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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 589 - (show annotations) (download)
Sat Jan 15 11:31:39 2011 UTC (3 years, 9 months ago) by ph10
File size: 198279 byte(s)
Tidies for Release 8.12

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.15 12-Jan-2011
5 ------------------------
6
7 1. Fixed some typos in the markup of the man pages, and wrote a script that
8 checks for such things as part of the documentation building process.
9
10 2. On a big-endian 64-bit system, pcregrep did not correctly process the
11 --match-limit and --recursion-limit options (added for 8.11). In
12 particular, this made one of the standard tests fail. (The integer value
13 went into the wrong half of a long int.)
14
15 3. If the --colour option was given to pcregrep with -v (invert match), it
16 did strange things, either producing crazy output, or crashing. It should,
17 of course, ignore a request for colour when reporting lines that do not
18 match.
19
20 4. Another pcregrep bug caused similar problems if --colour was specified with
21 -M (multiline) and the pattern match finished with a line ending.
22
23 5. In pcregrep, when a pattern that ended with a literal newline sequence was
24 matched in multiline mode, the following line was shown as part of the
25 match. This seems wrong, so I have changed it.
26
27 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
28 the check for further matches in the same line (so they could be coloured)
29 to overrun the end of the current line. If another match was found, it was
30 incorrectly shown (and then shown again when found in the next line).
31
32 7. If pcregrep was compiled under Windows, there was a reference to the
33 function pcregrep_exit() before it was defined. I am assuming this was
34 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
35 reported by a user. I've moved the definition above the reference.
36
37
38 Version 8.11 10-Dec-2010
39 ------------------------
40
41 1. (*THEN) was not working properly if there were untried alternatives prior
42 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
43 backtracked to try for "b" instead of moving to the next alternative branch
44 at the same level (in this case, to look for "c"). The Perl documentation
45 is clear that when (*THEN) is backtracked onto, it goes to the "next
46 alternative in the innermost enclosing group".
47
48 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
49 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
50 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
51 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
52 (*THEN).
53
54 3. If \s appeared in a character class, it removed the VT character from
55 the class, even if it had been included by some previous item, for example
56 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
57 of \s, but is part of the POSIX "space" class.)
58
59 4. A partial match never returns an empty string (because you can always
60 match an empty string at the end of the subject); however the checking for
61 an empty string was starting at the "start of match" point. This has been
62 changed to the "earliest inspected character" point, because the returned
63 data for a partial match starts at this character. This means that, for
64 example, /(?<=abc)def/ gives a partial match for the subject "abc"
65 (previously it gave "no match").
66
67 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
68 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
69 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
70 has an implication that the given string is incomplete (because a partial
71 match is preferred over a full match). For this reason, these items now
72 give a partial match in this situation. [Aside: previously, the one case
73 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
74 match rather than a full match, which was wrong by the old rules, but is
75 now correct.]
76
77 6. There was a bug in the handling of #-introduced comments, recognized when
78 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
79 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
80 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
81 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
82 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
83 places in pcre_compile().
84
85 7. Related to (6) above, when pcre_compile() was skipping #-introduced
86 comments when looking ahead for named forward references to subpatterns,
87 the only newline sequence it recognized was NL. It now handles newlines
88 according to the set newline convention.
89
90 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
91 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
92 cater for a lack of strerror(). These oversights have been fixed.
93
94 9. Added --match-limit and --recursion-limit to pcregrep.
95
96 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
97
98 11. When the -o option was used, pcregrep was setting a return code of 1, even
99 when matches were found, and --line-buffered was not being honoured.
100
101 12. Added an optional parentheses number to the -o and --only-matching options
102 of pcregrep.
103
104 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
105 can match an empty string. The code to do it in pcretest and pcredemo
106 needed fixing:
107
108 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
109 only one byte after an empty string match just before CRLF (this case
110 just got forgotten; "any" and "anycrlf" were OK).
111
112 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
113 mode when an empty string match preceded an ASCII character followed by
114 a non-ASCII character. (The code for advancing by one character rather
115 than one byte was nonsense.)
116
117 (c) The pcredemo.c sample program did not have any code at all to handle
118 the cases when CRLF is a valid newline sequence.
119
120 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
121 as a starting offset was within the subject string. There is now a new
122 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
123 negative or greater than the length of the string. In order to test this,
124 pcretest is extended to allow the setting of negative starting offsets.
125
126 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
127 starting offset points to the beginning of a UTF-8 character was
128 unnecessarily clumsy. I tidied it up.
129
130 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
131 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
132
133 17. Nobody had reported that the --include_dir option, which was added in
134 release 7.7 should have been called --include-dir (hyphen, not underscore)
135 for compatibility with GNU grep. I have changed it to --include-dir, but
136 left --include_dir as an undocumented synonym, and the same for
137 --exclude-dir, though that is not available in GNU grep, at least as of
138 release 2.5.4.
139
140 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
141 characters from a string of bytes) have been redefined so as not to use
142 loops, in order to improve performance in some environments. At the same
143 time, I abstracted some of the common code into auxiliary macros to save
144 repetition (this should not affect the compiled code).
145
146 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
147 compile-time error is now given if \c is not followed by an ASCII
148 character, that is, a byte less than 128. (In EBCDIC mode, the code is
149 different, and any byte value is allowed.)
150
151 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
152 START_OPTIMIZE option, which is now allowed at compile time - but just
153 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
154 to pcregrep and other applications that have no direct access to PCRE
155 options. The new /Y option in pcretest sets this option when calling
156 pcre_compile().
157
158 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
159 back references. Groups containing recursive back references were forced to
160 be atomic by that change, but in the case of named groups, the amount of
161 memory required was incorrectly computed, leading to "Failed: internal
162 error: code overflow". This has been fixed.
163
164 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
165 pcretest.c, to avoid build problems in some Borland environments.
166
167
168 Version 8.10 25-Jun-2010
169 ------------------------
170
171 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
172 THEN.
173
174 2. (*ACCEPT) was not working when inside an atomic group.
175
176 3. Inside a character class, \B is treated as a literal by default, but
177 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
178 causes the error). The code is unchanged, but I tidied the documentation.
179
180 4. Inside a character class, PCRE always treated \R and \X as literals,
181 whereas Perl faults them if its -w option is set. I have changed PCRE so
182 that it faults them when PCRE_EXTRA is set.
183
184 5. Added support for \N, which always matches any character other than
185 newline. (It is the same as "." when PCRE_DOTALL is not set.)
186
187 6. When compiling pcregrep with newer versions of gcc which may have
188 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
189 declared with attribute warn_unused_result" were given. Just casting the
190 result to (void) does not stop the warnings; a more elaborate fudge is
191 needed. I've used a macro to implement this.
192
193 7. Minor change to pcretest.c to avoid a compiler warning.
194
195 8. Added four artifical Unicode properties to help with an option to make
196 \s etc use properties (see next item). The new properties are: Xan
197 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
198
199 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
200 use Unicode properties. (*UCP) at the start of a pattern can be used to set
201 this option. Modified pcretest to add /W to test this facility. Added
202 REG_UCP to make it available via the POSIX interface.
203
204 10. Added --line-buffered to pcregrep.
205
206 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
207 studied, and the match started with a letter with a code point greater than
208 127 whose first byte was different to the first byte of the other case of
209 the letter, the other case of this starting letter was not recognized
210 (#976).
211
212 12. If a pattern that was studied started with a repeated Unicode property
213 test, for example, \p{Nd}+, there was the theoretical possibility of
214 setting up an incorrect bitmap of starting bytes, but fortunately it could
215 not have actually happened in practice until change 8 above was made (it
216 added property types that matched character-matching opcodes).
217
218 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
219 possible starting bytes for non-anchored patterns.
220
221 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
222 \R, and also a number of cases that involve Unicode properties, both
223 explicit and implicit when PCRE_UCP is set.
224
225 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
226 input, it could crash or give wrong results if characters with values
227 greater than 0xc0 were present in the subject string. (Detail: it assumed
228 UTF-8 input when processing these items.)
229
230 16. Added a lot of (int) casts to avoid compiler warnings in systems where
231 size_t is 64-bit (#991).
232
233 17. Added a check for running out of memory when PCRE is compiled with
234 --disable-stack-for-recursion (#990).
235
236 18. If the last data line in a file for pcretest does not have a newline on
237 the end, a newline was missing in the output.
238
239 19. The default pcre_chartables.c file recognizes only ASCII characters (values
240 less than 128) in its various bitmaps. However, there is a facility for
241 generating tables according to the current locale when PCRE is compiled. It
242 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
243 characters, are recognized by isspace() and therefore were getting set in
244 these tables, and indeed these tables seem to approximate to ISO 8859. This
245 caused a problem in UTF-8 mode when pcre_study() was used to create a list
246 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
247 which of course cannot start UTF-8 characters. I have changed the code so
248 that only real ASCII characters (less than 128) and the correct starting
249 bytes for UTF-8 encodings are set for characters greater than 127 when in
250 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
251 altogether.)
252
253 20. Added the /T option to pcretest so as to be able to run tests with non-
254 standard character tables, thus making it possible to include the tests
255 used for 19 above in the standard set of tests.
256
257 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
258 reference to a subpattern the other side of a comment that contains an
259 opening parenthesis caused either an internal compiling error, or a
260 reference to the wrong subpattern.
261
262
263 Version 8.02 19-Mar-2010
264 ------------------------
265
266 1. The Unicode data tables have been updated to Unicode 5.2.0.
267
268 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
269 configured.
270
271 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
272 original author of that file, following a query about its status.
273
274 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
275 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
276
277 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
278 quantifier applied to a forward-referencing subroutine call, could compile
279 incorrect code or give the error "internal error: previously-checked
280 referenced subpattern not found".
281
282 6. Both MS Visual Studio and Symbian OS have problems with initializing
283 variables to point to external functions. For these systems, therefore,
284 pcre_malloc etc. are now initialized to local functions that call the
285 relevant global functions.
286
287 7. There were two entries missing in the vectors called coptable and poptable
288 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
289 I've fixed the data, and added a kludgy way of testing at compile time that
290 the lengths are correct (equal to the number of opcodes).
291
292 8. Following on from 7, I added a similar kludge to check the length of the
293 eint vector in pcreposix.c.
294
295 9. Error texts for pcre_compile() are held as one long string to avoid too
296 much relocation at load time. To find a text, the string is searched,
297 counting zeros. There was no check for running off the end of the string,
298 which could happen if a new error number was added without updating the
299 string.
300
301 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
302
303 11. \K was not working if it appeared in an atomic group or in a group that
304 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
305 \K is "not well defined" if used in an assertion. PCRE now accepts it if
306 the assertion is positive, but not if it is negative.
307
308 12. Change 11 fortuitously reduced the size of the stack frame used in the
309 "match()" function of pcre_exec.c by one pointer. Forthcoming
310 implementation of support for (*MARK) will need an extra pointer on the
311 stack; I have reserved it now, so that the stack frame size does not
312 decrease.
313
314 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
315 item in branch that calls a recursion is a subroutine call - as in the
316 second branch in the above example - was incorrectly given the compile-
317 time error "recursive call could loop indefinitely" because pcre_compile()
318 was not correctly checking the subroutine for matching a non-empty string.
319
320 14. The checks for overrunning compiling workspace could trigger after an
321 overrun had occurred. This is a "should never occur" error, but it can be
322 triggered by pathological patterns such as hundreds of nested parentheses.
323 The checks now trigger 100 bytes before the end of the workspace.
324
325 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
326
327
328 Version 8.01 19-Jan-2010
329 ------------------------
330
331 1. If a pattern contained a conditional subpattern with only one branch (in
332 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
333 computed the wrong minimum data length (which is of course zero for such
334 subpatterns). This could cause incorrect "no match" results.
335
336 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
337 the pattern is reset in the first branch, pcre_compile() failed with
338 "internal error: code overflow at offset...". This happened only when
339 the reset was to the original external option setting. (An optimization
340 abstracts leading options settings into an external setting, which was the
341 cause of this.)
342
343 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
344 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
345 assertion pattern did not match (meaning that the assertion was true), it
346 was incorrectly treated as false if the SKIP had been reached during the
347 matching. This also applied to assertions used as conditions.
348
349 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
350 assertion subpattern, including such a pattern used as a condition,
351 unpredictable results occurred, instead of the error return
352 PCRE_ERROR_DFA_UITEM.
353
354 5. The C++ GlobalReplace function was not working like Perl for the special
355 situation when an empty string is matched. It now does the fancy magic
356 stuff that is necessary.
357
358 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
359 removed. (These were left over from very, very early versions of PCRE.)
360
361 7. Some cosmetic changes to the code to make life easier when compiling it
362 as part of something else:
363
364 (a) Change DEBUG to PCRE_DEBUG.
365
366 (b) In pcre_compile(), rename the member of the "branch_chain" structure
367 called "current" as "current_branch", to prevent a collision with the
368 Linux macro when compiled as a kernel module.
369
370 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
371 prevent a collision with the Linux macro when compiled as a kernel
372 module.
373
374 8. In pcre_compile() there are some checks for integer overflows that used to
375 cast potentially large values to (double). This has been changed to that
376 when building, a check for int64_t is made, and if it is found, it is used
377 instead, thus avoiding the use of floating point arithmetic. (There is no
378 other use of FP in PCRE.) If int64_t is not found, the fallback is to
379 double.
380
381 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
382 2005 (difference between two addresses compared to an unsigned value).
383
384 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
385 custom one, because of the following reported problem in Windows:
386
387 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
388 under Win32.
389 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
390 therefore missing the function definition.
391 - The compiler thus generates a "C" signature for the test function.
392 - The linker fails to find the "C" function.
393 - PCRE fails to configure if asked to do so against libbz2.
394
395 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
396 messages were output:
397
398 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
399 rerunning libtoolize, to keep the correct libtool macros in-tree.
400 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
401
402 I have done both of these things.
403
404 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
405 most of the time, it *can* run out if it is given a pattern that contains a
406 runaway infinite recursion. I updated the discussion in the pcrestack man
407 page.
408
409 13. Now that we have gone to the x.xx style of version numbers, the minor
410 version may start with zero. Using 08 or 09 is a bad idea because users
411 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
412 interpreted as invalid octal numbers. I've updated the previous comment in
413 configure.ac, and also added a check that gives an error if 08 or 09 are
414 used.
415
416 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
417 causing partial matching to fail when the end of the subject matched \W
418 in a UTF-8 pattern where \W was quantified with a minimum of 3.
419
420 15. There were some discrepancies between the declarations in pcre_internal.h
421 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
422 their definitions. The declarations used "const uschar *" and the
423 definitions used USPTR. Even though USPTR is normally defined as "const
424 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
425 reported that: "This difference in casting confuses some C++ compilers, for
426 example, SunCC recognizes above declarations as different functions and
427 generates broken code for hbpcre." I have changed the declarations to use
428 USPTR.
429
430 16. GNU libtool is named differently on some systems. The autogen.sh script now
431 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
432 (FreeBSD).
433
434 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
435 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
436 comment: "Figure out how to create a longlong from a string: strtoll and
437 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
438 instance, but it only takes 2 args instead of 3!"
439
440 18. A subtle bug concerned with back references has been fixed by a change of
441 specification, with a corresponding code fix. A pattern such as
442 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
443 refers, was giving matches when it shouldn't. For example, xa=xaaa would
444 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
445 same bug. Such groups have to be quantified to be useful, or contained
446 inside another quantified group. (If there's no repetition, the reference
447 can never match.) The problem arises because, having left the group and
448 moved on to the rest of the pattern, a later failure that backtracks into
449 the group uses the captured value from the final iteration of the group
450 rather than the correct earlier one. I have fixed this in PCRE by forcing
451 any group that contains a reference to itself to be an atomic group; that
452 is, there cannot be any backtracking into it once it has completed. This is
453 similar to recursive and subroutine calls.
454
455
456 Version 8.00 19-Oct-09
457 ----------------------
458
459 1. The table for translating pcre_compile() error codes into POSIX error codes
460 was out-of-date, and there was no check on the pcre_compile() error code
461 being within the table. This could lead to an OK return being given in
462 error.
463
464 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
465 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
466 in a Windows environment.
467
468 3. The pcregrep --count option prints the count for each file even when it is
469 zero, as does GNU grep. However, pcregrep was also printing all files when
470 --files-with-matches was added. Now, when both options are given, it prints
471 counts only for those files that have at least one match. (GNU grep just
472 prints the file name in this circumstance, but including the count seems
473 more useful - otherwise, why use --count?) Also ensured that the
474 combination -clh just lists non-zero counts, with no names.
475
476 4. The long form of the pcregrep -F option was incorrectly implemented as
477 --fixed_strings instead of --fixed-strings. This is an incompatible change,
478 but it seems right to fix it, and I didn't think it was worth preserving
479 the old behaviour.
480
481 5. The command line items --regex=pattern and --regexp=pattern were not
482 recognized by pcregrep, which required --regex pattern or --regexp pattern
483 (with a space rather than an '='). The man page documented the '=' forms,
484 which are compatible with GNU grep; these now work.
485
486 6. No libpcreposix.pc file was created for pkg-config; there was just
487 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
488
489 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
490 when UCP support is not needed, by modifying the Python script that
491 generates it from Unicode data files. This should not matter if the module
492 is correctly used as a library, but I received one complaint about 50K of
493 unwanted data. My guess is that the person linked everything into his
494 program rather than using a library. Anyway, it does no harm.
495
496 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
497 was a minimum greater than 1 for a wide character in a possessive
498 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
499 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
500 character. Chaos in the form of incorrect output or a compiling loop could
501 result.
502
503 9. The restrictions on what a pattern can contain when partial matching is
504 requested for pcre_exec() have been removed. All patterns can now be
505 partially matched by this function. In addition, if there are at least two
506 slots in the offset vector, the offset of the earliest inspected character
507 for the match and the offset of the end of the subject are set in them when
508 PCRE_ERROR_PARTIAL is returned.
509
510 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
511 synonymous with PCRE_PARTIAL, for backwards compatibility, and
512 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
513 and may be more useful for multi-segment matching.
514
515 11. Partial matching with pcre_exec() is now more intuitive. A partial match
516 used to be given if ever the end of the subject was reached; now it is
517 given only if matching could not proceed because another character was
518 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
519 string "Z", which now yields "no match" instead of "partial match". In the
520 case of pcre_dfa_exec(), "no match" is given if every matching path for the
521 final character ended with (*FAIL).
522
523 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
524 if the pattern had a "must contain" character that was already found in the
525 earlier partial match, unless partial matching was again requested. For
526 example, with the pattern /dog.(body)?/, the "must contain" character is
527 "g". If the first part-match was for the string "dog", restarting with
528 "sbody" failed. This bug has been fixed.
529
530 13. The string returned by pcre_dfa_exec() after a partial match has been
531 changed so that it starts at the first inspected character rather than the
532 first character of the match. This makes a difference only if the pattern
533 starts with a lookbehind assertion or \b or \B (\K is not supported by
534 pcre_dfa_exec()). It's an incompatible change, but it makes the two
535 matching functions compatible, and I think it's the right thing to do.
536
537 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
538 so that the demonstration program is easily available in environments where
539 PCRE has not been installed from source.
540
541 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
542 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
543 library.
544
545 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
546 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
547 is not the first non-POSIX option to be added. Clearly some people find
548 these options useful.
549
550 17. If a caller to the POSIX matching function regexec() passes a non-zero
551 value for nmatch with a NULL value for pmatch, the value of
552 nmatch is forced to zero.
553
554 18. RunGrepTest did not have a test for the availability of the -u option of
555 the diff command, as RunTest does. It now checks in the same way as
556 RunTest, and also checks for the -b option.
557
558 19. If an odd number of negated classes containing just a single character
559 interposed, within parentheses, between a forward reference to a named
560 subpattern and the definition of the subpattern, compilation crashed with
561 an internal error, complaining that it could not find the referenced
562 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
563 [The bug was that it was starting one character too far in when skipping
564 over the character class, thus treating the ] as data rather than
565 terminating the class. This meant it could skip too much.]
566
567 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
568 /g option in pcretest when the pattern contains \K, which makes it possible
569 to have an empty string match not at the start, even when the pattern is
570 anchored. Updated pcretest and pcredemo to use this option.
571
572 21. If the maximum number of capturing subpatterns in a recursion was greater
573 than the maximum at the outer level, the higher number was returned, but
574 with unset values at the outer level. The correct (outer level) value is
575 now given.
576
577 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
578 PCRE did not set those parentheses (unlike Perl). I have now found a way to
579 make it do so. The string so far is captured, making this feature
580 compatible with Perl.
581
582 23. The tests have been re-organized, adding tests 11 and 12, to make it
583 possible to check the Perl 5.10 features against Perl 5.10.
584
585 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
586 pattern matches a fixed length string. PCRE did not allow this; now it
587 does. Neither allows recursion.
588
589 25. I finally figured out how to implement a request to provide the minimum
590 length of subject string that was needed in order to match a given pattern.
591 (It was back references and recursion that I had previously got hung up
592 on.) This code has now been added to pcre_study(); it finds a lower bound
593 to the length of subject needed. It is not necessarily the greatest lower
594 bound, but using it to avoid searching strings that are too short does give
595 some useful speed-ups. The value is available to calling programs via
596 pcre_fullinfo().
597
598 26. While implementing 25, I discovered to my embarrassment that pcretest had
599 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
600 study optimizations had never been tested with that matching function.
601 Oops. What is worse, even when it was passed study data, there was a bug in
602 pcre_dfa_exec() that meant it never actually used it. Double oops. There
603 were also very few tests of studied patterns with pcre_dfa_exec().
604
605 27. If (?| is used to create subpatterns with duplicate numbers, they are now
606 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
607 on the other side of the coin, they are no longer allowed to have different
608 names, because these cannot be distinguished in PCRE, and this has caused
609 confusion. (This is a difference from Perl.)
610
611 28. When duplicate subpattern names are present (necessarily with different
612 numbers, as required by 27 above), and a test is made by name in a
613 conditional pattern, either for a subpattern having been matched, or for
614 recursion in such a pattern, all the associated numbered subpatterns are
615 tested, and the overall condition is true if the condition is true for any
616 one of them. This is the way Perl works, and is also more like the way
617 testing by number works.
618
619
620 Version 7.9 11-Apr-09
621 ---------------------
622
623 1. When building with support for bzlib/zlib (pcregrep) and/or readline
624 (pcretest), all targets were linked against these libraries. This included
625 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
626 libraries. This caused unwanted dependencies to be created. This problem
627 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
628 pcretest is linked with readline.
629
630 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
631 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
632 moved outside it again, because FALSE and TRUE are already defined in AIX,
633 but BOOL is not.
634
635 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
636 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
637
638 4. The pcregrep documentation said spaces were inserted as well as colons (or
639 hyphens) following file names and line numbers when outputting matching
640 lines. This is not true; no spaces are inserted. I have also clarified the
641 wording for the --colour (or --color) option.
642
643 5. In pcregrep, when --colour was used with -o, the list of matching strings
644 was not coloured; this is different to GNU grep, so I have changed it to be
645 the same.
646
647 6. When --colo(u)r was used in pcregrep, only the first matching substring in
648 each matching line was coloured. Now it goes on to look for further matches
649 of any of the test patterns, which is the same behaviour as GNU grep.
650
651 7. A pattern that could match an empty string could cause pcregrep to loop; it
652 doesn't make sense to accept an empty string match in pcregrep, so I have
653 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
654 seems to be how GNU grep behaves.
655
656 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
657 start or after a newline", because the conditional assertion was not being
658 correctly handled. The rule now is that both the assertion and what follows
659 in the first alternative must satisfy the test.
660
661 9. If auto-callout was enabled in a pattern with a conditional group whose
662 condition was an assertion, PCRE could crash during matching, both with
663 pcre_exec() and pcre_dfa_exec().
664
665 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
666 used for matching.
667
668 11. Unicode property support in character classes was not working for
669 characters (bytes) greater than 127 when not in UTF-8 mode.
670
671 12. Added the -M command line option to pcretest.
672
673 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
674
675 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
676
677 16. Added comments and documentation about mis-use of no_arg in the C++
678 wrapper.
679
680 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
681 from Martin Jerabek that uses macro names for all relevant character and
682 string constants.
683
684 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
685 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
686 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
687 these, but not everybody uses configure.
688
689 19. A conditional group that had only one branch was not being correctly
690 recognized as an item that could match an empty string. This meant that an
691 enclosing group might also not be so recognized, causing infinite looping
692 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
693 with the subject "ab", where knowledge that the repeated group can match
694 nothing is needed in order to break the loop.
695
696 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
697 exec(), but without supplying a callout function, matching went wrong.
698
699 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
700 leak if the size of the offset vector was greater than 30. When the vector
701 is smaller, the saved offsets during recursion go onto a local stack
702 vector, but for larger vectors malloc() is used. It was failing to free
703 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
704 error, in fact).
705
706 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
707 heapframe that is used only when UTF-8 support is enabled. This caused no
708 problem, but was untidy.
709
710 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
711 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
712 included within another project.
713
714 24. Steven Van Ingelgem's patches to add more options to the CMake support,
715 slightly modified by me:
716
717 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
718 not building pcregrep.
719
720 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
721 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
722
723 25. Forward references, both numeric and by name, in patterns that made use of
724 duplicate group numbers, could behave incorrectly or give incorrect errors,
725 because when scanning forward to find the reference group, PCRE was not
726 taking into account the duplicate group numbers. A pattern such as
727 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
728
729 26. Changed a few more instances of "const unsigned char *" to USPTR, making
730 the feature of a custom pointer more persuasive (as requested by a user).
731
732 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
733 pcretest.c, inside #ifndefs, because it seems they are sometimes already
734 pre-defined.
735
736 28. Added support for (*UTF8) at the start of a pattern.
737
738 29. Arrange for flags added by the "release type" setting in CMake to be shown
739 in the configuration summary.
740
741
742 Version 7.8 05-Sep-08
743 ---------------------
744
745 1. Replaced UCP searching code with optimized version as implemented for Ad
746 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
747 stage table and inline lookup instead of a function, giving speed ups of 2
748 to 5 times on some simple patterns that I tested. Permission was given to
749 distribute the MultiStage2.py script that generates the tables (it's not in
750 the tarball, but is in the Subversion repository).
751
752 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
753 scripts.
754
755 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
756 a group with a zero qualifier. The result of the study could be incorrect,
757 or the function might crash, depending on the pattern.
758
759 4. Caseless matching was not working for non-ASCII characters in back
760 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
761 It now works when Unicode Property Support is available.
762
763 5. In pcretest, an escape such as \x{de} in the data was always generating
764 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
765 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
766 truncation.
767
768 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
769
770 7. Added two (int) casts to pcregrep when printing the difference of two
771 pointers, in case they are 64-bit values.
772
773 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
774 test 2 if it fails.
775
776 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
777 and a #define of that name to empty if it is not externally set. This is to
778 allow users of MSVC to set it if necessary.
779
780 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
781 the convenience functions in the pcre_get.c source file.
782
783 11. An option change at the start of a pattern that had top-level alternatives
784 could cause overwriting and/or a crash. This command provoked a crash in
785 some environments:
786
787 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
788
789 This potential security problem was recorded as CVE-2008-2371.
790
791 12. For a pattern where the match had to start at the beginning or immediately
792 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
793 pcre_dfa_exec() could read past the end of the passed subject if there was
794 no match. To help with detecting such bugs (e.g. with valgrind), I modified
795 pcretest so that it places the subject at the end of its malloc-ed buffer.
796
797 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
798 exec() might read past the end of the data buffer in UTF-8 mode.
799
800 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
801 the data contained the byte 0x85 as part of a UTF-8 character within its
802 first line. This applied both to normal and DFA matching.
803
804 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
805 /^[^d]*?$/8 failed to match "abc".
806
807 16. Added a missing copyright notice to pcrecpp_internal.h.
808
809 17. Make it more clear in the documentation that values returned from
810 pcre_exec() in ovector are byte offsets, not character counts.
811
812 18. Tidied a few places to stop certain compilers from issuing warnings.
813
814 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
815 supplied by Stefan Weber. I made a further small update for 7.8 because
816 there is a change of source arrangements: the pcre_searchfuncs.c module is
817 replaced by pcre_ucd.c.
818
819
820 Version 7.7 07-May-08
821 ---------------------
822
823 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
824 a string to a long long, pretend we don't even have a long long." This is
825 done by checking for the strtoq, strtoll, and _strtoi64 functions.
826
827 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
828 pre-7.6 versions, which defined a global no_arg variable instead of putting
829 it in the RE class. (See also #8 below.)
830
831 3. Remove a line of dead code, identified by coverity and reported by Nuno
832 Lopes.
833
834 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
835
836 (1) The include/exclude patterns were being applied to the whole pathnames
837 of files, instead of just to the final components.
838
839 (2) If there was more than one level of directory, the subdirectories were
840 skipped unless they satisfied the include/exclude conditions. This is
841 inconsistent with GNU grep (and could even be seen as contrary to the
842 pcregrep specification - which I improved to make it absolutely clear).
843 The action now is always to scan all levels of directory, and just
844 apply the include/exclude patterns to regular files.
845
846 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
847 --exclude_dir in the tests to avoid scanning .svn directories.
848
849 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
850 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
851 doesn't support NULs in patterns.
852
853 7. Added some missing "const"s to declarations of static tables in
854 pcre_compile.c and pcre_dfa_exec.c.
855
856 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
857 caused by fix #2 above. (Subsequently also a second patch to fix the
858 first patch. And a third patch - this was a messy problem.)
859
860 9. Applied Craig's patch to remove the use of push_back().
861
862 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
863 matching function regexec().
864
865 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
866 which, however, unlike Perl's \g{...}, are subroutine calls, not back
867 references. PCRE supports relative numbers with this syntax (I don't think
868 Oniguruma does).
869
870 12. Previously, a group with a zero repeat such as (...){0} was completely
871 omitted from the compiled regex. However, this means that if the group
872 was called as a subroutine from elsewhere in the pattern, things went wrong
873 (an internal error was given). Such groups are now left in the compiled
874 pattern, with a new opcode that causes them to be skipped at execution
875 time.
876
877 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
878 to the way PCRE behaves:
879
880 (a) A lone ] character is dis-allowed (Perl treats it as data).
881
882 (b) A back reference to an unmatched subpattern matches an empty string
883 (Perl fails the current match path).
884
885 (c) A data ] in a character class must be notated as \] because if the
886 first data character in a class is ], it defines an empty class. (In
887 Perl it is not possible to have an empty class.) The empty class []
888 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
889 The negative empty class [^] matches any one character, independently
890 of the DOTALL setting.
891
892 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
893 non-existent subpattern following a character class starting with ']' and
894 containing () gave an internal compiling error instead of "reference to
895 non-existent subpattern". Fortunately, when the pattern did exist, the
896 compiled code was correct. (When scanning forwards to check for the
897 existencd of the subpattern, it was treating the data ']' as terminating
898 the class, so got the count wrong. When actually compiling, the reference
899 was subsequently set up correctly.)
900
901 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
902 it was being rejected as not supported by pcre_dfa_exec(), even though
903 other assertions are supported. I have made pcre_dfa_exec() support
904 (*FAIL).
905
906 16. The implementation of 13c above involved the invention of a new opcode,
907 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
908 cannot be changed at match time, I realized I could make a small
909 improvement to matching performance by compiling OP_ALLANY instead of
910 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
911 on the OP_ANY path.
912
913 17. Compiling pcretest on Windows with readline support failed without the
914 following two fixes: (1) Make the unistd.h include conditional on
915 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
916
917 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
918 ncurses library to be included for pcretest when ReadLine support is
919 requested, but also to allow for it to be overridden. This patch came from
920 Daniel Bergström.
921
922 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
923 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
924 any errors with the current Unicode tables. Thanks to Peter Kankowski for
925 spotting this.
926
927
928 Version 7.6 28-Jan-08
929 ---------------------
930
931 1. A character class containing a very large number of characters with
932 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
933 overflow.
934
935 2. Patch to cut out the "long long" test in pcrecpp_unittest when
936 HAVE_LONG_LONG is not defined.
937
938 3. Applied Christian Ehrlicher's patch to update the CMake build files to
939 bring them up to date and include new features. This patch includes:
940
941 - Fixed PH's badly added libz and libbz2 support.
942 - Fixed a problem with static linking.
943 - Added pcredemo. [But later removed - see 7 below.]
944 - Fixed dftables problem and added an option.
945 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
946 HAVE_LONG_LONG.
947 - Added readline support for pcretest.
948 - Added an listing of the option settings after cmake has run.
949
950 4. A user submitted a patch to Makefile that makes it easy to create
951 "pcre.dll" under mingw when using Configure/Make. I added stuff to
952 Makefile.am that cause it to include this special target, without
953 affecting anything else. Note that the same mingw target plus all
954 the other distribution libraries and programs are now supported
955 when configuring with CMake (see 6 below) instead of with
956 Configure/Make.
957
958 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
959 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
960 exported in the Windows port". It has not yet been confirmed that the patch
961 solves the problem, but it does no harm.
962
963 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
964 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
965 with CMake, and also correct the comment about stack recursion.
966
967 7. Remove the automatic building of pcredemo from the ./configure system and
968 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
969 of a program that users should build themselves after PCRE is installed, so
970 building it automatically is not really right. What is more, it gave
971 trouble in some build environments.
972
973 8. Further tidies to CMakeLists.txt from Sheri and Christian.
974
975
976 Version 7.5 10-Jan-08
977 ---------------------
978
979 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
980 values in parens when parsing an RE using the C++ wrapper."
981
982 2. Negative specials like \S did not work in character classes in UTF-8 mode.
983 Characters greater than 255 were excluded from the class instead of being
984 included.
985
986 3. The same bug as (2) above applied to negated POSIX classes such as
987 [:^space:].
988
989 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
990 defined or documented. It seems to have been a typo for PCRE_STATIC, so
991 I have changed it.
992
993 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
994 first named subpattern) and a construct such as (?&a) would reference the
995 first named subpattern whose name started with "a" (in other words, the
996 length check was missing). Both these problems are fixed. "Subpattern name
997 expected" is now given for (?&) (a zero-length name), and this patch also
998 makes it give the same error for \k'' (previously it complained that that
999 was a reference to a non-existent subpattern).
1000
1001 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1002 this is right because (?- can be followed by option settings as well as by
1003 digits. I have, however, made the messages clearer.
1004
1005 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1006 than the number used in the conditional) now cause a compile-time error.
1007 This is actually not compatible with Perl, which accepts such patterns, but
1008 treats the conditional as always being FALSE (as PCRE used to), but it
1009 seems to me that giving a diagnostic is better.
1010
1011 8. Change "alphameric" to the more common word "alphanumeric" in comments
1012 and messages.
1013
1014 9. Fix two occurrences of "backslash" in comments that should have been
1015 "backspace".
1016
1017 10. Remove two redundant lines of code that can never be obeyed (their function
1018 was moved elsewhere).
1019
1020 11. The program that makes PCRE's Unicode character property table had a bug
1021 which caused it to generate incorrect table entries for sequences of
1022 characters that have the same character type, but are in different scripts.
1023 It amalgamated them into a single range, with the script of the first of
1024 them. In other words, some characters were in the wrong script. There were
1025 thirteen such cases, affecting characters in the following ranges:
1026
1027 U+002b0 - U+002c1
1028 U+0060c - U+0060d
1029 U+0061e - U+00612
1030 U+0064b - U+0065e
1031 U+0074d - U+0076d
1032 U+01800 - U+01805
1033 U+01d00 - U+01d77
1034 U+01d9b - U+01dbf
1035 U+0200b - U+0200f
1036 U+030fc - U+030fe
1037 U+03260 - U+0327f
1038 U+0fb46 - U+0fbb1
1039 U+10450 - U+1049d
1040
1041 12. The -o option (show only the matching part of a line) for pcregrep was not
1042 compatible with GNU grep in that, if there was more than one match in a
1043 line, it showed only the first of them. It now behaves in the same way as
1044 GNU grep.
1045
1046 13. If the -o and -v options were combined for pcregrep, it printed a blank
1047 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1048 does the same. The return code can be used to tell if there were any
1049 non-matching lines.
1050
1051 14. Added --file-offsets and --line-offsets to pcregrep.
1052
1053 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1054 infinitely looping recursion. The bug was that positive lookaheads were not
1055 being skipped when checking for a possible empty match (negative lookaheads
1056 and both kinds of lookbehind were skipped).
1057
1058 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1059 inclusion of <windows.h> to before rather than after the definition of
1060 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1061
1062 17. Specifying a possessive quantifier with a specific limit for a Unicode
1063 character property caused pcre_compile() to compile bad code, which led at
1064 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1065 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1066 caused the error; without that there was no problem.
1067
1068 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1069
1070 19. Added --enable-pcretest-libreadline.
1071
1072 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1073 RE::GlobalReplace(). As a result, the number of replacements returned was
1074 double what it should be. I removed one of the increments, but Craig sent a
1075 later patch that removed the other one (the right fix) and added unit tests
1076 that check the return values (which was not done before).
1077
1078 21. Several CMake things:
1079
1080 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1081 the names libpcre and libpcreposix, not just pcre and pcreposix.
1082
1083 (2) The above change means that pcretest and pcregrep are now correctly
1084 linked with the newly-built libraries, not previously installed ones.
1085
1086 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1087
1088 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1089 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1090 UTF-8 newline character). The key issue is that the pattern starts .*;
1091 this means that the match must be either at the beginning, or after a
1092 newline. The bug was in the code for advancing after a failed match and
1093 checking that the new position followed a newline. It was not taking
1094 account of UTF-8 characters correctly.
1095
1096 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1097 character classes. PCRE was not treating the sequence [:...:] as a
1098 character class unless the ... were all letters. Perl, however, seems to
1099 allow any characters between [: and :], though of course it rejects as
1100 unknown any "names" that contain non-letters, because all the known class
1101 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1102 for example, whereas PCRE did not - it did not recognize a POSIX character
1103 class. This seemed a bit dangerous, so the code has been changed to be
1104 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1105 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1106 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1107 Perl does, and where it didn't before.
1108
1109 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1110 Windows environments %n is disabled by default.
1111
1112
1113 Version 7.4 21-Sep-07
1114 ---------------------
1115
1116 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1117 means that a class such as [\s] counted as "explicit reference to CR or
1118 LF". That isn't really right - the whole point of the change was to try to
1119 help when there was an actual mention of one of the two characters. So now
1120 the change happens only if \r or \n (or a literal CR or LF) character is
1121 encountered.
1122
1123 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1124 of both had grown to the point where there were only 3 bits left.
1125 Fortunately, there was spare space in the data structure, and so I have
1126 moved the internal flags into a new 16-bit field to free up more option
1127 bits.
1128
1129 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1130 but did not set the internal JCHANGED flag - either of these is enough to
1131 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1132 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1133 start sets both bits.
1134
1135 4. Added options (at build time, compile time, exec time) to change \R from
1136 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1137
1138 5. doc/pcresyntax.html was missing from the distribution.
1139
1140 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1141 compatibility, even though it is no longer used.
1142
1143 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1144 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1145 windows.h file is present (where different names are used). [This was
1146 reversed later after testing - see 16 below.]
1147
1148 8. Changed all #include <config.h> to #include "config.h". There were also
1149 some further <pcre.h> cases that I changed to "pcre.h".
1150
1151 9. When pcregrep was used with the --colour option, it missed the line ending
1152 sequence off the lines that it output.
1153
1154 10. It was pointed out to me that arrays of string pointers cause lots of
1155 relocations when a shared library is dynamically loaded. A technique of
1156 using a single long string with a table of offsets can drastically reduce
1157 these. I have refactored PCRE in four places to do this. The result is
1158 dramatic:
1159
1160 Originally: 290
1161 After changing UCP table: 187
1162 After changing error message table: 43
1163 After changing table of "verbs" 36
1164 After changing table of Posix names 22
1165
1166 Thanks to the folks working on Gregex for glib for this insight.
1167
1168 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1169 unicode-properties was also set.
1170
1171 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1172
1173 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1174 checked only for CRLF.
1175
1176 14. Added casts to pcretest.c to avoid compiler warnings.
1177
1178 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1179
1180 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1181 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1182 entirely. This removes changes made in 7 above.
1183
1184 17. The CMake files have been updated, and there is now more information about
1185 building with CMake in the NON-UNIX-USE document.
1186
1187
1188 Version 7.3 28-Aug-07
1189 ---------------------
1190
1191 1. In the rejigging of the build system that eventually resulted in 7.1, the
1192 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1193 brackets there is not right, since it causes compilers to look for an
1194 installed pcre.h, not the version that is in the source that is being
1195 compiled (which of course may be different). I have changed it back to:
1196
1197 #include "pcre.h"
1198
1199 I have a vague recollection that the change was concerned with compiling in
1200 different directories, but in the new build system, that is taken care of
1201 by the VPATH setting the Makefile.
1202
1203 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1204 when the subject happened to end in the byte 0x85 (e.g. if the last
1205 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1206 characters but of course it shouldn't be taken as a newline when it is part
1207 of another character. The bug was that, for an unlimited repeat of . in
1208 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1209 characters when looking for a newline.
1210
1211 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1212
1213 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1214 in debug output.
1215
1216 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1217 long printing in the pcrecpp unittest when running under MinGW.
1218
1219 6. ESC_K was left out of the EBCDIC table.
1220
1221 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1222 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1223 limit also applies to "virtual nesting" when a pattern is recursive, and in
1224 this case 1000 isn't so big. I have been able to remove this limit at the
1225 expense of backing off one optimization in certain circumstances. Normally,
1226 when pcre_exec() would call its internal match() function recursively and
1227 immediately return the result unconditionally, it uses a "tail recursion"
1228 feature to save stack. However, when a subpattern that can match an empty
1229 string has an unlimited repetition quantifier, it no longer makes this
1230 optimization. That gives it a stack frame in which to save the data for
1231 checking that an empty string has been matched. Previously this was taken
1232 from the 1000-entry workspace that had been reserved. So now there is no
1233 explicit limit, but more stack is used.
1234
1235 8. Applied Daniel's patches to solve problems with the import/export magic
1236 syntax that is required for Windows, and which was going wrong for the
1237 pcreposix and pcrecpp parts of the library. These were overlooked when this
1238 problem was solved for the main library.
1239
1240 9. There were some crude static tests to avoid integer overflow when computing
1241 the size of patterns that contain repeated groups with explicit upper
1242 limits. As the maximum quantifier is 65535, the maximum group length was
1243 set at 30,000 so that the product of these two numbers did not overflow a
1244 32-bit integer. However, it turns out that people want to use groups that
1245 are longer than 30,000 bytes (though not repeat them that many times).
1246 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1247 made it possible to implement the integer overflow checks in a much more
1248 dynamic way, which I have now done. The artificial limitation on group
1249 length has been removed - we now have only the limit on the total length of
1250 the compiled pattern, which depends on the LINK_SIZE setting.
1251
1252 10. Fixed a bug in the documentation for get/copy named substring when
1253 duplicate names are permitted. If none of the named substrings are set, the
1254 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1255 empty string.
1256
1257 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1258 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1259 because the ] is interpreted as the first data character and the
1260 terminating ] is not found. PCRE has been made compatible with Perl in this
1261 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1262 cause memory overwriting.
1263
1264 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1265 string has been matched (to stop an infinite loop). It was not recognizing
1266 a conditional subpattern that could match an empty string if that
1267 subpattern was within another subpattern. For example, it looped when
1268 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1269 condition was not nested. This bug has been fixed.
1270
1271 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1272 past the start of the subject in the presence of bytes with the top bit
1273 set, for example "\x8aBCD".
1274
1275 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1276 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1277
1278 14. Optimized (?!) to (*FAIL).
1279
1280 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1281 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1282 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1283 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1284 does: it's just the validity check that is more restrictive.
1285
1286 16. Inserted checks for integer overflows during escape sequence (backslash)
1287 processing, and also fixed erroneous offset values for syntax errors during
1288 backslash processing.
1289
1290 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1291 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1292
1293 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1294 caused an overrun.
1295
1296 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1297 something other than just ASCII characters) inside a group that had an
1298 unlimited repeat caused a loop at compile time (while checking to see
1299 whether the group could match an empty string).
1300
1301 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1302 [\P{Any}] did so. (Error in the code for printing property names.)
1303
1304 21. An orphan \E inside a character class could cause a crash.
1305
1306 22. A repeated capturing bracket such as (A)? could cause a wild memory
1307 reference during compilation.
1308
1309 23. There are several functions in pcre_compile() that scan along a compiled
1310 expression for various reasons (e.g. to see if it's fixed length for look
1311 behind). There were bugs in these functions when a repeated \p or \P was
1312 present in the pattern. These operators have additional parameters compared
1313 with \d, etc, and these were not being taken into account when moving along
1314 the compiled data. Specifically:
1315
1316 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1317 length.
1318
1319 (b) An item such as \pL+ within a repeated group could cause crashes or
1320 loops.
1321
1322 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1323 "reference to non-existent subpattern" error.
1324
1325 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1326
1327 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1328 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1329
1330 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1331
1332 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1333 character were causing crashes (broken optimization).
1334
1335 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1336 \p or \P) caused a compile-time loop.
1337
1338 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1339 break. For example, the unstudied pattern [\r\n]A does not match the string
1340 "\r\nA" because change 7.0/46 below moves the current point on by two
1341 characters after failing to match at the start. However, the pattern \nA
1342 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1343 the same is true. There doesn't seem any very clean way out of this, but
1344 what I have chosen to do makes the common cases work: PCRE now takes note
1345 of whether there can be an explicit match for \r or \n anywhere in the
1346 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1347 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1348 pattern has explicit CR or LF references.
1349
1350 29. Added (*CR) etc for changing newline setting at start of pattern.
1351
1352
1353 Version 7.2 19-Jun-07
1354 ---------------------
1355
1356 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1357 which is apparently normally available under Windows.
1358
1359 2. Re-jig the pcregrep tests with different newline settings in an attempt
1360 to make them independent of the local environment's newline setting.
1361
1362 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1363
1364 4. Some of the "internals" tests were previously cut out when the link size
1365 was not 2, because the output contained actual offsets. The recent new
1366 "Z" feature of pcretest means that these can be cut out, making the tests
1367 usable with all link sizes.
1368
1369 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1370 stack recursion. This gives a massive performance boost under BSD, but just
1371 a small improvement under Linux. However, it saves one field in the frame
1372 in all cases.
1373
1374 6. Added more features from the forthcoming Perl 5.10:
1375
1376 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1377 recursion call. It refers to the nth most recently opened parentheses.
1378
1379 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1380 to be opened parentheses.
1381
1382 (c) Conditions that refer to capturing parentheses can be specified
1383 relatively, for example, (?(-2)... or (?(+3)...
1384
1385 (d) \K resets the start of the current match so that everything before
1386 is not part of it.
1387
1388 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1389
1390 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1391 reference syntax.
1392
1393 (g) (?| introduces a group in which the numbering of parentheses in each
1394 alternative starts with the same number.
1395
1396 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1397
1398 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1399 PCRE_INFO_JCHANGED.
1400
1401 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1402 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1403 for detecting groups that can match an empty string.
1404
1405 9. A pattern with a very large number of alternatives (more than several
1406 hundred) was running out of internal workspace during the pre-compile
1407 phase, where pcre_compile() figures out how much memory will be needed. A
1408 bit of new cunning has reduced the workspace needed for groups with
1409 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1410 workspace instead of running out of the 4096 that are available.
1411
1412 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1413
1414 11. Applied patch from Google to remove an optimization that didn't quite work.
1415 The report of the bug said:
1416
1417 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1418 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1419 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1420
1421 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1422 it matched the wrong number of bytes.
1423
1424
1425 Version 7.1 24-Apr-07
1426 ---------------------
1427
1428 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1429 that is more "standard", making use of automake and other Autotools. There
1430 is some re-arrangement of the files and adjustment of comments consequent
1431 on this.
1432
1433 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1434 for recursive directory scanning broke on some systems because the files
1435 are not scanned in any specific order and on different systems the order
1436 was different. A call to "sort" has been inserted into RunGrepTest for the
1437 approprate test as a short-term fix. In the longer term there may be an
1438 alternative.
1439
1440 3. I had an email from Eric Raymond about problems translating some of PCRE's
1441 man pages to HTML (despite the fact that I distribute HTML pages, some
1442 people do their own conversions for various reasons). The problems
1443 concerned the use of low-level troff macros .br and .in. I have therefore
1444 removed all such uses from the man pages (some were redundant, some could
1445 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1446 HTML has been updated to handle .nf/.fi and to complain if it encounters
1447 .br or .in.
1448
1449 4. Updated comments in configure.ac that get placed in config.h.in and also
1450 arranged for config.h to be included in the distribution, with the name
1451 config.h.generic, for the benefit of those who have to compile without
1452 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1453
1454 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1455 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1456 makevp.bat for the current PCRE, using the additional files
1457 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1458
1459 6. A Windows user reported a minor discrepancy with test 2, which turned out
1460 to be caused by a trailing space on an input line that had got lost in his
1461 copy. The trailing space was an accident, so I've just removed it.
1462
1463 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1464 that is needed.
1465
1466 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1467 as "const" (a) because they are and (b) because it helps the PHP
1468 maintainers who have recently made a script to detect big data structures
1469 in the php code that should be moved to the .rodata section. I remembered
1470 to update Builducptable as well, so it won't revert if ucptable.h is ever
1471 re-created.
1472
1473 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1474 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1475 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1476 support is not required. This saves 1.5-2K of code, which is important in
1477 some applications.
1478
1479 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1480 so as not to refer to the tables, even though these functions will never be
1481 called when UTF-8 support is disabled. Otherwise there are problems with a
1482 shared library.
1483
1484 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1485
1486 (a) It was defining its arguments as char * instead of void *.
1487
1488 (b) It was assuming that all moves were upwards in memory; this was true
1489 a long time ago when I wrote it, but is no longer the case.
1490
1491 The emulated memove() is provided for those environments that have neither
1492 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1493 is clearly not the case, as these two bugs were recently reported.
1494
1495 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1496 and Detrail to create the HTML documentation, the .txt form of the man
1497 pages, and it removes trailing spaces from listed files. It also creates
1498 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1499 case, it wraps all the #defines with #ifndefs. This script should be run
1500 before "make dist".
1501
1502 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1503 with Unicode property support.
1504
1505 (a) For a maximizing quantifier, if the two different cases of the
1506 character were of different lengths in their UTF-8 codings (there are
1507 some cases like this - I found 11), and the matching function had to
1508 back up over a mixture of the two cases, it incorrectly assumed they
1509 were both the same length.
1510
1511 (b) When PCRE was configured to use the heap rather than the stack for
1512 recursion during matching, it was not correctly preserving the data for
1513 the other case of a UTF-8 character when checking ahead for a match
1514 while processing a minimizing repeat. If the check also involved
1515 matching a wide character, but failed, corruption could cause an
1516 erroneous result when trying to check for a repeat of the original
1517 character.
1518
1519 13. Some tidying changes to the testing mechanism:
1520
1521 (a) The RunTest script now detects the internal link size and whether there
1522 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1523 values substituted by "configure". (The RunGrepTest script already did
1524 this for UTF-8.) The configure.ac script no longer substitutes the
1525 relevant variables.
1526
1527 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1528 with length and offset values. This means that the output is different
1529 for different internal link sizes. Test 2 is skipped for link sizes
1530 other than 2 because of this, bypassing the problem. Unfortunately,
1531 there was also a test in test 3 (the locale tests) that used /B and
1532 failed for link sizes other than 2. Rather than cut the whole test out,
1533 I have added a new /Z option to pcretest that replaces the length and
1534 offset values with spaces. This is now used to make test 3 independent
1535 of link size. (Test 2 will be tidied up later.)
1536
1537 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1538 segmentation fault instead of returning the appropriate error message.
1539
1540 15. In multiline mode when the newline sequence was set to "any", the pattern
1541 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1542 This doesn't seem right; it now treats the CRLF combination as the line
1543 ending, and so does not match in that case. It's only a pattern such as ^$
1544 that would hit this one: something like ^ABC$ would have failed after \r
1545 and then tried again after \r\n.
1546
1547 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1548 in an attempt to make files that differ only in their line terminators
1549 compare equal. This works on Linux.
1550
1551 17. Under certain error circumstances pcregrep might try to free random memory
1552 as it exited. This is now fixed, thanks to valgrind.
1553
1554 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1555 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1556 was because its rules for how to advance for /g after matching an empty
1557 string at the end of a line did not allow for this case. They now check for
1558 it specially.
1559
1560 20. pcretest is supposed to handle patterns and data of any length, by
1561 extending its buffers when necessary. It was getting this wrong when the
1562 buffer for a data line had to be extended.
1563
1564 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1565 CRLF as a newline sequence.
1566
1567 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1568 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1569 I have nevertheless tidied it up.
1570
1571 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1572
1573 24. Added a man page for pcre-config.
1574
1575
1576 Version 7.0 19-Dec-06
1577 ---------------------
1578
1579 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1580 moving to gcc 4.1.1.
1581
1582 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1583 sys/time.h, which is documented as needed for this function. It doesn't
1584 seem to matter on Linux, but it showed up on some releases of OS X.
1585
1586 3. It seems that there are systems where bytes whose values are greater than
1587 127 match isprint() in the "C" locale. The "C" locale should be the
1588 default when a C program starts up. In most systems, only ASCII printing
1589 characters match isprint(). This difference caused the output from pcretest
1590 to vary, making some of the tests fail. I have changed pcretest so that:
1591
1592 (a) When it is outputting text in the compiled version of a pattern, bytes
1593 other than 32-126 are always shown as hex escapes.
1594
1595 (b) When it is outputting text that is a matched part of a subject string,
1596 it does the same, unless a different locale has been set for the match
1597 (using the /L modifier). In this case, it uses isprint() to decide.
1598
1599 4. Fixed a major bug that caused incorrect computation of the amount of memory
1600 required for a compiled pattern when options that changed within the
1601 pattern affected the logic of the preliminary scan that determines the
1602 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1603 that the computed length was too small. The symptoms of this bug were
1604 either the PCRE error "internal error: code overflow" from pcre_compile(),
1605 or a glibc crash with a message such as "pcretest: free(): invalid next
1606 size (fast)". Examples of patterns that provoked this bug (shown in
1607 pcretest format) are:
1608
1609 /(?-x: )/x
1610 /(?x)(?-x: \s*#\s*)/
1611 /((?i)[\x{c0}])/8
1612 /(?i:[\x{c0}])/8
1613
1614 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1615 is now done differently.
1616
1617 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1618 wrapper classes; (b) implement a new function in the C++ scanner that is
1619 more efficient than the old way of doing things because it avoids levels of
1620 recursion in the regex matching; (c) add a paragraph to the documentation
1621 for the FullMatch() function.
1622
1623 6. The escape sequence \n was being treated as whatever was defined as
1624 "newline". Not only was this contrary to the documentation, which states
1625 that \n is character 10 (hex 0A), but it also went horribly wrong when
1626 "newline" was defined as CRLF. This has been fixed.
1627
1628 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1629 was being set to -1 for the "end of line" case (supposedly a value that no
1630 character can have). Though this value is never used (the check for end of
1631 line is "zero bytes in current character"), it caused compiler complaints.
1632 I've changed it to 0xffffffff.
1633
1634 8. In pcre_version.c, the version string was being built by a sequence of
1635 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1636 string (as it is for production releases) called a macro with an empty
1637 argument. The C standard says the result of this is undefined. The gcc
1638 compiler treats it as an empty string (which was what was wanted) but it is
1639 reported that Visual C gives an error. The source has been hacked around to
1640 avoid this problem.
1641
1642 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1643 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1644 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1645 of them did).
1646
1647 10. Originally, pcretest opened its input and output without "b"; then I was
1648 told that "b" was needed in some environments, so it was added for release
1649 5.0 to both the input and output. (It makes no difference on Unix-like
1650 systems.) Later I was told that it is wrong for the input on Windows. I've
1651 now abstracted the modes into two macros, to make it easier to fiddle with
1652 them, and removed "b" from the input mode under Windows.
1653
1654 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1655
1656 12. Added -help and --help to pcretest as an official way of being reminded
1657 of the options.
1658
1659 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1660 and pcrecpp.cc because they annoy compilers at high warning levels.
1661
1662 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1663
1664 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1665 scripts are not C programs :-) and which was not noticed because it works
1666 on Linux.
1667
1668 16. pcretest is supposed to handle any length of pattern and data line (as one
1669 line or as a continued sequence of lines) by extending its input buffer if
1670 necessary. This feature was broken for very long pattern lines, leading to
1671 a string of junk being passed to pcre_compile() if the pattern was longer
1672 than about 50K.
1673
1674 17. I have done a major re-factoring of the way pcre_compile() computes the
1675 amount of memory needed for a compiled pattern. Previously, there was code
1676 that made a preliminary scan of the pattern in order to do this. That was
1677 OK when PCRE was new, but as the facilities have expanded, it has become
1678 harder and harder to keep it in step with the real compile phase, and there
1679 have been a number of bugs (see for example, 4 above). I have now found a
1680 cunning way of running the real compile function in a "fake" mode that
1681 enables it to compute how much memory it would need, while actually only
1682 ever using a few hundred bytes of working memory and without too many
1683 tests of the mode. This should make future maintenance and development
1684 easier. A side effect of this work is that the limit of 200 on the nesting
1685 depth of parentheses has been removed (though this was never a serious
1686 limitation, I suspect). However, there is a downside: pcre_compile() now
1687 runs more slowly than before (30% or more, depending on the pattern). I
1688 hope this isn't a big issue. There is no effect on runtime performance.
1689
1690 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1691 newline (only possible for the last line of a file) and it was a
1692 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1693
1694 19. Added additional timing features to pcretest. (1) The -tm option now times
1695 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1696 separate command line item, by a number that specifies the number of
1697 repeats to use when timing. The default is 50000; this gives better
1698 precision, but takes uncomfortably long for very large patterns.
1699
1700 20. Extended pcre_study() to be more clever in cases where a branch of a
1701 subpattern has no definite first character. For example, (a*|b*)[cd] would
1702 previously give no result from pcre_study(). Now it recognizes that the
1703 first character must be a, b, c, or d.
1704
1705 21. There was an incorrect error "recursive call could loop indefinitely" if
1706 a subpattern (or the entire pattern) that was being tested for matching an
1707 empty string contained only one non-empty item after a nested subpattern.
1708 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1709 incorrectly, because the \d was being skipped in the check.
1710
1711 22. The pcretest program now has a new pattern option /B and a command line
1712 option -b, which is equivalent to adding /B to every pattern. This causes
1713 it to show the compiled bytecode, without the additional information that
1714 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1715 is the same as /B/I).
1716
1717 23. A new optimization is now able automatically to treat some sequences such
1718 as a*b as a*+b. More specifically, if something simple (such as a character
1719 or a simple class like \d) has an unlimited quantifier, and is followed by
1720 something that cannot possibly match the quantified thing, the quantifier
1721 is automatically "possessified".
1722
1723 24. A recursive reference to a subpattern whose number was greater than 39
1724 went wrong under certain circumstances in UTF-8 mode. This bug could also
1725 have affected the operation of pcre_study().
1726
1727 25. Realized that a little bit of performance could be had by replacing
1728 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1729
1730 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1731
1732 27. Possessive quantifiers such as a++ were previously implemented by turning
1733 them into atomic groups such as ($>a+). Now they have their own opcodes,
1734 which improves performance. This includes the automatically created ones
1735 from 23 above.
1736
1737 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1738 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1739 the first matched character to be a colon. This applied both to named and
1740 numbered groups.
1741
1742 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1743
1744 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1745 building PCRE on Windows easier, so I have included it in the distribution.
1746
1747 31. There is now a check in pcretest against a ridiculously large number being
1748 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1749 loop, the loop is abandoned.
1750
1751 32. Forward references to subpatterns in conditions such as (?(2)...) where
1752 subpattern 2 is defined later cause pcre_compile() to search forwards in
1753 the pattern for the relevant set of parentheses. This search went wrong
1754 when there were unescaped parentheses in a character class, parentheses
1755 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1756
1757 33. "Subroutine" calls and backreferences were previously restricted to
1758 referencing subpatterns earlier in the regex. This restriction has now
1759 been removed.
1760
1761 34. Added a number of extra features that are going to be in Perl 5.10. On the
1762 whole, these are just syntactic alternatives for features that PCRE had
1763 previously implemented using the Python syntax or my own invention. The
1764 other formats are all retained for compatibility.
1765
1766 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1767 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1768 also .NET compatible.
1769
1770 (b) A recursion or subroutine call to a named group can now be defined as
1771 (?&name) as well as (?P>name).
1772
1773 (c) A backreference to a named group can now be defined as \k<name> or
1774 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1775 5.10, are also .NET compatible.
1776
1777 (d) A conditional reference to a named group can now use the syntax
1778 (?(<name>) or (?('name') as well as (?(name).
1779
1780 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1781 groups (named and numbered) that are never evaluated inline, but can be
1782 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1783 is always false. There may be only one alternative in such a group.
1784
1785 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1786 as the simple (?(R). The condition is true only if the most recent
1787 recursion is that of the given number or name. It does not search out
1788 through the entire recursion stack.
1789
1790 (g) The escape \gN or \g{N} has been added, where N is a positive or
1791 negative number, specifying an absolute or relative reference.
1792
1793 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1794 some "unreachable code" warnings.
1795
1796 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1797 things, this adds five new scripts.
1798
1799 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1800 There were also incompatibilities regarding the handling of \Q..\E inside
1801 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1802 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1803
1804 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1805 matches an empty string, and forcibly breaks the loop. There were bugs in
1806 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1807 against aaaa the result was just "a" rather than "aaaa", for example. Two
1808 separate and independent bugs (that affected different cases) have been
1809 fixed.
1810
1811 39. Refactored the code to abolish the use of different opcodes for small
1812 capturing bracket numbers. This is a tidy that I avoided doing when I
1813 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1814 The new approach is not only tidier, it makes it possible to reduce the
1815 memory needed to fix the previous bug (38).
1816
1817 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1818 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1819 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1820 mode.
1821
1822 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1823 report.
1824
1825 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1826 copy construction and assignment in the C++ wrapper.
1827
1828 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1829 couple of bugs that could have given wrong results in the "--newline=crlf"
1830 case.
1831
1832 44. Added a number of casts and did some reorganization of signed/unsigned int
1833 variables following suggestions from Dair Grant. Also renamed the variable
1834 "this" as "item" because it is a C++ keyword.
1835
1836 45. Arranged for dftables to add
1837
1838 #include "pcre_internal.h"
1839
1840 to pcre_chartables.c because without it, gcc 4.x may remove the array
1841 definition from the final binary if PCRE is built into a static library and
1842 dead code stripping is activated.
1843
1844 46. For an unanchored pattern, if a match attempt fails at the start of a
1845 newline sequence, and the newline setting is CRLF or ANY, and the next two
1846 characters are CRLF, advance by two characters instead of one.
1847
1848
1849 Version 6.7 04-Jul-06
1850 ---------------------
1851
1852 1. In order to handle tests when input lines are enormously long, pcretest has
1853 been re-factored so that it automatically extends its buffers when
1854 necessary. The code is crude, but this _is_ just a test program. The
1855 default size has been increased from 32K to 50K.
1856
1857 2. The code in pcre_study() was using the value of the re argument before
1858 testing it for NULL. (Of course, in any sensible call of the function, it
1859 won't be NULL.)
1860
1861 3. The memmove() emulation function in pcre_internal.h, which is used on
1862 systems that lack both memmove() and bcopy() - that is, hardly ever -
1863 was missing a "static" storage class specifier.
1864
1865 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1866 containing an extended class (one that cannot be represented by a bitmap
1867 because it contains high-valued characters or Unicode property items, e.g.
1868 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1869 pattern, but PCRE should not loop if you do not (it no longer does).
1870 [Detail: two cases were found: (a) a repeated subpattern containing an
1871 extended class; (b) a recursive reference to a subpattern that followed a
1872 previous extended class. It wasn't skipping over the extended class
1873 correctly when UTF-8 mode was not set.]
1874
1875 5. A negated single-character class was not being recognized as fixed-length
1876 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1877 compile error "lookbehind assertion is not fixed length".
1878
1879 6. The RunPerlTest auxiliary script was showing an unexpected difference
1880 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1881 write a Perl script that can interpret lines of an input file either as
1882 byte characters or as UTF-8, which is what "perltest" was being required to
1883 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1884 can't do is switch easily at run time between having the "use utf8;" pragma
1885 or not. In the end, I fudged it by using the RunPerlTest script to insert
1886 "use utf8;" explicitly for the UTF-8 tests.
1887
1888 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1889 the end of the subject string, contrary to the documentation and to what
1890 Perl does. This was true of both matching functions. Now it matches only at
1891 the start of the subject and immediately after *internal* newlines.
1892
1893 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1894 a pointer to an int instead of a pointer to an unsigned long int. This
1895 caused problems on 64-bit systems.
1896
1897 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1898 instance of the 'standard' template library not being so standard".
1899
1900 10. There was no check on the number of named subpatterns nor the maximum
1901 length of a subpattern name. The product of these values is used to compute
1902 the size of the memory block for a compiled pattern. By supplying a very
1903 long subpattern name and a large number of named subpatterns, the size
1904 computation could be caused to overflow. This is now prevented by limiting
1905 the length of names to 32 characters, and the number of named subpatterns
1906 to 10,000.
1907
1908 11. Subpatterns that are repeated with specific counts have to be replicated in
1909 the compiled pattern. The size of memory for this was computed from the
1910 length of the subpattern and the repeat count. The latter is limited to
1911 65535, but there was no limit on the former, meaning that integer overflow
1912 could in principle occur. The compiled length of a repeated subpattern is
1913 now limited to 30,000 bytes in order to prevent this.
1914
1915 12. Added the optional facility to have named substrings with the same name.
1916
1917 13. Added the ability to use a named substring as a condition, using the
1918 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1919 are numbers (not recommended). Forward references are permitted.
1920
1921 14. Added forward references in named backreferences (if you see what I mean).
1922
1923 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1924 pattern could run off the end of the subject. For example, the pattern
1925 "(?s)(.{1,5})"8 did this with the subject "ab".
1926
1927 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1928 PCRE_CASELESS was set when matching characters that were quantified with ?
1929 or *.
1930
1931 17. A character class other than a single negated character that had a minimum
1932 but no maximum quantifier - for example [ab]{6,} - was not handled
1933 correctly by pce_dfa_exec(). It would match only one character.
1934
1935 18. A valid (though odd) pattern that looked like a POSIX character
1936 class but used an invalid character after [ (for example [[,abc,]]) caused
1937 pcre_compile() to give the error "Failed: internal error: code overflow" or
1938 in some cases to crash with a glibc free() error. This could even happen if
1939 the pattern terminated after [[ but there just happened to be a sequence of
1940 letters, a binary zero, and a closing ] in the memory that followed.
1941
1942 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1943 over the years. Originally (before any Unicode support), just the bottom 8
1944 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1945 output from "man perlunicode" includes this:
1946
1947 The regular expression compiler produces polymorphic opcodes. That
1948 is, the pattern adapts to the data and automatically switches to
1949 the Unicode character scheme when presented with Unicode data--or
1950 instead uses a traditional byte scheme when presented with byte
1951 data.
1952
1953 Sadly, a wide octal escape does not cause a switch, and in a string with
1954 no other multibyte characters, these octal escapes are treated as before.
1955 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1956 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1957 Unicode string.
1958
1959 I have not perpetrated such confusion in PCRE. Up till now, it took just
1960 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1961 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1962 translate to the appropriate multibyte character.
1963
1964 29. Applied some refactoring to reduce the number of warnings from Microsoft
1965 and Borland compilers. This has included removing the fudge introduced
1966 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1967 a warning about an unused variable.
1968
1969 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1970 characters since release 4.0, because Perl (from release 5.004) does not.
1971 [Or at least, is documented not to: some releases seem to be in conflict
1972 with the documentation.] However, when a pattern was studied with
1973 pcre_study() and all its branches started with \s, PCRE still included VT
1974 as a possible starting character. Of course, this did no harm; it just
1975 caused an unnecessary match attempt.
1976
1977 22. Removed a now-redundant internal flag bit that recorded the fact that case
1978 dependency changed within the pattern. This was once needed for "required
1979 byte" processing, but is no longer used. This recovers a now-scarce options
1980 bit. Also moved the least significant internal flag bit to the most-
1981 significant bit of the word, which was not previously used (hangover from
1982 the days when it was an int rather than a uint) to free up another bit for
1983 the future.
1984
1985 23. Added support for CRLF line endings as well as CR and LF. As well as the
1986 default being selectable at build time, it can now be changed at runtime
1987 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1988 specify that it is scanning data with non-default line endings.
1989
1990 24. Changed the definition of CXXLINK to make it agree with the definition of
1991 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1992
1993 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1994 recursions. This makes a big different to stack usage for some patterns.
1995
1996 26. If a subpattern containing a named recursion or subroutine reference such
1997 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1998 the space required for the compiled pattern went wrong and gave too small a
1999 value. Depending on the environment, this could lead to "Failed: internal
2000 error: code overflow at offset 49" or "glibc detected double free or
2001 corruption" errors.
2002
2003 27. Applied patches from Google (a) to support the new newline modes and (b) to
2004 advance over multibyte UTF-8 characters in GlobalReplace.
2005
2006 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2007 difference for some implementation of PCRE in some Windows version.
2008
2009 29. Added some extra testing facilities to pcretest:
2010
2011 \q<number> in a data line sets the "match limit" value
2012 \Q<number> in a data line sets the "match recursion limt" value
2013 -S <number> sets the stack size, where <number> is in megabytes
2014
2015 The -S option isn't available for Windows.
2016
2017
2018 Version 6.6 06-Feb-06
2019 ---------------------
2020
2021 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2022 in pcreposix.h. I have copied the definition from pcre.h.
2023
2024 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2025 because pcre.h is no longer a built file.
2026
2027 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2028 not normally included in the compiled code.
2029
2030
2031 Version 6.5 01-Feb-06
2032 ---------------------
2033
2034 1. When using the partial match feature with pcre_dfa_exec(), it was not
2035 anchoring the second and subsequent partial matches at the new starting
2036 point. This could lead to incorrect results. For example, with the pattern
2037 /1234/, partially matching against "123" and then "a4" gave a match.
2038
2039 2. Changes to pcregrep:
2040
2041 (a) All non-match returns from pcre_exec() were being treated as failures
2042 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2043 error message is output. Some extra information is given for the
2044 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2045 probably the only errors that are likely to be caused by users (by
2046 specifying a regex that has nested indefinite repeats, for instance).
2047 If there are more than 20 of these errors, pcregrep is abandoned.
2048
2049 (b) A binary zero was treated as data while matching, but terminated the
2050 output line if it was written out. This has been fixed: binary zeroes
2051 are now no different to any other data bytes.
2052
2053 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2054 used to set a locale for matching. The --locale=xxxx long option has
2055 been added (no short equivalent) to specify a locale explicitly on the
2056 pcregrep command, overriding the environment variables.
2057
2058 (d) When -B was used with -n, some line numbers in the output were one less
2059 than they should have been.
2060
2061 (e) Added the -o (--only-matching) option.
2062
2063 (f) If -A or -C was used with -c (count only), some lines of context were
2064 accidentally printed for the final match.
2065
2066 (g) Added the -H (--with-filename) option.
2067
2068 (h) The combination of options -rh failed to suppress file names for files
2069 that were found from directory arguments.
2070
2071 (i) Added the -D (--devices) and -d (--directories) options.
2072
2073 (j) Added the -F (--fixed-strings) option.
2074
2075 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2076
2077 (l) Added the --colo(u)r option.
2078
2079 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2080 is not present by default.
2081
2082 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2083 items such as (?R) or (?1), when the recursion could match a number of
2084 alternatives. If it matched one of the alternatives, but subsequently,
2085 outside the recursion, there was a failure, the code tried to back up into
2086 the recursion. However, because of the way PCRE is implemented, this is not
2087 possible, and the result was an incorrect result from the match.
2088
2089 In order to prevent this happening, the specification of recursion has
2090 been changed so that all such subpatterns are automatically treated as
2091 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2092
2093 4. I had overlooked the fact that, in some locales, there are characters for
2094 which isalpha() is true but neither isupper() nor islower() are true. In
2095 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2096 and ordfeminine) are like this. This affected the treatment of \w and \W
2097 when they appeared in character classes, but not when they appeared outside
2098 a character class. The bit map for "word" characters is now created
2099 separately from the results of isalnum() instead of just taking it from the
2100 upper, lower, and digit maps. (Plus the underscore character, of course.)
2101
2102 5. The above bug also affected the handling of POSIX character classes such as
2103 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2104 permanent tables. Instead, the bit maps for such a class were previously
2105 created as the appropriate unions of the upper, lower, and digit bitmaps.
2106 Now they are created by subtraction from the [[:word:]] class, which has
2107 its own bitmap.
2108
2109 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2110 It is created by subtracting the vertical space characters (\x09, \x0a,
2111 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2112 subtraction was done in the overall bitmap for a character class, meaning
2113 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2114 be recognized. This bug has been fixed.
2115
2116 7. Patches from the folks at Google:
2117
2118 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2119 real life, but is still worth protecting against".
2120
2121 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2122 regular expressions".
2123
2124 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2125 have it.
2126
2127 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2128 "configure" and the latter not, in order to fix a problem somebody had
2129 with compiling the Arg class on HP-UX.
2130
2131 (e) Improve the error-handling of the C++ wrapper a little bit.
2132
2133 (f) New tests for checking recursion limiting.
2134
2135 8. The pcre_memmove() function, which is used only if the environment does not
2136 have a standard memmove() function (and is therefore rarely compiled),
2137 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2138 returning a result (though PCRE never actually uses the result).
2139
2140 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2141 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2142 returned instead of calling malloc() with an overflowing number that would
2143 most likely cause subsequent chaos.
2144
2145 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2146
2147 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2148 with this option is matched, the nmatch and pmatch options of regexec() are
2149 ignored.
2150
2151 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2152 provided in case anyone wants to the the POSIX interface with UTF-8
2153 strings.
2154
2155 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2156 C++ linking (needed for some HP-UX environments).
2157
2158 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2159 (unused parameter) and in the pcre_printint() function (omitted "default"
2160 switch label when the default is to do nothing).
2161
2162 15. Added some code to make it possible, when PCRE is compiled as a C++
2163 library, to replace subject pointers for pcre_exec() with a smart pointer
2164 class, thus making it possible to process discontinuous strings.
2165
2166 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2167 much the same function. They were added by different people who were trying
2168 to make PCRE easy to compile on non-Unix systems. It has been suggested
2169 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2170 for compiling on Windows systems. I have therefore replaced it with
2171 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2172 defaults to "extern" for C or "extern C" for C++, which works fine on
2173 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2174 SCOPE with something explicit in config.h. In addition:
2175
2176 (a) pcreposix.h still had just "extern" instead of either of these macros;
2177 I have replaced it with PCRE_DATA_SCOPE.
2178
2179 (b) Functions such as _pcre_xclass(), which are internal to the library,
2180 but external in the C sense, all had PCRE_EXPORT in their definitions.
2181 This is apparently wrong for the Windows case, so I have removed it.
2182 (It makes no difference on Unix-like systems.)
2183
2184 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2185 of recursive calls to match(). This is different to MATCH_LIMIT because
2186 that limits the total number of calls to match(), not all of which increase
2187 the depth of recursion. Limiting the recursion depth limits the amount of
2188 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2189 when PCRE is compiled, and changed at run time. A patch from Google adds
2190 this functionality to the C++ interface.
2191
2192 18. Changes to the handling of Unicode character properties:
2193
2194 (a) Updated the table to Unicode 4.1.0.
2195
2196 (b) Recognize characters that are not in the table as "Cn" (undefined).
2197
2198 (c) I revised the way the table is implemented to a much improved format
2199 which includes recognition of ranges. It now supports the ranges that
2200 are defined in UnicodeData.txt, and it also amalgamates other
2201 characters into ranges. This has reduced the number of entries in the
2202 table from around 16,000 to around 3,000, thus reducing its size
2203 considerably. I realized I did not need to use a tree structure after
2204 all - a binary chop search is just as efficient. Having reduced the
2205 number of entries, I extended their size from 6 bytes to 8 bytes to
2206 allow for more data.
2207
2208 (d) Added support for Unicode script names via properties such as \p{Han}.
2209
2210 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2211 matching that character.
2212
2213 20. When matching a repeated Unicode property with a minimum greater than zero,
2214 (for example \pL{2,}), PCRE could look past the end of the subject if it
2215 reached it while seeking the minimum number of characters. This could
2216 happen only if some of the characters were more than one byte long, because
2217 there is a check for at least the minimum number of bytes.
2218
2219 21. Refactored the implementation of \p and \P so as to be more general, to
2220 allow for more different types of property in future. This has changed the
2221 compiled form incompatibly. Anybody with saved compiled patterns that use
2222 \p or \P will have to recompile them.
2223
2224 22. Added "Any" and "L&" to the supported property types.
2225
2226 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2227 but give a compile time error if the value is greater than 0xff.
2228
2229 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2230 accidentally not being installed or uninstalled.
2231
2232 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2233 made were to insert the current release number. This seemed silly, because
2234 it made things harder for people building PCRE on systems that don't run
2235 "configure". I have turned pcre.h into a distributed file, no longer built
2236 by "configure", with the version identification directly included. There is
2237 no longer a pcre.h.in file.
2238
2239 However, this change necessitated a change to the pcre-config script as
2240 well. It is built from pcre-config.in, and one of the substitutions was the
2241 release number. I have updated configure.ac so that ./configure now finds
2242 the release number by grepping pcre.h.
2243
2244 26. Added the ability to run the tests under valgrind.
2245
2246
2247 Version 6.4 05-Sep-05
2248 ---------------------
2249
2250 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2251 "--" to be printed when multiple files were scanned, even when none of the
2252 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2253 consider it to be a bug, and have restored the previous behaviour.
2254
2255 2. A couple of code tidies to get rid of compiler warnings.
2256
2257 3. The pcretest program used to cheat by referring to symbols in the library
2258 whose names begin with _pcre_. These are internal symbols that are not
2259 really supposed to be visible externally, and in some environments it is
2260 possible to suppress them. The cheating is now confined to including
2261 certain files from the library's source, which is a bit cleaner.
2262
2263 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2264 file's purpose clearer.
2265
2266 5. Reorganized pcre_ucp_findchar().
2267
2268
2269 Version 6.3 15-Aug-05
2270 ---------------------
2271
2272 1. The file libpcre.pc.in did not have general read permission in the tarball.
2273
2274 2. There were some problems when building without C++ support:
2275
2276 (a) If C++ support was not built, "make install" and "make test" still
2277 tried to test it.
2278
2279 (b) There were problems when the value of CXX was explicitly set. Some
2280 changes have been made to try to fix these, and ...
2281
2282 (c) --disable-cpp can now be used to explicitly disable C++ support.
2283
2284 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2285 backslash in a target when C++ was disabled. This confuses some
2286 versions of "make", apparently. Using an intermediate variable solves
2287 this. (Same for CPP_LOBJ.)
2288
2289 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2290 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2291 necessary on certain architectures.
2292
2293 4. Added a setting of -export-symbols-regex to the link command to remove
2294 those symbols that are exported in the C sense, but actually are local
2295 within the library, and not documented. Their names all begin with
2296 "_pcre_". This is not a perfect job, because (a) we have to except some
2297 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2298 available (and never for static libraries). I have made a note to try to
2299 find a way round (a) in the future.
2300
2301
2302 Version 6.2 01-Aug-05
2303 ---------------------
2304
2305 1. There was no test for integer overflow of quantifier values. A construction
2306 such as {1111111111111111} would give undefined results. What is worse, if
2307 a minimum quantifier for a parenthesized subpattern overflowed and became
2308 negative, the calculation of the memory size went wrong. This could have
2309 led to memory overwriting.
2310
2311 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2312
2313 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2314 operating environments where this matters.
2315
2316 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2317 PCRE options from within the C++ wrapper.
2318
2319 5. Named capturing subpatterns were not being correctly counted when a pattern
2320 was compiled. This caused two problems: (a) If there were more than 100
2321 such subpatterns, the calculation of the memory needed for the whole
2322 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2323 back references of the form \12, where the number was greater than 9, were
2324 not recognized as back references, even though there were sufficient
2325 previous subpatterns.
2326
2327 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2328 versions of gcc, e.g. 2.95.4.
2329
2330
2331 Version 6.1 21-Jun-05
2332 ---------------------
2333
2334 1. There was one reference to the variable "posix" in pcretest.c that was not
2335 surrounded by "#if !defined NOPOSIX".
2336
2337 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2338 the cross-check on the old pcre_info() function, for the benefit of the
2339 cut-down version of PCRE that is currently imported into Exim.
2340
2341 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2342 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2343 patterns that start (?i) but I don't think that matters. The use of (?i) is
2344 just an example; this all applies to the other options as well.
2345
2346 4. Since libtool seems to echo the compile commands it is issuing, the output
2347 from "make" can be reduced a bit by putting "@" in front of each libtool
2348 compile command.
2349
2350 5. Patch from the folks at Google for configure.in to be a bit more thorough
2351 in checking for a suitable C++ installation before trying to compile the
2352 C++ stuff. This should fix a reported problem when a compiler was present,
2353 but no suitable headers.
2354
2355 6. The man pages all had just "PCRE" as their title. I have changed them to
2356 be the relevant file name. I have also arranged that these names are
2357 retained in the file doc/pcre.txt, which is a concatenation in text format
2358 of all the man pages except the little individual ones for each function.
2359
2360 7. The NON-UNIX-USE file had not been updated for the different set of source
2361 files that come with release 6. I also added a few comments about the C++
2362 wrapper.
2363
2364
2365 Version 6.0 07-Jun-05
2366 ---------------------
2367
2368 1. Some minor internal re-organization to help with my DFA experiments.
2369
2370 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2371 didn't matter for the library itself when fully configured, but did matter
2372 when compiling without UCP support, or within Exim, where the ucp files are
2373 not imported.
2374
2375 3. Refactoring of the library code to split up the various functions into
2376 different source modules. The addition of the new DFA matching code (see
2377 below) to a single monolithic source would have made it really too
2378 unwieldy, quite apart from causing all the code to be include in a
2379 statically linked application, when only some functions are used. This is
2380 relevant even without the DFA addition now that patterns can be compiled in
2381 one application and matched in another.
2382
2383 The downside of splitting up is that there have to be some external
2384 functions and data tables that are used internally in different modules of
2385 the library but which are not part of the API. These have all had their
2386 names changed to start with "_pcre_" so that they are unlikely to clash
2387 with other external names.
2388
2389 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2390 a different (DFA) algorithm. Although it is slower than the original
2391 function, it does have some advantages for certain types of matching
2392 problem.
2393
2394 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2395 including restarting after a partial match.
2396
2397 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2398 defined when compiling for Windows was sent to me. I have put it into the
2399 code, though I have no means of testing or verifying it.
2400
2401 7. Added the pcre_refcount() auxiliary function.
2402
2403 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2404 match before or at the first newline in the subject string. In pcretest,
2405 the /f option on a pattern can be used to set this.
2406
2407 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2408 would behave wrongly. This has been present in PCRE since release 4.0.
2409
2410 10. A number of changes to the pcregrep command:
2411
2412 (a) Refactored how -x works; insert ^(...)$ instead of setting
2413 PCRE_ANCHORED and checking the length, in preparation for adding
2414 something similar for -w.
2415
2416 (b) Added the -w (match as a word) option.
2417
2418 (c) Refactored the way lines are read and buffered so as to have more
2419 than one at a time available.
2420
2421 (d) Implemented a pcregrep test script.
2422
2423 (e) Added the -M (multiline match) option. This allows patterns to match
2424 over several lines of the subject. The buffering ensures that at least
2425 8K, or the rest of the document (whichever is the shorter) is available
2426 for matching (and similarly the previous 8K for lookbehind assertions).
2427
2428 (f) Changed the --help output so that it now says
2429
2430 -w, --word-regex(p)
2431
2432 instead of two lines, one with "regex" and the other with "regexp"
2433 because that confused at least one person since the short forms are the
2434 same. (This required a bit of code, as the output is generated
2435 automatically from a table. It wasn't just a text change.)
2436
2437 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2438 option but starts with a hyphen. Could be a pattern or a path name
2439 starting with a hyphen, for instance.
2440
2441 (h) "-" can be given as a file name to represent stdin.
2442
2443 (i) When file names are being printed, "(standard input)" is used for
2444 the standard input, for compatibility with GNU grep. Previously
2445 "<stdin>" was used.
2446
2447 (j) The option --label=xxx can be used to supply a name to be used for
2448 stdin when file names are being printed. There is no short form.
2449
2450 (k) Re-factored the options decoding logic because we are going to add
2451 two more options that take data. Such options can now be given in four
2452 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2453
2454 (l) Added the -A, -B, and -C options for requesting that lines of context
2455 around matches be printed.
2456
2457 (m) Added the -L option to print the names of files that do not contain
2458 any matching lines, that is, the complement of -l.
2459
2460 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2461 continue to scan other files.
2462
2463 (o) The -s option was incorrectly implemented. For compatibility with other
2464 greps, it now suppresses the error message for a non-existent or non-
2465 accessible file (but not the return code). There is a new option called
2466 -q that suppresses the output of matching lines, which was what -s was
2467 previously doing.
2468
2469 (p) Added --include and --exclude options to specify files for inclusion
2470 and exclusion when recursing.
2471
2472 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2473 Hopefully, it now does.
2474
2475 12. Missing cast in pcre_study().
2476
2477 13. Added an "uninstall" target to the makefile.
2478
2479 14. Replaced "extern" in the function prototypes in Makefile.in with
2480 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2481 world, but is set differently for Windows.
2482
2483 15. Added a second compiling function called pcre_compile2(). The only
2484 difference is that it has an extra argument, which is a pointer to an
2485 integer error code. When there is a compile-time failure, this is set
2486 non-zero, in addition to the error test pointer being set to point to an
2487 error message. The new argument may be NULL if no error number is required
2488 (but then you may as well call pcre_compile(), which is now just a
2489 wrapper). This facility is provided because some applications need a
2490 numeric error indication, but it has also enabled me to tidy up the way
2491 compile-time errors are handled in the POSIX wrapper.
2492
2493 16. Added VPATH=.libs to the makefile; this should help when building with one
2494 prefix path and installing with another. (Or so I'm told by someone who
2495 knows more about this stuff than I do.)
2496
2497 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2498 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2499 match everything, including newlines. This is not POSIX-compatible, but
2500 somebody wanted the feature. From pcretest it can be activated by using
2501 both the P and the s flags.
2502
2503 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2504
2505 19. libpcre.pc was being incorrectly installed as executable.
2506
2507 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2508 it now also looks for '\r' so that it will work unmodified on Windows.
2509
2510 21. Added Google's contributed C++ wrapper to the distribution.
2511
2512 22. Added some untidy missing memory free() calls in pcretest, to keep
2513 Electric Fence happy when testing.
2514
2515
2516
2517 Version 5.0 13-Sep-04
2518 ---------------------
2519
2520 1. Internal change: literal characters are no longer packed up into items
2521 containing multiple characters in a single byte-string. Each character
2522 is now matched using a separate opcode. However, there may be more than one
2523 byte in the character in UTF-8 mode.
2524
2525 2. The pcre_callout_block structure has two new fields: pattern_position and
2526 next_item_length. These contain the offset in the pattern to the next match
2527 item, and its length, respectively.
2528
2529 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2530 insertion of callouts before each pattern item. Added the /C option to
2531 pcretest to make use of this.
2532
2533 4. On the advice of a Windows user, the lines
2534
2535 #if defined(_WIN32) || defined(WIN32)
2536 _setmode( _fileno( stdout ), 0x8000 );
2537 #endif /* defined(_WIN32) || defined(WIN32) */
2538
2539 have been added to the source of pcretest. This apparently does useful
2540 magic in relation to line terminators.
2541
2542 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2543 for the benefit of those environments where the "b" makes a difference.
2544
2545 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2546 to know about it. I have put a hack into configure.in that adds in code
2547 to set GCC=yes if CC=icc. This seems to end up at a point in the
2548 generated configure script that is early enough to affect the setting of
2549 compiler options, which is what is needed, but I have no means of testing
2550 whether it really works. (The user who reported this had patched the
2551 generated configure script, which of course I cannot do.)
2552
2553 LATER: After change 22 below (new libtool files), the configure script
2554 seems to know about icc (and also ecc). Therefore, I have commented out
2555 this hack in configure.in.
2556
2557 7. Added support for pkg-config (2 patches were sent in).
2558
2559 8. Negated POSIX character classes that used a combination of internal tables
2560 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2561 [[:^ascii]]. Typically, they would match almost any characters. The other
2562 POSIX classes were not broken in this way.
2563
2564 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2565 to find the match, as PCRE was deluded into thinking that the match had to
2566 start at the start point or following a newline. The same bug applied to
2567 patterns with negative forward assertions or any backward assertions
2568 preceding ".*" at the start, unless the pattern required a fixed first
2569 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2570
2571 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2572 starting at the last subject character, bytes beyond the end of the subject
2573 string were read.
2574
2575 11. Renamed the variable "class" as "classbits" to make life easier for C++
2576 users. (Previously there was a macro definition, but it apparently wasn't
2577 enough.)
2578
2579 12. Added the new field "tables" to the extra data so that tables can be passed
2580 in at exec time, or the internal tables can be re-selected. This allows
2581 a compiled regex to be saved and re-used at a later time by a different
2582 program that might have everything at different addresses.
2583
2584 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2585 -R library as well as a -L library.
2586
2587 14. The debugging options of pcretest (-d on the command line or D on a
2588 pattern) showed incorrect output for anything following an extended class
2589 that contained multibyte characters and which was followed by a quantifier.
2590
2591 15. Added optional support for general category Unicode character properties
2592 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2593 support. It adds about 90K to the size of the library. The meanings of the
2594 inbuilt class escapes such as \d and \s have NOT been changed.
2595
2596 16. Updated pcredemo.c to include calls to free() to release the memory for the
2597 compiled pattern.
2598
2599 17. The generated file chartables.c was being created in the source directory
2600 instead of in the building directory. This caused the build to fail if the
2601 source directory was different from the building directory, and was
2602 read-only.
2603
2604 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2605 file. No doubt somebody will tell me if they don't make sense... Also added
2606 Dan Mooney's comments about building on OpenVMS.
2607
2608 19. Added support for partial matching via the PCRE_PARTIAL option for
2609 pcre_exec() and the \P data escape in pcretest.
2610
2611 20. Extended pcretest with 3 new pattern features:
2612
2613 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2614 write the compiled pattern to the file whose name is "rest-of-line".
2615 This is a straight binary dump of the data, with the saved pointer to
2616 the character tables forced to be NULL. The study data, if any, is
2617 written too. After writing, pcretest reads a new pattern.
2618
2619 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2620 compiled pattern from the given file. There must not be any
2621 occurrences of "<" in the file name (pretty unlikely); if there are,
2622 pcretest will instead treat the initial "<" as a pattern delimiter.
2623 After reading in the pattern, pcretest goes on to read data lines as
2624 usual.
2625
2626 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2627 and 16-bit fields in a compiled pattern, to simulate a pattern that
2628 was compiled on a host of opposite endianness.
2629
2630 21. The pcre-exec() function can now cope with patterns that were compiled on
2631 hosts of opposite endianness, with this restriction:
2632
2633 As for any compiled expression that is saved and used later, the tables
2634 pointer field cannot be preserved; the extra_data field in the arguments
2635 to pcre_exec() should be used to pass in a tables address if a value
2636 other than the default internal tables were used at compile time.
2637
2638 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2639 now diagnosed as an error. Previously, most of the time, a negative number
2640 would have been treated as zero, but if in addition "ovector" was passed as
2641 NULL, a crash could occur.
2642
2643 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2644 new versions from the libtool 1.5 distribution (the last one is a copy of
2645 a file called libtool.m4). This seems to have fixed the need to patch
2646 "configure" to support Darwin 1.3 (which I used to do). However, I still
2647 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2648 workstation).
2649
2650 24. Changed the PCRE licence to be the more standard "BSD" licence.
2651
2652
2653 Version 4.5 01-Dec-03
2654 ---------------------
2655
2656 1. There has been some re-arrangement of the code for the match() function so
2657 that it can be compiled in a version that does not call itself recursively.
2658 Instead, it keeps those local variables that need separate instances for
2659 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2660 needs to "recurse". Keeping track of where control must go is done by means
2661 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2662 hide most of the details from the main code, and operates only if
2663 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2664 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2665 operating.
2666
2667 To make it easier for callers to provide specially tailored get/free
2668 functions for this usage, two new functions, pcre_stack_malloc, and
2669 pcre_stack_free, are used. They are always called in strict stacking order,
2670 and the size of block requested is always the same.
2671
2672 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2673 PCRE has been compiled to use the stack or the heap for recursion. The
2674 -C option of pcretest uses this to show which version is compiled.
2675
2676 A new data escape \S, is added to pcretest; it causes the amounts of store
2677 obtained and freed by both kinds of malloc/free at match time to be added
2678 to the output.
2679
2680 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2681 what's available on my current Linux desktop machine.
2682
2683 3. When matching a UTF-8 string, the test for a valid string at the start has
2684 been extended. If start_offset is not zero, PCRE now checks that it points
2685 to a byte that is the start of a UTF-8 character. If not, it returns
2686 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2687 this is necessary because there may be backward assertions in the pattern.
2688 When matching the same subject several times, it may save resources to use
2689 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2690
2691 4. The code for checking the validity of UTF-8 strings has been tightened so
2692 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2693 containing "overlong sequences".
2694
2695 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2696 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2697 should have been "&", but it just so happened that all the cases this let
2698 through by mistake were picked up later in the function.
2699
2700 6. I had used a variable called "isblank" - this is a C99 function, causing
2701 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2702
2703 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2704 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2705 so the version is shown; (c) stop "make test" echoing "./RunTest".
2706
2707 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2708
2709 9. The prototype for memmove() for systems that don't have it was using
2710 size_t, but the inclusion of the header that defines size_t was later. I've
2711 moved the #includes for the C headers earlier to avoid this.
2712
2713 10. Added some adjustments to the code to make it easier to compiler on certain
2714 special systems:
2715
2716 (a) Some "const" qualifiers were missing.
2717 (b) Added the macro EXPORT before all exported functions; by default this
2718 is defined to be empty.
2719 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2720 that it reads its output file name as an argument instead of writing
2721 to the standard output and assuming this can be redirected.
2722
2723 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2724 class containing characters with values greater than 255, PCRE compilation
2725 went into a loop.
2726
2727 12. A recursive reference to a subpattern that was within another subpattern
2728 that had a minimum quantifier of zero caused PCRE to crash. For example,
2729 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2730 recursion. If the recursively-called subpattern itself had a zero repeat,
2731 that was OK.
2732
2733 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2734 buffer into which it was copied (for escape processing) was still set at
2735 1024, so long lines caused crashes.
2736
2737 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2738 "internal error: code overflow...". This applied to any character class
2739 that was followed by a possessive quantifier.
2740
2741 15. Modified the Makefile to add libpcre.la as a prerequisite for
2742 libpcreposix.la because I was told this is needed for a parallel build to
2743 work.
2744
2745 16. If a pattern that contained .* following optional items at the start was
2746 studied, the wrong optimizing data was generated, leading to matching
2747 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2748 matching string must start with a or b or c. The correct conclusion for
2749 this pattern is that a match can start with any character.
2750
2751
2752 Version 4.4 13-Aug-03
2753 ---------------------
2754
2755 1. In UTF-8 mode, a character class containing characters with values between
2756 127 and 255 was not handled correctly if the compiled pattern was studied.
2757 In fixing this, I have also improved the studying algorithm for such
2758 classes (slightly).
2759
2760 2. Three internal functions had redundant arguments passed to them. Removal
2761 might give a very teeny performance improvement.
2762
2763 3. Documentation bug: the value of the capture_top field in a callout is *one
2764 more than* the number of the hightest numbered captured substring.
2765
2766 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2767 in incorrectly linking with a previously installed version. They now link
2768 explicitly with libpcre.la.
2769
2770 5. configure.in no longer needs to recognize Cygwin specially.
2771
2772 6. A problem in pcre.in for Windows platforms is fixed.
2773
2774 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2775 pcretest, it used to include the size of the study block as part of its
2776 output. Unfortunately, the structure contains a field that has a different
2777 size on different hardware architectures. This meant that the tests that
2778 showed this size failed. As the block is currently always of a fixed size,
2779 this information isn't actually particularly useful in pcretest output, so
2780 I have just removed it.
2781
2782 8. Three pre-processor statements accidentally did not start in column 1.
2783 Sadly, there are *still* compilers around that complain, even though
2784 standard C has not required this for well over a decade. Sigh.
2785
2786 9. In pcretest, the code for checking callouts passed small integers in the
2787 callout_data field, which is a void * field. However, some picky compilers
2788 complained about the casts involved for this on 64-bit systems. Now
2789 pcretest passes the address of the small integer instead, which should get
2790 rid of the warnings.
2791
2792 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2793 both compile and run time, and gives an error if an invalid UTF-8 sequence
2794 is found. There is a option for disabling this check in cases where the
2795 string is known to be correct and/or the maximum performance is wanted.
2796
2797 11. In response to a bug report, I changed one line in Makefile.in from
2798
2799 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2800 to
2801 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2802
2803 to look similar to other lines, but I have no way of telling whether this
2804 is the right thing to do, as I do not use Windows. No doubt I'll get told
2805 if it's wrong...
2806
2807
2808 Version 4.3 21-May-03
2809 ---------------------
2810
2811 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2812 Makefile.
2813
2814 2. Some refactoring to improve the quality of the code:
2815
2816 (i) The utf8_table... variables are now declared "const".
2817
2818 (ii) The code for \cx, which used the "case flipping" table to upper case
2819 lower case letters, now just substracts 32. This is ASCII-specific,
2820 but the whole concept of \cx is ASCII-specific, so it seems
2821 reasonable.
2822
2823 (iii) PCRE was using its character types table to recognize decimal and
2824 hexadecimal digits in the pattern. This is silly, because it handles
2825 only 0-9, a-f, and A-F, but the character types table is locale-
2826 specific, which means strange things might happen. A private
2827 table is now used for this - though it costs 256 bytes, a table is
2828 much faster than multiple explicit tests. Of course, the standard
2829 character types table is still used for matching digits in subject
2830 strings against \d.
2831
2832 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2833 ending in _t are). So I've renamed it as ESC_tee.
2834
2835 3. The first argument for regexec() in the POSIX wrapper should have been
2836 defined as "const".
2837
2838 4. Changed pcretest to use malloc() for its buffers so that they can be
2839 Electric Fenced for debugging.
2840
2841 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2842 to read one or more bytes before the start of the subject string. Often this
2843 had no effect on PCRE's behaviour, but in some circumstances it could
2844 provoke a segmentation fault.
2845
2846 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2847 to try to read one or more bytes before the start of the subject string.
2848
2849 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2850 UTF-8 support could misbehave in various ways if the subject string
2851 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2852 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2853 back over UTF-8 characters.)
2854
2855
2856 Version 4.2 14-Apr-03
2857 ---------------------
2858
2859 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2860
2861 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2862 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2863 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2864 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2865 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2866 and BUILD_EXEEXT
2867 Note: automatic setting of the BUILD variables is not yet working
2868 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2869 compile-time but not at link-time
2870 [LINK]: use for linking executables only
2871 make different versions for Windows and non-Windows
2872 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2873 libraries
2874 [LINK_FOR_BUILD]: new variable
2875 [OBJEXT]: use throughout
2876 [EXEEXT]: use throughout
2877 <winshared>: new target
2878 <wininstall>: new target
2879 <dftables.o>: use native compiler
2880 <dftables>: use native linker
2881 <install>: handle Windows platform correctly
2882 <clean>: ditto
2883 <check>: ditto
2884 copy DLL to top builddir before testing
2885
2886 As part of these changes, -no-undefined was removed again. This was reported
2887 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2888 in any case.
2889
2890 3. Some tidies to get rid of compiler warnings:
2891
2892 . In the match_data structure, match_limit was an unsigned long int, whereas
2893 match_call_count was an int. I've made them both unsigned long ints.
2894
2895 . In pcretest the fact that a const uschar * doesn't automatically cast to
2896 a void * provoked a warning.
2897
2898 . Turning on some more compiler warnings threw up some "shadow" variables
2899 and a few more missing casts.
2900
2901 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2902 option, a class that contained a single character with a value between 128
2903 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2904
2905 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2906 option, a class that contained several characters, but with at least one
2907 whose value was between 128 and 255 caused PCRE to crash.
2908
2909
2910 Version 4.1 12-Mar-03
2911 ---------------------
2912
2913 1. Compiling with gcc -pedantic found a couple of places where casts were
2914 needed, and a string in dftables.c that was longer than standard compilers are
2915 required to support.
2916
2917 2. Compiling with Sun's compiler found a few more places where the code could
2918 be tidied up in order to avoid warnings.
2919
2920 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2921 first of these names is deprecated in the latest Autoconf in favour of the name
2922 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2923 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2924 analogy I have changed it to CFLAGS_FOR_BUILD.
2925
2926 4. Added -no-undefined to the linking command in the Makefile, because this is
2927 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2928 linking step for the pcreposix library.
2929
2930 5. PCRE was failing to diagnose the case of two named groups with the same
2931 name.
2932
2933 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2934 literal character that is needed in the subject for a match, and scans along to
2935 ensure that it is present before embarking on the full matching process. This
2936 saves time in cases of nested unlimited repeats that are never going to match.
2937 Problem: the scan can take a lot of time if the subject is very long (e.g.
2938 megabytes), thus penalizing straightforward matches. It is now done only if the
2939 amount of subject to be scanned is less than 1000 bytes.
2940
2941 7. A lesser problem with the same optimization is that it was recording the
2942 first character of an anchored pattern as "needed", thus provoking a search
2943 right along the subject, even when the first match of the pattern was going to
2944 fail. The "needed" character is now not set for anchored patterns, unless it
2945 follows something in the pattern that is of non-fixed length. Thus, it still
2946 fulfils its original purpose of finding quick non-matches in cases of nested
2947 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2948
2949
2950 Version 4.0 17-Feb-03
2951 ---------------------
2952
2953 1. If a comment in an extended regex that started immediately after a meta-item
2954 extended to the end of string, PCRE compiled incorrect data. This could lead to
2955 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2956
2957 2. Moved to autoconf 2.53 and libtool 1.4.2.
2958
2959 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2960 the special perltest8 script is no longer needed - all the tests can be run
2961 from a single perltest script.
2962
2963 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2964 by \s. It has now been removed in PCRE. This means it isn't recognized as
2965 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2966 class [:space:] *does* include VT, thereby creating a mess.
2967
2968 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2969 space and tab.
2970
2971 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2972 its new features into the main test script, reducing the number of scripts.
2973
2974 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2975 were backward compatible, and made the (?i) apply to the whole pattern, as if
2976 /i were given. Now it behaves more logically, and applies the option setting
2977 only to what follows. PCRE has been changed to follow suit. However, if it
2978 finds options settings right at the start of the pattern, it extracts them into
2979 the global options, as before. Thus, they show up in the info data.
2980
2981 8. Added support for the \Q...\E escape sequence. Characters in between are
2982 treated as literals. This is slightly different from Perl in that $ and @ are
2983 also handled as literals inside the quotes. In Perl, they will cause variable
2984 interpolation. Note the following examples:
2985
2986 Pattern PCRE matches Perl matches
2987
2988 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2989 \Qabc\$xyz\E abc\$xyz abc\$xyz
2990 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2991
2992 For compatibility with Perl, \Q...\E sequences are recognized inside character
2993 classes as well as outside them.
2994
2995 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2996 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2997 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2998 signed/unsigned warnings.
2999
3000 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3001 option for pcretest, so I've replaced it by a simple function that does just
3002 that job.
3003
3004 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3005 "pcregrep -".
3006
3007 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3008 Java package. This provides some syntactic sugar for simple cases of what my
3009 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3010 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3011 item, you can use this simplified notation. Note that only makes sense with
3012 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3013 greediness, whatever the setting of the PCRE_UNGREEDY option.
3014
3015 13. A change of greediness default within a pattern was not taking effect at
3016 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3017 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3018 was abstracted outside.
3019
3020 14. PCRE now supports the \G assertion. It is true when the current matching
3021 position is at the start point of the match. This differs from \A when the
3022 starting offset is non-zero. Used with the /g option of pcretest (or similar
3023 code), it works in the same way as it does for Perl's /g option. If all
3024 alternatives of a regex begin with \G, the expression is anchored to the start
3025 match position, and the "anchored" flag is set in the compiled expression.
3026
3027 15. Some bugs concerning the handling of certain option changes within patterns
3028 have been fixed. These applied to options other than (?ims). For example,
3029 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3030 been the other way round. Some of this was related to change 7 above.
3031
3032 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3033 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3034 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3035 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3036
3037 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3038 mode. Unlike ".", it always matches newline, whatever the setting of
3039 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3040 assertions. Perl allows it, but it doesn't (in general) work because it can't
3041 calculate the length of the lookbehind. At least, that's the case for Perl
3042 5.8.0 - I've been told they are going to document that it doesn't work in
3043 future.
3044
3045 18. Added an error diagnosis for escapes that PCRE does not support: these are
3046 \L, \l, \N, \P, \p, \U, \u, and \X.
3047
3048 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3049 reading past the end of the pattern in cases such as /[abcd/.
3050
3051 20. PCRE was getting more memory than necessary for patterns with classes that
3052 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3053
3054 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3055 compiling PCRE for use with Virtual Pascal.
3056
3057 22. Small fix to the Makefile to make it work properly if the build is done
3058 outside the source tree.
3059
3060 23. Added a new extension: a condition to go with recursion. If a conditional
3061 subpattern starts with (?(R) the "true" branch is used if recursion has
3062 happened, whereas the "false" branch is used only at the top level.
3063
3064 24. When there was a very long string of literal characters (over 255 bytes
3065 without UTF support, over 250 bytes with UTF support), the computation of how
3066 much memory was required could be incorrect, leading to segfaults or other
3067 strange effects.
3068
3069 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3070 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3071 there was a subsequent back reference to those brackets. This meant that, for
3072 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3073 possible to check for precisely this case. All we can do is abandon the
3074 optimization if .* occurs inside capturing brackets when there are any back
3075 references whatsoever. (See below for a better fix that came later.)
3076
3077 26. The handling of the optimization for finding the first character of a
3078 non-anchored pattern, and for finding a character that is required later in the
3079 match were failing in some cases. This didn't break the matching; it just
3080 failed to optimize when it could. The way this is done has been re-implemented.
3081
3082 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3083
3084 28. Added a new feature that provides some of the functionality that Perl
3085 provides with (?{...}). The facility is termed a "callout". The way it is done
3086 in PCRE is for the caller to provide an optional function, by setting
3087 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3088 global variable. By default it is unset, which disables all calling out. To get
3089 the function called, the regex must include (?C) at appropriate points. This
3090 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3091 This provides a means of identifying different callout points. When PCRE
3092 reaches such a point in the regex, if pcre_callout has been set, the external
3093 function is called. It is provided with data in a structure called
3094 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3095 matching continues; if it returns a non-zero value, the match at the current
3096 point fails. However, backtracking will occur if possible. [This was changed
3097 later and other features added - see item 49 below.]
3098
3099 29. pcretest is upgraded to test the callout functionality. It provides a
3100 callout function that displays information. By default, it shows the start of
3101 the match and the current position in the text. There are some new data escapes
3102 to vary what happens:
3103
3104 \C+ in addition, show current contents of captured substrings
3105 \C- do not supply a callout function
3106 \C!n return 1 when callout number n is reached
3107 \C!n!m return 1 when callout number n is reached for the mth time
3108
3109 30. If pcregrep was called with the -l option and just a single file name, it
3110 output "<stdin>" if a match was found, instead of the file name.
3111
3112 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3113 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3114 pcre_exec(). This saves a malloc/free per call. The default value of
3115 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3116 when configuring.
3117
3118 32. The default maximum size of a compiled pattern is 64K. There have been a
3119 few cases of people hitting this limit. The code now uses macros to handle the
3120 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3121 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3122 configuring. Tests 2 and 5 work only with 2-byte links because they output
3123 debugging information about compiled patterns.
3124
3125 33. Internal code re-arrangements:
3126
3127 (a) Moved the debugging function for printing out a compiled regex into
3128 its own source file (printint.c) and used #include to pull it into
3129 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3130 separate copies.
3131
3132 (b) Defined the list of op-code names for debugging as a macro in
3133 internal.h so that it is next to the definition of the opcodes.
3134
3135 (c) Defined a table of op-code lengths for simpler skipping along compiled
3136 code. This is again a macro in internal.h so that it is next to the
3137 definition of the opcodes.
3138
3139 34. Added support for recursive calls to individual subpatterns, along the
3140 lines of Robin Houston's patch (but implemented somewhat differently).
3141
3142 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3143 allow it to read and process whole directories in Win32. This code was
3144 contributed by Lionel Fourquaux; it has not been tested by me.
3145
3146 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3147 used to name a group. Names consist of alphanumerics and underscores, and must
3148 be unique. Back references use the syntax (?P=name) and recursive calls use
3149 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3150 numbers. The function pcre_fullinfo() can be used after compilation to extract
3151 a name/number map. There are three relevant calls:
3152
3153 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3154 PCRE_INFO_NAMECOUNT yields the number of entries
3155 PCRE_INFO_NAMETABLE yields a pointer to the map.
3156
3157 The map is a vector of fixed-size entries. The size of each entry depends on
3158 the length of the longest name used. The first two bytes of each entry are the
3159 group number, most significant byte first. There follows the corresponding
3160 name, zero terminated. The names are in alphabetical order.
3161
3162 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3163 case instead of 255. Making it the same both with and without UTF-8 support
3164 means that the same test output works with both.
3165
3166 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3167 calling malloc() with a zero argument.
3168
3169 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3170 optimization. I've improved things by keeping a bitmap of backreferences with
3171 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3172 fact referenced, the optimization can be applied. It is unlikely that a
3173 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3174 the match to follow \n) will appear inside brackets with a number greater than
3175 31, but if it does, any back reference > 31 suppresses the optimization.
3176
3177 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3178 of disabling numbered capturing parentheses. Any opening parenthesis that is
3179 not followed by ? behaves as if it were followed by ?: but named parentheses
3180 can still be used for capturing (and they will acquire numbers in the usual
3181 way).
3182
3183 41. Redesigned the return codes from the match() function into yes/no/error so
3184 that errors can be passed back from deep inside the nested calls. A malloc
3185 failure while inside a recursive subpattern call now causes the
3186 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3187
3188 42. It is now possible to set a limit on the number of times the match()
3189 function is called in a call to pcre_exec(). This facility makes it possible to
3190 limit the amount of recursion and backtracking, though not in a directly
3191 obvious way, because the match() function is used in a number of different
3192 circumstances. The count starts from zero for each position in the subject
3193 string (for non-anchored patterns). The default limit is, for compatibility, a
3194 large number, namely 10 000 000. You can change this in two ways:
3195
3196 (a) When configuring PCRE before making, you can use --with-match-limit=n
3197 to set a default value for the compiled library.
3198
3199 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3200 a different value is set. See 45 below.
3201
3202 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3203
3204 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3205 of things that can be changed at compile time. The first argument specifies
3206 what is wanted and the second points to where the information is to be placed.
3207 The current list of available information is:
3208
3209 PCRE_CONFIG_UTF8
3210
3211 The output is an integer that is set to one if UTF-8 support is available;
3212 otherwise it is set to zero.
3213
3214 PCRE_CONFIG_NEWLINE
3215
3216 The output is an integer that it set to the value of the code that is used for
3217 newline. It is either LF (10) or CR (13).
3218
3219 PCRE_CONFIG_LINK_SIZE
3220
3221 The output is an integer that contains the number of bytes used for internal
3222 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3223
3224 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3225
3226 The output is an integer that contains the threshold above which the POSIX
3227 interface uses malloc() for output vectors. See item 31 above.
3228
3229 PCRE_CONFIG_MATCH_LIMIT
3230
3231 The output is an unsigned integer that contains the default limit of the number
3232 of match() calls in a pcre_exec() execution. See 42 above.
3233
3234 44. pcretest has been upgraded by the addition of the -C option. This causes it
3235 to extract all the available output from the new pcre_config() function, and to
3236 output it. The program then exits immediately.
3237
3238 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3239 order to support additional features. One way would have been to define
3240 pcre_exec2() (for example) with extra arguments, but this would not have been
3241 extensible, and would also have required all calls to the original function to
3242 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3243 is used for passing in "extra" data from pcre_study().
3244
3245 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3246 contains the following fields:
3247
3248 flags a bitmap indicating which of the following fields are set
3249 study_data opaque data from pcre_study()
3250 match_limit a way of specifying a limit on match() calls for a specific
3251 call to pcre_exec()
3252 callout_data data for callouts (see 49 below)
3253
3254 The flag bits are also defined in pcre.h, and are
3255
3256 PCRE_EXTRA_STUDY_DATA
3257 PCRE_EXTRA_MATCH_LIMIT
3258 PCRE_EXTRA_CALLOUT_DATA
3259
3260 The pcre_study() function now returns one of these new pcre_extra blocks, with
3261 the actual study data pointed to by the study_data field, and the
3262 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3263 before. That is, this change is entirely upwards-compatible and requires no
3264 change to existing code.
3265
3266 If you want to pass in additional data to pcre_exec(), you can either place it
3267 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3268 block.
3269
3270 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3271 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3272 times with different match limits, until it finds the minimum value needed for
3273 pcre_exec() to complete. The value is then output. This can be instructive; for
3274 most simple matches the number is quite small, but for pathological cases it
3275 gets very large very quickly.
3276
3277 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3278 returns the size of the data block pointed to by the study_data field in a
3279 pcre_extra block, that is, the value that was passed as the argument to
3280 pcre_malloc() when PCRE was getting memory in which to place the information
3281 created by pcre_study(). The fourth argument should point to a size_t variable.
3282 pcretest has been extended so that this information is shown after a successful
3283 pcre_study() call when information about the compiled regex is being displayed.
3284
3285 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3286 because what follows is always an absolute path. (Later: it turns out that this
3287 is more than cosmetic for MinGW, because it doesn't like empty path
3288 components.)
3289
3290 49. Some changes have been made to the callout feature (see 28 above):
3291
3292 (i) A callout function now has three choices for what it returns:
3293
3294 0 => success, carry on matching
3295 > 0 => failure at this point, but backtrack if possible
3296 < 0 => serious error, return this value from pcre_exec()
3297
3298 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3299 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3300 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3301 use by callout functions. It will never be used by PCRE itself.
3302
3303 (ii) The pcre_extra structure (see 45 above) has a void * field called
3304 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3305 pcre_callout_block structure has a field of the same name. The contents of
3306 the field passed in the pcre_extra structure are passed to the callout
3307 function in the corresponding field in the callout block. This makes it
3308 easier to use the same callout-containing regex from multiple threads. For
3309 testing, the pcretest program has a new data escape
3310
3311 \C*n pass the number n (may be negative) as callout_data
3312
3313 If the callout function in pcretest receives a non-zero value as
3314 callout_data, it returns that value.
3315
3316 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3317 there were some redundant $(CFLAGS) in commands that are now specified as
3318 $(LINK), which already includes $(CFLAGS).
3319
3320 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3321 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3322 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3323 one-byte characters throughout. Note that case-insensitive matching applies
3324 only to characters whose values are less than 256. PCRE doesn't support the
3325 notion of cases for higher-valued characters.
3326
3327 (i) A character class whose characters are all within 0-255 is handled as
3328 a bit map, and the map is inverted for negative classes. Previously, a
3329 character > 255 always failed to match such a class; however it should
3330 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3331
3332 (ii) A negated character class with a single character < 255 is coded as
3333 "not this character" (OP_NOT). This wasn't working properly when the test
3334 character was multibyte, either singly or repeated.
3335
3336 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3337 mode, for example: \x{100}{2,3}.
3338
3339 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3340 singly or repeated) now correctly test multibyte characters. However,
3341 PCRE doesn't recognize any characters with values greater than 255 as
3342 digits, spaces, or word characters. Such characters always match \D, \S,
3343 and \W, and never match \d, \s, or \w.
3344
3345 (v) Classes may now contain characters and character ranges with values
3346 greater than 255. For example: [ab\x{100}-\x{400}].
3347
3348 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3349 PCRE in UTF-8 mode.
3350
3351 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3352 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3353 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3354 value.)
3355
3356 53. The single man page has become too large. I have therefore split it up into
3357 a number of separate man pages. These also give rise to individual HTML pages;
3358 these are now put in a separate directory, and there is an index.html page that
3359 lists them all. Some hyperlinking between the pages has been installed.
3360
3361 54. Added convenience functions for handling named capturing parentheses.
3362
3363 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3364 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3365 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3366 are faulted.
3367
3368 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3369 calling configure. These values are used when compiling the dftables.c program
3370 which is run to generate the source of the default character tables. They
3371 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3372 you will need to set these values.
3373
3374 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3375
3376
3377 Version 3.9 02-Jan-02
3378 ---------------------
3379
3380 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3381
3382 2. If --disable-static was given, the building process failed when trying to
3383 build pcretest and pcregrep. (For some reason it was using libtool to compile
3384 them, which is not right, as they aren't part of the library.)
3385
3386
3387 Version 3.8 18-Dec-01
3388 ---------------------
3389
3390 1. The experimental UTF-8 code was completely screwed up. It was packing the
3391 bytes in the wrong order. How dumb can you get?
3392
3393
3394 Version 3.7 29-Oct-01
3395 ---------------------
3396
3397 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3398 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3399 this didn't happen under Solaris 8, where I normally test things.
3400
3401 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3402 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3403 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3404 this fix an hour or so after the initial 3.7 release.)
3405
3406
3407 Version 3.6 23-Oct-01
3408 ---------------------
3409
3410 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3411 offsets passed as NULL with zero offset count.
3412
3413 2. The config.guess and config.sub files had not been updated when I moved to
3414 the latest autoconf.
3415
3416
3417 Version 3.5 15-Aug-01
3418 ---------------------
3419
3420 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3421 had been forgotten.
3422
3423 2. By using declared but undefined structures, we can avoid using "void"
3424 definitions in pcre.h while keeping the internal definitions of the structures
3425 private.
3426
3427 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3428 user point of view, this means that both static and shared libraries are built
3429 by default, but this can be individually controlled. More of the work of
3430 handling this static/shared cases is now inside libtool instead of PCRE's make
3431 file.
3432
3433 4. The pcretest utility is now installed along with pcregrep because it is
3434 useful for users (to test regexs) and by doing this, it automatically gets
3435 relinked by libtool. The documentation has been turned into a man page, so
3436 there are now .1, .txt, and .html versions in /doc.
3437
3438 5. Upgrades to pcregrep:
3439 (i) Added long-form option names like gnu grep.
3440 (ii) Added --help to list all options with an explanatory phrase.
3441 (iii) Added -r, --recursive to recurse into sub-directories.
3442 (iv) Added -f, --file to read patterns from a file.
3443
3444 6. pcre_exec() was referring to its "code" argument before testing that
3445 argument for NULL (and giving an error if it was NULL).
3446
3447 7. Upgraded Makefile.in to allow for compiling in a different directory from
3448 the source directory.
3449
3450 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3451 options bits, the pointer it was passed was to an int instead of to an unsigned
3452 long int. This mattered only on 64-bit systems.
3453
3454 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3455 generated) instead of pcre.in, which it its source. Also made the same change
3456 in several of the .c files.
3457
3458 10. A new release of gcc defines printf() as a macro, which broke pcretest
3459 because it had an ifdef in the middle of a string argument for printf(). Fixed
3460 by using separate calls to printf().
3461
3462 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3463 script, to force use of CR or LF instead of \n in the source. On non-Unix
3464 systems, the value can be set in config.h.
3465
3466 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3467 absolute limit. Changed the text of the error message to make this clear, and
3468 likewise updated the man page.
3469
3470 13. The limit of 99 on the number of capturing subpatterns has been removed.
3471 The new limit is 65535, which I hope will not be a "real" limit.
3472
3473
3474 Version 3.4 22-Aug-00
3475 ---------------------
3476
3477 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3478
3479 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3480
3481
3482 Version 3.3 01-Aug-00
3483 ---------------------
3484
3485 1. If an octal character was given, but the value was greater than \377, it
3486 was not getting masked to the least significant bits, as documented. This could
3487 lead to crashes in some systems.
3488
3489 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3490 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3491
3492 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3493 These just pass their arguments on to (pcre_free)(), but they are provided
3494 because some uses of PCRE bind it to non-C systems that can call its functions,
3495 but cannot call free() or pcre_free() directly.
3496
3497 4. Add "make test" as a synonym for "make check". Corrected some comments in
3498 the Makefile.
3499
3500 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3501 Makefile.
3502
3503 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3504 command called pgrep for grepping around the active processes.
3505
3506 7. Added the beginnings of support for UTF-8 character strings.
3507
3508 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3509 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3510 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3511 out for the ar command.)
3512
3513
3514 Version 3.2 12-May-00
3515 ---------------------
3516
3517 This is purely a bug fixing release.
3518
3519 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3520 of ZA. This was just one example of several cases that could provoke this bug,
3521 which was introduced by change 9 of version 2.00. The code for breaking
3522 infinite loops after an iteration that matches an empty string was't working
3523 correctly.
3524
3525 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3526 when matched against abbab (for example). After matching an empty string, it
3527 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3528 caused it to match further down the string than it should.
3529
3530 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3531 was there because it doesn't seem to be needed, and it causes trouble on some
3532 systems, as it is not a Standard C header. It has been removed.
3533
3534 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3535 were reported on the Macintosh. The changes were from
3536
3537 while ((c = *(++ptr)) != 0 && c != '\n');
3538 to
3539 while ((c = *(++ptr)) != 0 && c != '\n') ;
3540
3541 Totally extraordinary, but if that's what it takes...
3542
3543 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3544 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3545 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3546 assumes the way PCRE uses memmove() (always moving upwards).
3547
3548 6. PCRE is being used in one environment where strchr() is not available. There
3549 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3550 faster code anyway.
3551
3552
3553 Version 3.1 09-Feb-00
3554 ---------------------
3555
3556 The only change in this release is the fixing of some bugs in Makefile.in for
3557 the "install" target:
3558
3559 (1) It was failing to install pcreposix.h.
3560
3561 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3562
3563
3564 Version 3.0 01-Feb-00
3565 ---------------------
3566
3567 1. Add support for the /+ modifier to perltest (to output $` like it does in
3568 pcretest).
3569
3570 2. Add support for the /g modifier to perltest.
3571
3572 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3573 matches null strings.
3574
3575 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3576 pattern. Perl treats empty patterns specially - it reuses the most recent
3577 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3578 effect.
3579
3580 5. The POSIX interface was broken in that it was just handing over the POSIX
3581 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3582 required a bigger vector, with some working space on the end. This means that
3583 the POSIX wrapper now has to get and free some memory, and copy the results.
3584
3585 6. Added some simple autoconf support, placing the test data and the
3586 documentation in separate directories, re-organizing some of the
3587 information files, and making it build pcre-config (a GNU standard). Also added
3588 libtool support for building PCRE as a shared library, which is now the
3589 default.
3590
3591 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3592 09 are not valid octal constants. Single digits will be used for minor values
3593 less than 10.
3594
3595 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3596 existing programs that set these in the POSIX interface can use PCRE without
3597 modification.
3598
3599 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3600 return all that pcre_info() returns, plus additional data. The pcre_info()
3601 function is retained for compatibility, but is considered to be obsolete.
3602
3603 10. Added experimental recursion feature (?R) to handle one common case that
3604 Perl 5.6 will be able to do with (?p{...}).
3605
3606 11. Added support for POSIX character classes like [:alpha:], which Perl is
3607 adopting.
3608
3609
3610 Version 2.08 31-Aug-99
3611 ----------------------
3612
3613 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3614 trying to match at the startoffset position, but instead was moving forward to
3615 the next newline as if a previous match had failed.
3616
3617 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3618 and could get into a loop if a null string was matched other than at the start
3619 of the subject.
3620
3621 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3622 be distinguished at compile time, and for completeness also added PCRE_DATE.
3623
3624 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3625 in GnuWin32 environments.
3626
3627
3628 Version 2.07 29-Jul-99
3629 ----------------------
3630
3631 1. The documentation is now supplied in plain text form and HTML as well as in
3632 the form of man page sources.
3633
3634 2. C++ compilers don't like assigning (void *) values to other pointer types.
3635 In particular this affects malloc(). Although there is no problem in Standard
3636 C, I've put in casts to keep C++ compilers happy.
3637
3638 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3639 should be (const char *).
3640
3641 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3642 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3643 However, I haven't made this a standard facility. The documentation doesn't
3644 mention it, and the Makefile doesn't support it.
3645
3646 5. The Makefile now contains an "install" target, with editable destinations at
3647 the top of the file. The pcretest program is not installed.
3648
3649 6. pgrep -V now gives the PCRE version number and date.
3650
3651 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3652 causing the entire string to be ignored, instead of just the last character.
3653
3654 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3655 non-matching string, it can take a very, very long time, even for strings of
3656 quite modest length, because of the nested recursion. PCRE now does better in
3657 some of these cases. It does this by remembering the last required literal
3658 character in the pattern, and pre-searching the subject to ensure it is present
3659 before running the real match. In other words, it applies a heuristic to detect
3660 some types of certain failure quickly, and in the above example, if presented
3661 with a string that has no trailing " it gives "no match" very quickly.
3662
3663 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3664 other alternatives are tried instead.
3665
3666
3667 Version 2.06 09-Jun-99
3668 ----------------------
3669
3670 1. Change pcretest's output for amount of store used to show just the code
3671 space, because the remainder (the data block) varies in size between 32-bit and
3672 64-bit systems.
3673
3674 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3675 start matching at. This allows lookbehinds to work when searching for multiple
3676 occurrences in a string.
3677
3678 3. Added additional options to pcretest for testing multiple occurrences:
3679
3680 /+ outputs the rest of the string that follows a match
3681 /g loops for multiple occurrences, using the new startoffset argument
3682 /G loops for multiple occurrences by passing an incremented pointer
3683
3684 4. PCRE wasn't doing the "first character" optimization for patterns starting
3685 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3686 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3687 the letter 'x'. On long subject strings, this gives a significant speed-up.
3688
3689
3690 Version 2.05 21-Apr-99
3691 ----------------------
3692
3693 1. Changed the type of magic_number from int to long int so that it works
3694 properly on 16-bit systems.
3695
3696 2. Fixed a bug which caused patterns starting with .* not to work correctly
3697 when the subject string contained newline characters. PCRE was assuming
3698 anchoring for such patterns in all cases, which is not correct because .* will
3699 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3700 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3701 must be retried after every newline in the subject.
3702
3703
3704 Version 2.04 18-Feb-99
3705 ----------------------
3706
3707 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3708 computation of the store needed to hold the pattern was incorrect (too large).
3709 If such patterns were nested a few deep, this could multiply and become a real
3710 problem.
3711
3712 2. Added /M option to pcretest to show the memory requirement of a specific
3713 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3714
3715 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3716 compiled in such a way that the backtracking after subsequent failure was
3717 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3718 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3719
3720
3721 Version 2.03 02-Feb-99
3722 ----------------------
3723
3724 1. Fixed typo and small mistake in man page.
3725
3726 2. Added 4th condition (GPL supersedes if conflict) and created separate
3727 LICENCE file containing the conditions.
3728
3729 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3730 Perl, that is the internal \ allows the delimiter to be included in the
3731 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3732 the final delimiter, add \ to the end of the pattern (to test the error).
3733
3734 4. Added the convenience functions for extracting substrings after a successful
3735 match. Updated pcretest to make it able to test these functions.
3736
3737
3738 Version 2.02 14-Jan-99
3739 ----------------------
3740
3741 1. Initialized the working variables associated with each extraction so that
3742 their saving and restoring doesn't refer to uninitialized store.
3743
3744 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3745 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3746 fix the problem.
3747
3748 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3749 calls, and wasn't printing the correct value for compiling calls. Increased the
3750 default value of LOOPREPEAT, and the number of significant figures in the
3751 times.
3752
3753 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3754
3755 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3756 a building problem on Windows NT with a FAT file system.
3757
3758
3759 Version 2.01 21-Oct-98
3760 ----------------------
3761
3762 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3763 to character tables built by pcre_maketables() in the current locale. If NULL
3764 is passed, the default tables are used.
3765
3766
3767 Version 2.00 24-Sep-98
3768 ----------------------
3769
3770 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3771 it any more.
3772
3773 2. Allow quantification of (?>) groups, and make it work correctly.
3774
3775 3. The first character computation wasn't working for (?>) groups.
3776
3777 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3778 end of the subject) and add 5.005's \z, which really does match only at the
3779 very end of the subject.
3780
3781 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3782
3783 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3784 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3785 localized options. All options to pcre_study() were also removed.
3786
3787 7. Add other new features from 5.005:
3788
3789 $(?<= positive lookbehind
3790 $(?<! negative lookbehind
3791 (?imsx-imsx) added the unsetting capability
3792 such a setting is global if at outer level; local otherwise
3793 (?imsx-imsx:) non-capturing groups with option setting
3794 (?(cond)re|re) conditional pattern matching
3795
3796 A backreference to itself in a repeated group matches the previous
3797 captured string.
3798
3799 8. General tidying up of studying (both automatic and via "study")
3800 consequential on the addition of new assertions.
3801
3802 9. As in 5.005, unlimited repeated groups that could match an empty substring
3803 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3804 runtime if any iteration does actually match an empty substring.
3805
3806 10. Include the RunTest script in the distribution.
3807
3808 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3809 discrepancies, some of which were old and were also with respect to 5.004. They
3810 have now been fixed.
3811
3812
3813 Version 1.09 28-Apr-98
3814 ----------------------
3815
3816 1. A negated single character class followed by a quantifier with a minimum
3817 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3818 program crashes, or just wrong answers. This did not apply to negated classes
3819 containing more than one character, or to minima other than one.
3820
3821
3822 Version 1.08 27-Mar-98
3823 ----------------------
3824
3825 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3826
3827 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3828 latter must appear before anything that relies on it in the pattern.
3829
3830
3831 Version 1.07 16-Feb-98
3832 ----------------------
3833
3834 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3835 repeat of a potentially empty string).
3836
3837
3838 Version 1.06 23-Jan-98
3839 ----------------------
3840
3841 1. Added Markus Oberhumer's little patches for C++.
3842
3843 2. Literal strings longer than 255 characters were broken.
3844
3845
3846 Version 1.05 23-Dec-97
3847 ----------------------
3848
3849 1. Negated character classes containing more than one character were failing if
3850 PCRE_CASELESS was set at run time.
3851
3852
3853 Version 1.04 19-Dec-97
3854 ----------------------
3855
3856 1. Corrected the man page, where some "const" qualifiers had been omitted.
3857
3858 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3859 input syntax.
3860
3861 3. Fixed memory leak which occurred when a regex with back references was
3862 matched with an offsets vector that wasn't big enough. The temporary memory
3863 that is used in this case wasn't being freed if the match failed.
3864
3865 4. Tidied pcretest to ensure it frees memory that it gets.
3866
3867 5. Temporary memory was being obtained in the case where the passed offsets
3868 vector was exactly big enough.
3869
3870 6. Corrected definition of offsetof() from change 5 below.
3871
3872 7. I had screwed up change 6 below and broken the rules for the use of
3873 setjmp(). Now fixed.
3874
3875
3876 Version 1.03 18-Dec-97
3877 ----------------------
3878
3879 1. A erroneous regex with a missing opening parenthesis was correctly
3880 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3881 on some systems.
3882
3883 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3884 it was reported that one broken compiler failed on the former because "code" is
3885 also an independent variable.
3886
3887 3. The erroneous regex a[]b caused an array overrun reference.
3888
3889 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3890 fail on data ending with that character. (It was going on too far, and checking
3891 the next character, typically a binary zero.) This was specific to the
3892 optimized code for single-character negative classes.
3893
3894 5. Added a contributed patch from the TIN world which does the following:
3895
3896 + Add an undef for memmove, in case the the system defines a macro for it.
3897
3898 + Add a definition of offsetof(), in case there isn't one. (I don't know
3899 the reason behind this - offsetof() is part of the ANSI standard - but
3900 it does no harm).
3901
3902 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3903 most of the places where whitespace preceded '#'. I have given up and
3904 allowed the remaining 2 cases to be at the margin.
3905
3906 + Rename some variables in pcre to eliminate shadowing. This seems very
3907 pedantic, but does no harm, of course.
3908
3909 6. Moved the call to setjmp() into its own function, to get rid of warnings
3910 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3911
3912 7. Constructs such as \d{8,} were compiling into the equivalent of
3913 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3914 outcome, but in this particular case used more store than had been allocated,
3915 which caused the bug to be discovered because it threw up an internal error.
3916
3917 8. The debugging code in both pcre and pcretest for outputting the compiled
3918 form of a regex was going wrong in the case of back references followed by
3919 curly-bracketed repeats.
3920
3921
3922 Version 1.02 12-Dec-97
3923 ----------------------
3924
3925 1. Typos in pcre.3 and comments in the source fixed.
3926
3927 2. Applied a contributed patch to get rid of places where it used to remove
3928 'const' from variables, and fixed some signed/unsigned and uninitialized
3929 variable warnings.
3930
3931 3. Added the "runtest" target to Makefile.
3932
3933 4. Set default compiler flag to -O2 rather than just -O.
3934
3935
3936 Version 1.01 19-Nov-97
3937 ----------------------
3938
3939 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3940 like /([ab]*)*/, that is, for classes with more than one character in them.
3941
3942 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3943 as /((?>a*))*/ (a PCRE_EXTRA facility).
3944
3945
3946 Version 1.00 18-Nov-97
3947 ----------------------
3948
3949 1. Added compile-time macros to support systems such as SunOS4 which don't have
3950 memmove() or strerror() but have other things that can be used instead.
3951
3952 2. Arranged that "make clean" removes the executables.
3953
3954
3955 Version 0.99 27-Oct-97
3956 ----------------------
3957
3958 1. Fixed bug in code for optimizing classes with only one character. It was
3959 initializing a 32-byte map regardless, which could cause it to run off the end
3960 of the memory it had got.
3961
3962 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3963
3964
3965 Version 0.98 22-Oct-97
3966 ----------------------
3967
3968 1. Fixed bug in code for handling temporary memory usage when there are more
3969 back references than supplied space in the ovector. This could cause segfaults.
3970
3971
3972 Version 0.97 21-Oct-97
3973 ----------------------
3974
3975 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3976
3977 2. Optimized negated single characters not to use a bit map.
3978
3979 3. Brought error texts together as macro definitions; clarified some of them;
3980 fixed one that was wrong - it said "range out of order" when it meant "invalid
3981 escape sequence".
3982
3983 4. Changed some char * arguments to const char *.
3984
3985 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3986
3987 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3988 pcretest.
3989
3990
3991 Version 0.96 16-Oct-97
3992 ----------------------
3993
3994 1. Added a simple "pgrep" utility to the distribution.
3995
3996 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3997 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3998 where "ddd" means "one or more decimal digits".
3999
4000 3. Fixed serious bug. If a pattern had a back reference, but the call to
4001 pcre_exec() didn't supply a large enough ovector to record the related
4002 identifying subpattern, the match always failed. PCRE now remembers the number
4003 of the largest back reference, and gets some temporary memory in which to save
4004 the offsets during matching if necessary, in order to ensure that
4005 backreferences always work.
4006
4007 4. Increased the compatibility with Perl in a number of ways:
4008
4009 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4010 to request this handling. The option can be set at compile or exec time.
4011
4012 (b) $ matches before a terminating newline by default; an option
4013 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4014 mode). The option can be set at compile or exec time.
4015
4016 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4017 the same as Perl's. If the decimal number it represents is less than 10
4018 or there aren't that many previous left capturing parentheses, an octal
4019 escape is read. Inside a character class, it's always an octal escape,
4020 even if it is a single digit.
4021
4022 (d) An escaped but undefined alphabetic character is taken as a literal,
4023 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4024 escapes.
4025
4026 (e) {0} is now permitted. (The previous item is removed from the compiled
4027 pattern).
4028
4029 5. Changed all the names of code files so that the basic parts are no longer
4030 than 10 characters, and abolished the teeny "globals.c" file.
4031
4032 6. Changed the handling of character classes; they are now done with a 32-byte
4033 bit map always.
4034
4035 7. Added the -d and /D options to pcretest to make it possible to look at the
4036 internals of compilation without having to recompile pcre.
4037
4038
4039 Version 0.95 23-Sep-97
4040 ----------------------
4041
4042 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4043 \x20 at the start of a run of normal characters. These were being treated as
4044 real characters, instead of the source characters being re-checked.
4045
4046
4047 Version 0.94 18-Sep-97
4048 ----------------------
4049
4050 1. The functions are now thread-safe, with the caveat that the global variables
4051 containing pointers to malloc() and free() or alternative functions are the
4052 same for all threads.
4053
4054 2. Get pcre_study() to generate a bitmap of initial characters for non-
4055 anchored patterns when this is possible, and use it if passed to pcre_exec().
4056
4057
4058 Version 0.93 15-Sep-97
4059 ----------------------
4060
4061 1. /(b)|(:+)/ was computing an incorrect first character.
4062
4063 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4064 but not actually doing anything yet.
4065
4066 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4067 as Perl does (e.g. [-az] or [az-]).
4068
4069 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4070 all possible positions.
4071
4072 5. Split up into different modules to avoid including unneeded functions in a
4073 compiled binary. However, compile and exec are still in one module. The "study"
4074 function is split off.
4075
4076 6. The character tables are now in a separate module whose source is generated
4077 by an auxiliary program - but can then be edited by hand if required. There are
4078 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4079 toupper() in the code.
4080
4081 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4082 make them global. Abolish the function for setting them, as the caller can now
4083 set them directly.
4084
4085
4086 Version 0.92 11-Sep-97
4087 ----------------------
4088
4089 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4090 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4091
4092 2. Caseless matching was not working in character classes if the characters in
4093 the pattern were in upper case.
4094
4095 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4096
4097 4. Make PCRE_ANCHORED public and accept as a compile option.
4098
4099 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4100 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4101 pass them.
4102
4103 6. Give an error if bad option bits passed at compile or run time.
4104
4105 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4106 pcretest to cause it to pass that flag.
4107
4108 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4109 options, and the first character, if set.
4110
4111 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4112
4113
4114 Version 0.91 10-Sep-97
4115 ----------------------
4116
4117 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4118 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4119
4120 2. PCRE was looping on encountering an indefinitely repeated back reference to
4121 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4122 Perl does - treats the match as successful.
4123
4124 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12