elsewhere).
1025
1026 11. The program that makes PCRE's Unicode character property table had a bug
1027 which caused it to generate incorrect table entries for sequences of
1028 characters that have the same character type, but are in different scripts.
1029 It amalgamated them into a single range, with the script of the first of
1030 them. In other words, some characters were in the wrong script. There were
1031 thirteen such cases, affecting characters in the following ranges:
1032
1033 U+002b0 - U+002c1
1034 U+0060c - U+0060d
1035 U+0061e - U+00612
1036 U+0064b - U+0065e
1037 U+0074d - U+0076d
1038 U+01800 - U+01805
1039 U+01d00 - U+01d77
1040 U+01d9b - U+01dbf
1041 U+0200b - U+0200f
1042 U+030fc - U+030fe
1043 U+03260 - U+0327f
1044 U+0fb46 - U+0fbb1
1045 U+10450 - U+1049d
1046
1047 12. The -o option (show only the matching part of a line) for pcregrep was not
1048 compatible with GNU grep in that, if there was more than one match in a
1049 line, it showed only the first of them. It now behaves in the same way as
1050 GNU grep.
1051
1052 13. If the -o and -v options were combined for pcregrep, it printed a blank
1053 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1054 does the same. The return code can be used to tell if there were any
1055 non-matching lines.
1056
1057 14. Added --file-offsets and --line-offsets to pcregrep.
1058
1059 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1060 infinitely looping recursion. The bug was that positive lookaheads were not
1061 being skipped when checking for a possible empty match (negative lookaheads
1062 and both kinds of lookbehind were skipped).
1063
1064 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1065 inclusion of <windows.h> to before rather than after the definition of
1066 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1067
1068 17. Specifying a possessive quantifier with a specific limit for a Unicode
1069 character property caused pcre_compile() to compile bad code, which led at
1070 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1071 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1072 caused the error; without that there was no problem.
1073
1074 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1075
1076 19. Added --enable-pcretest-libreadline.
1077
1078 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1079 RE::GlobalReplace(). As a result, the number of replacements returned was
1080 double what it should be. I removed one of the increments, but Craig sent a
1081 later patch that removed the other one (the right fix) and added unit tests
1082 that check the return values (which was not done before).
1083
1084 21. Several CMake things:
1085
1086 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1087 the names libpcre and libpcreposix, not just pcre and pcreposix.
1088
1089 (2) The above change means that pcretest and pcregrep are now correctly
1090 linked with the newly-built libraries, not previously installed ones.
1091
1092 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1093
1094 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1095 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1096 UTF-8 newline character). The key issue is that the pattern starts .*;
1097 this means that the match must be either at the beginning, or after a
1098 newline. The bug was in the code for advancing after a failed match and
1099 checking that the new position followed a newline. It was not taking
1100 account of UTF-8 characters correctly.
1101
1102 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1103 character classes. PCRE was not treating the sequence [:...:] as a
1104 character class unless the ... were all letters. Perl, however, seems to
1105 allow any characters between [: and :], though of course it rejects as
1106 unknown any "names" that contain non-letters, because all the known class
1107 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1108 for example, whereas PCRE did not - it did not recognize a POSIX character
1109 class. This seemed a bit dangerous, so the code has been changed to be
1110 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1111 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1112 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1113 Perl does, and where it didn't before.
1114
1115 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1116 Windows environments %n is disabled by default.
1117
1118
1119 Version 7.4 21-Sep-07
1120 ---------------------
1121
1122 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1123 means that a class such as [\s] counted as "explicit reference to CR or
1124 LF". That isn't really right - the whole point of the change was to try to
1125 help when there was an actual mention of one of the two characters. So now
1126 the change happens only if \r or \n (or a literal CR or LF) character is
1127 encountered.
1128
1129 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1130 of both had grown to the point where there were only 3 bits left.
1131 Fortunately, there was spare space in the data structure, and so I have
1132 moved the internal flags into a new 16-bit field to free up more option
1133 bits.
1134
1135 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1136 but did not set the internal JCHANGED flag - either of these is enough to
1137 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1138 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1139 start sets both bits.
1140
1141 4. Added options (at build time, compile time, exec time) to change \R from
1142 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1143
1144 5. doc/pcresyntax.html was missing from the distribution.
1145
1146 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1147 compatibility, even though it is no longer used.
1148
1149 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1150 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1151 windows.h file is present (where different names are used). [This was
1152 reversed later after testing - see 16 below.]
1153
1154 8. Changed all #include <config.h> to #include "config.h". There were also
1155 some further <pcre.h> cases that I changed to "pcre.h".
1156
1157 9. When pcregrep was used with the --colour option, it missed the line ending
1158 sequence off the lines that it output.
1159
1160 10. It was pointed out to me that arrays of string pointers cause lots of
1161 relocations when a shared library is dynamically loaded. A technique of
1162 using a single long string with a table of offsets can drastically reduce
1163 these. I have refactored PCRE in four places to do this. The result is
1164 dramatic:
1165
1166 Originally: 290
1167 After changing UCP table: 187
1168 After changing error message table: 43
1169 After changing table of "verbs" 36
1170 After changing table of Posix names 22
1171
1172 Thanks to the folks working on Gregex for glib for this insight.
1173
1174 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1175 unicode-properties was also set.
1176
1177 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1178
1179 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1180 checked only for CRLF.
1181
1182 14. Added casts to pcretest.c to avoid compiler warnings.
1183
1184 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1185
1186 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1187 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1188 entirely. This removes changes made in 7 above.
1189
1190 17. The CMake files have been updated, and there is now more information about
1191 building with CMake in the NON-UNIX-USE document.
1192
1193
1194 Version 7.3 28-Aug-07
1195 ---------------------
1196
1197 1. In the rejigging of the build system that eventually resulted in 7.1, the
1198 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1199 brackets there is not right, since it causes compilers to look for an
1200 installed pcre.h, not the version that is in the source that is being
1201 compiled (which of course may be different). I have changed it back to:
1202
1203 #include "pcre.h"
1204
1205 I have a vague recollection that the change was concerned with compiling in
1206 different directories, but in the new build system, that is taken care of
1207 by the VPATH setting the Makefile.
1208
1209 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1210 when the subject happened to end in the byte 0x85 (e.g. if the last
1211 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1212 characters but of course it shouldn't be taken as a newline when it is part
1213 of another character. The bug was that, for an unlimited repeat of . in
1214 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1215 characters when looking for a newline.
1216
1217 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1218
1219 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1220 in debug output.
1221
1222 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1223 long printing in the pcrecpp unittest when running under MinGW.
1224
1225 6. ESC_K was left out of the EBCDIC table.
1226
1227 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1228 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1229 limit also applies to "virtual nesting" when a pattern is recursive, and in
1230 this case 1000 isn't so big. I have been able to remove this limit at the
1231 expense of backing off one optimization in certain circumstances. Normally,
1232 when pcre_exec() would call its internal match() function recursively and
1233 immediately return the result unconditionally, it uses a "tail recursion"
1234 feature to save stack. However, when a subpattern that can match an empty
1235 string has an unlimited repetition quantifier, it no longer makes this
1236 optimization. That gives it a stack frame in which to save the data for
1237 checking that an empty string has been matched. Previously this was taken
1238 from the 1000-entry workspace that had been reserved. So now there is no
1239 explicit limit, but more stack is used.
1240
1241 8. Applied Daniel's patches to solve problems with the import/export magic
1242 syntax that is required for Windows, and which was going wrong for the
1243 pcreposix and pcrecpp parts of the library. These were overlooked when this
1244 problem was solved for the main library.
1245
1246 9. There were some crude static tests to avoid integer overflow when computing
1247 the size of patterns that contain repeated groups with explicit upper
1248 limits. As the maximum quantifier is 65535, the maximum group length was
1249 set at 30,000 so that the product of these two numbers did not overflow a
1250 32-bit integer. However, it turns out that people want to use groups that
1251 are longer than 30,000 bytes (though not repeat them that many times).
1252 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1253 made it possible to implement the integer overflow checks in a much more
1254 dynamic way, which I have now done. The artificial limitation on group
1255 length has been removed - we now have only the limit on the total length of
1256 the compiled pattern, which depends on the LINK_SIZE setting.
1257
1258 10. Fixed a bug in the documentation for get/copy named substring when
1259 duplicate names are permitted. If none of the named substrings are set, the
1260 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1261 empty string.
1262
1263 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1264 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1265 because the ] is interpreted as the first data character and the
1266 terminating ] is not found. PCRE has been made compatible with Perl in this
1267 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1268 cause memory overwriting.
1269
1270 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1271 string has been matched (to stop an infinite loop). It was not recognizing
1272 a conditional subpattern that could match an empty string if that
1273 subpattern was within another subpattern. For example, it looped when
1274 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1275 condition was not nested. This bug has been fixed.
1276
1277 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1278 past the start of the subject in the presence of bytes with the top bit
1279 set, for example "\x8aBCD".
1280
1281 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1282 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1283
1284 14. Optimized (?!) to (*FAIL).
1285
1286 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1287 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1288 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1289 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1290 does: it's just the validity check that is more restrictive.
1291
1292 16. Inserted checks for integer overflows during escape sequence (backslash)
1293 processing, and also fixed erroneous offset values for syntax errors during
1294 backslash processing.
1295
1296 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1297 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1298
1299 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1300 caused an overrun.
1301
1302 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1303 something other than just ASCII characters) inside a group that had an
1304 unlimited repeat caused a loop at compile time (while checking to see
1305 whether the group could match an empty string).
1306
1307 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1308 [\P{Any}] did so. (Error in the code for printing property names.)
1309
1310 21. An orphan \E inside a character class could cause a crash.
1311
1312 22. A repeated capturing bracket such as (A)? could cause a wild memory
1313 reference during compilation.
1314
1315 23. There are several functions in pcre_compile() that scan along a compiled
1316 expression for various reasons (e.g. to see if it's fixed length for look
1317 behind). There were bugs in these functions when a repeated \p or \P was
1318 present in the pattern. These operators have additional parameters compared
1319 with \d, etc, and these were not being taken into account when moving along
1320 the compiled data. Specifically:
1321
1322 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1323 length.
1324
1325 (b) An item such as \pL+ within a repeated group could cause crashes or
1326 loops.
1327
1328 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1329 "reference to non-existent subpattern" error.
1330
1331 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1332
1333 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1334 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1335
1336 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1337
1338 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1339 character were causing crashes (broken optimization).
1340
1341 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1342 \p or \P) caused a compile-time loop.
1343
1344 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1345 break. For example, the unstudied pattern [\r\n]A does not match the string
1346 "\r\nA" because change 7.0/46 below moves the current point on by two
1347 characters after failing to match at the start. However, the pattern \nA
1348 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1349 the same is true. There doesn't seem any very clean way out of this, but
1350 what I have chosen to do makes the common cases work: PCRE now takes note
1351 of whether there can be an explicit match for \r or \n anywhere in the
1352 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1353 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1354 pattern has explicit CR or LF references.
1355
1356 29. Added (*CR) etc for changing newline setting at start of pattern.
1357
1358
1359 Version 7.2 19-Jun-07
1360 ---------------------
1361
1362 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1363 which is apparently normally available under Windows.
1364
1365 2. Re-jig the pcregrep tests with different newline settings in an attempt
1366 to make them independent of the local environment's newline setting.
1367
1368 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1369
1370 4. Some of the "internals" tests were previously cut out when the link size
1371 was not 2, because the output contained actual offsets. The recent new
1372 "Z" feature of pcretest means that these can be cut out, making the tests
1373 usable with all link sizes.
1374
1375 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1376 stack recursion. This gives a massive performance boost under BSD, but just
1377 a small improvement under Linux. However, it saves one field in the frame
1378 in all cases.
1379
1380 6. Added more features from the forthcoming Perl 5.10:
1381
1382 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1383 recursion call. It refers to the nth most recently opened parentheses.
1384
1385 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1386 to be opened parentheses.
1387
1388 (c) Conditions that refer to capturing parentheses can be specified
1389 relatively, for example, (?(-2)... or (?(+3)...
1390
1391 (d) \K resets the start of the current match so that everything before
1392 is not part of it.
1393
1394 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1395
1396 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1397 reference syntax.
1398
1399 (g) (?| introduces a group in which the numbering of parentheses in each
1400 alternative starts with the same number.
1401
1402 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1403
1404 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1405 PCRE_INFO_JCHANGED.
1406
1407 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1408 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1409 for detecting groups that can match an empty string.
1410
1411 9. A pattern with a very large number of alternatives (more than several
1412 hundred) was running out of internal workspace during the pre-compile
1413 phase, where pcre_compile() figures out how much memory will be needed. A
1414 bit of new cunning has reduced the workspace needed for groups with
1415 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1416 workspace instead of running out of the 4096 that are available.
1417
1418 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1419
1420 11. Applied patch from Google to remove an optimization that didn't quite work.
1421 The report of the bug said:
1422
1423 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1424 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1425 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1426
1427 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1428 it matched the wrong number of bytes.
1429
1430
1431 Version 7.1 24-Apr-07
1432 ---------------------
1433
1434 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1435 that is more "standard", making use of automake and other Autotools. There
1436 is some re-arrangement of the files and adjustment of comments consequent
1437 on this.
1438
1439 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1440 for recursive directory scanning broke on some systems because the files
1441 are not scanned in any specific order and on different systems the order
1442 was different. A call to "sort" has been inserted into RunGrepTest for the
1443 approprate test as a short-term fix. In the longer term there may be an
1444 alternative.
1445
1446 3. I had an email from Eric Raymond about problems translating some of PCRE's
1447 man pages to HTML (despite the fact that I distribute HTML pages, some
1448 people do their own conversions for various reasons). The problems
1449 concerned the use of low-level troff macros .br and .in. I have therefore
1450 removed all such uses from the man pages (some were redundant, some could
1451 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1452 HTML has been updated to handle .nf/.fi and to complain if it encounters
1453 .br or .in.
1454
1455 4. Updated comments in configure.ac that get placed in config.h.in and also
1456 arranged for config.h to be included in the distribution, with the name
1457 config.h.generic, for the benefit of those who have to compile without
1458 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1459
1460 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1461 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1462 makevp.bat for the current PCRE, using the additional files
1463 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1464
1465 6. A Windows user reported a minor discrepancy with test 2, which turned out
1466 to be caused by a trailing space on an input line that had got lost in his
1467 copy. The trailing space was an accident, so I've just removed it.
1468
1469 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1470 that is needed.
1471
1472 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1473 as "const" (a) because they are and (b) because it helps the PHP
1474 maintainers who have recently made a script to detect big data structures
1475 in the php code that should be moved to the .rodata section. I remembered
1476 to update Builducptable as well, so it won't revert if ucptable.h is ever
1477 re-created.
1478
1479 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1480 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1481 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1482 support is not required. This saves 1.5-2K of code, which is important in
1483 some applications.
1484
1485 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1486 so as not to refer to the tables, even though these functions will never be
1487 called when UTF-8 support is disabled. Otherwise there are problems with a
1488 shared library.
1489
1490 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1491
1492 (a) It was defining its arguments as char * instead of void *.
1493
1494 (b) It was assuming that all moves were upwards in memory; this was true
1495 a long time ago when I wrote it, but is no longer the case.
1496
1497 The emulated memove() is provided for those environments that have neither
1498 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1499 is clearly not the case, as these two bugs were recently reported.
1500
1501 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1502 and Detrail to create the HTML documentation, the .txt form of the man
1503 pages, and it removes trailing spaces from listed files. It also creates
1504 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1505 case, it wraps all the #defines with #ifndefs. This script should be run
1506 before "make dist".
1507
1508 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1509 with Unicode property support.
1510
1511 (a) For a maximizing quantifier, if the two different cases of the
1512 character were of different lengths in their UTF-8 codings (there are
1513 some cases like this - I found 11), and the matching function had to
1514 back up over a mixture of the two cases, it incorrectly assumed they
1515 were both the same length.
1516
1517 (b) When PCRE was configured to use the heap rather than the stack for
1518 recursion during matching, it was not correctly preserving the data for
1519 the other case of a UTF-8 character when checking ahead for a match
1520 while processing a minimizing repeat. If the check also involved
1521 matching a wide character, but failed, corruption could cause an
1522 erroneous result when trying to check for a repeat of the original
1523 character.
1524
1525 13. Some tidying changes to the testing mechanism:
1526
1527 (a) The RunTest script now detects the internal link size and whether there
1528 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1529 values substituted by "configure". (The RunGrepTest script already did
1530 this for UTF-8.) The configure.ac script no longer substitutes the
1531 relevant variables.
1532
1533 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1534 with length and offset values. This means that the output is different
1535 for different internal link sizes. Test 2 is skipped for link sizes
1536 other than 2 because of this, bypassing the problem. Unfortunately,
1537 there was also a test in test 3 (the locale tests) that used /B and
1538 failed for link sizes other than 2. Rather than cut the whole test out,
1539 I have added a new /Z option to pcretest that replaces the length and
1540 offset values with spaces. This is now used to make test 3 independent
1541 of link size. (Test 2 will be tidied up later.)
1542
1543 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1544 segmentation fault instead of returning the appropriate error message.
1545
1546 15. In multiline mode when the newline sequence was set to "any", the pattern
1547 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1548 This doesn't seem right; it now treats the CRLF combination as the line
1549 ending, and so does not match in that case. It's only a pattern such as ^$
1550 that would hit this one: something like ^ABC$ would have failed after \r
1551 and then tried again after \r\n.
1552
1553 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1554 in an attempt to make files that differ only in their line terminators
1555 compare equal. This works on Linux.
1556
1557 17. Under certain error circumstances pcregrep might try to free random memory
1558 as it exited. This is now fixed, thanks to valgrind.
1559
1560 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1561 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1562 was because its rules for how to advance for /g after matching an empty
1563 string at the end of a line did not allow for this case. They now check for
1564 it specially.
1565
1566 20. pcretest is supposed to handle patterns and data of any length, by
1567 extending its buffers when necessary. It was getting this wrong when the
1568 buffer for a data line had to be extended.
1569
1570 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1571 CRLF as a newline sequence.
1572
1573 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1574 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1575 I have nevertheless tidied it up.
1576
1577 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1578
1579 24. Added a man page for pcre-config.
1580
1581
1582 Version 7.0 19-Dec-06
1583 ---------------------
1584
1585 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1586 moving to gcc 4.1.1.
1587
1588 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1589 sys/time.h, which is documented as needed for this function. It doesn't
1590 seem to matter on Linux, but it showed up on some releases of OS X.
1591
1592 3. It seems that there are systems where bytes whose values are greater than
1593 127 match isprint() in the "C" locale. The "C" locale should be the
1594 default when a C program starts up. In most systems, only ASCII printing
1595 characters match isprint(). This difference caused the output from pcretest
1596 to vary, making some of the tests fail. I have changed pcretest so that:
1597
1598 (a) When it is outputting text in the compiled version of a pattern, bytes
1599 other than 32-126 are always shown as hex escapes.
1600
1601 (b) When it is outputting text that is a matched part of a subject string,
1602 it does the same, unless a different locale has been set for the match
1603 (using the /L modifier). In this case, it uses isprint() to decide.
1604
1605 4. Fixed a major bug that caused incorrect computation of the amount of memory
1606 required for a compiled pattern when options that changed within the
1607 pattern affected the logic of the preliminary scan that determines the
1608 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1609 that the computed length was too small. The symptoms of this bug were
1610 either the PCRE error "internal error: code overflow" from pcre_compile(),
1611 or a glibc crash with a message such as "pcretest: free(): invalid next
1612 size (fast)". Examples of patterns that provoked this bug (shown in
1613 pcretest format) are:
1614
1615 /(?-x: )/x
1616 /(?x)(?-x: \s*#\s*)/
1617 /((?i)[\x{c0}])/8
1618 /(?i:[\x{c0}])/8
1619
1620 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1621 is now done differently.
1622
1623 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1624 wrapper classes; (b) implement a new function in the C++ scanner that is
1625 more efficient than the old way of doing things because it avoids levels of
1626 recursion in the regex matching; (c) add a paragraph to the documentation
1627 for the FullMatch() function.
1628
1629 6. The escape sequence \n was being treated as whatever was defined as
1630 "newline". Not only was this contrary to the documentation, which states
1631 that \n is character 10 (hex 0A), but it also went horribly wrong when
1632 "newline" was defined as CRLF. This has been fixed.
1633
1634 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1635 was being set to -1 for the "end of line" case (supposedly a value that no
1636 character can have). Though this value is never used (the check for end of
1637 line is "zero bytes in current character"), it caused compiler complaints.
1638 I've changed it to 0xffffffff.
1639
1640 8. In pcre_version.c, the version string was being built by a sequence of
1641 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1642 string (as it is for production releases) called a macro with an empty
1643 argument. The C standard says the result of this is undefined. The gcc
1644 compiler treats it as an empty string (which was what was wanted) but it is
1645 reported that Visual C gives an error. The source has been hacked around to
1646 avoid this problem.
1647
1648 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1649 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1650 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1651 of them did).
1652
1653 10. Originally, pcretest opened its input and output without "b"; then I was
1654 told that "b" was needed in some environments, so it was added for release
1655 5.0 to both the input and output. (It makes no difference on Unix-like
1656 systems.) Later I was told that it is wrong for the input on Windows. I've
1657 now abstracted the modes into two macros, to make it easier to fiddle with
1658 them, and removed "b" from the input mode under Windows.
1659
1660 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1661
1662 12. Added -help and --help to pcretest as an official way of being reminded
1663 of the options.
1664
1665 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1666 and pcrecpp.cc because they annoy compilers at high warning levels.
1667
1668 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1669
1670 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1671 scripts are not C programs :-) and which was not noticed because it works
1672 on Linux.
1673
1674 16. pcretest is supposed to handle any length of pattern and data line (as one
1675 line or as a continued sequence of lines) by extending its input buffer if
1676 necessary. This feature was broken for very long pattern lines, leading to
1677 a string of junk being passed to pcre_compile() if the pattern was longer
1678 than about 50K.
1679
1680 17. I have done a major re-factoring of the way pcre_compile() computes the
1681 amount of memory needed for a compiled pattern. Previously, there was code
1682 that made a preliminary scan of the pattern in order to do this. That was
1683 OK when PCRE was new, but as the facilities have expanded, it has become
1684 harder and harder to keep it in step with the real compile phase, and there
1685 have been a number of bugs (see for example, 4 above). I have now found a
1686 cunning way of running the real compile function in a "fake" mode that
1687 enables it to compute how much memory it would need, while actually only
1688 ever using a few hundred bytes of working memory and without too many
1689 tests of the mode. This should make future maintenance and development
1690 easier. A side effect of this work is that the limit of 200 on the nesting
1691 depth of parentheses has been removed (though this was never a serious
1692 limitation, I suspect). However, there is a downside: pcre_compile() now
1693 runs more slowly than before (30% or more, depending on the pattern). I
1694 hope this isn't a big issue. There is no effect on runtime performance.
1695
1696 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1697 newline (only possible for the last line of a file) and it was a
1698 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1699
1700 19. Added additional timing features to pcretest. (1) The -tm option now times
1701 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1702 separate command line item, by a number that specifies the number of
1703 repeats to use when timing. The default is 50000; this gives better
1704 precision, but takes uncomfortably long for very large patterns.
1705
1706 20. Extended pcre_study() to be more clever in cases where a branch of a
1707 subpattern has no definite first character. For example, (a*|b*)[cd] would
1708 previously give no result from pcre_study(). Now it recognizes that the
1709 first character must be a, b, c, or d.
1710
1711 21. There was an incorrect error "recursive call could loop indefinitely" if
1712 a subpattern (or the entire pattern) that was being tested for matching an
1713 empty string contained only one non-empty item after a nested subpattern.
1714 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1715 incorrectly, because the \d was being skipped in the check.
1716
1717 22. The pcretest program now has a new pattern option /B and a command line
1718 option -b, which is equivalent to adding /B to every pattern. This causes
1719 it to show the compiled bytecode, without the additional information that
1720 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1721 is the same as /B/I).
1722
1723 23. A new optimization is now able automatically to treat some sequences such
1724 as a*b as a*+b. More specifically, if something simple (such as a character
1725 or a simple class like \d) has an unlimited quantifier, and is followed by
1726 something that cannot possibly match the quantified thing, the quantifier
1727 is automatically "possessified".
1728
1729 24. A recursive reference to a subpattern whose number was greater than 39
1730 went wrong under certain circumstances in UTF-8 mode. This bug could also
1731 have affected the operation of pcre_study().
1732
1733 25. Realized that a little bit of performance could be had by replacing
1734 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1735
1736 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1737
1738 27. Possessive quantifiers such as a++ were previously implemented by turning
1739 them into atomic groups such as ($>a+). Now they have their own opcodes,
1740 which improves performance. This includes the automatically created ones
1741 from 23 above.
1742
1743 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1744 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1745 the first matched character to be a colon. This applied both to named and
1746 numbered groups.
1747
1748 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1749
1750 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1751 building PCRE on Windows easier, so I have included it in the distribution.
1752
1753 31. There is now a check in pcretest against a ridiculously large number being
1754 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1755 loop, the loop is abandoned.
1756
1757 32. Forward references to subpatterns in conditions such as (?(2)...) where
1758 subpattern 2 is defined later cause pcre_compile() to search forwards in
1759 the pattern for the relevant set of parentheses. This search went wrong
1760 when there were unescaped parentheses in a character class, parentheses
1761 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1762
1763 33. "Subroutine" calls and backreferences were previously restricted to
1764 referencing subpatterns earlier in the regex. This restriction has now
1765 been removed.
1766
1767 34. Added a number of extra features that are going to be in Perl 5.10. On the
1768 whole, these are just syntactic alternatives for features that PCRE had
1769 previously implemented using the Python syntax or my own invention. The
1770 other formats are all retained for compatibility.
1771
1772 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1773 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1774 also .NET compatible.
1775
1776 (b) A recursion or subroutine call to a named group can now be defined as
1777 (?&name) as well as (?P>name).
1778
1779 (c) A backreference to a named group can now be defined as \k<name> or
1780 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1781 5.10, are also .NET compatible.
1782
1783 (d) A conditional reference to a named group can now use the syntax
1784 (?(<name>) or (?('name') as well as (?(name).
1785
1786 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1787 groups (named and numbered) that are never evaluated inline, but can be
1788 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1789 is always false. There may be only one alternative in such a group.
1790
1791 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1792 as the simple (?(R). The condition is true only if the most recent
1793 recursion is that of the given number or name. It does not search out
1794 through the entire recursion stack.
1795
1796 (g) The escape \gN or \g{N} has been added, where N is a positive or
1797 negative number, specifying an absolute or relative reference.
1798
1799 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1800 some "unreachable code" warnings.
1801
1802 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1803 things, this adds five new scripts.
1804
1805 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1806 There were also incompatibilities regarding the handling of \Q..\E inside
1807 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1808 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1809
1810 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1811 matches an empty string, and forcibly breaks the loop. There were bugs in
1812 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1813 against aaaa the result was just "a" rather than "aaaa", for example. Two
1814 separate and independent bugs (that affected different cases) have been
1815 fixed.
1816
1817 39. Refactored the code to abolish the use of different opcodes for small
1818 capturing bracket numbers. This is a tidy that I avoided doing when I
1819 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1820 The new approach is not only tidier, it makes it possible to reduce the
1821 memory needed to fix the previous bug (38).
1822
1823 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1824 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1825 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1826 mode.
1827
1828 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1829 report.
1830
1831 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1832 copy construction and assignment in the C++ wrapper.
1833
1834 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1835 couple of bugs that could have given wrong results in the "--newline=crlf"
1836 case.
1837
1838 44. Added a number of casts and did some reorganization of signed/unsigned int
1839 variables following suggestions from Dair Grant. Also renamed the variable
1840 "this" as "item" because it is a C++ keyword.
1841
1842 45. Arranged for dftables to add
1843
1844 #include "pcre_internal.h"
1845
1846 to pcre_chartables.c because without it, gcc 4.x may remove the array
1847 definition from the final binary if PCRE is built into a static library and
1848 dead code stripping is activated.
1849
1850 46. For an unanchored pattern, if a match attempt fails at the start of a
1851 newline sequence, and the newline setting is CRLF or ANY, and the next two
1852 characters are CRLF, advance by two characters instead of one.
1853
1854
1855 Version 6.7 04-Jul-06
1856 ---------------------
1857
1858 1. In order to handle tests when input lines are enormously long, pcretest has
1859 been re-factored so that it automatically extends its buffers when
1860 necessary. The code is crude, but this _is_ just a test program. The
1861 default size has been increased from 32K to 50K.
1862
1863 2. The code in pcre_study() was using the value of the re argument before
1864 testing it for NULL. (Of course, in any sensible call of the function, it
1865 won't be NULL.)
1866
1867 3. The memmove() emulation function in pcre_internal.h, which is used on
1868 systems that lack both memmove() and bcopy() - that is, hardly ever -
1869 was missing a "static" storage class specifier.
1870
1871 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1872 containing an extended class (one that cannot be represented by a bitmap
1873 because it contains high-valued characters or Unicode property items, e.g.
1874 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1875 pattern, but PCRE should not loop if you do not (it no longer does).
1876 [Detail: two cases were found: (a) a repeated subpattern containing an
1877 extended class; (b) a recursive reference to a subpattern that followed a
1878 previous extended class. It wasn't skipping over the extended class
1879 correctly when UTF-8 mode was not set.]
1880
1881 5. A negated single-character class was not being recognized as fixed-length
1882 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1883 compile error "lookbehind assertion is not fixed length".
1884
1885 6. The RunPerlTest auxiliary script was showing an unexpected difference
1886 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1887 write a Perl script that can interpret lines of an input file either as
1888 byte characters or as UTF-8, which is what "perltest" was being required to
1889 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1890 can't do is switch easily at run time between having the "use utf8;" pragma
1891 or not. In the end, I fudged it by using the RunPerlTest script to insert
1892 "use utf8;" explicitly for the UTF-8 tests.
1893
1894 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1895 the end of the subject string, contrary to the documentation and to what
1896 Perl does. This was true of both matching functions. Now it matches only at
1897 the start of the subject and immediately after *internal* newlines.
1898
1899 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1900 a pointer to an int instead of a pointer to an unsigned long int. This
1901 caused problems on 64-bit systems.
1902
1903 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1904 instance of the 'standard' template library not being so standard".
1905
1906 10. There was no check on the number of named subpatterns nor the maximum
1907 length of a subpattern name. The product of these values is used to compute
1908 the size of the memory block for a compiled pattern. By supplying a very
1909 long subpattern name and a large number of named subpatterns, the size
1910 computation could be caused to overflow. This is now prevented by limiting
1911 the length of names to 32 characters, and the number of named subpatterns
1912 to 10,000.
1913
1914 11. Subpatterns that are repeated with specific counts have to be replicated in
1915 the compiled pattern. The size of memory for this was computed from the
1916 length of the subpattern and the repeat count. The latter is limited to
1917 65535, but there was no limit on the former, meaning that integer overflow
1918 could in principle occur. The compiled length of a repeated subpattern is
1919 now limited to 30,000 bytes in order to prevent this.
1920
1921 12. Added the optional facility to have named substrings with the same name.
1922
1923 13. Added the ability to use a named substring as a condition, using the
1924 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1925 are numbers (not recommended). Forward references are permitted.
1926
1927 14. Added forward references in named backreferences (if you see what I mean).
1928
1929 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1930 pattern could run off the end of the subject. For example, the pattern
1931 "(?s)(.{1,5})"8 did this with the subject "ab".
1932
1933 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1934 PCRE_CASELESS was set when matching characters that were quantified with ?
1935 or *.
1936
1937 17. A character class other than a single negated character that had a minimum
1938 but no maximum quantifier - for example [ab]{6,} - was not handled
1939 correctly by pce_dfa_exec(). It would match only one character.
1940
1941 18. A valid (though odd) pattern that looked like a POSIX character
1942 class but used an invalid character after [ (for example [[,abc,]]) caused
1943 pcre_compile() to give the error "Failed: internal error: code overflow" or
1944 in some cases to crash with a glibc free() error. This could even happen if
1945 the pattern terminated after [[ but there just happened to be a sequence of
1946 letters, a binary zero, and a closing ] in the memory that followed.
1947
1948 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1949 over the years. Originally (before any Unicode support), just the bottom 8
1950 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1951 output from "man perlunicode" includes this:
1952
1953 The regular expression compiler produces polymorphic opcodes. That
1954 is, the pattern adapts to the data and automatically switches to
1955 the Unicode character scheme when presented with Unicode data--or
1956 instead uses a traditional byte scheme when presented with byte
1957 data.
1958
1959 Sadly, a wide octal escape does not cause a switch, and in a string with
1960 no other multibyte characters, these octal escapes are treated as before.
1961 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1962 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1963 Unicode string.
1964
1965 I have not perpetrated such confusion in PCRE. Up till now, it took just
1966 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1967 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1968 translate to the appropriate multibyte character.
1969
1970 29. Applied some refactoring to reduce the number of warnings from Microsoft
1971 and Borland compilers. This has included removing the fudge introduced
1972 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1973 a warning about an unused variable.
1974
1975 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1976 characters since release 4.0, because Perl (from release 5.004) does not.
1977 [Or at least, is documented not to: some releases seem to be in conflict
1978 with the documentation.] However, when a pattern was studied with
1979 pcre_study() and all its branches started with \s, PCRE still included VT
1980 as a possible starting character. Of course, this did no harm; it just
1981 caused an unnecessary match attempt.
1982
1983 22. Removed a now-redundant internal flag bit that recorded the fact that case
1984 dependency changed within the pattern. This was once needed for "required
1985 byte" processing, but is no longer used. This recovers a now-scarce options
1986 bit. Also moved the least significant internal flag bit to the most-
1987 significant bit of the word, which was not previously used (hangover from
1988 the days when it was an int rather than a uint) to free up another bit for
1989 the future.
1990
1991 23. Added support for CRLF line endings as well as CR and LF. As well as the
1992 default being selectable at build time, it can now be changed at runtime
1993 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1994 specify that it is scanning data with non-default line endings.
1995
1996 24. Changed the definition of CXXLINK to make it agree with the definition of
1997 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1998
1999 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2000 recursions. This makes a big different to stack usage for some patterns.
2001
2002 26. If a subpattern containing a named recursion or subroutine reference such
2003 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2004 the space required for the compiled pattern went wrong and gave too small a
2005 value. Depending on the environment, this could lead to "Failed: internal
2006 error: code overflow at offset 49" or "glibc detected double free or
2007 corruption" errors.
2008
2009 27. Applied patches from Google (a) to support the new newline modes and (b) to
2010 advance over multibyte UTF-8 characters in GlobalReplace.
2011
2012 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2013 difference for some implementation of PCRE in some Windows version.
2014
2015 29. Added some extra testing facilities to pcretest:
2016
2017 \q<number> in a data line sets the "match limit" value
2018 \Q<number> in a data line sets the "match recursion limt" value
2019 -S <number> sets the stack size, where <number> is in megabytes
2020
2021 The -S option isn't available for Windows.
2022
2023
2024 Version 6.6 06-Feb-06
2025 ---------------------
2026
2027 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2028 in pcreposix.h. I have copied the definition from pcre.h.
2029
2030 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2031 because pcre.h is no longer a built file.
2032
2033 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2034 not normally included in the compiled code.
2035
2036
2037 Version 6.5 01-Feb-06
2038 ---------------------
2039
2040 1. When using the partial match feature with pcre_dfa_exec(), it was not
2041 anchoring the second and subsequent partial matches at the new starting
2042 point. This could lead to incorrect results. For example, with the pattern
2043 /1234/, partially matching against "123" and then "a4" gave a match.
2044
2045 2. Changes to pcregrep:
2046
2047 (a) All non-match returns from pcre_exec() were being treated as failures
2048 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2049 error message is output. Some extra information is given for the
2050 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2051 probably the only errors that are likely to be caused by users (by
2052 specifying a regex that has nested indefinite repeats, for instance).
2053 If there are more than 20 of these errors, pcregrep is abandoned.
2054
2055 (b) A binary zero was treated as data while matching, but terminated the
2056 output line if it was written out. This has been fixed: binary zeroes
2057 are now no different to any other data bytes.
2058
2059 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2060 used to set a locale for matching. The --locale=xxxx long option has
2061 been added (no short equivalent) to specify a locale explicitly on the
2062 pcregrep command, overriding the environment variables.
2063
2064 (d) When -B was used with -n, some line numbers in the output were one less
2065 than they should have been.
2066
2067 (e) Added the -o (--only-matching) option.
2068
2069 (f) If -A or -C was used with -c (count only), some lines of context were
2070 accidentally printed for the final match.
2071
2072 (g) Added the -H (--with-filename) option.
2073
2074 (h) The combination of options -rh failed to suppress file names for files
2075 that were found from directory arguments.
2076
2077 (i) Added the -D (--devices) and -d (--directories) options.
2078
2079 (j) Added the -F (--fixed-strings) option.
2080
2081 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2082
2083 (l) Added the --colo(u)r option.
2084
2085 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2086 is not present by default.
2087
2088 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2089 items such as (?R) or (?1), when the recursion could match a number of
2090 alternatives. If it matched one of the alternatives, but subsequently,
2091 outside the recursion, there was a failure, the code tried to back up into
2092 the recursion. However, because of the way PCRE is implemented, this is not
2093 possible, and the result was an incorrect result from the match.
2094
2095 In order to prevent this happening, the specification of recursion has
2096 been changed so that all such subpatterns are automatically treated as
2097 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2098
2099 4. I had overlooked the fact that, in some locales, there are characters for
2100 which isalpha() is true but neither isupper() nor islower() are true. In
2101 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2102 and ordfeminine) are like this. This affected the treatment of \w and \W
2103 when they appeared in character classes, but not when they appeared outside
2104 a character class. The bit map for "word" characters is now created
2105 separately from the results of isalnum() instead of just taking it from the
2106 upper, lower, and digit maps. (Plus the underscore character, of course.)
2107
2108 5. The above bug also affected the handling of POSIX character classes such as
2109 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2110 permanent tables. Instead, the bit maps for such a class were previously
2111 created as the appropriate unions of the upper, lower, and digit bitmaps.
2112 Now they are created by subtraction from the [[:word:]] class, which has
2113 its own bitmap.
2114
2115 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2116 It is created by subtracting the vertical space characters (\x09, \x0a,
2117 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2118 subtraction was done in the overall bitmap for a character class, meaning
2119 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2120 be recognized. This bug has been fixed.
2121
2122 7. Patches from the folks at Google:
2123
2124 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2125 real life, but is still worth protecting against".
2126
2127 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2128 regular expressions".
2129
2130 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2131 have it.
2132
2133 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2134 "configure" and the latter not, in order to fix a problem somebody had
2135 with compiling the Arg class on HP-UX.
2136
2137 (e) Improve the error-handling of the C++ wrapper a little bit.
2138
2139 (f) New tests for checking recursion limiting.
2140
2141 8. The pcre_memmove() function, which is used only if the environment does not
2142 have a standard memmove() function (and is therefore rarely compiled),
2143 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2144 returning a result (though PCRE never actually uses the result).
2145
2146 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2147 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2148 returned instead of calling malloc() with an overflowing number that would
2149 most likely cause subsequent chaos.
2150
2151 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2152
2153 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2154 with this option is matched, the nmatch and pmatch options of regexec() are
2155 ignored.
2156
2157 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2158 provided in case anyone wants to the the POSIX interface with UTF-8
2159 strings.
2160
2161 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2162 C++ linking (needed for some HP-UX environments).
2163
2164 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2165 (unused parameter) and in the pcre_printint() function (omitted "default"
2166 switch label when the default is to do nothing).
2167
2168 15. Added some code to make it possible, when PCRE is compiled as a C++
2169 library, to replace subject pointers for pcre_exec() with a smart pointer
2170 class, thus making it possible to process discontinuous strings.
2171
2172 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2173 much the same function. They were added by different people who were trying
2174 to make PCRE easy to compile on non-Unix systems. It has been suggested
2175 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2176 for compiling on Windows systems. I have therefore replaced it with
2177 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2178 defaults to "extern" for C or "extern C" for C++, which works fine on
2179 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2180 SCOPE with something explicit in config.h. In addition:
2181
2182 (a) pcreposix.h still had just "extern" instead of either of these macros;
2183 I have replaced it with PCRE_DATA_SCOPE.
2184
2185 (b) Functions such as _pcre_xclass(), which are internal to the library,
2186 but external in the C sense, all had PCRE_EXPORT in their definitions.
2187 This is apparently wrong for the Windows case, so I have removed it.
2188 (It makes no difference on Unix-like systems.)
2189
2190 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2191 of recursive calls to match(). This is different to MATCH_LIMIT because
2192 that limits the total number of calls to match(), not all of which increase
2193 the depth of recursion. Limiting the recursion depth limits the amount of
2194 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2195 when PCRE is compiled, and changed at run time. A patch from Google adds
2196 this functionality to the C++ interface.
2197
2198 18. Changes to the handling of Unicode character properties:
2199
2200 (a) Updated the table to Unicode 4.1.0.
2201
2202 (b) Recognize characters that are not in the table as "Cn" (undefined).
2203
2204 (c) I revised the way the table is implemented to a much improved format
2205 which includes recognition of ranges. It now supports the ranges that
2206 are defined in UnicodeData.txt, and it also amalgamates other
2207 characters into ranges. This has reduced the number of entries in the
2208 table from around 16,000 to around 3,000, thus reducing its size
2209 considerably. I realized I did not need to use a tree structure after
2210 all - a binary chop search is just as efficient. Having reduced the
2211 number of entries, I extended their size from 6 bytes to 8 bytes to
2212 allow for more data.
2213
2214 (d) Added support for Unicode script names via properties such as \p{Han}.
2215
2216 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2217 matching that character.
2218
2219 20. When matching a repeated Unicode property with a minimum greater than zero,
2220 (for example \pL{2,}), PCRE could look past the end of the subject if it
2221 reached it while seeking the minimum number of characters. This could
2222 happen only if some of the characters were more than one byte long, because
2223 there is a check for at least the minimum number of bytes.
2224
2225 21. Refactored the implementation of \p and \P so as to be more general, to
2226 allow for more different types of property in future. This has changed the
2227 compiled form incompatibly. Anybody with saved compiled patterns that use
2228 \p or \P will have to recompile them.
2229
2230 22. Added "Any" and "L&" to the supported property types.
2231
2232 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2233 but give a compile time error if the value is greater than 0xff.
2234
2235 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2236 accidentally not being installed or uninstalled.
2237
2238 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2239 made were to insert the current release number. This seemed silly, because
2240 it made things harder for people building PCRE on systems that don't run
2241 "configure". I have turned pcre.h into a distributed file, no longer built
2242 by "configure", with the version identification directly included. There is
2243 no longer a pcre.h.in file.
2244
2245 However, this change necessitated a change to the pcre-config script as
2246 well. It is built from pcre-config.in, and one of the substitutions was the
2247 release number. I have updated configure.ac so that ./configure now finds
2248 the release number by grepping pcre.h.
2249
2250 26. Added the ability to run the tests under valgrind.
2251
2252
2253 Version 6.4 05-Sep-05
2254 ---------------------
2255
2256 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2257 "--" to be printed when multiple files were scanned, even when none of the
2258 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2259 consider it to be a bug, and have restored the previous behaviour.
2260
2261 2. A couple of code tidies to get rid of compiler warnings.
2262
2263 3. The pcretest program used to cheat by referring to symbols in the library
2264 whose names begin with _pcre_. These are internal symbols that are not
2265 really supposed to be visible externally, and in some environments it is
2266 possible to suppress them. The cheating is now confined to including
2267 certain files from the library's source, which is a bit cleaner.
2268
2269 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2270 file's purpose clearer.
2271
2272 5. Reorganized pcre_ucp_findchar().
2273
2274
2275 Version 6.3 15-Aug-05
2276 ---------------------
2277
2278 1. The file libpcre.pc.in did not have general read permission in the tarball.
2279
2280 2. There were some problems when building without C++ support:
2281
2282 (a) If C++ support was not built, "make install" and "make test" still
2283 tried to test it.
2284
2285 (b) There were problems when the value of CXX was explicitly set. Some
2286 changes have been made to try to fix these, and ...
2287
2288 (c) --disable-cpp can now be used to explicitly disable C++ support.
2289
2290 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2291 backslash in a target when C++ was disabled. This confuses some
2292 versions of "make", apparently. Using an intermediate variable solves
2293 this. (Same for CPP_LOBJ.)
2294
2295 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2296 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2297 necessary on certain architectures.
2298
2299 4. Added a setting of -export-symbols-regex to the link command to remove
2300 those symbols that are exported in the C sense, but actually are local
2301 within the library, and not documented. Their names all begin with
2302 "_pcre_". This is not a perfect job, because (a) we have to except some
2303 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2304 available (and never for static libraries). I have made a note to try to
2305 find a way round (a) in the future.
2306
2307
2308 Version 6.2 01-Aug-05
2309 ---------------------
2310
2311 1. There was no test for integer overflow of quantifier values. A construction
2312 such as {1111111111111111} would give undefined results. What is worse, if
2313 a minimum quantifier for a parenthesized subpattern overflowed and became
2314 negative, the calculation of the memory size went wrong. This could have
2315 led to memory overwriting.
2316
2317 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2318
2319 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2320 operating environments where this matters.
2321
2322 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2323 PCRE options from within the C++ wrapper.
2324
2325 5. Named capturing subpatterns were not being correctly counted when a pattern
2326 was compiled. This caused two problems: (a) If there were more than 100
2327 such subpatterns, the calculation of the memory needed for the whole
2328 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2329 back references of the form \12, where the number was greater than 9, were
2330 not recognized as back references, even though there were sufficient
2331 previous subpatterns.
2332
2333 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2334 versions of gcc, e.g. 2.95.4.
2335
2336
2337 Version 6.1 21-Jun-05
2338 ---------------------
2339
2340 1. There was one reference to the variable "posix" in pcretest.c that was not
2341 surrounded by "#if !defined NOPOSIX".
2342
2343 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2344 the cross-check on the old pcre_info() function, for the benefit of the
2345 cut-down version of PCRE that is currently imported into Exim.
2346
2347 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2348 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2349 patterns that start (?i) but I don't think that matters. The use of (?i) is
2350 just an example; this all applies to the other options as well.
2351
2352 4. Since libtool seems to echo the compile commands it is issuing, the output
2353 from "make" can be reduced a bit by putting "@" in front of each libtool
2354 compile command.
2355
2356 5. Patch from the folks at Google for configure.in to be a bit more thorough
2357 in checking for a suitable C++ installation before trying to compile the
2358 C++ stuff. This should fix a reported problem when a compiler was present,
2359 but no suitable headers.
2360
2361 6. The man pages all had just "PCRE" as their title. I have changed them to
2362 be the relevant file name. I have also arranged that these names are
2363 retained in the file doc/pcre.txt, which is a concatenation in text format
2364 of all the man pages except the little individual ones for each function.
2365
2366 7. The NON-UNIX-USE file had not been updated for the different set of source
2367 files that come with release 6. I also added a few comments about the C++
2368 wrapper.
2369
2370
2371 Version 6.0 07-Jun-05
2372 ---------------------
2373
2374 1. Some minor internal re-organization to help with my DFA experiments.
2375
2376 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2377 didn't matter for the library itself when fully configured, but did matter
2378 when compiling without UCP support, or within Exim, where the ucp files are
2379 not imported.
2380
2381 3. Refactoring of the library code to split up the various functions into
2382 different source modules. The addition of the new DFA matching code (see
2383 below) to a single monolithic source would have made it really too
2384 unwieldy, quite apart from causing all the code to be include in a
2385 statically linked application, when only some functions are used. This is
2386 relevant even without the DFA addition now that patterns can be compiled in
2387 one application and matched in another.
2388
2389 The downside of splitting up is that there have to be some external
2390 functions and data tables that are used internally in different modules of
2391 the library but which are not part of the API. These have all had their
2392 names changed to start with "_pcre_" so that they are unlikely to clash
2393 with other external names.
2394
2395 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2396 a different (DFA) algorithm. Although it is slower than the original
2397 function, it does have some advantages for certain types of matching
2398 problem.
2399
2400 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2401 including restarting after a partial match.
2402
2403 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2404 defined when compiling for Windows was sent to me. I have put it into the
2405 code, though I have no means of testing or verifying it.
2406
2407 7. Added the pcre_refcount() auxiliary function.
2408
2409 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2410 match before or at the first newline in the subject string. In pcretest,
2411 the /f option on a pattern can be used to set this.
2412
2413 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2414 would behave wrongly. This has been present in PCRE since release 4.0.
2415
2416 10. A number of changes to the pcregrep command:
2417
2418 (a) Refactored how -x works; insert ^(...)$ instead of setting
2419 PCRE_ANCHORED and checking the length, in preparation for adding
2420 something similar for -w.
2421
2422 (b) Added the -w (match as a word) option.
2423
2424 (c) Refactored the way lines are read and buffered so as to have more
2425 than one at a time available.
2426
2427 (d) Implemented a pcregrep test script.
2428
2429 (e) Added the -M (multiline match) option. This allows patterns to match
2430 over several lines of the subject. The buffering ensures that at least
2431 8K, or the rest of the document (whichever is the shorter) is available
2432 for matching (and similarly the previous 8K for lookbehind assertions).
2433
2434 (f) Changed the --help output so that it now says
2435
2436 -w, --word-regex(p)
2437
2438 instead of two lines, one with "regex" and the other with "regexp"
2439 because that confused at least one person since the short forms are the
2440 same. (This required a bit of code, as the output is generated
2441 automatically from a table. It wasn't just a text change.)
2442
2443 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2444 option but starts with a hyphen. Could be a pattern or a path name
2445 starting with a hyphen, for instance.
2446
2447 (h) "-" can be given as a file name to represent stdin.
2448
2449 (i) When file names are being printed, "(standard input)" is used for
2450 the standard input, for compatibility with GNU grep. Previously
2451 "<stdin>" was used.
2452
2453 (j) The option --label=xxx can be used to supply a name to be used for
2454 stdin when file names are being printed. There is no short form.
2455
2456 (k) Re-factored the options decoding logic because we are going to add
2457 two more options that take data. Such options can now be given in four
2458 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2459
2460 (l) Added the -A, -B, and -C options for requesting that lines of context
2461 around matches be printed.
2462
2463 (m) Added the -L option to print the names of files that do not contain
2464 any matching lines, that is, the complement of -l.
2465
2466 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2467 continue to scan other files.
2468
2469 (o) The -s option was incorrectly implemented. For compatibility with other
2470 greps, it now suppresses the error message for a non-existent or non-
2471 accessible file (but not the return code). There is a new option called
2472 -q that suppresses the output of matching lines, which was what -s was
2473 previously doing.
2474
2475 (p) Added --include and --exclude options to specify files for inclusion
2476 and exclusion when recursing.
2477
2478 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2479 Hopefully, it now does.
2480
2481 12. Missing cast in pcre_study().
2482
2483 13. Added an "uninstall" target to the makefile.
2484
2485 14. Replaced "extern" in the function prototypes in Makefile.in with
2486 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2487 world, but is set differently for Windows.
2488
2489 15. Added a second compiling function called pcre_compile2(). The only
2490 difference is that it has an extra argument, which is a pointer to an
2491 integer error code. When there is a compile-time failure, this is set
2492 non-zero, in addition to the error test pointer being set to point to an
2493 error message. The new argument may be NULL if no error number is required
2494 (but then you may as well call pcre_compile(), which is now just a
2495 wrapper). This facility is provided because some applications need a
2496 numeric error indication, but it has also enabled me to tidy up the way
2497 compile-time errors are handled in the POSIX wrapper.
2498
2499 16. Added VPATH=.libs to the makefile; this should help when building with one
2500 prefix path and installing with another. (Or so I'm told by someone who
2501 knows more about this stuff than I do.)
2502
2503 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2504 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2505 match everything, including newlines. This is not POSIX-compatible, but
2506 somebody wanted the feature. From pcretest it can be activated by using
2507 both the P and the s flags.
2508
2509 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2510
2511 19. libpcre.pc was being incorrectly installed as executable.
2512
2513 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2514 it now also looks for '\r' so that it will work unmodified on Windows.
2515
2516 21. Added Google's contributed C++ wrapper to the distribution.
2517
2518 22. Added some untidy missing memory free() calls in pcretest, to keep
2519 Electric Fence happy when testing.
2520
2521
2522
2523 Version 5.0 13-Sep-04
2524 ---------------------
2525
2526 1. Internal change: literal characters are no longer packed up into items
2527 containing multiple characters in a single byte-string. Each character
2528 is now matched using a separate opcode. However, there may be more than one
2529 byte in the character in UTF-8 mode.
2530
2531 2. The pcre_callout_block structure has two new fields: pattern_position and
2532 next_item_length. These contain the offset in the pattern to the next match
2533 item, and its length, respectively.
2534
2535 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2536 insertion of callouts before each pattern item. Added the /C option to
2537 pcretest to make use of this.
2538
2539 4. On the advice of a Windows user, the lines
2540
2541 #if defined(_WIN32) || defined(WIN32)
2542 _setmode( _fileno( stdout ), 0x8000 );
2543 #endif /* defined(_WIN32) || defined(WIN32) */
2544
2545 have been added to the source of pcretest. This apparently does useful
2546 magic in relation to line terminators.
2547
2548 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2549 for the benefit of those environments where the "b" makes a difference.
2550
2551 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2552 to know about it. I have put a hack into configure.in that adds in code
2553 to set GCC=yes if CC=icc. This seems to end up at a point in the
2554 generated configure script that is early enough to affect the setting of
2555 compiler options, which is what is needed, but I have no means of testing
2556 whether it really works. (The user who reported this had patched the
2557 generated configure script, which of course I cannot do.)
2558
2559 LATER: After change 22 below (new libtool files), the configure script
2560 seems to know about icc (and also ecc). Therefore, I have commented out
2561 this hack in configure.in.
2562
2563 7. Added support for pkg-config (2 patches were sent in).
2564
2565 8. Negated POSIX character classes that used a combination of internal tables
2566 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2567 [[:^ascii]]. Typically, they would match almost any characters. The other
2568 POSIX classes were not broken in this way.
2569
2570 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2571 to find the match, as PCRE was deluded into thinking that the match had to
2572 start at the start point or following a newline. The same bug applied to
2573 patterns with negative forward assertions or any backward assertions
2574 preceding ".*" at the start, unless the pattern required a fixed first
2575 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2576
2577 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2578 starting at the last subject character, bytes beyond the end of the subject
2579 string were read.
2580
2581 11. Renamed the variable "class" as "classbits" to make life easier for C++
2582 users. (Previously there was a macro definition, but it apparently wasn't
2583 enough.)
2584
2585 12. Added the new field "tables" to the extra data so that tables can be passed
2586 in at exec time, or the internal tables can be re-selected. This allows
2587 a compiled regex to be saved and re-used at a later time by a different
2588 program that might have everything at different addresses.
2589
2590 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2591 -R library as well as a -L library.
2592
2593 14. The debugging options of pcretest (-d on the command line or D on a
2594 pattern) showed incorrect output for anything following an extended class
2595 that contained multibyte characters and which was followed by a quantifier.
2596
2597 15. Added optional support for general category Unicode character properties
2598 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2599 support. It adds about 90K to the size of the library. The meanings of the
2600 inbuilt class escapes such as \d and \s have NOT been changed.
2601
2602 16. Updated pcredemo.c to include calls to free() to release the memory for the
2603 compiled pattern.
2604
2605 17. The generated file chartables.c was being created in the source directory
2606 instead of in the building directory. This caused the build to fail if the
2607 source directory was different from the building directory, and was
2608 read-only.
2609
2610 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2611 file. No doubt somebody will tell me if they don't make sense... Also added
2612 Dan Mooney's comments about building on OpenVMS.
2613
2614 19. Added support for partial matching via the PCRE_PARTIAL option for
2615 pcre_exec() and the \P data escape in pcretest.
2616
2617 20. Extended pcretest with 3 new pattern features:
2618
2619 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2620 write the compiled pattern to the file whose name is "rest-of-line".
2621 This is a straight binary dump of the data, with the saved pointer to
2622 the character tables forced to be NULL. The study data, if any, is
2623 written too. After writing, pcretest reads a new pattern.
2624
2625 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2626 compiled pattern from the given file. There must not be any
2627 occurrences of "<" in the file name (pretty unlikely); if there are,
2628 pcretest will instead treat the initial "<" as a pattern delimiter.
2629 After reading in the pattern, pcretest goes on to read data lines as
2630 usual.
2631
2632 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2633 and 16-bit fields in a compiled pattern, to simulate a pattern that
2634 was compiled on a host of opposite endianness.
2635
2636 21. The pcre-exec() function can now cope with patterns that were compiled on
2637 hosts of opposite endianness, with this restriction:
2638
2639 As for any compiled expression that is saved and used later, the tables
2640 pointer field cannot be preserved; the extra_data field in the arguments
2641 to pcre_exec() should be used to pass in a tables address if a value
2642 other than the default internal tables were used at compile time.
2643
2644 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2645 now diagnosed as an error. Previously, most of the time, a negative number
2646 would have been treated as zero, but if in addition "ovector" was passed as
2647 NULL, a crash could occur.
2648
2649 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2650 new versions from the libtool 1.5 distribution (the last one is a copy of
2651 a file called libtool.m4). This seems to have fixed the need to patch
2652 "configure" to support Darwin 1.3 (which I used to do). However, I still
2653 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2654 workstation).
2655
2656 24. Changed the PCRE licence to be the more standard "BSD" licence.
2657
2658
2659 Version 4.5 01-Dec-03
2660 ---------------------
2661
2662 1. There has been some re-arrangement of the code for the match() function so
2663 that it can be compiled in a version that does not call itself recursively.
2664 Instead, it keeps those local variables that need separate instances for
2665 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2666 needs to "recurse". Keeping track of where control must go is done by means
2667 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2668 hide most of the details from the main code, and operates only if
2669 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2670 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2671 operating.
2672
2673 To make it easier for callers to provide specially tailored get/free
2674 functions for this usage, two new functions, pcre_stack_malloc, and
2675 pcre_stack_free, are used. They are always called in strict stacking order,
2676 and the size of block requested is always the same.
2677
2678 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2679 PCRE has been compiled to use the stack or the heap for recursion. The
2680 -C option of pcretest uses this to show which version is compiled.
2681
2682 A new data escape \S, is added to pcretest; it causes the amounts of store
2683 obtained and freed by both kinds of malloc/free at match time to be added
2684 to the output.
2685
2686 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2687 what's available on my current Linux desktop machine.
2688
2689 3. When matching a UTF-8 string, the test for a valid string at the start has
2690 been extended. If start_offset is not zero, PCRE now checks that it points
2691 to a byte that is the start of a UTF-8 character. If not, it returns
2692 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2693 this is necessary because there may be backward assertions in the pattern.
2694 When matching the same subject several times, it may save resources to use
2695 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2696
2697 4. The code for checking the validity of UTF-8 strings has been tightened so
2698 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2699 containing "overlong sequences".
2700
2701 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2702 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2703 should have been "&", but it just so happened that all the cases this let
2704 through by mistake were picked up later in the function.
2705
2706 6. I had used a variable called "isblank" - this is a C99 function, causing
2707 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2708
2709 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2710 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2711 so the version is shown; (c) stop "make test" echoing "./RunTest".
2712
2713 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2714
2715 9. The prototype for memmove() for systems that don't have it was using
2716 size_t, but the inclusion of the header that defines size_t was later. I've
2717 moved the #includes for the C headers earlier to avoid this.
2718
2719 10. Added some adjustments to the code to make it easier to compiler on certain
2720 special systems:
2721
2722 (a) Some "const" qualifiers were missing.
2723 (b) Added the macro EXPORT before all exported functions; by default this
2724 is defined to be empty.
2725 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2726 that it reads its output file name as an argument instead of writing
2727 to the standard output and assuming this can be redirected.
2728
2729 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2730 class containing characters with values greater than 255, PCRE compilation
2731 went into a loop.
2732
2733 12. A recursive reference to a subpattern that was within another subpattern
2734 that had a minimum quantifier of zero caused PCRE to crash. For example,
2735 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2736 recursion. If the recursively-called subpattern itself had a zero repeat,
2737 that was OK.
2738
2739 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2740 buffer into which it was copied (for escape processing) was still set at
2741 1024, so long lines caused crashes.
2742
2743 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2744 "internal error: code overflow...". This applied to any character class
2745 that was followed by a possessive quantifier.
2746
2747 15. Modified the Makefile to add libpcre.la as a prerequisite for
2748 libpcreposix.la because I was told this is needed for a parallel build to
2749 work.
2750
2751 16. If a pattern that contained .* following optional items at the start was
2752 studied, the wrong optimizing data was generated, leading to matching
2753 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2754 matching string must start with a or b or c. The correct conclusion for
2755 this pattern is that a match can start with any character.
2756
2757
2758 Version 4.4 13-Aug-03
2759 ---------------------
2760
2761 1. In UTF-8 mode, a character class containing characters with values between
2762 127 and 255 was not handled correctly if the compiled pattern was studied.
2763 In fixing this, I have also improved the studying algorithm for such
2764 classes (slightly).
2765
2766 2. Three internal functions had redundant arguments passed to them. Removal
2767 might give a very teeny performance improvement.
2768
2769 3. Documentation bug: the value of the capture_top field in a callout is *one
2770 more than* the number of the hightest numbered captured substring.
2771
2772 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2773 in incorrectly linking with a previously installed version. They now link
2774 explicitly with libpcre.la.
2775
2776 5. configure.in no longer needs to recognize Cygwin specially.
2777
2778 6. A problem in pcre.in for Windows platforms is fixed.
2779
2780 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2781 pcretest, it used to include the size of the study block as part of its
2782 output. Unfortunately, the structure contains a field that has a different
2783 size on different hardware architectures. This meant that the tests that
2784 showed this size failed. As the block is currently always of a fixed size,
2785 this information isn't actually particularly useful in pcretest output, so
2786 I have just removed it.
2787
2788 8. Three pre-processor statements accidentally did not start in column 1.
2789 Sadly, there are *still* compilers around that complain, even though
2790 standard C has not required this for well over a decade. Sigh.
2791
2792 9. In pcretest, the code for checking callouts passed small integers in the
2793 callout_data field, which is a void * field. However, some picky compilers
2794 complained about the casts involved for this on 64-bit systems. Now
2795 pcretest passes the address of the small integer instead, which should get
2796 rid of the warnings.
2797
2798 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2799 both compile and run time, and gives an error if an invalid UTF-8 sequence
2800 is found. There is a option for disabling this check in cases where the
2801 string is known to be correct and/or the maximum performance is wanted.
2802
2803 11. In response to a bug report, I changed one line in Makefile.in from
2804
2805 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2806 to
2807 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2808
2809 to look similar to other lines, but I have no way of telling whether this
2810 is the right thing to do, as I do not use Windows. No doubt I'll get told
2811 if it's wrong...
2812
2813
2814 Version 4.3 21-May-03
2815 ---------------------
2816
2817 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2818 Makefile.
2819
2820 2. Some refactoring to improve the quality of the code:
2821
2822 (i) The utf8_table... variables are now declared "const".
2823
2824 (ii) The code for \cx, which used the "case flipping" table to upper case
2825 lower case letters, now just substracts 32. This is ASCII-specific,
2826 but the whole concept of \cx is ASCII-specific, so it seems
2827 reasonable.
2828
2829 (iii) PCRE was using its character types table to recognize decimal and
2830 hexadecimal digits in the pattern. This is silly, because it handles
2831 only 0-9, a-f, and A-F, but the character types table is locale-
2832 specific, which means strange things might happen. A private
2833 table is now used for this - though it costs 256 bytes, a table is
2834 much faster than multiple explicit tests. Of course, the standard
2835 character types table is still used for matching digits in subject
2836 strings against \d.
2837
2838 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2839 ending in _t are). So I've renamed it as ESC_tee.
2840
2841 3. The first argument for regexec() in the POSIX wrapper should have been
2842 defined as "const".
2843
2844 4. Changed pcretest to use malloc() for its buffers so that they can be
2845 Electric Fenced for debugging.
2846
2847 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2848 to read one or more bytes before the start of the subject string. Often this
2849 had no effect on PCRE's behaviour, but in some circumstances it could
2850 provoke a segmentation fault.
2851
2852 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2853 to try to read one or more bytes before the start of the subject string.
2854
2855 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2856 UTF-8 support could misbehave in various ways if the subject string
2857 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2858 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2859 back over UTF-8 characters.)
2860
2861
2862 Version 4.2 14-Apr-03
2863 ---------------------
2864
2865 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2866
2867 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2868 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2869 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2870 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2871 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2872 and BUILD_EXEEXT
2873 Note: automatic setting of the BUILD variables is not yet working
2874 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2875 compile-time but not at link-time
2876 [LINK]: use for linking executables only
2877 make different versions for Windows and non-Windows
2878 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2879 libraries
2880 [LINK_FOR_BUILD]: new variable
2881 [OBJEXT]: use throughout
2882 [EXEEXT]: use throughout
2883 <winshared>: new target
2884 <wininstall>: new target
2885 <dftables.o>: use native compiler
2886 <dftables>: use native linker
2887 <install>: handle Windows platform correctly
2888 <clean>: ditto
2889 <check>: ditto
2890 copy DLL to top builddir before testing
2891
2892 As part of these changes, -no-undefined was removed again. This was reported
2893 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2894 in any case.
2895
2896 3. Some tidies to get rid of compiler warnings:
2897
2898 . In the match_data structure, match_limit was an unsigned long int, whereas
2899 match_call_count was an int. I've made them both unsigned long ints.
2900
2901 . In pcretest the fact that a const uschar * doesn't automatically cast to
2902 a void * provoked a warning.
2903
2904 . Turning on some more compiler warnings threw up some "shadow" variables
2905 and a few more missing casts.
2906
2907 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2908 option, a class that contained a single character with a value between 128
2909 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2910
2911 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2912 option, a class that contained several characters, but with at least one
2913 whose value was between 128 and 255 caused PCRE to crash.
2914
2915
2916 Version 4.1 12-Mar-03
2917 ---------------------
2918
2919 1. Compiling with gcc -pedantic found a couple of places where casts were
2920 needed, and a string in dftables.c that was longer than standard compilers are
2921 required to support.
2922
2923 2. Compiling with Sun's compiler found a few more places where the code could
2924 be tidied up in order to avoid warnings.
2925
2926 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2927 first of these names is deprecated in the latest Autoconf in favour of the name
2928 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2929 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2930 analogy I have changed it to CFLAGS_FOR_BUILD.
2931
2932 4. Added -no-undefined to the linking command in the Makefile, because this is
2933 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2934 linking step for the pcreposix library.
2935
2936 5. PCRE was failing to diagnose the case of two named groups with the same
2937 name.
2938
2939 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2940 literal character that is needed in the subject for a match, and scans along to
2941 ensure that it is present before embarking on the full matching process. This
2942 saves time in cases of nested unlimited repeats that are never going to match.
2943 Problem: the scan can take a lot of time if the subject is very long (e.g.
2944 megabytes), thus penalizing straightforward matches. It is now done only if the
2945 amount of subject to be scanned is less than 1000 bytes.
2946
2947 7. A lesser problem with the same optimization is that it was recording the
2948 first character of an anchored pattern as "needed", thus provoking a search
2949 right along the subject, even when the first match of the pattern was going to
2950 fail. The "needed" character is now not set for anchored patterns, unless it
2951 follows something in the pattern that is of non-fixed length. Thus, it still
2952 fulfils its original purpose of finding quick non-matches in cases of nested
2953 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2954
2955
2956 Version 4.0 17-Feb-03
2957 ---------------------
2958
2959 1. If a comment in an extended regex that started immediately after a meta-item
2960 extended to the end of string, PCRE compiled incorrect data. This could lead to
2961 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2962
2963 2. Moved to autoconf 2.53 and libtool 1.4.2.
2964
2965 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2966 the special perltest8 script is no longer needed - all the tests can be run
2967 from a single perltest script.
2968
2969 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2970 by \s. It has now been removed in PCRE. This means it isn't recognized as
2971 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2972 class [:space:] *does* include VT, thereby creating a mess.
2973
2974 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2975 space and tab.
2976
2977 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2978 its new features into the main test script, reducing the number of scripts.
2979
2980 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2981 were backward compatible, and made the (?i) apply to the whole pattern, as if
2982 /i were given. Now it behaves more logically, and applies the option setting
2983 only to what follows. PCRE has been changed to follow suit. However, if it
2984 finds options settings right at the start of the pattern, it extracts them into
2985 the global options, as before. Thus, they show up in the info data.
2986
2987 8. Added support for the \Q...\E escape sequence. Characters in between are
2988 treated as literals. This is slightly different from Perl in that $ and @ are
2989 also handled as literals inside the quotes. In Perl, they will cause variable
2990 interpolation. Note the following examples:
2991
2992 Pattern PCRE matches Perl matches
2993
2994 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2995 \Qabc\$xyz\E abc\$xyz abc\$xyz
2996 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2997
2998 For compatibility with Perl, \Q...\E sequences are recognized inside character
2999 classes as well as outside them.
3000
3001 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3002 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3003 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3004 signed/unsigned warnings.
3005
3006 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3007 option for pcretest, so I've replaced it by a simple function that does just
3008 that job.
3009
3010 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3011 "pcregrep -".
3012
3013 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3014 Java package. This provides some syntactic sugar for simple cases of what my
3015 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3016 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3017 item, you can use this simplified notation. Note that only makes sense with
3018 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3019 greediness, whatever the setting of the PCRE_UNGREEDY option.
3020
3021 13. A change of greediness default within a pattern was not taking effect at
3022 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3023 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3024 was abstracted outside.
3025
3026 14. PCRE now supports the \G assertion. It is true when the current matching
3027 position is at the start point of the match. This differs from \A when the
3028 starting offset is non-zero. Used with the /g option of pcretest (or similar
3029 code), it works in the same way as it does for Perl's /g option. If all
3030 alternatives of a regex begin with \G, the expression is anchored to the start
3031 match position, and the "anchored" flag is set in the compiled expression.
3032
3033 15. Some bugs concerning the handling of certain option changes within patterns
3034 have been fixed. These applied to options other than (?ims). For example,
3035 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3036 been the other way round. Some of this was related to change 7 above.
3037
3038 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3039 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3040 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3041 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3042
3043 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3044 mode. Unlike ".", it always matches newline, whatever the setting of
3045 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3046 assertions. Perl allows it, but it doesn't (in general) work because it can't
3047 calculate the length of the lookbehind. At least, that's the case for Perl
3048 5.8.0 - I've been told they are going to document that it doesn't work in
3049 future.
3050
3051 18. Added an error diagnosis for escapes that PCRE does not support: these are
3052 \L, \l, \N, \P, \p, \U, \u, and \X.
3053
3054 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3055 reading past the end of the pattern in cases such as /[abcd/.
3056
3057 20. PCRE was getting more memory than necessary for patterns with classes that
3058 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3059
3060 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3061 compiling PCRE for use with Virtual Pascal.
3062
3063 22. Small fix to the Makefile to make it work properly if the build is done
3064 outside the source tree.
3065
3066 23. Added a new extension: a condition to go with recursion. If a conditional
3067 subpattern starts with (?(R) the "true" branch is used if recursion has
3068 happened, whereas the "false" branch is used only at the top level.
3069
3070 24. When there was a very long string of literal characters (over 255 bytes
3071 without UTF support, over 250 bytes with UTF support), the computation of how
3072 much memory was required could be incorrect, leading to segfaults or other
3073 strange effects.
3074
3075 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3076 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3077 there was a subsequent back reference to those brackets. This meant that, for
3078 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3079 possible to check for precisely this case. All we can do is abandon the
3080 optimization if .* occurs inside capturing brackets when there are any back
3081 references whatsoever. (See below for a better fix that came later.)
3082
3083 26. The handling of the optimization for finding the first character of a
3084 non-anchored pattern, and for finding a character that is required later in the
3085 match were failing in some cases. This didn't break the matching; it just
3086 failed to optimize when it could. The way this is done has been re-implemented.
3087
3088 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3089
3090 28. Added a new feature that provides some of the functionality that Perl
3091 provides with (?{...}). The facility is termed a "callout". The way it is done
3092 in PCRE is for the caller to provide an optional function, by setting
3093 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3094 global variable. By default it is unset, which disables all calling out. To get
3095 the function called, the regex must include (?C) at appropriate points. This
3096 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3097 This provides a means of identifying different callout points. When PCRE
3098 reaches such a point in the regex, if pcre_callout has been set, the external
3099 function is called. It is provided with data in a structure called
3100 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3101 matching continues; if it returns a non-zero value, the match at the current
3102 point fails. However, backtracking will occur if possible. [This was changed
3103 later and other features added - see item 49 below.]
3104
3105 29. pcretest is upgraded to test the callout functionality. It provides a
3106 callout function that displays information. By default, it shows the start of
3107 the match and the current position in the text. There are some new data escapes
3108 to vary what happens:
3109
3110 \C+ in addition, show current contents of captured substrings
3111 \C- do not supply a callout function
3112 \C!n return 1 when callout number n is reached
3113 \C!n!m return 1 when callout number n is reached for the mth time
3114
3115 30. If pcregrep was called with the -l option and just a single file name, it
3116 output "<stdin>" if a match was found, instead of the file name.
3117
3118 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3119 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3120 pcre_exec(). This saves a malloc/free per call. The default value of
3121 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3122 when configuring.
3123
3124 32. The default maximum size of a compiled pattern is 64K. There have been a
3125 few cases of people hitting this limit. The code now uses macros to handle the
3126 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3127 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3128 configuring. Tests 2 and 5 work only with 2-byte links because they output
3129 debugging information about compiled patterns.
3130
3131 33. Internal code re-arrangements:
3132
3133 (a) Moved the debugging function for printing out a compiled regex into
3134 its own source file (printint.c) and used #include to pull it into
3135 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3136 separate copies.
3137
3138 (b) Defined the list of op-code names for debugging as a macro in
3139 internal.h so that it is next to the definition of the opcodes.
3140
3141 (c) Defined a table of op-code lengths for simpler skipping along compiled
3142 code. This is again a macro in internal.h so that it is next to the
3143 definition of the opcodes.
3144
3145 34. Added support for recursive calls to individual subpatterns, along the
3146 lines of Robin Houston's patch (but implemented somewhat differently).
3147
3148 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3149 allow it to read and process whole directories in Win32. This code was
3150 contributed by Lionel Fourquaux; it has not been tested by me.
3151
3152 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3153 used to name a group. Names consist of alphanumerics and underscores, and must
3154 be unique. Back references use the syntax (?P=name) and recursive calls use
3155 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3156 numbers. The function pcre_fullinfo() can be used after compilation to extract
3157 a name/number map. There are three relevant calls:
3158
3159 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3160 PCRE_INFO_NAMECOUNT yields the number of entries
3161 PCRE_INFO_NAMETABLE yields a pointer to the map.
3162
3163 The map is a vector of fixed-size entries. The size of each entry depends on
3164 the length of the longest name used. The first two bytes of each entry are the
3165 group number, most significant byte first. There follows the corresponding
3166 name, zero terminated. The names are in alphabetical order.
3167
3168 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3169 case instead of 255. Making it the same both with and without UTF-8 support
3170 means that the same test output works with both.
3171
3172 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3173 calling malloc() with a zero argument.
3174
3175 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3176 optimization. I've improved things by keeping a bitmap of backreferences with
3177 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3178 fact referenced, the optimization can be applied. It is unlikely that a
3179 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3180 the match to follow \n) will appear inside brackets with a number greater than
3181 31, but if it does, any back reference > 31 suppresses the optimization.
3182
3183 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3184 of disabling numbered capturing parentheses. Any opening parenthesis that is
3185 not followed by ? behaves as if it were followed by ?: but named parentheses
3186 can still be used for capturing (and they will acquire numbers in the usual
3187 way).
3188
3189 41. Redesigned the return codes from the match() function into yes/no/error so
3190 that errors can be passed back from deep inside the nested calls. A malloc
3191 failure while inside a recursive subpattern call now causes the
3192 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3193
3194 42. It is now possible to set a limit on the number of times the match()
3195 function is called in a call to pcre_exec(). This facility makes it possible to
3196 limit the amount of recursion and backtracking, though not in a directly
3197 obvious way, because the match() function is used in a number of different
3198 circumstances. The count starts from zero for each position in the subject
3199 string (for non-anchored patterns). The default limit is, for compatibility, a
3200 large number, namely 10 000 000. You can change this in two ways:
3201
3202 (a) When configuring PCRE before making, you can use --with-match-limit=n
3203 to set a default value for the compiled library.
3204
3205 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3206 a different value is set. See 45 below.
3207
3208 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3209
3210 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3211 of things that can be changed at compile time. The first argument specifies
3212 what is wanted and the second points to where the information is to be placed.
3213 The current list of available information is:
3214
3215 PCRE_CONFIG_UTF8
3216
3217 The output is an integer that is set to one if UTF-8 support is available;
3218 otherwise it is set to zero.
3219
3220 PCRE_CONFIG_NEWLINE
3221
3222 The output is an integer that it set to the value of the code that is used for
3223 newline. It is either LF (10) or CR (13).
3224
3225 PCRE_CONFIG_LINK_SIZE
3226
3227 The output is an integer that contains the number of bytes used for internal
3228 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3229
3230 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3231
3232 The output is an integer that contains the threshold above which the POSIX
3233 interface uses malloc() for output vectors. See item 31 above.
3234
3235 PCRE_CONFIG_MATCH_LIMIT
3236
3237 The output is an unsigned integer that contains the default limit of the number
3238 of match() calls in a pcre_exec() execution. See 42 above.
3239
3240 44. pcretest has been upgraded by the addition of the -C option. This causes it
3241 to extract all the available output from the new pcre_config() function, and to
3242 output it. The program then exits immediately.
3243
3244 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3245 order to support additional features. One way would have been to define
3246 pcre_exec2() (for example) with extra arguments, but this would not have been
3247 extensible, and would also have required all calls to the original function to
3248 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3249 is used for passing in "extra" data from pcre_study().
3250
3251 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3252 contains the following fields:
3253
3254 flags a bitmap indicating which of the following fields are set
3255 study_data opaque data from pcre_study()
3256 match_limit a way of specifying a limit on match() calls for a specific
3257 call to pcre_exec()
3258 callout_data data for callouts (see 49 below)
3259
3260 The flag bits are also defined in pcre.h, and are
3261
3262 PCRE_EXTRA_STUDY_DATA
3263 PCRE_EXTRA_MATCH_LIMIT
3264 PCRE_EXTRA_CALLOUT_DATA
3265
3266 The pcre_study() function now returns one of these new pcre_extra blocks, with
3267 the actual study data pointed to by the study_data field, and the
3268 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3269 before. That is, this change is entirely upwards-compatible and requires no
3270 change to existing code.
3271
3272 If you want to pass in additional data to pcre_exec(), you can either place it
3273 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3274 block.
3275
3276 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3277 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3278 times with different match limits, until it finds the minimum value needed for
3279 pcre_exec() to complete. The value is then output. This can be instructive; for
3280 most simple matches the number is quite small, but for pathological cases it
3281 gets very large very quickly.
3282
3283 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3284 returns the size of the data block pointed to by the study_data field in a
3285 pcre_extra block, that is, the value that was passed as the argument to
3286 pcre_malloc() when PCRE was getting memory in which to place the information
3287 created by pcre_study(). The fourth argument should point to a size_t variable.
3288 pcretest has been extended so that this information is shown after a successful
3289 pcre_study() call when information about the compiled regex is being displayed.
3290
3291 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3292 because what follows is always an absolute path. (Later: it turns out that this
3293 is more than cosmetic for MinGW, because it doesn't like empty path
3294 components.)
3295
3296 49. Some changes have been made to the callout feature (see 28 above):
3297
3298 (i) A callout function now has three choices for what it returns:
3299
3300 0 => success, carry on matching
3301 > 0 => failure at this point, but backtrack if possible
3302 < 0 => serious error, return this value from pcre_exec()
3303
3304 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3305 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3306 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3307 use by callout functions. It will never be used by PCRE itself.
3308
3309 (ii) The pcre_extra structure (see 45 above) has a void * field called
3310 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3311 pcre_callout_block structure has a field of the same name. The contents of
3312 the field passed in the pcre_extra structure are passed to the callout
3313 function in the corresponding field in the callout block. This makes it
3314 easier to use the same callout-containing regex from multiple threads. For
3315 testing, the pcretest program has a new data escape
3316
3317 \C*n pass the number n (may be negative) as callout_data
3318
3319 If the callout function in pcretest receives a non-zero value as
3320 callout_data, it returns that value.
3321
3322 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3323 there were some redundant $(CFLAGS) in commands that are now specified as
3324 $(LINK), which already includes $(CFLAGS).
3325
3326 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3327 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3328 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3329 one-byte characters throughout. Note that case-insensitive matching applies
3330 only to characters whose values are less than 256. PCRE doesn't support the
3331 notion of cases for higher-valued characters.
3332
3333 (i) A character class whose characters are all within 0-255 is handled as
3334 a bit map, and the map is inverted for negative classes. Previously, a
3335 character > 255 always failed to match such a class; however it should
3336 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3337
3338 (ii) A negated character class with a single character < 255 is coded as
3339 "not this character" (OP_NOT). This wasn't working properly when the test
3340 character was multibyte, either singly or repeated.
3341
3342 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3343 mode, for example: \x{100}{2,3}.
3344
3345 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3346 singly or repeated) now correctly test multibyte characters. However,
3347 PCRE doesn't recognize any characters with values greater than 255 as
3348 digits, spaces, or word characters. Such characters always match \D, \S,
3349 and \W, and never match \d, \s, or \w.
3350
3351 (v) Classes may now contain characters and character ranges with values
3352 greater than 255. For example: [ab\x{100}-\x{400}].
3353
3354 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3355 PCRE in UTF-8 mode.
3356
3357 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3358 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3359 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3360 value.)
3361
3362 53. The single man page has become too large. I have therefore split it up into
3363 a number of separate man pages. These also give rise to individual HTML pages;
3364 these are now put in a separate directory, and there is an index.html page that
3365 lists them all. Some hyperlinking between the pages has been installed.
3366
3367 54. Added convenience functions for handling named capturing parentheses.
3368
3369 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3370 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3371 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3372 are faulted.
3373
3374 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3375 calling configure. These values are used when compiling the dftables.c program
3376 which is run to generate the source of the default character tables. They
3377 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3378 you will need to set these values.
3379
3380 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3381
3382
3383 Version 3.9 02-Jan-02
3384 ---------------------
3385
3386 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3387
3388 2. If --disable-static was given, the building process failed when trying to
3389 build pcretest and pcregrep. (For some reason it was using libtool to compile
3390 them, which is not right, as they aren't part of the library.)
3391
3392
3393 Version 3.8 18-Dec-01
3394 ---------------------
3395
3396 1. The experimental UTF-8 code was completely screwed up. It was packing the
3397 bytes in the wrong order. How dumb can you get?
3398
3399
3400 Version 3.7 29-Oct-01
3401 ---------------------
3402
3403 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3404 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3405 this didn't happen under Solaris 8, where I normally test things.
3406
3407 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3408 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3409 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3410 this fix an hour or so after the initial 3.7 release.)
3411
3412
3413 Version 3.6 23-Oct-01
3414 ---------------------
3415
3416 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3417 offsets passed as NULL with zero offset count.
3418
3419 2. The config.guess and config.sub files had not been updated when I moved to
3420 the latest autoconf.
3421
3422
3423 Version 3.5 15-Aug-01
3424 ---------------------
3425
3426 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3427 had been forgotten.
3428
3429 2. By using declared but undefined structures, we can avoid using "void"
3430 definitions in pcre.h while keeping the internal definitions of the structures
3431 private.
3432
3433 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3434 user point of view, this means that both static and shared libraries are built
3435 by default, but this can be individually controlled. More of the work of
3436 handling this static/shared cases is now inside libtool instead of PCRE's make
3437 file.
3438
3439 4. The pcretest utility is now installed along with pcregrep because it is
3440 useful for users (to test regexs) and by doing this, it automatically gets
3441 relinked by libtool. The documentation has been turned into a man page, so
3442 there are now .1, .txt, and .html versions in /doc.
3443
3444 5. Upgrades to pcregrep:
3445 (i) Added long-form option names like gnu grep.
3446 (ii) Added --help to list all options with an explanatory phrase.
3447 (iii) Added -r, --recursive to recurse into sub-directories.
3448 (iv) Added -f, --file to read patterns from a file.
3449
3450 6. pcre_exec() was referring to its "code" argument before testing that
3451 argument for NULL (and giving an error if it was NULL).
3452
3453 7. Upgraded Makefile.in to allow for compiling in a different directory from
3454 the source directory.
3455
3456 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3457 options bits, the pointer it was passed was to an int instead of to an unsigned
3458 long int. This mattered only on 64-bit systems.
3459
3460 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3461 generated) instead of pcre.in, which it its source. Also made the same change
3462 in several of the .c files.
3463
3464 10. A new release of gcc defines printf() as a macro, which broke pcretest
3465 because it had an ifdef in the middle of a string argument for printf(). Fixed
3466 by using separate calls to printf().
3467
3468 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3469 script, to force use of CR or LF instead of \n in the source. On non-Unix
3470 systems, the value can be set in config.h.
3471
3472 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3473 absolute limit. Changed the text of the error message to make this clear, and
3474 likewise updated the man page.
3475
3476 13. The limit of 99 on the number of capturing subpatterns has been removed.
3477 The new limit is 65535, which I hope will not be a "real" limit.
3478
3479
3480 Version 3.4 22-Aug-00
3481 ---------------------
3482
3483 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3484
3485 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3486
3487
3488 Version 3.3 01-Aug-00
3489 ---------------------
3490
3491 1. If an octal character was given, but the value was greater than \377, it
3492 was not getting masked to the least significant bits, as documented. This could
3493 lead to crashes in some systems.
3494
3495 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3496 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3497
3498 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3499 These just pass their arguments on to (pcre_free)(), but they are provided
3500 because some uses of PCRE bind it to non-C systems that can call its functions,
3501 but cannot call free() or pcre_free() directly.
3502
3503 4. Add "make test" as a synonym for "make check". Corrected some comments in
3504 the Makefile.
3505
3506 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3507 Makefile.
3508
3509 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3510 command called pgrep for grepping around the active processes.
3511
3512 7. Added the beginnings of support for UTF-8 character strings.
3513
3514 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3515 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3516 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3517 out for the ar command.)
3518
3519
3520 Version 3.2 12-May-00
3521 ---------------------
3522
3523 This is purely a bug fixing release.
3524
3525 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3526 of ZA. This was just one example of several cases that could provoke this bug,
3527 which was introduced by change 9 of version 2.00. The code for breaking
3528 infinite loops after an iteration that matches an empty string was't working
3529 correctly.
3530
3531 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3532 when matched against abbab (for example). After matching an empty string, it
3533 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3534 caused it to match further down the string than it should.
3535
3536 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3537 was there because it doesn't seem to be needed, and it causes trouble on some
3538 systems, as it is not a Standard C header. It has been removed.
3539
3540 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3541 were reported on the Macintosh. The changes were from
3542
3543 while ((c = *(++ptr)) != 0 && c != '\n');
3544 to
3545 while ((c = *(++ptr)) != 0 && c != '\n') ;
3546
3547 Totally extraordinary, but if that's what it takes...
3548
3549 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3550 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3551 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3552 assumes the way PCRE uses memmove() (always moving upwards).
3553
3554 6. PCRE is being used in one environment where strchr() is not available. There
3555 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3556 faster code anyway.
3557
3558
3559 Version 3.1 09-Feb-00
3560 ---------------------
3561
3562 The only change in this release is the fixing of some bugs in Makefile.in for
3563 the "install" target:
3564
3565 (1) It was failing to install pcreposix.h.
3566
3567 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3568
3569
3570 Version 3.0 01-Feb-00
3571 ---------------------
3572
3573 1. Add support for the /+ modifier to perltest (to output $` like it does in
3574 pcretest).
3575
3576 2. Add support for the /g modifier to perltest.
3577
3578 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3579 matches null strings.
3580
3581 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3582 pattern. Perl treats empty patterns specially - it reuses the most recent
3583 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3584 effect.
3585
3586 5. The POSIX interface was broken in that it was just handing over the POSIX
3587 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3588 required a bigger vector, with some working space on the end. This means that
3589 the POSIX wrapper now has to get and free some memory, and copy the results.
3590
3591 6. Added some simple autoconf support, placing the test data and the
3592 documentation in separate directories, re-organizing some of the
3593 information files, and making it build pcre-config (a GNU standard). Also added
3594 libtool support for building PCRE as a shared library, which is now the
3595 default.
3596
3597 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3598 09 are not valid octal constants. Single digits will be used for minor values
3599 less than 10.
3600
3601 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3602 existing programs that set these in the POSIX interface can use PCRE without
3603 modification.
3604
3605 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3606 return all that pcre_info() returns, plus additional data. The pcre_info()
3607 function is retained for compatibility, but is considered to be obsolete.
3608
3609 10. Added experimental recursion feature (?R) to handle one common case that
3610 Perl 5.6 will be able to do with (?p{...}).
3611
3612 11. Added support for POSIX character classes like [:alpha:], which Perl is
3613 adopting.
3614
3615
3616 Version 2.08 31-Aug-99
3617 ----------------------
3618
3619 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3620 trying to match at the startoffset position, but instead was moving forward to
3621 the next newline as if a previous match had failed.
3622
3623 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3624 and could get into a loop if a null string was matched other than at the start
3625 of the subject.
3626
3627 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3628 be distinguished at compile time, and for completeness also added PCRE_DATE.
3629
3630 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3631 in GnuWin32 environments.
3632
3633
3634 Version 2.07 29-Jul-99
3635 ----------------------
3636
3637 1. The documentation is now supplied in plain text form and HTML as well as in
3638 the form of man page sources.
3639
3640 2. C++ compilers don't like assigning (void *) values to other pointer types.
3641 In particular this affects malloc(). Although there is no problem in Standard
3642 C, I've put in casts to keep C++ compilers happy.
3643
3644 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3645 should be (const char *).
3646
3647 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3648 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3649 However, I haven't made this a standard facility. The documentation doesn't
3650 mention it, and the Makefile doesn't support it.
3651
3652 5. The Makefile now contains an "install" target, with editable destinations at
3653 the top of the file. The pcretest program is not installed.
3654
3655 6. pgrep -V now gives the PCRE version number and date.
3656
3657 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3658 causing the entire string to be ignored, instead of just the last character.
3659
3660 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3661 non-matching string, it can take a very, very long time, even for strings of
3662 quite modest length, because of the nested recursion. PCRE now does better in
3663 some of these cases. It does this by remembering the last required literal
3664 character in the pattern, and pre-searching the subject to ensure it is present
3665 before running the real match. In other words, it applies a heuristic to detect
3666 some types of certain failure quickly, and in the above example, if presented
3667 with a string that has no trailing " it gives "no match" very quickly.
3668
3669 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3670 other alternatives are tried instead.
3671
3672
3673 Version 2.06 09-Jun-99
3674 ----------------------
3675
3676 1. Change pcretest's output for amount of store used to show just the code
3677 space, because the remainder (the data block) varies in size between 32-bit and
3678 64-bit systems.
3679
3680 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3681 start matching at. This allows lookbehinds to work when searching for multiple
3682 occurrences in a string.
3683
3684 3. Added additional options to pcretest for testing multiple occurrences:
3685
3686 /+ outputs the rest of the string that follows a match
3687 /g loops for multiple occurrences, using the new startoffset argument
3688 /G loops for multiple occurrences by passing an incremented pointer
3689
3690 4. PCRE wasn't doing the "first character" optimization for patterns starting
3691 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3692 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3693 the letter 'x'. On long subject strings, this gives a significant speed-up.
3694
3695
3696 Version 2.05 21-Apr-99
3697 ----------------------
3698
3699 1. Changed the type of magic_number from int to long int so that it works
3700 properly on 16-bit systems.
3701
3702 2. Fixed a bug which caused patterns starting with .* not to work correctly
3703 when the subject string contained newline characters. PCRE was assuming
3704 anchoring for such patterns in all cases, which is not correct because .* will
3705 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3706 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3707 must be retried after every newline in the subject.
3708
3709
3710 Version 2.04 18-Feb-99
3711 ----------------------
3712
3713 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3714 computation of the store needed to hold the pattern was incorrect (too large).
3715 If such patterns were nested a few deep, this could multiply and become a real
3716 problem.
3717
3718 2. Added /M option to pcretest to show the memory requirement of a specific
3719 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3720
3721 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3722 compiled in such a way that the backtracking after subsequent failure was
3723 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3724 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3725
3726
3727 Version 2.03 02-Feb-99
3728 ----------------------
3729
3730 1. Fixed typo and small mistake in man page.
3731
3732 2. Added 4th condition (GPL supersedes if conflict) and created separate
3733 LICENCE file containing the conditions.
3734
3735 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3736 Perl, that is the internal \ allows the delimiter to be included in the
3737 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3738 the final delimiter, add \ to the end of the pattern (to test the error).
3739
3740 4. Added the convenience functions for extracting substrings after a successful
3741 match. Updated pcretest to make it able to test these functions.
3742
3743
3744 Version 2.02 14-Jan-99
3745 ----------------------
3746
3747 1. Initialized the working variables associated with each extraction so that
3748 their saving and restoring doesn't refer to uninitialized store.
3749
3750 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3751 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3752 fix the problem.
3753
3754 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3755 calls, and wasn't printing the correct value for compiling calls. Increased the
3756 default value of LOOPREPEAT, and the number of significant figures in the
3757 times.
3758
3759 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3760
3761 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3762 a building problem on Windows NT with a FAT file system.
3763
3764
3765 Version 2.01 21-Oct-98
3766 ----------------------
3767
3768 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3769 to character tables built by pcre_maketables() in the current locale. If NULL
3770 is passed, the default tables are used.
3771
3772
3773 Version 2.00 24-Sep-98
3774 ----------------------
3775
3776 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3777 it any more.
3778
3779 2. Allow quantification of (?>) groups, and make it work correctly.
3780
3781 3. The first character computation wasn't working for (?>) groups.
3782
3783 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3784 end of the subject) and add 5.005's \z, which really does match only at the
3785 very end of the subject.
3786
3787 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3788
3789 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3790 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3791 localized options. All options to pcre_study() were also removed.
3792
3793 7. Add other new features from 5.005:
3794
3795 $(?<= positive lookbehind
3796 $(?<! negative lookbehind
3797 (?imsx-imsx) added the unsetting capability
3798 such a setting is global if at outer level; local otherwise
3799 (?imsx-imsx:) non-capturing groups with option setting
3800 (?(cond)re|re) conditional pattern matching
3801
3802 A backreference to itself in a repeated group matches the previous
3803 captured string.
3804
3805 8. General tidying up of studying (both automatic and via "study")
3806 consequential on the addition of new assertions.
3807
3808 9. As in 5.005, unlimited repeated groups that could match an empty substring
3809 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3810 runtime if any iteration does actually match an empty substring.
3811
3812 10. Include the RunTest script in the distribution.
3813
3814 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3815 discrepancies, some of which were old and were also with respect to 5.004. They
3816 have now been fixed.
3817
3818
3819 Version 1.09 28-Apr-98
3820 ----------------------
3821
3822 1. A negated single character class followed by a quantifier with a minimum
3823 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3824 program crashes, or just wrong answers. This did not apply to negated classes
3825 containing more than one character, or to minima other than one.
3826
3827
3828 Version 1.08 27-Mar-98
3829 ----------------------
3830
3831 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3832
3833 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3834 latter must appear before anything that relies on it in the pattern.
3835
3836
3837 Version 1.07 16-Feb-98
3838 ----------------------
3839
3840 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3841 repeat of a potentially empty string).
3842
3843
3844 Version 1.06 23-Jan-98
3845 ----------------------
3846
3847 1. Added Markus Oberhumer's little patches for C++.
3848
3849 2. Literal strings longer than 255 characters were broken.
3850
3851
3852 Version 1.05 23-Dec-97
3853 ----------------------
3854
3855 1. Negated character classes containing more than one character were failing if
3856 PCRE_CASELESS was set at run time.
3857
3858
3859 Version 1.04 19-Dec-97
3860 ----------------------
3861
3862 1. Corrected the man page, where some "const" qualifiers had been omitted.
3863
3864 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3865 input syntax.
3866
3867 3. Fixed memory leak which occurred when a regex with back references was
3868 matched with an offsets vector that wasn't big enough. The temporary memory
3869 that is used in this case wasn't being freed if the match failed.
3870
3871 4. Tidied pcretest to ensure it frees memory that it gets.
3872
3873 5. Temporary memory was being obtained in the case where the passed offsets
3874 vector was exactly big enough.
3875
3876 6. Corrected definition of offsetof() from change 5 below.
3877
3878 7. I had screwed up change 6 below and broken the rules for the use of
3879 setjmp(). Now fixed.
3880
3881
3882 Version 1.03 18-Dec-97
3883 ----------------------
3884
3885 1. A erroneous regex with a missing opening parenthesis was correctly
3886 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3887 on some systems.
3888
3889 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3890 it was reported that one broken compiler failed on the former because "code" is
3891 also an independent variable.
3892
3893 3. The erroneous regex a[]b caused an array overrun reference.
3894
3895 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3896 fail on data ending with that character. (It was going on too far, and checking
3897 the next character, typically a binary zero.) This was specific to the
3898 optimized code for single-character negative classes.
3899
3900 5. Added a contributed patch from the TIN world which does the following:
3901
3902 + Add an undef for memmove, in case the the system defines a macro for it.
3903
3904 + Add a definition of offsetof(), in case there isn't one. (I don't know
3905 the reason behind this - offsetof() is part of the ANSI standard - but
3906 it does no harm).
3907
3908 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3909 most of the places where whitespace preceded '#'. I have given up and
3910 allowed the remaining 2 cases to be at the margin.
3911
3912 + Rename some variables in pcre to eliminate shadowing. This seems very
3913 pedantic, but does no harm, of course.
3914
3915 6. Moved the call to setjmp() into its own function, to get rid of warnings
3916 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3917
3918 7. Constructs such as \d{8,} were compiling into the equivalent of
3919 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3920 outcome, but in this particular case used more store than had been allocated,
3921 which caused the bug to be discovered because it threw up an internal error.
3922
3923 8. The debugging code in both pcre and pcretest for outputting the compiled
3924 form of a regex was going wrong in the case of back references followed by
3925 curly-bracketed repeats.
3926
3927
3928 Version 1.02 12-Dec-97
3929 ----------------------
3930
3931 1. Typos in pcre.3 and comments in the source fixed.
3932
3933 2. Applied a contributed patch to get rid of places where it used to remove
3934 'const' from variables, and fixed some signed/unsigned and uninitialized
3935 variable warnings.
3936
3937 3. Added the "runtest" target to Makefile.
3938
3939 4. Set default compiler flag to -O2 rather than just -O.
3940
3941
3942 Version 1.01 19-Nov-97
3943 ----------------------
3944
3945 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3946 like /([ab]*)*/, that is, for classes with more than one character in them.
3947
3948 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3949 as /((?>a*))*/ (a PCRE_EXTRA facility).
3950
3951
3952 Version 1.00 18-Nov-97
3953 ----------------------
3954
3955 1. Added compile-time macros to support systems such as SunOS4 which don't have
3956 memmove() or strerror() but have other things that can be used instead.
3957
3958 2. Arranged that "make clean" removes the executables.
3959
3960
3961 Version 0.99 27-Oct-97
3962 ----------------------
3963
3964 1. Fixed bug in code for optimizing classes with only one character. It was
3965 initializing a 32-byte map regardless, which could cause it to run off the end
3966 of the memory it had got.
3967
3968 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3969
3970
3971 Version 0.98 22-Oct-97
3972 ----------------------
3973
3974 1. Fixed bug in code for handling temporary memory usage when there are more
3975 back references than supplied space in the ovector. This could cause segfaults.
3976
3977
3978 Version 0.97 21-Oct-97
3979 ----------------------
3980
3981 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3982
3983 2. Optimized negated single characters not to use a bit map.
3984
3985 3. Brought error texts together as macro definitions; clarified some of them;
3986 fixed one that was wrong - it said "range out of order" when it meant "invalid
3987 escape sequence".
3988
3989 4. Changed some char * arguments to const char *.
3990
3991 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3992
3993 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3994 pcretest.
3995
3996
3997 Version 0.96 16-Oct-97
3998 ----------------------
3999
4000 1. Added a simple "pgrep" utility to the distribution.
4001
4002 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4003 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4004 where "ddd" means "one or more decimal digits".
4005
4006 3. Fixed serious bug. If a pattern had a back reference, but the call to
4007 pcre_exec() didn't supply a large enough ovector to record the related
4008 identifying subpattern, the match always failed. PCRE now remembers the number
4009 of the largest back reference, and gets some temporary memory in which to save
4010 the offsets during matching if necessary, in order to ensure that
4011 backreferences always work.
4012
4013 4. Increased the compatibility with Perl in a number of ways:
4014
4015 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4016 to request this handling. The option can be set at compile or exec time.
4017
4018 (b) $ matches before a terminating newline by default; an option
4019 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4020 mode). The option can be set at compile or exec time.
4021
4022 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4023 the same as Perl's. If the decimal number it represents is less than 10
4024 or there aren't that many previous left capturing parentheses, an octal
4025 escape is read. Inside a character class, it's always an octal escape,
4026 even if it is a single digit.
4027
4028 (d) An escaped but undefined alphabetic character is taken as a literal,
4029 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4030 escapes.
4031
4032 (e) {0} is now permitted. (The previous item is removed from the compiled
4033 pattern).
4034
4035 5. Changed all the names of code files so that the basic parts are no longer
4036 than 10 characters, and abolished the teeny "globals.c" file.
4037
4038 6. Changed the handling of character classes; they are now done with a 32-byte
4039 bit map always.
4040
4041 7. Added the -d and /D options to pcretest to make it possible to look at the
4042 internals of compilation without having to recompile pcre.
4043
4044
4045 Version 0.95 23-Sep-97
4046 ----------------------
4047
4048 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4049 \x20 at the start of a run of normal characters. These were being treated as
4050 real characters, instead of the source characters being re-checked.
4051
4052
4053 Version 0.94 18-Sep-97
4054 ----------------------
4055
4056 1. The functions are now thread-safe, with the caveat that the global variables
4057 containing pointers to malloc() and free() or alternative functions are the
4058 same for all threads.
4059
4060 2. Get pcre_study() to generate a bitmap of initial characters for non-
4061 anchored patterns when this is possible, and use it if passed to pcre_exec().
4062
4063
4064 Version 0.93 15-Sep-97
4065 ----------------------
4066
4067 1. /(b)|(:+)/ was computing an incorrect first character.
4068
4069 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4070 but not actually doing anything yet.
4071
4072 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4073 as Perl does (e.g. [-az] or [az-]).
4074
4075 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4076 all possible positions.
4077
4078 5. Split up into different modules to avoid including unneeded functions in a
4079 compiled binary. However, compile and exec are still in one module. The "study"
4080 function is split off.
4081
4082 6. The character tables are now in a separate module whose source is generated
4083 by an auxiliary program - but can then be edited by hand if required. There are
4084 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4085 toupper() in the code.
4086
4087 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4088 make them global. Abolish the function for setting them, as the caller can now
4089 set them directly.
4090
4091
4092 Version 0.92 11-Sep-97
4093 ----------------------
4094
4095 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4096 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4097
4098 2. Caseless matching was not working in character classes if the characters in
4099 the pattern were in upper case.
4100
4101 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4102
4103 4. Make PCRE_ANCHORED public and accept as a compile option.
4104
4105 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4106 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4107 pass them.
4108
4109 6. Give an error if bad option bits passed at compile or run time.
4110
4111 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4112 pcretest to cause it to pass that flag.
4113
4114 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4115 options, and the first character, if set.
4116
4117 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4118
4119
4120 Version 0.91 10-Sep-97
4121 ----------------------
4122
4123 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4124 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4125
4126 2. PCRE was looping on encountering an indefinitely repeated back reference to
4127 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4128 Perl does - treats the match as successful.
4129
4130 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5