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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 594 - (show annotations) (download)
Sun May 1 08:22:12 2011 UTC (2 years, 11 months ago) by ph10
File size: 198682 byte(s)
Added #include <string.h> to some .cc files (needed in some environments).

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.13 30-Apr-2011
5 ------------------------
6
7 1. The Unicode data tables have been updated to Unicode 6.0.0.
8
9 2. Two minor typos in pcre_internal.h have been fixed.
10
11 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
12 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
13 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
14
15
16 Version 8.12 15-Jan-2011
17 ------------------------
18
19 1. Fixed some typos in the markup of the man pages, and wrote a script that
20 checks for such things as part of the documentation building process.
21
22 2. On a big-endian 64-bit system, pcregrep did not correctly process the
23 --match-limit and --recursion-limit options (added for 8.11). In
24 particular, this made one of the standard tests fail. (The integer value
25 went into the wrong half of a long int.)
26
27 3. If the --colour option was given to pcregrep with -v (invert match), it
28 did strange things, either producing crazy output, or crashing. It should,
29 of course, ignore a request for colour when reporting lines that do not
30 match.
31
32 4. Another pcregrep bug caused similar problems if --colour was specified with
33 -M (multiline) and the pattern match finished with a line ending.
34
35 5. In pcregrep, when a pattern that ended with a literal newline sequence was
36 matched in multiline mode, the following line was shown as part of the
37 match. This seems wrong, so I have changed it.
38
39 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
40 the check for further matches in the same line (so they could be coloured)
41 to overrun the end of the current line. If another match was found, it was
42 incorrectly shown (and then shown again when found in the next line).
43
44 7. If pcregrep was compiled under Windows, there was a reference to the
45 function pcregrep_exit() before it was defined. I am assuming this was
46 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
47 reported by a user. I've moved the definition above the reference.
48
49
50 Version 8.11 10-Dec-2010
51 ------------------------
52
53 1. (*THEN) was not working properly if there were untried alternatives prior
54 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
55 backtracked to try for "b" instead of moving to the next alternative branch
56 at the same level (in this case, to look for "c"). The Perl documentation
57 is clear that when (*THEN) is backtracked onto, it goes to the "next
58 alternative in the innermost enclosing group".
59
60 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
61 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
62 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
63 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
64 (*THEN).
65
66 3. If \s appeared in a character class, it removed the VT character from
67 the class, even if it had been included by some previous item, for example
68 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
69 of \s, but is part of the POSIX "space" class.)
70
71 4. A partial match never returns an empty string (because you can always
72 match an empty string at the end of the subject); however the checking for
73 an empty string was starting at the "start of match" point. This has been
74 changed to the "earliest inspected character" point, because the returned
75 data for a partial match starts at this character. This means that, for
76 example, /(?<=abc)def/ gives a partial match for the subject "abc"
77 (previously it gave "no match").
78
79 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
80 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
81 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
82 has an implication that the given string is incomplete (because a partial
83 match is preferred over a full match). For this reason, these items now
84 give a partial match in this situation. [Aside: previously, the one case
85 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
86 match rather than a full match, which was wrong by the old rules, but is
87 now correct.]
88
89 6. There was a bug in the handling of #-introduced comments, recognized when
90 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
91 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
92 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
93 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
94 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
95 places in pcre_compile().
96
97 7. Related to (6) above, when pcre_compile() was skipping #-introduced
98 comments when looking ahead for named forward references to subpatterns,
99 the only newline sequence it recognized was NL. It now handles newlines
100 according to the set newline convention.
101
102 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
103 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
104 cater for a lack of strerror(). These oversights have been fixed.
105
106 9. Added --match-limit and --recursion-limit to pcregrep.
107
108 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
109
110 11. When the -o option was used, pcregrep was setting a return code of 1, even
111 when matches were found, and --line-buffered was not being honoured.
112
113 12. Added an optional parentheses number to the -o and --only-matching options
114 of pcregrep.
115
116 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
117 can match an empty string. The code to do it in pcretest and pcredemo
118 needed fixing:
119
120 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
121 only one byte after an empty string match just before CRLF (this case
122 just got forgotten; "any" and "anycrlf" were OK).
123
124 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
125 mode when an empty string match preceded an ASCII character followed by
126 a non-ASCII character. (The code for advancing by one character rather
127 than one byte was nonsense.)
128
129 (c) The pcredemo.c sample program did not have any code at all to handle
130 the cases when CRLF is a valid newline sequence.
131
132 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
133 as a starting offset was within the subject string. There is now a new
134 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
135 negative or greater than the length of the string. In order to test this,
136 pcretest is extended to allow the setting of negative starting offsets.
137
138 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
139 starting offset points to the beginning of a UTF-8 character was
140 unnecessarily clumsy. I tidied it up.
141
142 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
143 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
144
145 17. Nobody had reported that the --include_dir option, which was added in
146 release 7.7 should have been called --include-dir (hyphen, not underscore)
147 for compatibility with GNU grep. I have changed it to --include-dir, but
148 left --include_dir as an undocumented synonym, and the same for
149 --exclude-dir, though that is not available in GNU grep, at least as of
150 release 2.5.4.
151
152 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
153 characters from a string of bytes) have been redefined so as not to use
154 loops, in order to improve performance in some environments. At the same
155 time, I abstracted some of the common code into auxiliary macros to save
156 repetition (this should not affect the compiled code).
157
158 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
159 compile-time error is now given if \c is not followed by an ASCII
160 character, that is, a byte less than 128. (In EBCDIC mode, the code is
161 different, and any byte value is allowed.)
162
163 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
164 START_OPTIMIZE option, which is now allowed at compile time - but just
165 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
166 to pcregrep and other applications that have no direct access to PCRE
167 options. The new /Y option in pcretest sets this option when calling
168 pcre_compile().
169
170 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
171 back references. Groups containing recursive back references were forced to
172 be atomic by that change, but in the case of named groups, the amount of
173 memory required was incorrectly computed, leading to "Failed: internal
174 error: code overflow". This has been fixed.
175
176 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
177 pcretest.c, to avoid build problems in some Borland environments.
178
179
180 Version 8.10 25-Jun-2010
181 ------------------------
182
183 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
184 THEN.
185
186 2. (*ACCEPT) was not working when inside an atomic group.
187
188 3. Inside a character class, \B is treated as a literal by default, but
189 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
190 causes the error). The code is unchanged, but I tidied the documentation.
191
192 4. Inside a character class, PCRE always treated \R and \X as literals,
193 whereas Perl faults them if its -w option is set. I have changed PCRE so
194 that it faults them when PCRE_EXTRA is set.
195
196 5. Added support for \N, which always matches any character other than
197 newline. (It is the same as "." when PCRE_DOTALL is not set.)
198
199 6. When compiling pcregrep with newer versions of gcc which may have
200 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
201 declared with attribute warn_unused_result" were given. Just casting the
202 result to (void) does not stop the warnings; a more elaborate fudge is
203 needed. I've used a macro to implement this.
204
205 7. Minor change to pcretest.c to avoid a compiler warning.
206
207 8. Added four artifical Unicode properties to help with an option to make
208 \s etc use properties (see next item). The new properties are: Xan
209 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
210
211 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
212 use Unicode properties. (*UCP) at the start of a pattern can be used to set
213 this option. Modified pcretest to add /W to test this facility. Added
214 REG_UCP to make it available via the POSIX interface.
215
216 10. Added --line-buffered to pcregrep.
217
218 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
219 studied, and the match started with a letter with a code point greater than
220 127 whose first byte was different to the first byte of the other case of
221 the letter, the other case of this starting letter was not recognized
222 (#976).
223
224 12. If a pattern that was studied started with a repeated Unicode property
225 test, for example, \p{Nd}+, there was the theoretical possibility of
226 setting up an incorrect bitmap of starting bytes, but fortunately it could
227 not have actually happened in practice until change 8 above was made (it
228 added property types that matched character-matching opcodes).
229
230 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
231 possible starting bytes for non-anchored patterns.
232
233 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
234 \R, and also a number of cases that involve Unicode properties, both
235 explicit and implicit when PCRE_UCP is set.
236
237 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
238 input, it could crash or give wrong results if characters with values
239 greater than 0xc0 were present in the subject string. (Detail: it assumed
240 UTF-8 input when processing these items.)
241
242 16. Added a lot of (int) casts to avoid compiler warnings in systems where
243 size_t is 64-bit (#991).
244
245 17. Added a check for running out of memory when PCRE is compiled with
246 --disable-stack-for-recursion (#990).
247
248 18. If the last data line in a file for pcretest does not have a newline on
249 the end, a newline was missing in the output.
250
251 19. The default pcre_chartables.c file recognizes only ASCII characters (values
252 less than 128) in its various bitmaps. However, there is a facility for
253 generating tables according to the current locale when PCRE is compiled. It
254 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
255 characters, are recognized by isspace() and therefore were getting set in
256 these tables, and indeed these tables seem to approximate to ISO 8859. This
257 caused a problem in UTF-8 mode when pcre_study() was used to create a list
258 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
259 which of course cannot start UTF-8 characters. I have changed the code so
260 that only real ASCII characters (less than 128) and the correct starting
261 bytes for UTF-8 encodings are set for characters greater than 127 when in
262 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
263 altogether.)
264
265 20. Added the /T option to pcretest so as to be able to run tests with non-
266 standard character tables, thus making it possible to include the tests
267 used for 19 above in the standard set of tests.
268
269 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
270 reference to a subpattern the other side of a comment that contains an
271 opening parenthesis caused either an internal compiling error, or a
272 reference to the wrong subpattern.
273
274
275 Version 8.02 19-Mar-2010
276 ------------------------
277
278 1. The Unicode data tables have been updated to Unicode 5.2.0.
279
280 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
281 configured.
282
283 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
284 original author of that file, following a query about its status.
285
286 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
287 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
288
289 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
290 quantifier applied to a forward-referencing subroutine call, could compile
291 incorrect code or give the error "internal error: previously-checked
292 referenced subpattern not found".
293
294 6. Both MS Visual Studio and Symbian OS have problems with initializing
295 variables to point to external functions. For these systems, therefore,
296 pcre_malloc etc. are now initialized to local functions that call the
297 relevant global functions.
298
299 7. There were two entries missing in the vectors called coptable and poptable
300 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
301 I've fixed the data, and added a kludgy way of testing at compile time that
302 the lengths are correct (equal to the number of opcodes).
303
304 8. Following on from 7, I added a similar kludge to check the length of the
305 eint vector in pcreposix.c.
306
307 9. Error texts for pcre_compile() are held as one long string to avoid too
308 much relocation at load time. To find a text, the string is searched,
309 counting zeros. There was no check for running off the end of the string,
310 which could happen if a new error number was added without updating the
311 string.
312
313 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
314
315 11. \K was not working if it appeared in an atomic group or in a group that
316 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
317 \K is "not well defined" if used in an assertion. PCRE now accepts it if
318 the assertion is positive, but not if it is negative.
319
320 12. Change 11 fortuitously reduced the size of the stack frame used in the
321 "match()" function of pcre_exec.c by one pointer. Forthcoming
322 implementation of support for (*MARK) will need an extra pointer on the
323 stack; I have reserved it now, so that the stack frame size does not
324 decrease.
325
326 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
327 item in branch that calls a recursion is a subroutine call - as in the
328 second branch in the above example - was incorrectly given the compile-
329 time error "recursive call could loop indefinitely" because pcre_compile()
330 was not correctly checking the subroutine for matching a non-empty string.
331
332 14. The checks for overrunning compiling workspace could trigger after an
333 overrun had occurred. This is a "should never occur" error, but it can be
334 triggered by pathological patterns such as hundreds of nested parentheses.
335 The checks now trigger 100 bytes before the end of the workspace.
336
337 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
338
339
340 Version 8.01 19-Jan-2010
341 ------------------------
342
343 1. If a pattern contained a conditional subpattern with only one branch (in
344 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
345 computed the wrong minimum data length (which is of course zero for such
346 subpatterns). This could cause incorrect "no match" results.
347
348 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
349 the pattern is reset in the first branch, pcre_compile() failed with
350 "internal error: code overflow at offset...". This happened only when
351 the reset was to the original external option setting. (An optimization
352 abstracts leading options settings into an external setting, which was the
353 cause of this.)
354
355 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
356 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
357 assertion pattern did not match (meaning that the assertion was true), it
358 was incorrectly treated as false if the SKIP had been reached during the
359 matching. This also applied to assertions used as conditions.
360
361 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
362 assertion subpattern, including such a pattern used as a condition,
363 unpredictable results occurred, instead of the error return
364 PCRE_ERROR_DFA_UITEM.
365
366 5. The C++ GlobalReplace function was not working like Perl for the special
367 situation when an empty string is matched. It now does the fancy magic
368 stuff that is necessary.
369
370 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
371 removed. (These were left over from very, very early versions of PCRE.)
372
373 7. Some cosmetic changes to the code to make life easier when compiling it
374 as part of something else:
375
376 (a) Change DEBUG to PCRE_DEBUG.
377
378 (b) In pcre_compile(), rename the member of the "branch_chain" structure
379 called "current" as "current_branch", to prevent a collision with the
380 Linux macro when compiled as a kernel module.
381
382 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
383 prevent a collision with the Linux macro when compiled as a kernel
384 module.
385
386 8. In pcre_compile() there are some checks for integer overflows that used to
387 cast potentially large values to (double). This has been changed to that
388 when building, a check for int64_t is made, and if it is found, it is used
389 instead, thus avoiding the use of floating point arithmetic. (There is no
390 other use of FP in PCRE.) If int64_t is not found, the fallback is to
391 double.
392
393 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
394 2005 (difference between two addresses compared to an unsigned value).
395
396 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
397 custom one, because of the following reported problem in Windows:
398
399 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
400 under Win32.
401 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
402 therefore missing the function definition.
403 - The compiler thus generates a "C" signature for the test function.
404 - The linker fails to find the "C" function.
405 - PCRE fails to configure if asked to do so against libbz2.
406
407 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
408 messages were output:
409
410 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
411 rerunning libtoolize, to keep the correct libtool macros in-tree.
412 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
413
414 I have done both of these things.
415
416 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
417 most of the time, it *can* run out if it is given a pattern that contains a
418 runaway infinite recursion. I updated the discussion in the pcrestack man
419 page.
420
421 13. Now that we have gone to the x.xx style of version numbers, the minor
422 version may start with zero. Using 08 or 09 is a bad idea because users
423 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
424 interpreted as invalid octal numbers. I've updated the previous comment in
425 configure.ac, and also added a check that gives an error if 08 or 09 are
426 used.
427
428 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
429 causing partial matching to fail when the end of the subject matched \W
430 in a UTF-8 pattern where \W was quantified with a minimum of 3.
431
432 15. There were some discrepancies between the declarations in pcre_internal.h
433 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
434 their definitions. The declarations used "const uschar *" and the
435 definitions used USPTR. Even though USPTR is normally defined as "const
436 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
437 reported that: "This difference in casting confuses some C++ compilers, for
438 example, SunCC recognizes above declarations as different functions and
439 generates broken code for hbpcre." I have changed the declarations to use
440 USPTR.
441
442 16. GNU libtool is named differently on some systems. The autogen.sh script now
443 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
444 (FreeBSD).
445
446 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
447 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
448 comment: "Figure out how to create a longlong from a string: strtoll and
449 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
450 instance, but it only takes 2 args instead of 3!"
451
452 18. A subtle bug concerned with back references has been fixed by a change of
453 specification, with a corresponding code fix. A pattern such as
454 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
455 refers, was giving matches when it shouldn't. For example, xa=xaaa would
456 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
457 same bug. Such groups have to be quantified to be useful, or contained
458 inside another quantified group. (If there's no repetition, the reference
459 can never match.) The problem arises because, having left the group and
460 moved on to the rest of the pattern, a later failure that backtracks into
461 the group uses the captured value from the final iteration of the group
462 rather than the correct earlier one. I have fixed this in PCRE by forcing
463 any group that contains a reference to itself to be an atomic group; that
464 is, there cannot be any backtracking into it once it has completed. This is
465 similar to recursive and subroutine calls.
466
467
468 Version 8.00 19-Oct-09
469 ----------------------
470
471 1. The table for translating pcre_compile() error codes into POSIX error codes
472 was out-of-date, and there was no check on the pcre_compile() error code
473 being within the table. This could lead to an OK return being given in
474 error.
475
476 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
477 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
478 in a Windows environment.
479
480 3. The pcregrep --count option prints the count for each file even when it is
481 zero, as does GNU grep. However, pcregrep was also printing all files when
482 --files-with-matches was added. Now, when both options are given, it prints
483 counts only for those files that have at least one match. (GNU grep just
484 prints the file name in this circumstance, but including the count seems
485 more useful - otherwise, why use --count?) Also ensured that the
486 combination -clh just lists non-zero counts, with no names.
487
488 4. The long form of the pcregrep -F option was incorrectly implemented as
489 --fixed_strings instead of --fixed-strings. This is an incompatible change,
490 but it seems right to fix it, and I didn't think it was worth preserving
491 the old behaviour.
492
493 5. The command line items --regex=pattern and --regexp=pattern were not
494 recognized by pcregrep, which required --regex pattern or --regexp pattern
495 (with a space rather than an '='). The man page documented the '=' forms,
496 which are compatible with GNU grep; these now work.
497
498 6. No libpcreposix.pc file was created for pkg-config; there was just
499 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
500
501 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
502 when UCP support is not needed, by modifying the Python script that
503 generates it from Unicode data files. This should not matter if the module
504 is correctly used as a library, but I received one complaint about 50K of
505 unwanted data. My guess is that the person linked everything into his
506 program rather than using a library. Anyway, it does no harm.
507
508 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
509 was a minimum greater than 1 for a wide character in a possessive
510 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
511 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
512 character. Chaos in the form of incorrect output or a compiling loop could
513 result.
514
515 9. The restrictions on what a pattern can contain when partial matching is
516 requested for pcre_exec() have been removed. All patterns can now be
517 partially matched by this function. In addition, if there are at least two
518 slots in the offset vector, the offset of the earliest inspected character
519 for the match and the offset of the end of the subject are set in them when
520 PCRE_ERROR_PARTIAL is returned.
521
522 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
523 synonymous with PCRE_PARTIAL, for backwards compatibility, and
524 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
525 and may be more useful for multi-segment matching.
526
527 11. Partial matching with pcre_exec() is now more intuitive. A partial match
528 used to be given if ever the end of the subject was reached; now it is
529 given only if matching could not proceed because another character was
530 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
531 string "Z", which now yields "no match" instead of "partial match". In the
532 case of pcre_dfa_exec(), "no match" is given if every matching path for the
533 final character ended with (*FAIL).
534
535 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
536 if the pattern had a "must contain" character that was already found in the
537 earlier partial match, unless partial matching was again requested. For
538 example, with the pattern /dog.(body)?/, the "must contain" character is
539 "g". If the first part-match was for the string "dog", restarting with
540 "sbody" failed. This bug has been fixed.
541
542 13. The string returned by pcre_dfa_exec() after a partial match has been
543 changed so that it starts at the first inspected character rather than the
544 first character of the match. This makes a difference only if the pattern
545 starts with a lookbehind assertion or \b or \B (\K is not supported by
546 pcre_dfa_exec()). It's an incompatible change, but it makes the two
547 matching functions compatible, and I think it's the right thing to do.
548
549 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
550 so that the demonstration program is easily available in environments where
551 PCRE has not been installed from source.
552
553 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
554 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
555 library.
556
557 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
558 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
559 is not the first non-POSIX option to be added. Clearly some people find
560 these options useful.
561
562 17. If a caller to the POSIX matching function regexec() passes a non-zero
563 value for nmatch with a NULL value for pmatch, the value of
564 nmatch is forced to zero.
565
566 18. RunGrepTest did not have a test for the availability of the -u option of
567 the diff command, as RunTest does. It now checks in the same way as
568 RunTest, and also checks for the -b option.
569
570 19. If an odd number of negated classes containing just a single character
571 interposed, within parentheses, between a forward reference to a named
572 subpattern and the definition of the subpattern, compilation crashed with
573 an internal error, complaining that it could not find the referenced
574 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
575 [The bug was that it was starting one character too far in when skipping
576 over the character class, thus treating the ] as data rather than
577 terminating the class. This meant it could skip too much.]
578
579 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
580 /g option in pcretest when the pattern contains \K, which makes it possible
581 to have an empty string match not at the start, even when the pattern is
582 anchored. Updated pcretest and pcredemo to use this option.
583
584 21. If the maximum number of capturing subpatterns in a recursion was greater
585 than the maximum at the outer level, the higher number was returned, but
586 with unset values at the outer level. The correct (outer level) value is
587 now given.
588
589 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
590 PCRE did not set those parentheses (unlike Perl). I have now found a way to
591 make it do so. The string so far is captured, making this feature
592 compatible with Perl.
593
594 23. The tests have been re-organized, adding tests 11 and 12, to make it
595 possible to check the Perl 5.10 features against Perl 5.10.
596
597 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
598 pattern matches a fixed length string. PCRE did not allow this; now it
599 does. Neither allows recursion.
600
601 25. I finally figured out how to implement a request to provide the minimum
602 length of subject string that was needed in order to match a given pattern.
603 (It was back references and recursion that I had previously got hung up
604 on.) This code has now been added to pcre_study(); it finds a lower bound
605 to the length of subject needed. It is not necessarily the greatest lower
606 bound, but using it to avoid searching strings that are too short does give
607 some useful speed-ups. The value is available to calling programs via
608 pcre_fullinfo().
609
610 26. While implementing 25, I discovered to my embarrassment that pcretest had
611 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
612 study optimizations had never been tested with that matching function.
613 Oops. What is worse, even when it was passed study data, there was a bug in
614 pcre_dfa_exec() that meant it never actually used it. Double oops. There
615 were also very few tests of studied patterns with pcre_dfa_exec().
616
617 27. If (?| is used to create subpatterns with duplicate numbers, they are now
618 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
619 on the other side of the coin, they are no longer allowed to have different
620 names, because these cannot be distinguished in PCRE, and this has caused
621 confusion. (This is a difference from Perl.)
622
623 28. When duplicate subpattern names are present (necessarily with different
624 numbers, as required by 27 above), and a test is made by name in a
625 conditional pattern, either for a subpattern having been matched, or for
626 recursion in such a pattern, all the associated numbered subpatterns are
627 tested, and the overall condition is true if the condition is true for any
628 one of them. This is the way Perl works, and is also more like the way
629 testing by number works.
630
631
632 Version 7.9 11-Apr-09
633 ---------------------
634
635 1. When building with support for bzlib/zlib (pcregrep) and/or readline
636 (pcretest), all targets were linked against these libraries. This included
637 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
638 libraries. This caused unwanted dependencies to be created. This problem
639 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
640 pcretest is linked with readline.
641
642 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
643 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
644 moved outside it again, because FALSE and TRUE are already defined in AIX,
645 but BOOL is not.
646
647 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
648 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
649
650 4. The pcregrep documentation said spaces were inserted as well as colons (or
651 hyphens) following file names and line numbers when outputting matching
652 lines. This is not true; no spaces are inserted. I have also clarified the
653 wording for the --colour (or --color) option.
654
655 5. In pcregrep, when --colour was used with -o, the list of matching strings
656 was not coloured; this is different to GNU grep, so I have changed it to be
657 the same.
658
659 6. When --colo(u)r was used in pcregrep, only the first matching substring in
660 each matching line was coloured. Now it goes on to look for further matches
661 of any of the test patterns, which is the same behaviour as GNU grep.
662
663 7. A pattern that could match an empty string could cause pcregrep to loop; it
664 doesn't make sense to accept an empty string match in pcregrep, so I have
665 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
666 seems to be how GNU grep behaves.
667
668 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
669 start or after a newline", because the conditional assertion was not being
670 correctly handled. The rule now is that both the assertion and what follows
671 in the first alternative must satisfy the test.
672
673 9. If auto-callout was enabled in a pattern with a conditional group whose
674 condition was an assertion, PCRE could crash during matching, both with
675 pcre_exec() and pcre_dfa_exec().
676
677 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
678 used for matching.
679
680 11. Unicode property support in character classes was not working for
681 characters (bytes) greater than 127 when not in UTF-8 mode.
682
683 12. Added the -M command line option to pcretest.
684
685 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
686
687 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
688
689 16. Added comments and documentation about mis-use of no_arg in the C++
690 wrapper.
691
692 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
693 from Martin Jerabek that uses macro names for all relevant character and
694 string constants.
695
696 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
697 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
698 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
699 these, but not everybody uses configure.
700
701 19. A conditional group that had only one branch was not being correctly
702 recognized as an item that could match an empty string. This meant that an
703 enclosing group might also not be so recognized, causing infinite looping
704 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
705 with the subject "ab", where knowledge that the repeated group can match
706 nothing is needed in order to break the loop.
707
708 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
709 exec(), but without supplying a callout function, matching went wrong.
710
711 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
712 leak if the size of the offset vector was greater than 30. When the vector
713 is smaller, the saved offsets during recursion go onto a local stack
714 vector, but for larger vectors malloc() is used. It was failing to free
715 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
716 error, in fact).
717
718 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
719 heapframe that is used only when UTF-8 support is enabled. This caused no
720 problem, but was untidy.
721
722 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
723 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
724 included within another project.
725
726 24. Steven Van Ingelgem's patches to add more options to the CMake support,
727 slightly modified by me:
728
729 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
730 not building pcregrep.
731
732 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
733 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
734
735 25. Forward references, both numeric and by name, in patterns that made use of
736 duplicate group numbers, could behave incorrectly or give incorrect errors,
737 because when scanning forward to find the reference group, PCRE was not
738 taking into account the duplicate group numbers. A pattern such as
739 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
740
741 26. Changed a few more instances of "const unsigned char *" to USPTR, making
742 the feature of a custom pointer more persuasive (as requested by a user).
743
744 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
745 pcretest.c, inside #ifndefs, because it seems they are sometimes already
746 pre-defined.
747
748 28. Added support for (*UTF8) at the start of a pattern.
749
750 29. Arrange for flags added by the "release type" setting in CMake to be shown
751 in the configuration summary.
752
753
754 Version 7.8 05-Sep-08
755 ---------------------
756
757 1. Replaced UCP searching code with optimized version as implemented for Ad
758 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
759 stage table and inline lookup instead of a function, giving speed ups of 2
760 to 5 times on some simple patterns that I tested. Permission was given to
761 distribute the MultiStage2.py script that generates the tables (it's not in
762 the tarball, but is in the Subversion repository).
763
764 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
765 scripts.
766
767 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
768 a group with a zero qualifier. The result of the study could be incorrect,
769 or the function might crash, depending on the pattern.
770
771 4. Caseless matching was not working for non-ASCII characters in back
772 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
773 It now works when Unicode Property Support is available.
774
775 5. In pcretest, an escape such as \x{de} in the data was always generating
776 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
777 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
778 truncation.
779
780 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
781
782 7. Added two (int) casts to pcregrep when printing the difference of two
783 pointers, in case they are 64-bit values.
784
785 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
786 test 2 if it fails.
787
788 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
789 and a #define of that name to empty if it is not externally set. This is to
790 allow users of MSVC to set it if necessary.
791
792 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
793 the convenience functions in the pcre_get.c source file.
794
795 11. An option change at the start of a pattern that had top-level alternatives
796 could cause overwriting and/or a crash. This command provoked a crash in
797 some environments:
798
799 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
800
801 This potential security problem was recorded as CVE-2008-2371.
802
803 12. For a pattern where the match had to start at the beginning or immediately
804 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
805 pcre_dfa_exec() could read past the end of the passed subject if there was
806 no match. To help with detecting such bugs (e.g. with valgrind), I modified
807 pcretest so that it places the subject at the end of its malloc-ed buffer.
808
809 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
810 exec() might read past the end of the data buffer in UTF-8 mode.
811
812 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
813 the data contained the byte 0x85 as part of a UTF-8 character within its
814 first line. This applied both to normal and DFA matching.
815
816 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
817 /^[^d]*?$/8 failed to match "abc".
818
819 16. Added a missing copyright notice to pcrecpp_internal.h.
820
821 17. Make it more clear in the documentation that values returned from
822 pcre_exec() in ovector are byte offsets, not character counts.
823
824 18. Tidied a few places to stop certain compilers from issuing warnings.
825
826 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
827 supplied by Stefan Weber. I made a further small update for 7.8 because
828 there is a change of source arrangements: the pcre_searchfuncs.c module is
829 replaced by pcre_ucd.c.
830
831
832 Version 7.7 07-May-08
833 ---------------------
834
835 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
836 a string to a long long, pretend we don't even have a long long." This is
837 done by checking for the strtoq, strtoll, and _strtoi64 functions.
838
839 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
840 pre-7.6 versions, which defined a global no_arg variable instead of putting
841 it in the RE class. (See also #8 below.)
842
843 3. Remove a line of dead code, identified by coverity and reported by Nuno
844 Lopes.
845
846 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
847
848 (1) The include/exclude patterns were being applied to the whole pathnames
849 of files, instead of just to the final components.
850
851 (2) If there was more than one level of directory, the subdirectories were
852 skipped unless they satisfied the include/exclude conditions. This is
853 inconsistent with GNU grep (and could even be seen as contrary to the
854 pcregrep specification - which I improved to make it absolutely clear).
855 The action now is always to scan all levels of directory, and just
856 apply the include/exclude patterns to regular files.
857
858 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
859 --exclude_dir in the tests to avoid scanning .svn directories.
860
861 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
862 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
863 doesn't support NULs in patterns.
864
865 7. Added some missing "const"s to declarations of static tables in
866 pcre_compile.c and pcre_dfa_exec.c.
867
868 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
869 caused by fix #2 above. (Subsequently also a second patch to fix the
870 first patch. And a third patch - this was a messy problem.)
871
872 9. Applied Craig's patch to remove the use of push_back().
873
874 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
875 matching function regexec().
876
877 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
878 which, however, unlike Perl's \g{...}, are subroutine calls, not back
879 references. PCRE supports relative numbers with this syntax (I don't think
880 Oniguruma does).
881
882 12. Previously, a group with a zero repeat such as (...){0} was completely
883 omitted from the compiled regex. However, this means that if the group
884 was called as a subroutine from elsewhere in the pattern, things went wrong
885 (an internal error was given). Such groups are now left in the compiled
886 pattern, with a new opcode that causes them to be skipped at execution
887 time.
888
889 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
890 to the way PCRE behaves:
891
892 (a) A lone ] character is dis-allowed (Perl treats it as data).
893
894 (b) A back reference to an unmatched subpattern matches an empty string
895 (Perl fails the current match path).
896
897 (c) A data ] in a character class must be notated as \] because if the
898 first data character in a class is ], it defines an empty class. (In
899 Perl it is not possible to have an empty class.) The empty class []
900 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
901 The negative empty class [^] matches any one character, independently
902 of the DOTALL setting.
903
904 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
905 non-existent subpattern following a character class starting with ']' and
906 containing () gave an internal compiling error instead of "reference to
907 non-existent subpattern". Fortunately, when the pattern did exist, the
908 compiled code was correct. (When scanning forwards to check for the
909 existencd of the subpattern, it was treating the data ']' as terminating
910 the class, so got the count wrong. When actually compiling, the reference
911 was subsequently set up correctly.)
912
913 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
914 it was being rejected as not supported by pcre_dfa_exec(), even though
915 other assertions are supported. I have made pcre_dfa_exec() support
916 (*FAIL).
917
918 16. The implementation of 13c above involved the invention of a new opcode,
919 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
920 cannot be changed at match time, I realized I could make a small
921 improvement to matching performance by compiling OP_ALLANY instead of
922 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
923 on the OP_ANY path.
924
925 17. Compiling pcretest on Windows with readline support failed without the
926 following two fixes: (1) Make the unistd.h include conditional on
927 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
928
929 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
930 ncurses library to be included for pcretest when ReadLine support is
931 requested, but also to allow for it to be overridden. This patch came from
932 Daniel Bergström.
933
934 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
935 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
936 any errors with the current Unicode tables. Thanks to Peter Kankowski for
937 spotting this.
938
939
940 Version 7.6 28-Jan-08
941 ---------------------
942
943 1. A character class containing a very large number of characters with
944 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
945 overflow.
946
947 2. Patch to cut out the "long long" test in pcrecpp_unittest when
948 HAVE_LONG_LONG is not defined.
949
950 3. Applied Christian Ehrlicher's patch to update the CMake build files to
951 bring them up to date and include new features. This patch includes:
952
953 - Fixed PH's badly added libz and libbz2 support.
954 - Fixed a problem with static linking.
955 - Added pcredemo. [But later removed - see 7 below.]
956 - Fixed dftables problem and added an option.
957 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
958 HAVE_LONG_LONG.
959 - Added readline support for pcretest.
960 - Added an listing of the option settings after cmake has run.
961
962 4. A user submitted a patch to Makefile that makes it easy to create
963 "pcre.dll" under mingw when using Configure/Make. I added stuff to
964 Makefile.am that cause it to include this special target, without
965 affecting anything else. Note that the same mingw target plus all
966 the other distribution libraries and programs are now supported
967 when configuring with CMake (see 6 below) instead of with
968 Configure/Make.
969
970 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
971 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
972 exported in the Windows port". It has not yet been confirmed that the patch
973 solves the problem, but it does no harm.
974
975 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
976 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
977 with CMake, and also correct the comment about stack recursion.
978
979 7. Remove the automatic building of pcredemo from the ./configure system and
980 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
981 of a program that users should build themselves after PCRE is installed, so
982 building it automatically is not really right. What is more, it gave
983 trouble in some build environments.
984
985 8. Further tidies to CMakeLists.txt from Sheri and Christian.
986
987
988 Version 7.5 10-Jan-08
989 ---------------------
990
991 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
992 values in parens when parsing an RE using the C++ wrapper."
993
994 2. Negative specials like \S did not work in character classes in UTF-8 mode.
995 Characters greater than 255 were excluded from the class instead of being
996 included.
997
998 3. The same bug as (2) above applied to negated POSIX classes such as
999 [:^space:].
1000
1001 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1002 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1003 I have changed it.
1004
1005 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1006 first named subpattern) and a construct such as (?&a) would reference the
1007 first named subpattern whose name started with "a" (in other words, the
1008 length check was missing). Both these problems are fixed. "Subpattern name
1009 expected" is now given for (?&) (a zero-length name), and this patch also
1010 makes it give the same error for \k'' (previously it complained that that
1011 was a reference to a non-existent subpattern).
1012
1013 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1014 this is right because (?- can be followed by option settings as well as by
1015 digits. I have, however, made the messages clearer.
1016
1017 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1018 than the number used in the conditional) now cause a compile-time error.
1019 This is actually not compatible with Perl, which accepts such patterns, but
1020 treats the conditional as always being FALSE (as PCRE used to), but it
1021 seems to me that giving a diagnostic is better.
1022
1023 8. Change "alphameric" to the more common word "alphanumeric" in comments
1024 and messages.
1025
1026 9. Fix two occurrences of "backslash" in comments that should have been
1027 "backspace".
1028
1029 10. Remove two redundant lines of code that can never be obeyed (their function
1030 was moved elsewhere).
1031
1032 11. The program that makes PCRE's Unicode character property table had a bug
1033 which caused it to generate incorrect table entries for sequences of
1034 characters that have the same character type, but are in different scripts.
1035 It amalgamated them into a single range, with the script of the first of
1036 them. In other words, some characters were in the wrong script. There were
1037 thirteen such cases, affecting characters in the following ranges:
1038
1039 U+002b0 - U+002c1
1040 U+0060c - U+0060d
1041 U+0061e - U+00612
1042 U+0064b - U+0065e
1043 U+0074d - U+0076d
1044 U+01800 - U+01805
1045 U+01d00 - U+01d77
1046 U+01d9b - U+01dbf
1047 U+0200b - U+0200f
1048 U+030fc - U+030fe
1049 U+03260 - U+0327f
1050 U+0fb46 - U+0fbb1
1051 U+10450 - U+1049d
1052
1053 12. The -o option (show only the matching part of a line) for pcregrep was not
1054 compatible with GNU grep in that, if there was more than one match in a
1055 line, it showed only the first of them. It now behaves in the same way as
1056 GNU grep.
1057
1058 13. If the -o and -v options were combined for pcregrep, it printed a blank
1059 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1060 does the same. The return code can be used to tell if there were any
1061 non-matching lines.
1062
1063 14. Added --file-offsets and --line-offsets to pcregrep.
1064
1065 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1066 infinitely looping recursion. The bug was that positive lookaheads were not
1067 being skipped when checking for a possible empty match (negative lookaheads
1068 and both kinds of lookbehind were skipped).
1069
1070 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1071 inclusion of <windows.h> to before rather than after the definition of
1072 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1073
1074 17. Specifying a possessive quantifier with a specific limit for a Unicode
1075 character property caused pcre_compile() to compile bad code, which led at
1076 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1077 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1078 caused the error; without that there was no problem.
1079
1080 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1081
1082 19. Added --enable-pcretest-libreadline.
1083
1084 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1085 RE::GlobalReplace(). As a result, the number of replacements returned was
1086 double what it should be. I removed one of the increments, but Craig sent a
1087 later patch that removed the other one (the right fix) and added unit tests
1088 that check the return values (which was not done before).
1089
1090 21. Several CMake things:
1091
1092 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1093 the names libpcre and libpcreposix, not just pcre and pcreposix.
1094
1095 (2) The above change means that pcretest and pcregrep are now correctly
1096 linked with the newly-built libraries, not previously installed ones.
1097
1098 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1099
1100 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1101 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1102 UTF-8 newline character). The key issue is that the pattern starts .*;
1103 this means that the match must be either at the beginning, or after a
1104 newline. The bug was in the code for advancing after a failed match and
1105 checking that the new position followed a newline. It was not taking
1106 account of UTF-8 characters correctly.
1107
1108 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1109 character classes. PCRE was not treating the sequence [:...:] as a
1110 character class unless the ... were all letters. Perl, however, seems to
1111 allow any characters between [: and :], though of course it rejects as
1112 unknown any "names" that contain non-letters, because all the known class
1113 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1114 for example, whereas PCRE did not - it did not recognize a POSIX character
1115 class. This seemed a bit dangerous, so the code has been changed to be
1116 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1117 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1118 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1119 Perl does, and where it didn't before.
1120
1121 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1122 Windows environments %n is disabled by default.
1123
1124
1125 Version 7.4 21-Sep-07
1126 ---------------------
1127
1128 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1129 means that a class such as [\s] counted as "explicit reference to CR or
1130 LF". That isn't really right - the whole point of the change was to try to
1131 help when there was an actual mention of one of the two characters. So now
1132 the change happens only if \r or \n (or a literal CR or LF) character is
1133 encountered.
1134
1135 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1136 of both had grown to the point where there were only 3 bits left.
1137 Fortunately, there was spare space in the data structure, and so I have
1138 moved the internal flags into a new 16-bit field to free up more option
1139 bits.
1140
1141 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1142 but did not set the internal JCHANGED flag - either of these is enough to
1143 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1144 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1145 start sets both bits.
1146
1147 4. Added options (at build time, compile time, exec time) to change \R from
1148 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1149
1150 5. doc/pcresyntax.html was missing from the distribution.
1151
1152 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1153 compatibility, even though it is no longer used.
1154
1155 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1156 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1157 windows.h file is present (where different names are used). [This was
1158 reversed later after testing - see 16 below.]
1159
1160 8. Changed all #include <config.h> to #include "config.h". There were also
1161 some further <pcre.h> cases that I changed to "pcre.h".
1162
1163 9. When pcregrep was used with the --colour option, it missed the line ending
1164 sequence off the lines that it output.
1165
1166 10. It was pointed out to me that arrays of string pointers cause lots of
1167 relocations when a shared library is dynamically loaded. A technique of
1168 using a single long string with a table of offsets can drastically reduce
1169 these. I have refactored PCRE in four places to do this. The result is
1170 dramatic:
1171
1172 Originally: 290
1173 After changing UCP table: 187
1174 After changing error message table: 43
1175 After changing table of "verbs" 36
1176 After changing table of Posix names 22
1177
1178 Thanks to the folks working on Gregex for glib for this insight.
1179
1180 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1181 unicode-properties was also set.
1182
1183 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1184
1185 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1186 checked only for CRLF.
1187
1188 14. Added casts to pcretest.c to avoid compiler warnings.
1189
1190 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1191
1192 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1193 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1194 entirely. This removes changes made in 7 above.
1195
1196 17. The CMake files have been updated, and there is now more information about
1197 building with CMake in the NON-UNIX-USE document.
1198
1199
1200 Version 7.3 28-Aug-07
1201 ---------------------
1202
1203 1. In the rejigging of the build system that eventually resulted in 7.1, the
1204 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1205 brackets there is not right, since it causes compilers to look for an
1206 installed pcre.h, not the version that is in the source that is being
1207 compiled (which of course may be different). I have changed it back to:
1208
1209 #include "pcre.h"
1210
1211 I have a vague recollection that the change was concerned with compiling in
1212 different directories, but in the new build system, that is taken care of
1213 by the VPATH setting the Makefile.
1214
1215 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1216 when the subject happened to end in the byte 0x85 (e.g. if the last
1217 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1218 characters but of course it shouldn't be taken as a newline when it is part
1219 of another character. The bug was that, for an unlimited repeat of . in
1220 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1221 characters when looking for a newline.
1222
1223 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1224
1225 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1226 in debug output.
1227
1228 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1229 long printing in the pcrecpp unittest when running under MinGW.
1230
1231 6. ESC_K was left out of the EBCDIC table.
1232
1233 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1234 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1235 limit also applies to "virtual nesting" when a pattern is recursive, and in
1236 this case 1000 isn't so big. I have been able to remove this limit at the
1237 expense of backing off one optimization in certain circumstances. Normally,
1238 when pcre_exec() would call its internal match() function recursively and
1239 immediately return the result unconditionally, it uses a "tail recursion"
1240 feature to save stack. However, when a subpattern that can match an empty
1241 string has an unlimited repetition quantifier, it no longer makes this
1242 optimization. That gives it a stack frame in which to save the data for
1243 checking that an empty string has been matched. Previously this was taken
1244 from the 1000-entry workspace that had been reserved. So now there is no
1245 explicit limit, but more stack is used.
1246
1247 8. Applied Daniel's patches to solve problems with the import/export magic
1248 syntax that is required for Windows, and which was going wrong for the
1249 pcreposix and pcrecpp parts of the library. These were overlooked when this
1250 problem was solved for the main library.
1251
1252 9. There were some crude static tests to avoid integer overflow when computing
1253 the size of patterns that contain repeated groups with explicit upper
1254 limits. As the maximum quantifier is 65535, the maximum group length was
1255 set at 30,000 so that the product of these two numbers did not overflow a
1256 32-bit integer. However, it turns out that people want to use groups that
1257 are longer than 30,000 bytes (though not repeat them that many times).
1258 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1259 made it possible to implement the integer overflow checks in a much more
1260 dynamic way, which I have now done. The artificial limitation on group
1261 length has been removed - we now have only the limit on the total length of
1262 the compiled pattern, which depends on the LINK_SIZE setting.
1263
1264 10. Fixed a bug in the documentation for get/copy named substring when
1265 duplicate names are permitted. If none of the named substrings are set, the
1266 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1267 empty string.
1268
1269 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1270 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1271 because the ] is interpreted as the first data character and the
1272 terminating ] is not found. PCRE has been made compatible with Perl in this
1273 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1274 cause memory overwriting.
1275
1276 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1277 string has been matched (to stop an infinite loop). It was not recognizing
1278 a conditional subpattern that could match an empty string if that
1279 subpattern was within another subpattern. For example, it looped when
1280 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1281 condition was not nested. This bug has been fixed.
1282
1283 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1284 past the start of the subject in the presence of bytes with the top bit
1285 set, for example "\x8aBCD".
1286
1287 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1288 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1289
1290 14. Optimized (?!) to (*FAIL).
1291
1292 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1293 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1294 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1295 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1296 does: it's just the validity check that is more restrictive.
1297
1298 16. Inserted checks for integer overflows during escape sequence (backslash)
1299 processing, and also fixed erroneous offset values for syntax errors during
1300 backslash processing.
1301
1302 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1303 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1304
1305 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1306 caused an overrun.
1307
1308 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1309 something other than just ASCII characters) inside a group that had an
1310 unlimited repeat caused a loop at compile time (while checking to see
1311 whether the group could match an empty string).
1312
1313 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1314 [\P{Any}] did so. (Error in the code for printing property names.)
1315
1316 21. An orphan \E inside a character class could cause a crash.
1317
1318 22. A repeated capturing bracket such as (A)? could cause a wild memory
1319 reference during compilation.
1320
1321 23. There are several functions in pcre_compile() that scan along a compiled
1322 expression for various reasons (e.g. to see if it's fixed length for look
1323 behind). There were bugs in these functions when a repeated \p or \P was
1324 present in the pattern. These operators have additional parameters compared
1325 with \d, etc, and these were not being taken into account when moving along
1326 the compiled data. Specifically:
1327
1328 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1329 length.
1330
1331 (b) An item such as \pL+ within a repeated group could cause crashes or
1332 loops.
1333
1334 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1335 "reference to non-existent subpattern" error.
1336
1337 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1338
1339 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1340 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1341
1342 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1343
1344 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1345 character were causing crashes (broken optimization).
1346
1347 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1348 \p or \P) caused a compile-time loop.
1349
1350 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1351 break. For example, the unstudied pattern [\r\n]A does not match the string
1352 "\r\nA" because change 7.0/46 below moves the current point on by two
1353 characters after failing to match at the start. However, the pattern \nA
1354 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1355 the same is true. There doesn't seem any very clean way out of this, but
1356 what I have chosen to do makes the common cases work: PCRE now takes note
1357 of whether there can be an explicit match for \r or \n anywhere in the
1358 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1359 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1360 pattern has explicit CR or LF references.
1361
1362 29. Added (*CR) etc for changing newline setting at start of pattern.
1363
1364
1365 Version 7.2 19-Jun-07
1366 ---------------------
1367
1368 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1369 which is apparently normally available under Windows.
1370
1371 2. Re-jig the pcregrep tests with different newline settings in an attempt
1372 to make them independent of the local environment's newline setting.
1373
1374 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1375
1376 4. Some of the "internals" tests were previously cut out when the link size
1377 was not 2, because the output contained actual offsets. The recent new
1378 "Z" feature of pcretest means that these can be cut out, making the tests
1379 usable with all link sizes.
1380
1381 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1382 stack recursion. This gives a massive performance boost under BSD, but just
1383 a small improvement under Linux. However, it saves one field in the frame
1384 in all cases.
1385
1386 6. Added more features from the forthcoming Perl 5.10:
1387
1388 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1389 recursion call. It refers to the nth most recently opened parentheses.
1390
1391 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1392 to be opened parentheses.
1393
1394 (c) Conditions that refer to capturing parentheses can be specified
1395 relatively, for example, (?(-2)... or (?(+3)...
1396
1397 (d) \K resets the start of the current match so that everything before
1398 is not part of it.
1399
1400 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1401
1402 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1403 reference syntax.
1404
1405 (g) (?| introduces a group in which the numbering of parentheses in each
1406 alternative starts with the same number.
1407
1408 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1409
1410 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1411 PCRE_INFO_JCHANGED.
1412
1413 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1414 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1415 for detecting groups that can match an empty string.
1416
1417 9. A pattern with a very large number of alternatives (more than several
1418 hundred) was running out of internal workspace during the pre-compile
1419 phase, where pcre_compile() figures out how much memory will be needed. A
1420 bit of new cunning has reduced the workspace needed for groups with
1421 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1422 workspace instead of running out of the 4096 that are available.
1423
1424 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1425
1426 11. Applied patch from Google to remove an optimization that didn't quite work.
1427 The report of the bug said:
1428
1429 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1430 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1431 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1432
1433 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1434 it matched the wrong number of bytes.
1435
1436
1437 Version 7.1 24-Apr-07
1438 ---------------------
1439
1440 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1441 that is more "standard", making use of automake and other Autotools. There
1442 is some re-arrangement of the files and adjustment of comments consequent
1443 on this.
1444
1445 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1446 for recursive directory scanning broke on some systems because the files
1447 are not scanned in any specific order and on different systems the order
1448 was different. A call to "sort" has been inserted into RunGrepTest for the
1449 approprate test as a short-term fix. In the longer term there may be an
1450 alternative.
1451
1452 3. I had an email from Eric Raymond about problems translating some of PCRE's
1453 man pages to HTML (despite the fact that I distribute HTML pages, some
1454 people do their own conversions for various reasons). The problems
1455 concerned the use of low-level troff macros .br and .in. I have therefore
1456 removed all such uses from the man pages (some were redundant, some could
1457 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1458 HTML has been updated to handle .nf/.fi and to complain if it encounters
1459 .br or .in.
1460
1461 4. Updated comments in configure.ac that get placed in config.h.in and also
1462 arranged for config.h to be included in the distribution, with the name
1463 config.h.generic, for the benefit of those who have to compile without
1464 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1465
1466 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1467 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1468 makevp.bat for the current PCRE, using the additional files
1469 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1470
1471 6. A Windows user reported a minor discrepancy with test 2, which turned out
1472 to be caused by a trailing space on an input line that had got lost in his
1473 copy. The trailing space was an accident, so I've just removed it.
1474
1475 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1476 that is needed.
1477
1478 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1479 as "const" (a) because they are and (b) because it helps the PHP
1480 maintainers who have recently made a script to detect big data structures
1481 in the php code that should be moved to the .rodata section. I remembered
1482 to update Builducptable as well, so it won't revert if ucptable.h is ever
1483 re-created.
1484
1485 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1486 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1487 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1488 support is not required. This saves 1.5-2K of code, which is important in
1489 some applications.
1490
1491 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1492 so as not to refer to the tables, even though these functions will never be
1493 called when UTF-8 support is disabled. Otherwise there are problems with a
1494 shared library.
1495
1496 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1497
1498 (a) It was defining its arguments as char * instead of void *.
1499
1500 (b) It was assuming that all moves were upwards in memory; this was true
1501 a long time ago when I wrote it, but is no longer the case.
1502
1503 The emulated memove() is provided for those environments that have neither
1504 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1505 is clearly not the case, as these two bugs were recently reported.
1506
1507 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1508 and Detrail to create the HTML documentation, the .txt form of the man
1509 pages, and it removes trailing spaces from listed files. It also creates
1510 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1511 case, it wraps all the #defines with #ifndefs. This script should be run
1512 before "make dist".
1513
1514 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1515 with Unicode property support.
1516
1517 (a) For a maximizing quantifier, if the two different cases of the
1518 character were of different lengths in their UTF-8 codings (there are
1519 some cases like this - I found 11), and the matching function had to
1520 back up over a mixture of the two cases, it incorrectly assumed they
1521 were both the same length.
1522
1523 (b) When PCRE was configured to use the heap rather than the stack for
1524 recursion during matching, it was not correctly preserving the data for
1525 the other case of a UTF-8 character when checking ahead for a match
1526 while processing a minimizing repeat. If the check also involved
1527 matching a wide character, but failed, corruption could cause an
1528 erroneous result when trying to check for a repeat of the original
1529 character.
1530
1531 13. Some tidying changes to the testing mechanism:
1532
1533 (a) The RunTest script now detects the internal link size and whether there
1534 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1535 values substituted by "configure". (The RunGrepTest script already did
1536 this for UTF-8.) The configure.ac script no longer substitutes the
1537 relevant variables.
1538
1539 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1540 with length and offset values. This means that the output is different
1541 for different internal link sizes. Test 2 is skipped for link sizes
1542 other than 2 because of this, bypassing the problem. Unfortunately,
1543 there was also a test in test 3 (the locale tests) that used /B and
1544 failed for link sizes other than 2. Rather than cut the whole test out,
1545 I have added a new /Z option to pcretest that replaces the length and
1546 offset values with spaces. This is now used to make test 3 independent
1547 of link size. (Test 2 will be tidied up later.)
1548
1549 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1550 segmentation fault instead of returning the appropriate error message.
1551
1552 15. In multiline mode when the newline sequence was set to "any", the pattern
1553 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1554 This doesn't seem right; it now treats the CRLF combination as the line
1555 ending, and so does not match in that case. It's only a pattern such as ^$
1556 that would hit this one: something like ^ABC$ would have failed after \r
1557 and then tried again after \r\n.
1558
1559 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1560 in an attempt to make files that differ only in their line terminators
1561 compare equal. This works on Linux.
1562
1563 17. Under certain error circumstances pcregrep might try to free random memory
1564 as it exited. This is now fixed, thanks to valgrind.
1565
1566 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1567 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1568 was because its rules for how to advance for /g after matching an empty
1569 string at the end of a line did not allow for this case. They now check for
1570 it specially.
1571
1572 20. pcretest is supposed to handle patterns and data of any length, by
1573 extending its buffers when necessary. It was getting this wrong when the
1574 buffer for a data line had to be extended.
1575
1576 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1577 CRLF as a newline sequence.
1578
1579 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1580 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1581 I have nevertheless tidied it up.
1582
1583 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1584
1585 24. Added a man page for pcre-config.
1586
1587
1588 Version 7.0 19-Dec-06
1589 ---------------------
1590
1591 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1592 moving to gcc 4.1.1.
1593
1594 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1595 sys/time.h, which is documented as needed for this function. It doesn't
1596 seem to matter on Linux, but it showed up on some releases of OS X.
1597
1598 3. It seems that there are systems where bytes whose values are greater than
1599 127 match isprint() in the "C" locale. The "C" locale should be the
1600 default when a C program starts up. In most systems, only ASCII printing
1601 characters match isprint(). This difference caused the output from pcretest
1602 to vary, making some of the tests fail. I have changed pcretest so that:
1603
1604 (a) When it is outputting text in the compiled version of a pattern, bytes
1605 other than 32-126 are always shown as hex escapes.
1606
1607 (b) When it is outputting text that is a matched part of a subject string,
1608 it does the same, unless a different locale has been set for the match
1609 (using the /L modifier). In this case, it uses isprint() to decide.
1610
1611 4. Fixed a major bug that caused incorrect computation of the amount of memory
1612 required for a compiled pattern when options that changed within the
1613 pattern affected the logic of the preliminary scan that determines the
1614 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1615 that the computed length was too small. The symptoms of this bug were
1616 either the PCRE error "internal error: code overflow" from pcre_compile(),
1617 or a glibc crash with a message such as "pcretest: free(): invalid next
1618 size (fast)". Examples of patterns that provoked this bug (shown in
1619 pcretest format) are:
1620
1621 /(?-x: )/x
1622 /(?x)(?-x: \s*#\s*)/
1623 /((?i)[\x{c0}])/8
1624 /(?i:[\x{c0}])/8
1625
1626 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1627 is now done differently.
1628
1629 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1630 wrapper classes; (b) implement a new function in the C++ scanner that is
1631 more efficient than the old way of doing things because it avoids levels of
1632 recursion in the regex matching; (c) add a paragraph to the documentation
1633 for the FullMatch() function.
1634
1635 6. The escape sequence \n was being treated as whatever was defined as
1636 "newline". Not only was this contrary to the documentation, which states
1637 that \n is character 10 (hex 0A), but it also went horribly wrong when
1638 "newline" was defined as CRLF. This has been fixed.
1639
1640 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1641 was being set to -1 for the "end of line" case (supposedly a value that no
1642 character can have). Though this value is never used (the check for end of
1643 line is "zero bytes in current character"), it caused compiler complaints.
1644 I've changed it to 0xffffffff.
1645
1646 8. In pcre_version.c, the version string was being built by a sequence of
1647 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1648 string (as it is for production releases) called a macro with an empty
1649 argument. The C standard says the result of this is undefined. The gcc
1650 compiler treats it as an empty string (which was what was wanted) but it is
1651 reported that Visual C gives an error. The source has been hacked around to
1652 avoid this problem.
1653
1654 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1655 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1656 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1657 of them did).
1658
1659 10. Originally, pcretest opened its input and output without "b"; then I was
1660 told that "b" was needed in some environments, so it was added for release
1661 5.0 to both the input and output. (It makes no difference on Unix-like
1662 systems.) Later I was told that it is wrong for the input on Windows. I've
1663 now abstracted the modes into two macros, to make it easier to fiddle with
1664 them, and removed "b" from the input mode under Windows.
1665
1666 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1667
1668 12. Added -help and --help to pcretest as an official way of being reminded
1669 of the options.
1670
1671 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1672 and pcrecpp.cc because they annoy compilers at high warning levels.
1673
1674 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1675
1676 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1677 scripts are not C programs :-) and which was not noticed because it works
1678 on Linux.
1679
1680 16. pcretest is supposed to handle any length of pattern and data line (as one
1681 line or as a continued sequence of lines) by extending its input buffer if
1682 necessary. This feature was broken for very long pattern lines, leading to
1683 a string of junk being passed to pcre_compile() if the pattern was longer
1684 than about 50K.
1685
1686 17. I have done a major re-factoring of the way pcre_compile() computes the
1687 amount of memory needed for a compiled pattern. Previously, there was code
1688 that made a preliminary scan of the pattern in order to do this. That was
1689 OK when PCRE was new, but as the facilities have expanded, it has become
1690 harder and harder to keep it in step with the real compile phase, and there
1691 have been a number of bugs (see for example, 4 above). I have now found a
1692 cunning way of running the real compile function in a "fake" mode that
1693 enables it to compute how much memory it would need, while actually only
1694 ever using a few hundred bytes of working memory and without too many
1695 tests of the mode. This should make future maintenance and development
1696 easier. A side effect of this work is that the limit of 200 on the nesting
1697 depth of parentheses has been removed (though this was never a serious
1698 limitation, I suspect). However, there is a downside: pcre_compile() now
1699 runs more slowly than before (30% or more, depending on the pattern). I
1700 hope this isn't a big issue. There is no effect on runtime performance.
1701
1702 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1703 newline (only possible for the last line of a file) and it was a
1704 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1705
1706 19. Added additional timing features to pcretest. (1) The -tm option now times
1707 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1708 separate command line item, by a number that specifies the number of
1709 repeats to use when timing. The default is 50000; this gives better
1710 precision, but takes uncomfortably long for very large patterns.
1711
1712 20. Extended pcre_study() to be more clever in cases where a branch of a
1713 subpattern has no definite first character. For example, (a*|b*)[cd] would
1714 previously give no result from pcre_study(). Now it recognizes that the
1715 first character must be a, b, c, or d.
1716
1717 21. There was an incorrect error "recursive call could loop indefinitely" if
1718 a subpattern (or the entire pattern) that was being tested for matching an
1719 empty string contained only one non-empty item after a nested subpattern.
1720 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1721 incorrectly, because the \d was being skipped in the check.
1722
1723 22. The pcretest program now has a new pattern option /B and a command line
1724 option -b, which is equivalent to adding /B to every pattern. This causes
1725 it to show the compiled bytecode, without the additional information that
1726 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1727 is the same as /B/I).
1728
1729 23. A new optimization is now able automatically to treat some sequences such
1730 as a*b as a*+b. More specifically, if something simple (such as a character
1731 or a simple class like \d) has an unlimited quantifier, and is followed by
1732 something that cannot possibly match the quantified thing, the quantifier
1733 is automatically "possessified".
1734
1735 24. A recursive reference to a subpattern whose number was greater than 39
1736 went wrong under certain circumstances in UTF-8 mode. This bug could also
1737 have affected the operation of pcre_study().
1738
1739 25. Realized that a little bit of performance could be had by replacing
1740 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1741
1742 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1743
1744 27. Possessive quantifiers such as a++ were previously implemented by turning
1745 them into atomic groups such as ($>a+). Now they have their own opcodes,
1746 which improves performance. This includes the automatically created ones
1747 from 23 above.
1748
1749 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1750 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1751 the first matched character to be a colon. This applied both to named and
1752 numbered groups.
1753
1754 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1755
1756 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1757 building PCRE on Windows easier, so I have included it in the distribution.
1758
1759 31. There is now a check in pcretest against a ridiculously large number being
1760 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1761 loop, the loop is abandoned.
1762
1763 32. Forward references to subpatterns in conditions such as (?(2)...) where
1764 subpattern 2 is defined later cause pcre_compile() to search forwards in
1765 the pattern for the relevant set of parentheses. This search went wrong
1766 when there were unescaped parentheses in a character class, parentheses
1767 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1768
1769 33. "Subroutine" calls and backreferences were previously restricted to
1770 referencing subpatterns earlier in the regex. This restriction has now
1771 been removed.
1772
1773 34. Added a number of extra features that are going to be in Perl 5.10. On the
1774 whole, these are just syntactic alternatives for features that PCRE had
1775 previously implemented using the Python syntax or my own invention. The
1776 other formats are all retained for compatibility.
1777
1778 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1779 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1780 also .NET compatible.
1781
1782 (b) A recursion or subroutine call to a named group can now be defined as
1783 (?&name) as well as (?P>name).
1784
1785 (c) A backreference to a named group can now be defined as \k<name> or
1786 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1787 5.10, are also .NET compatible.
1788
1789 (d) A conditional reference to a named group can now use the syntax
1790 (?(<name>) or (?('name') as well as (?(name).
1791
1792 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1793 groups (named and numbered) that are never evaluated inline, but can be
1794 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1795 is always false. There may be only one alternative in such a group.
1796
1797 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1798 as the simple (?(R). The condition is true only if the most recent
1799 recursion is that of the given number or name. It does not search out
1800 through the entire recursion stack.
1801
1802 (g) The escape \gN or \g{N} has been added, where N is a positive or
1803 negative number, specifying an absolute or relative reference.
1804
1805 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1806 some "unreachable code" warnings.
1807
1808 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1809 things, this adds five new scripts.
1810
1811 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1812 There were also incompatibilities regarding the handling of \Q..\E inside
1813 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1814 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1815
1816 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1817 matches an empty string, and forcibly breaks the loop. There were bugs in
1818 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1819 against aaaa the result was just "a" rather than "aaaa", for example. Two
1820 separate and independent bugs (that affected different cases) have been
1821 fixed.
1822
1823 39. Refactored the code to abolish the use of different opcodes for small
1824 capturing bracket numbers. This is a tidy that I avoided doing when I
1825 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1826 The new approach is not only tidier, it makes it possible to reduce the
1827 memory needed to fix the previous bug (38).
1828
1829 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1830 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1831 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1832 mode.
1833
1834 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1835 report.
1836
1837 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1838 copy construction and assignment in the C++ wrapper.
1839
1840 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1841 couple of bugs that could have given wrong results in the "--newline=crlf"
1842 case.
1843
1844 44. Added a number of casts and did some reorganization of signed/unsigned int
1845 variables following suggestions from Dair Grant. Also renamed the variable
1846 "this" as "item" because it is a C++ keyword.
1847
1848 45. Arranged for dftables to add
1849
1850 #include "pcre_internal.h"
1851
1852 to pcre_chartables.c because without it, gcc 4.x may remove the array
1853 definition from the final binary if PCRE is built into a static library and
1854 dead code stripping is activated.
1855
1856 46. For an unanchored pattern, if a match attempt fails at the start of a
1857 newline sequence, and the newline setting is CRLF or ANY, and the next two
1858 characters are CRLF, advance by two characters instead of one.
1859
1860
1861 Version 6.7 04-Jul-06
1862 ---------------------
1863
1864 1. In order to handle tests when input lines are enormously long, pcretest has
1865 been re-factored so that it automatically extends its buffers when
1866 necessary. The code is crude, but this _is_ just a test program. The
1867 default size has been increased from 32K to 50K.
1868
1869 2. The code in pcre_study() was using the value of the re argument before
1870 testing it for NULL. (Of course, in any sensible call of the function, it
1871 won't be NULL.)
1872
1873 3. The memmove() emulation function in pcre_internal.h, which is used on
1874 systems that lack both memmove() and bcopy() - that is, hardly ever -
1875 was missing a "static" storage class specifier.
1876
1877 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1878 containing an extended class (one that cannot be represented by a bitmap
1879 because it contains high-valued characters or Unicode property items, e.g.
1880 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1881 pattern, but PCRE should not loop if you do not (it no longer does).
1882 [Detail: two cases were found: (a) a repeated subpattern containing an
1883 extended class; (b) a recursive reference to a subpattern that followed a
1884 previous extended class. It wasn't skipping over the extended class
1885 correctly when UTF-8 mode was not set.]
1886
1887 5. A negated single-character class was not being recognized as fixed-length
1888 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1889 compile error "lookbehind assertion is not fixed length".
1890
1891 6. The RunPerlTest auxiliary script was showing an unexpected difference
1892 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1893 write a Perl script that can interpret lines of an input file either as
1894 byte characters or as UTF-8, which is what "perltest" was being required to
1895 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1896 can't do is switch easily at run time between having the "use utf8;" pragma
1897 or not. In the end, I fudged it by using the RunPerlTest script to insert
1898 "use utf8;" explicitly for the UTF-8 tests.
1899
1900 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1901 the end of the subject string, contrary to the documentation and to what
1902 Perl does. This was true of both matching functions. Now it matches only at
1903 the start of the subject and immediately after *internal* newlines.
1904
1905 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1906 a pointer to an int instead of a pointer to an unsigned long int. This
1907 caused problems on 64-bit systems.
1908
1909 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1910 instance of the 'standard' template library not being so standard".
1911
1912 10. There was no check on the number of named subpatterns nor the maximum
1913 length of a subpattern name. The product of these values is used to compute
1914 the size of the memory block for a compiled pattern. By supplying a very
1915 long subpattern name and a large number of named subpatterns, the size
1916 computation could be caused to overflow. This is now prevented by limiting
1917 the length of names to 32 characters, and the number of named subpatterns
1918 to 10,000.
1919
1920 11. Subpatterns that are repeated with specific counts have to be replicated in
1921 the compiled pattern. The size of memory for this was computed from the
1922 length of the subpattern and the repeat count. The latter is limited to
1923 65535, but there was no limit on the former, meaning that integer overflow
1924 could in principle occur. The compiled length of a repeated subpattern is
1925 now limited to 30,000 bytes in order to prevent this.
1926
1927 12. Added the optional facility to have named substrings with the same name.
1928
1929 13. Added the ability to use a named substring as a condition, using the
1930 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1931 are numbers (not recommended). Forward references are permitted.
1932
1933 14. Added forward references in named backreferences (if you see what I mean).
1934
1935 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1936 pattern could run off the end of the subject. For example, the pattern
1937 "(?s)(.{1,5})"8 did this with the subject "ab".
1938
1939 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1940 PCRE_CASELESS was set when matching characters that were quantified with ?
1941 or *.
1942
1943 17. A character class other than a single negated character that had a minimum
1944 but no maximum quantifier - for example [ab]{6,} - was not handled
1945 correctly by pce_dfa_exec(). It would match only one character.
1946
1947 18. A valid (though odd) pattern that looked like a POSIX character
1948 class but used an invalid character after [ (for example [[,abc,]]) caused
1949 pcre_compile() to give the error "Failed: internal error: code overflow" or
1950 in some cases to crash with a glibc free() error. This could even happen if
1951 the pattern terminated after [[ but there just happened to be a sequence of
1952 letters, a binary zero, and a closing ] in the memory that followed.
1953
1954 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1955 over the years. Originally (before any Unicode support), just the bottom 8
1956 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1957 output from "man perlunicode" includes this:
1958
1959 The regular expression compiler produces polymorphic opcodes. That
1960 is, the pattern adapts to the data and automatically switches to
1961 the Unicode character scheme when presented with Unicode data--or
1962 instead uses a traditional byte scheme when presented with byte
1963 data.
1964
1965 Sadly, a wide octal escape does not cause a switch, and in a string with
1966 no other multibyte characters, these octal escapes are treated as before.
1967 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1968 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1969 Unicode string.
1970
1971 I have not perpetrated such confusion in PCRE. Up till now, it took just
1972 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1973 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1974 translate to the appropriate multibyte character.
1975
1976 29. Applied some refactoring to reduce the number of warnings from Microsoft
1977 and Borland compilers. This has included removing the fudge introduced
1978 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1979 a warning about an unused variable.
1980
1981 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1982 characters since release 4.0, because Perl (from release 5.004) does not.
1983 [Or at least, is documented not to: some releases seem to be in conflict
1984 with the documentation.] However, when a pattern was studied with
1985 pcre_study() and all its branches started with \s, PCRE still included VT
1986 as a possible starting character. Of course, this did no harm; it just
1987 caused an unnecessary match attempt.
1988
1989 22. Removed a now-redundant internal flag bit that recorded the fact that case
1990 dependency changed within the pattern. This was once needed for "required
1991 byte" processing, but is no longer used. This recovers a now-scarce options
1992 bit. Also moved the least significant internal flag bit to the most-
1993 significant bit of the word, which was not previously used (hangover from
1994 the days when it was an int rather than a uint) to free up another bit for
1995 the future.
1996
1997 23. Added support for CRLF line endings as well as CR and LF. As well as the
1998 default being selectable at build time, it can now be changed at runtime
1999 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2000 specify that it is scanning data with non-default line endings.
2001
2002 24. Changed the definition of CXXLINK to make it agree with the definition of
2003 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2004
2005 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2006 recursions. This makes a big different to stack usage for some patterns.
2007
2008 26. If a subpattern containing a named recursion or subroutine reference such
2009 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2010 the space required for the compiled pattern went wrong and gave too small a
2011 value. Depending on the environment, this could lead to "Failed: internal
2012 error: code overflow at offset 49" or "glibc detected double free or
2013 corruption" errors.
2014
2015 27. Applied patches from Google (a) to support the new newline modes and (b) to
2016 advance over multibyte UTF-8 characters in GlobalReplace.
2017
2018 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2019 difference for some implementation of PCRE in some Windows version.
2020
2021 29. Added some extra testing facilities to pcretest:
2022
2023 \q<number> in a data line sets the "match limit" value
2024 \Q<number> in a data line sets the "match recursion limt" value
2025 -S <number> sets the stack size, where <number> is in megabytes
2026
2027 The -S option isn't available for Windows.
2028
2029
2030 Version 6.6 06-Feb-06
2031 ---------------------
2032
2033 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2034 in pcreposix.h. I have copied the definition from pcre.h.
2035
2036 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2037 because pcre.h is no longer a built file.
2038
2039 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2040 not normally included in the compiled code.
2041
2042
2043 Version 6.5 01-Feb-06
2044 ---------------------
2045
2046 1. When using the partial match feature with pcre_dfa_exec(), it was not
2047 anchoring the second and subsequent partial matches at the new starting
2048 point. This could lead to incorrect results. For example, with the pattern
2049 /1234/, partially matching against "123" and then "a4" gave a match.
2050
2051 2. Changes to pcregrep:
2052
2053 (a) All non-match returns from pcre_exec() were being treated as failures
2054 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2055 error message is output. Some extra information is given for the
2056 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2057 probably the only errors that are likely to be caused by users (by
2058 specifying a regex that has nested indefinite repeats, for instance).
2059 If there are more than 20 of these errors, pcregrep is abandoned.
2060
2061 (b) A binary zero was treated as data while matching, but terminated the
2062 output line if it was written out. This has been fixed: binary zeroes
2063 are now no different to any other data bytes.
2064
2065 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2066 used to set a locale for matching. The --locale=xxxx long option has
2067 been added (no short equivalent) to specify a locale explicitly on the
2068 pcregrep command, overriding the environment variables.
2069
2070 (d) When -B was used with -n, some line numbers in the output were one less
2071 than they should have been.
2072
2073 (e) Added the -o (--only-matching) option.
2074
2075 (f) If -A or -C was used with -c (count only), some lines of context were
2076 accidentally printed for the final match.
2077
2078 (g) Added the -H (--with-filename) option.
2079
2080 (h) The combination of options -rh failed to suppress file names for files
2081 that were found from directory arguments.
2082
2083 (i) Added the -D (--devices) and -d (--directories) options.
2084
2085 (j) Added the -F (--fixed-strings) option.
2086
2087 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2088
2089 (l) Added the --colo(u)r option.
2090
2091 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2092 is not present by default.
2093
2094 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2095 items such as (?R) or (?1), when the recursion could match a number of
2096 alternatives. If it matched one of the alternatives, but subsequently,
2097 outside the recursion, there was a failure, the code tried to back up into
2098 the recursion. However, because of the way PCRE is implemented, this is not
2099 possible, and the result was an incorrect result from the match.
2100
2101 In order to prevent this happening, the specification of recursion has
2102 been changed so that all such subpatterns are automatically treated as
2103 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2104
2105 4. I had overlooked the fact that, in some locales, there are characters for
2106 which isalpha() is true but neither isupper() nor islower() are true. In
2107 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2108 and ordfeminine) are like this. This affected the treatment of \w and \W
2109 when they appeared in character classes, but not when they appeared outside
2110 a character class. The bit map for "word" characters is now created
2111 separately from the results of isalnum() instead of just taking it from the
2112 upper, lower, and digit maps. (Plus the underscore character, of course.)
2113
2114 5. The above bug also affected the handling of POSIX character classes such as
2115 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2116 permanent tables. Instead, the bit maps for such a class were previously
2117 created as the appropriate unions of the upper, lower, and digit bitmaps.
2118 Now they are created by subtraction from the [[:word:]] class, which has
2119 its own bitmap.
2120
2121 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2122 It is created by subtracting the vertical space characters (\x09, \x0a,
2123 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2124 subtraction was done in the overall bitmap for a character class, meaning
2125 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2126 be recognized. This bug has been fixed.
2127
2128 7. Patches from the folks at Google:
2129
2130 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2131 real life, but is still worth protecting against".
2132
2133 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2134 regular expressions".
2135
2136 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2137 have it.
2138
2139 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2140 "configure" and the latter not, in order to fix a problem somebody had
2141 with compiling the Arg class on HP-UX.
2142
2143 (e) Improve the error-handling of the C++ wrapper a little bit.
2144
2145 (f) New tests for checking recursion limiting.
2146
2147 8. The pcre_memmove() function, which is used only if the environment does not
2148 have a standard memmove() function (and is therefore rarely compiled),
2149 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2150 returning a result (though PCRE never actually uses the result).
2151
2152 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2153 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2154 returned instead of calling malloc() with an overflowing number that would
2155 most likely cause subsequent chaos.
2156
2157 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2158
2159 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2160 with this option is matched, the nmatch and pmatch options of regexec() are
2161 ignored.
2162
2163 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2164 provided in case anyone wants to the the POSIX interface with UTF-8
2165 strings.
2166
2167 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2168 C++ linking (needed for some HP-UX environments).
2169
2170 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2171 (unused parameter) and in the pcre_printint() function (omitted "default"
2172 switch label when the default is to do nothing).
2173
2174 15. Added some code to make it possible, when PCRE is compiled as a C++
2175 library, to replace subject pointers for pcre_exec() with a smart pointer
2176 class, thus making it possible to process discontinuous strings.
2177
2178 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2179 much the same function. They were added by different people who were trying
2180 to make PCRE easy to compile on non-Unix systems. It has been suggested
2181 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2182 for compiling on Windows systems. I have therefore replaced it with
2183 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2184 defaults to "extern" for C or "extern C" for C++, which works fine on
2185 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2186 SCOPE with something explicit in config.h. In addition:
2187
2188 (a) pcreposix.h still had just "extern" instead of either of these macros;
2189 I have replaced it with PCRE_DATA_SCOPE.
2190
2191 (b) Functions such as _pcre_xclass(), which are internal to the library,
2192 but external in the C sense, all had PCRE_EXPORT in their definitions.
2193 This is apparently wrong for the Windows case, so I have removed it.
2194 (It makes no difference on Unix-like systems.)
2195
2196 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2197 of recursive calls to match(). This is different to MATCH_LIMIT because
2198 that limits the total number of calls to match(), not all of which increase
2199 the depth of recursion. Limiting the recursion depth limits the amount of
2200 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2201 when PCRE is compiled, and changed at run time. A patch from Google adds
2202 this functionality to the C++ interface.
2203
2204 18. Changes to the handling of Unicode character properties:
2205
2206 (a) Updated the table to Unicode 4.1.0.
2207
2208 (b) Recognize characters that are not in the table as "Cn" (undefined).
2209
2210 (c) I revised the way the table is implemented to a much improved format
2211 which includes recognition of ranges. It now supports the ranges that
2212 are defined in UnicodeData.txt, and it also amalgamates other
2213 characters into ranges. This has reduced the number of entries in the
2214 table from around 16,000 to around 3,000, thus reducing its size
2215 considerably. I realized I did not need to use a tree structure after
2216 all - a binary chop search is just as efficient. Having reduced the
2217 number of entries, I extended their size from 6 bytes to 8 bytes to
2218 allow for more data.
2219
2220 (d) Added support for Unicode script names via properties such as \p{Han}.
2221
2222 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2223 matching that character.
2224
2225 20. When matching a repeated Unicode property with a minimum greater than zero,
2226 (for example \pL{2,}), PCRE could look past the end of the subject if it
2227 reached it while seeking the minimum number of characters. This could
2228 happen only if some of the characters were more than one byte long, because
2229 there is a check for at least the minimum number of bytes.
2230
2231 21. Refactored the implementation of \p and \P so as to be more general, to
2232 allow for more different types of property in future. This has changed the
2233 compiled form incompatibly. Anybody with saved compiled patterns that use
2234 \p or \P will have to recompile them.
2235
2236 22. Added "Any" and "L&" to the supported property types.
2237
2238 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2239 but give a compile time error if the value is greater than 0xff.
2240
2241 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2242 accidentally not being installed or uninstalled.
2243
2244 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2245 made were to insert the current release number. This seemed silly, because
2246 it made things harder for people building PCRE on systems that don't run
2247 "configure". I have turned pcre.h into a distributed file, no longer built
2248 by "configure", with the version identification directly included. There is
2249 no longer a pcre.h.in file.
2250
2251 However, this change necessitated a change to the pcre-config script as
2252 well. It is built from pcre-config.in, and one of the substitutions was the
2253 release number. I have updated configure.ac so that ./configure now finds
2254 the release number by grepping pcre.h.
2255
2256 26. Added the ability to run the tests under valgrind.
2257
2258
2259 Version 6.4 05-Sep-05
2260 ---------------------
2261
2262 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2263 "--" to be printed when multiple files were scanned, even when none of the
2264 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2265 consider it to be a bug, and have restored the previous behaviour.
2266
2267 2. A couple of code tidies to get rid of compiler warnings.
2268
2269 3. The pcretest program used to cheat by referring to symbols in the library
2270 whose names begin with _pcre_. These are internal symbols that are not
2271 really supposed to be visible externally, and in some environments it is
2272 possible to suppress them. The cheating is now confined to including
2273 certain files from the library's source, which is a bit cleaner.
2274
2275 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2276 file's purpose clearer.
2277
2278 5. Reorganized pcre_ucp_findchar().
2279
2280
2281 Version 6.3 15-Aug-05
2282 ---------------------
2283
2284 1. The file libpcre.pc.in did not have general read permission in the tarball.
2285
2286 2. There were some problems when building without C++ support:
2287
2288 (a) If C++ support was not built, "make install" and "make test" still
2289 tried to test it.
2290
2291 (b) There were problems when the value of CXX was explicitly set. Some
2292 changes have been made to try to fix these, and ...
2293
2294 (c) --disable-cpp can now be used to explicitly disable C++ support.
2295
2296 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2297 backslash in a target when C++ was disabled. This confuses some
2298 versions of "make", apparently. Using an intermediate variable solves
2299 this. (Same for CPP_LOBJ.)
2300
2301 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2302 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2303 necessary on certain architectures.
2304
2305 4. Added a setting of -export-symbols-regex to the link command to remove
2306 those symbols that are exported in the C sense, but actually are local
2307 within the library, and not documented. Their names all begin with
2308 "_pcre_". This is not a perfect job, because (a) we have to except some
2309 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2310 available (and never for static libraries). I have made a note to try to
2311 find a way round (a) in the future.
2312
2313
2314 Version 6.2 01-Aug-05
2315 ---------------------
2316
2317 1. There was no test for integer overflow of quantifier values. A construction
2318 such as {1111111111111111} would give undefined results. What is worse, if
2319 a minimum quantifier for a parenthesized subpattern overflowed and became
2320 negative, the calculation of the memory size went wrong. This could have
2321 led to memory overwriting.
2322
2323 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2324
2325 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2326 operating environments where this matters.
2327
2328 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2329 PCRE options from within the C++ wrapper.
2330
2331 5. Named capturing subpatterns were not being correctly counted when a pattern
2332 was compiled. This caused two problems: (a) If there were more than 100
2333 such subpatterns, the calculation of the memory needed for the whole
2334 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2335 back references of the form \12, where the number was greater than 9, were
2336 not recognized as back references, even though there were sufficient
2337 previous subpatterns.
2338
2339 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2340 versions of gcc, e.g. 2.95.4.
2341
2342
2343 Version 6.1 21-Jun-05
2344 ---------------------
2345
2346 1. There was one reference to the variable "posix" in pcretest.c that was not
2347 surrounded by "#if !defined NOPOSIX".
2348
2349 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2350 the cross-check on the old pcre_info() function, for the benefit of the
2351 cut-down version of PCRE that is currently imported into Exim.
2352
2353 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2354 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2355 patterns that start (?i) but I don't think that matters. The use of (?i) is
2356 just an example; this all applies to the other options as well.
2357
2358 4. Since libtool seems to echo the compile commands it is issuing, the output
2359 from "make" can be reduced a bit by putting "@" in front of each libtool
2360 compile command.
2361
2362 5. Patch from the folks at Google for configure.in to be a bit more thorough
2363 in checking for a suitable C++ installation before trying to compile the
2364 C++ stuff. This should fix a reported problem when a compiler was present,
2365 but no suitable headers.
2366
2367 6. The man pages all had just "PCRE" as their title. I have changed them to
2368 be the relevant file name. I have also arranged that these names are
2369 retained in the file doc/pcre.txt, which is a concatenation in text format
2370 of all the man pages except the little individual ones for each function.
2371
2372 7. The NON-UNIX-USE file had not been updated for the different set of source
2373 files that come with release 6. I also added a few comments about the C++
2374 wrapper.
2375
2376
2377 Version 6.0 07-Jun-05
2378 ---------------------
2379
2380 1. Some minor internal re-organization to help with my DFA experiments.
2381
2382 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2383 didn't matter for the library itself when fully configured, but did matter
2384 when compiling without UCP support, or within Exim, where the ucp files are
2385 not imported.
2386
2387 3. Refactoring of the library code to split up the various functions into
2388 different source modules. The addition of the new DFA matching code (see
2389 below) to a single monolithic source would have made it really too
2390 unwieldy, quite apart from causing all the code to be include in a
2391 statically linked application, when only some functions are used. This is
2392 relevant even without the DFA addition now that patterns can be compiled in
2393 one application and matched in another.
2394
2395 The downside of splitting up is that there have to be some external
2396 functions and data tables that are used internally in different modules of
2397 the library but which are not part of the API. These have all had their
2398 names changed to start with "_pcre_" so that they are unlikely to clash
2399 with other external names.
2400
2401 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2402 a different (DFA) algorithm. Although it is slower than the original
2403 function, it does have some advantages for certain types of matching
2404 problem.
2405
2406 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2407 including restarting after a partial match.
2408
2409 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2410 defined when compiling for Windows was sent to me. I have put it into the
2411 code, though I have no means of testing or verifying it.
2412
2413 7. Added the pcre_refcount() auxiliary function.
2414
2415 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2416 match before or at the first newline in the subject string. In pcretest,
2417 the /f option on a pattern can be used to set this.
2418
2419 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2420 would behave wrongly. This has been present in PCRE since release 4.0.
2421
2422 10. A number of changes to the pcregrep command:
2423
2424 (a) Refactored how -x works; insert ^(...)$ instead of setting
2425 PCRE_ANCHORED and checking the length, in preparation for adding
2426 something similar for -w.
2427
2428 (b) Added the -w (match as a word) option.
2429
2430 (c) Refactored the way lines are read and buffered so as to have more
2431 than one at a time available.
2432
2433 (d) Implemented a pcregrep test script.
2434
2435 (e) Added the -M (multiline match) option. This allows patterns to match
2436 over several lines of the subject. The buffering ensures that at least
2437 8K, or the rest of the document (whichever is the shorter) is available
2438 for matching (and similarly the previous 8K for lookbehind assertions).
2439
2440 (f) Changed the --help output so that it now says
2441
2442 -w, --word-regex(p)
2443
2444 instead of two lines, one with "regex" and the other with "regexp"
2445 because that confused at least one person since the short forms are the
2446 same. (This required a bit of code, as the output is generated
2447 automatically from a table. It wasn't just a text change.)
2448
2449 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2450 option but starts with a hyphen. Could be a pattern or a path name
2451 starting with a hyphen, for instance.
2452
2453 (h) "-" can be given as a file name to represent stdin.
2454
2455 (i) When file names are being printed, "(standard input)" is used for
2456 the standard input, for compatibility with GNU grep. Previously
2457 "<stdin>" was used.
2458
2459 (j) The option --label=xxx can be used to supply a name to be used for
2460 stdin when file names are being printed. There is no short form.
2461
2462 (k) Re-factored the options decoding logic because we are going to add
2463 two more options that take data. Such options can now be given in four
2464 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2465
2466 (l) Added the -A, -B, and -C options for requesting that lines of context
2467 around matches be printed.
2468
2469 (m) Added the -L option to print the names of files that do not contain
2470 any matching lines, that is, the complement of -l.
2471
2472 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2473 continue to scan other files.
2474
2475 (o) The -s option was incorrectly implemented. For compatibility with other
2476 greps, it now suppresses the error message for a non-existent or non-
2477 accessible file (but not the return code). There is a new option called
2478 -q that suppresses the output of matching lines, which was what -s was
2479 previously doing.
2480
2481 (p) Added --include and --exclude options to specify files for inclusion
2482 and exclusion when recursing.
2483
2484 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2485 Hopefully, it now does.
2486
2487 12. Missing cast in pcre_study().
2488
2489 13. Added an "uninstall" target to the makefile.
2490
2491 14. Replaced "extern" in the function prototypes in Makefile.in with
2492 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2493 world, but is set differently for Windows.
2494
2495 15. Added a second compiling function called pcre_compile2(). The only
2496 difference is that it has an extra argument, which is a pointer to an
2497 integer error code. When there is a compile-time failure, this is set
2498 non-zero, in addition to the error test pointer being set to point to an
2499 error message. The new argument may be NULL if no error number is required
2500 (but then you may as well call pcre_compile(), which is now just a
2501 wrapper). This facility is provided because some applications need a
2502 numeric error indication, but it has also enabled me to tidy up the way
2503 compile-time errors are handled in the POSIX wrapper.
2504
2505 16. Added VPATH=.libs to the makefile; this should help when building with one
2506 prefix path and installing with another. (Or so I'm told by someone who
2507 knows more about this stuff than I do.)
2508
2509 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2510 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2511 match everything, including newlines. This is not POSIX-compatible, but
2512 somebody wanted the feature. From pcretest it can be activated by using
2513 both the P and the s flags.
2514
2515 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2516
2517 19. libpcre.pc was being incorrectly installed as executable.
2518
2519 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2520 it now also looks for '\r' so that it will work unmodified on Windows.
2521
2522 21. Added Google's contributed C++ wrapper to the distribution.
2523
2524 22. Added some untidy missing memory free() calls in pcretest, to keep
2525 Electric Fence happy when testing.
2526
2527
2528
2529 Version 5.0 13-Sep-04
2530 ---------------------
2531
2532 1. Internal change: literal characters are no longer packed up into items
2533 containing multiple characters in a single byte-string. Each character
2534 is now matched using a separate opcode. However, there may be more than one
2535 byte in the character in UTF-8 mode.
2536
2537 2. The pcre_callout_block structure has two new fields: pattern_position and
2538 next_item_length. These contain the offset in the pattern to the next match
2539 item, and its length, respectively.
2540
2541 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2542 insertion of callouts before each pattern item. Added the /C option to
2543 pcretest to make use of this.
2544
2545 4. On the advice of a Windows user, the lines
2546
2547 #if defined(_WIN32) || defined(WIN32)
2548 _setmode( _fileno( stdout ), 0x8000 );
2549 #endif /* defined(_WIN32) || defined(WIN32) */
2550
2551 have been added to the source of pcretest. This apparently does useful
2552 magic in relation to line terminators.
2553
2554 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2555 for the benefit of those environments where the "b" makes a difference.
2556
2557 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2558 to know about it. I have put a hack into configure.in that adds in code
2559 to set GCC=yes if CC=icc. This seems to end up at a point in the
2560 generated configure script that is early enough to affect the setting of
2561 compiler options, which is what is needed, but I have no means of testing
2562 whether it really works. (The user who reported this had patched the
2563 generated configure script, which of course I cannot do.)
2564
2565 LATER: After change 22 below (new libtool files), the configure script
2566 seems to know about icc (and also ecc). Therefore, I have commented out
2567 this hack in configure.in.
2568
2569 7. Added support for pkg-config (2 patches were sent in).
2570
2571 8. Negated POSIX character classes that used a combination of internal tables
2572 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2573 [[:^ascii]]. Typically, they would match almost any characters. The other
2574 POSIX classes were not broken in this way.
2575
2576 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2577 to find the match, as PCRE was deluded into thinking that the match had to
2578 start at the start point or following a newline. The same bug applied to
2579 patterns with negative forward assertions or any backward assertions
2580 preceding ".*" at the start, unless the pattern required a fixed first
2581 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2582
2583 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2584 starting at the last subject character, bytes beyond the end of the subject
2585 string were read.
2586
2587 11. Renamed the variable "class" as "classbits" to make life easier for C++
2588 users. (Previously there was a macro definition, but it apparently wasn't
2589 enough.)
2590
2591 12. Added the new field "tables" to the extra data so that tables can be passed
2592 in at exec time, or the internal tables can be re-selected. This allows
2593 a compiled regex to be saved and re-used at a later time by a different
2594 program that might have everything at different addresses.
2595
2596 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2597 -R library as well as a -L library.
2598
2599 14. The debugging options of pcretest (-d on the command line or D on a
2600 pattern) showed incorrect output for anything following an extended class
2601 that contained multibyte characters and which was followed by a quantifier.
2602
2603 15. Added optional support for general category Unicode character properties
2604 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2605 support. It adds about 90K to the size of the library. The meanings of the
2606 inbuilt class escapes such as \d and \s have NOT been changed.
2607
2608 16. Updated pcredemo.c to include calls to free() to release the memory for the
2609 compiled pattern.
2610
2611 17. The generated file chartables.c was being created in the source directory
2612 instead of in the building directory. This caused the build to fail if the
2613 source directory was different from the building directory, and was
2614 read-only.
2615
2616 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2617 file. No doubt somebody will tell me if they don't make sense... Also added
2618 Dan Mooney's comments about building on OpenVMS.
2619
2620 19. Added support for partial matching via the PCRE_PARTIAL option for
2621 pcre_exec() and the \P data escape in pcretest.
2622
2623 20. Extended pcretest with 3 new pattern features:
2624
2625 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2626 write the compiled pattern to the file whose name is "rest-of-line".
2627 This is a straight binary dump of the data, with the saved pointer to
2628 the character tables forced to be NULL. The study data, if any, is
2629 written too. After writing, pcretest reads a new pattern.
2630
2631 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2632 compiled pattern from the given file. There must not be any
2633 occurrences of "<" in the file name (pretty unlikely); if there are,
2634 pcretest will instead treat the initial "<" as a pattern delimiter.
2635 After reading in the pattern, pcretest goes on to read data lines as
2636 usual.
2637
2638 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2639 and 16-bit fields in a compiled pattern, to simulate a pattern that
2640 was compiled on a host of opposite endianness.
2641
2642 21. The pcre-exec() function can now cope with patterns that were compiled on
2643 hosts of opposite endianness, with this restriction:
2644
2645 As for any compiled expression that is saved and used later, the tables
2646 pointer field cannot be preserved; the extra_data field in the arguments
2647 to pcre_exec() should be used to pass in a tables address if a value
2648 other than the default internal tables were used at compile time.
2649
2650 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2651 now diagnosed as an error. Previously, most of the time, a negative number
2652 would have been treated as zero, but if in addition "ovector" was passed as
2653 NULL, a crash could occur.
2654
2655 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2656 new versions from the libtool 1.5 distribution (the last one is a copy of
2657 a file called libtool.m4). This seems to have fixed the need to patch
2658 "configure" to support Darwin 1.3 (which I used to do). However, I still
2659 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2660 workstation).
2661
2662 24. Changed the PCRE licence to be the more standard "BSD" licence.
2663
2664
2665 Version 4.5 01-Dec-03
2666 ---------------------
2667
2668 1. There has been some re-arrangement of the code for the match() function so
2669 that it can be compiled in a version that does not call itself recursively.
2670 Instead, it keeps those local variables that need separate instances for
2671 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2672 needs to "recurse". Keeping track of where control must go is done by means
2673 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2674 hide most of the details from the main code, and operates only if
2675 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2676 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2677 operating.
2678
2679 To make it easier for callers to provide specially tailored get/free
2680 functions for this usage, two new functions, pcre_stack_malloc, and
2681 pcre_stack_free, are used. They are always called in strict stacking order,
2682 and the size of block requested is always the same.
2683
2684 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2685 PCRE has been compiled to use the stack or the heap for recursion. The
2686 -C option of pcretest uses this to show which version is compiled.
2687
2688 A new data escape \S, is added to pcretest; it causes the amounts of store
2689 obtained and freed by both kinds of malloc/free at match time to be added
2690 to the output.
2691
2692 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2693 what's available on my current Linux desktop machine.
2694
2695 3. When matching a UTF-8 string, the test for a valid string at the start has
2696 been extended. If start_offset is not zero, PCRE now checks that it points
2697 to a byte that is the start of a UTF-8 character. If not, it returns
2698 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2699 this is necessary because there may be backward assertions in the pattern.
2700 When matching the same subject several times, it may save resources to use
2701 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2702
2703 4. The code for checking the validity of UTF-8 strings has been tightened so
2704 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2705 containing "overlong sequences".
2706
2707 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2708 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2709 should have been "&", but it just so happened that all the cases this let
2710 through by mistake were picked up later in the function.
2711
2712 6. I had used a variable called "isblank" - this is a C99 function, causing
2713 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2714
2715 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2716 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2717 so the version is shown; (c) stop "make test" echoing "./RunTest".
2718
2719 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2720
2721 9. The prototype for memmove() for systems that don't have it was using
2722 size_t, but the inclusion of the header that defines size_t was later. I've
2723 moved the #includes for the C headers earlier to avoid this.
2724
2725 10. Added some adjustments to the code to make it easier to compiler on certain
2726 special systems:
2727
2728 (a) Some "const" qualifiers were missing.
2729 (b) Added the macro EXPORT before all exported functions; by default this
2730 is defined to be empty.
2731 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2732 that it reads its output file name as an argument instead of writing
2733 to the standard output and assuming this can be redirected.
2734
2735 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2736 class containing characters with values greater than 255, PCRE compilation
2737 went into a loop.
2738
2739 12. A recursive reference to a subpattern that was within another subpattern
2740 that had a minimum quantifier of zero caused PCRE to crash. For example,
2741 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2742 recursion. If the recursively-called subpattern itself had a zero repeat,
2743 that was OK.
2744
2745 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2746 buffer into which it was copied (for escape processing) was still set at
2747 1024, so long lines caused crashes.
2748
2749 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2750 "internal error: code overflow...". This applied to any character class
2751 that was followed by a possessive quantifier.
2752
2753 15. Modified the Makefile to add libpcre.la as a prerequisite for
2754 libpcreposix.la because I was told this is needed for a parallel build to
2755 work.
2756
2757 16. If a pattern that contained .* following optional items at the start was
2758 studied, the wrong optimizing data was generated, leading to matching
2759 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2760 matching string must start with a or b or c. The correct conclusion for
2761 this pattern is that a match can start with any character.
2762
2763
2764 Version 4.4 13-Aug-03
2765 ---------------------
2766
2767 1. In UTF-8 mode, a character class containing characters with values between
2768 127 and 255 was not handled correctly if the compiled pattern was studied.
2769 In fixing this, I have also improved the studying algorithm for such
2770 classes (slightly).
2771
2772 2. Three internal functions had redundant arguments passed to them. Removal
2773 might give a very teeny performance improvement.
2774
2775 3. Documentation bug: the value of the capture_top field in a callout is *one
2776 more than* the number of the hightest numbered captured substring.
2777
2778 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2779 in incorrectly linking with a previously installed version. They now link
2780 explicitly with libpcre.la.
2781
2782 5. configure.in no longer needs to recognize Cygwin specially.
2783
2784 6. A problem in pcre.in for Windows platforms is fixed.
2785
2786 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2787 pcretest, it used to include the size of the study block as part of its
2788 output. Unfortunately, the structure contains a field that has a different
2789 size on different hardware architectures. This meant that the tests that
2790 showed this size failed. As the block is currently always of a fixed size,
2791 this information isn't actually particularly useful in pcretest output, so
2792 I have just removed it.
2793
2794 8. Three pre-processor statements accidentally did not start in column 1.
2795 Sadly, there are *still* compilers around that complain, even though
2796 standard C has not required this for well over a decade. Sigh.
2797
2798 9. In pcretest, the code for checking callouts passed small integers in the
2799 callout_data field, which is a void * field. However, some picky compilers
2800 complained about the casts involved for this on 64-bit systems. Now
2801 pcretest passes the address of the small integer instead, which should get
2802 rid of the warnings.
2803
2804 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2805 both compile and run time, and gives an error if an invalid UTF-8 sequence
2806 is found. There is a option for disabling this check in cases where the
2807 string is known to be correct and/or the maximum performance is wanted.
2808
2809 11. In response to a bug report, I changed one line in Makefile.in from
2810
2811 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2812 to
2813 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2814
2815 to look similar to other lines, but I have no way of telling whether this
2816 is the right thing to do, as I do not use Windows. No doubt I'll get told
2817 if it's wrong...
2818
2819
2820 Version 4.3 21-May-03
2821 ---------------------
2822
2823 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2824 Makefile.
2825
2826 2. Some refactoring to improve the quality of the code:
2827
2828 (i) The utf8_table... variables are now declared "const".
2829
2830 (ii) The code for \cx, which used the "case flipping" table to upper case
2831 lower case letters, now just substracts 32. This is ASCII-specific,
2832 but the whole concept of \cx is ASCII-specific, so it seems
2833 reasonable.
2834
2835 (iii) PCRE was using its character types table to recognize decimal and
2836 hexadecimal digits in the pattern. This is silly, because it handles
2837 only 0-9, a-f, and A-F, but the character types table is locale-
2838 specific, which means strange things might happen. A private
2839 table is now used for this - though it costs 256 bytes, a table is
2840 much faster than multiple explicit tests. Of course, the standard
2841 character types table is still used for matching digits in subject
2842 strings against \d.
2843
2844 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2845 ending in _t are). So I've renamed it as ESC_tee.
2846
2847 3. The first argument for regexec() in the POSIX wrapper should have been
2848 defined as "const".
2849
2850 4. Changed pcretest to use malloc() for its buffers so that they can be
2851 Electric Fenced for debugging.
2852
2853 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2854 to read one or more bytes before the start of the subject string. Often this
2855 had no effect on PCRE's behaviour, but in some circumstances it could
2856 provoke a segmentation fault.
2857
2858 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2859 to try to read one or more bytes before the start of the subject string.
2860
2861 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2862 UTF-8 support could misbehave in various ways if the subject string
2863 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2864 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2865 back over UTF-8 characters.)
2866
2867
2868 Version 4.2 14-Apr-03
2869 ---------------------
2870
2871 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2872
2873 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2874 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2875 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2876 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2877 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2878 and BUILD_EXEEXT
2879 Note: automatic setting of the BUILD variables is not yet working
2880 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2881 compile-time but not at link-time
2882 [LINK]: use for linking executables only
2883 make different versions for Windows and non-Windows
2884 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2885 libraries
2886 [LINK_FOR_BUILD]: new variable
2887 [OBJEXT]: use throughout
2888 [EXEEXT]: use throughout
2889 <winshared>: new target
2890 <wininstall>: new target
2891 <dftables.o>: use native compiler
2892 <dftables>: use native linker
2893 <install>: handle Windows platform correctly
2894 <clean>: ditto
2895 <check>: ditto
2896 copy DLL to top builddir before testing
2897
2898 As part of these changes, -no-undefined was removed again. This was reported
2899 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2900 in any case.
2901
2902 3. Some tidies to get rid of compiler warnings:
2903
2904 . In the match_data structure, match_limit was an unsigned long int, whereas
2905 match_call_count was an int. I've made them both unsigned long ints.
2906
2907 . In pcretest the fact that a const uschar * doesn't automatically cast to
2908 a void * provoked a warning.
2909
2910 . Turning on some more compiler warnings threw up some "shadow" variables
2911 and a few more missing casts.
2912
2913 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2914 option, a class that contained a single character with a value between 128
2915 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2916
2917 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2918 option, a class that contained several characters, but with at least one
2919 whose value was between 128 and 255 caused PCRE to crash.
2920
2921
2922 Version 4.1 12-Mar-03
2923 ---------------------
2924
2925 1. Compiling with gcc -pedantic found a couple of places where casts were
2926 needed, and a string in dftables.c that was longer than standard compilers are
2927 required to support.
2928
2929 2. Compiling with Sun's compiler found a few more places where the code could
2930 be tidied up in order to avoid warnings.
2931
2932 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2933 first of these names is deprecated in the latest Autoconf in favour of the name
2934 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2935 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2936 analogy I have changed it to CFLAGS_FOR_BUILD.
2937
2938 4. Added -no-undefined to the linking command in the Makefile, because this is
2939 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2940 linking step for the pcreposix library.
2941
2942 5. PCRE was failing to diagnose the case of two named groups with the same
2943 name.
2944
2945 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2946 literal character that is needed in the subject for a match, and scans along to
2947 ensure that it is present before embarking on the full matching process. This
2948 saves time in cases of nested unlimited repeats that are never going to match.
2949 Problem: the scan can take a lot of time if the subject is very long (e.g.
2950 megabytes), thus penalizing straightforward matches. It is now done only if the
2951 amount of subject to be scanned is less than 1000 bytes.
2952
2953 7. A lesser problem with the same optimization is that it was recording the
2954 first character of an anchored pattern as "needed", thus provoking a search
2955 right along the subject, even when the first match of the pattern was going to
2956 fail. The "needed" character is now not set for anchored patterns, unless it
2957 follows something in the pattern that is of non-fixed length. Thus, it still
2958 fulfils its original purpose of finding quick non-matches in cases of nested
2959 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2960
2961
2962 Version 4.0 17-Feb-03
2963 ---------------------
2964
2965 1. If a comment in an extended regex that started immediately after a meta-item
2966 extended to the end of string, PCRE compiled incorrect data. This could lead to
2967 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2968
2969 2. Moved to autoconf 2.53 and libtool 1.4.2.
2970
2971 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2972 the special perltest8 script is no longer needed - all the tests can be run
2973 from a single perltest script.
2974
2975 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2976 by \s. It has now been removed in PCRE. This means it isn't recognized as
2977 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2978 class [:space:] *does* include VT, thereby creating a mess.
2979
2980 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2981 space and tab.
2982
2983 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2984 its new features into the main test script, reducing the number of scripts.
2985
2986 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2987 were backward compatible, and made the (?i) apply to the whole pattern, as if
2988 /i were given. Now it behaves more logically, and applies the option setting
2989 only to what follows. PCRE has been changed to follow suit. However, if it
2990 finds options settings right at the start of the pattern, it extracts them into
2991 the global options, as before. Thus, they show up in the info data.
2992
2993 8. Added support for the \Q...\E escape sequence. Characters in between are
2994 treated as literals. This is slightly different from Perl in that $ and @ are
2995 also handled as literals inside the quotes. In Perl, they will cause variable
2996 interpolation. Note the following examples:
2997
2998 Pattern PCRE matches Perl matches
2999
3000 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3001 \Qabc\$xyz\E abc\$xyz abc\$xyz
3002 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3003
3004 For compatibility with Perl, \Q...\E sequences are recognized inside character
3005 classes as well as outside them.
3006
3007 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3008 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3009 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3010 signed/unsigned warnings.
3011
3012 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3013 option for pcretest, so I've replaced it by a simple function that does just
3014 that job.
3015
3016 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3017 "pcregrep -".
3018
3019 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3020 Java package. This provides some syntactic sugar for simple cases of what my
3021 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3022 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3023 item, you can use this simplified notation. Note that only makes sense with
3024 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3025 greediness, whatever the setting of the PCRE_UNGREEDY option.
3026
3027 13. A change of greediness default within a pattern was not taking effect at
3028 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3029 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3030 was abstracted outside.
3031
3032 14. PCRE now supports the \G assertion. It is true when the current matching
3033 position is at the start point of the match. This differs from \A when the
3034 starting offset is non-zero. Used with the /g option of pcretest (or similar
3035 code), it works in the same way as it does for Perl's /g option. If all
3036 alternatives of a regex begin with \G, the expression is anchored to the start
3037 match position, and the "anchored" flag is set in the compiled expression.
3038
3039 15. Some bugs concerning the handling of certain option changes within patterns
3040 have been fixed. These applied to options other than (?ims). For example,
3041 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3042 been the other way round. Some of this was related to change 7 above.
3043
3044 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3045 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3046 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3047 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3048
3049 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3050 mode. Unlike ".", it always matches newline, whatever the setting of
3051 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3052 assertions. Perl allows it, but it doesn't (in general) work because it can't
3053 calculate the length of the lookbehind. At least, that's the case for Perl
3054 5.8.0 - I've been told they are going to document that it doesn't work in
3055 future.
3056
3057 18. Added an error diagnosis for escapes that PCRE does not support: these are
3058 \L, \l, \N, \P, \p, \U, \u, and \X.
3059
3060 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3061 reading past the end of the pattern in cases such as /[abcd/.
3062
3063 20. PCRE was getting more memory than necessary for patterns with classes that
3064 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3065
3066 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3067 compiling PCRE for use with Virtual Pascal.
3068
3069 22. Small fix to the Makefile to make it work properly if the build is done
3070 outside the source tree.
3071
3072 23. Added a new extension: a condition to go with recursion. If a conditional
3073 subpattern starts with (?(R) the "true" branch is used if recursion has
3074 happened, whereas the "false" branch is used only at the top level.
3075
3076 24. When there was a very long string of literal characters (over 255 bytes
3077 without UTF support, over 250 bytes with UTF support), the computation of how
3078 much memory was required could be incorrect, leading to segfaults or other
3079 strange effects.
3080
3081 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3082 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3083 there was a subsequent back reference to those brackets. This meant that, for
3084 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3085 possible to check for precisely this case. All we can do is abandon the
3086 optimization if .* occurs inside capturing brackets when there are any back
3087 references whatsoever. (See below for a better fix that came later.)
3088
3089 26. The handling of the optimization for finding the first character of a
3090 non-anchored pattern, and for finding a character that is required later in the
3091 match were failing in some cases. This didn't break the matching; it just
3092 failed to optimize when it could. The way this is done has been re-implemented.
3093
3094 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3095
3096 28. Added a new feature that provides some of the functionality that Perl
3097 provides with (?{...}). The facility is termed a "callout". The way it is done
3098 in PCRE is for the caller to provide an optional function, by setting
3099 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3100 global variable. By default it is unset, which disables all calling out. To get
3101 the function called, the regex must include (?C) at appropriate points. This
3102 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3103 This provides a means of identifying different callout points. When PCRE
3104 reaches such a point in the regex, if pcre_callout has been set, the external
3105 function is called. It is provided with data in a structure called
3106 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3107 matching continues; if it returns a non-zero value, the match at the current
3108 point fails. However, backtracking will occur if possible. [This was changed
3109 later and other features added - see item 49 below.]
3110
3111 29. pcretest is upgraded to test the callout functionality. It provides a
3112 callout function that displays information. By default, it shows the start of
3113 the match and the current position in the text. There are some new data escapes
3114 to vary what happens:
3115
3116 \C+ in addition, show current contents of captured substrings
3117 \C- do not supply a callout function
3118 \C!n return 1 when callout number n is reached
3119 \C!n!m return 1 when callout number n is reached for the mth time
3120
3121 30. If pcregrep was called with the -l option and just a single file name, it
3122 output "<stdin>" if a match was found, instead of the file name.
3123
3124 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3125 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3126 pcre_exec(). This saves a malloc/free per call. The default value of
3127 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3128 when configuring.
3129
3130 32. The default maximum size of a compiled pattern is 64K. There have been a
3131 few cases of people hitting this limit. The code now uses macros to handle the
3132 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3133 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3134 configuring. Tests 2 and 5 work only with 2-byte links because they output
3135 debugging information about compiled patterns.
3136
3137 33. Internal code re-arrangements:
3138
3139 (a) Moved the debugging function for printing out a compiled regex into
3140 its own source file (printint.c) and used #include to pull it into
3141 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3142 separate copies.
3143
3144 (b) Defined the list of op-code names for debugging as a macro in
3145 internal.h so that it is next to the definition of the opcodes.
3146
3147 (c) Defined a table of op-code lengths for simpler skipping along compiled
3148 code. This is again a macro in internal.h so that it is next to the
3149 definition of the opcodes.
3150
3151 34. Added support for recursive calls to individual subpatterns, along the
3152 lines of Robin Houston's patch (but implemented somewhat differently).
3153
3154 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3155 allow it to read and process whole directories in Win32. This code was
3156 contributed by Lionel Fourquaux; it has not been tested by me.
3157
3158 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3159 used to name a group. Names consist of alphanumerics and underscores, and must
3160 be unique. Back references use the syntax (?P=name) and recursive calls use
3161 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3162 numbers. The function pcre_fullinfo() can be used after compilation to extract
3163 a name/number map. There are three relevant calls:
3164
3165 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3166 PCRE_INFO_NAMECOUNT yields the number of entries
3167 PCRE_INFO_NAMETABLE yields a pointer to the map.
3168
3169 The map is a vector of fixed-size entries. The size of each entry depends on
3170 the length of the longest name used. The first two bytes of each entry are the
3171 group number, most significant byte first. There follows the corresponding
3172 name, zero terminated. The names are in alphabetical order.
3173
3174 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3175 case instead of 255. Making it the same both with and without UTF-8 support
3176 means that the same test output works with both.
3177
3178 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3179 calling malloc() with a zero argument.
3180
3181 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3182 optimization. I've improved things by keeping a bitmap of backreferences with
3183 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3184 fact referenced, the optimization can be applied. It is unlikely that a
3185 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3186 the match to follow \n) will appear inside brackets with a number greater than
3187 31, but if it does, any back reference > 31 suppresses the optimization.
3188
3189 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3190 of disabling numbered capturing parentheses. Any opening parenthesis that is
3191 not followed by ? behaves as if it were followed by ?: but named parentheses
3192 can still be used for capturing (and they will acquire numbers in the usual
3193 way).
3194
3195 41. Redesigned the return codes from the match() function into yes/no/error so
3196 that errors can be passed back from deep inside the nested calls. A malloc
3197 failure while inside a recursive subpattern call now causes the
3198 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3199
3200 42. It is now possible to set a limit on the number of times the match()
3201 function is called in a call to pcre_exec(). This facility makes it possible to
3202 limit the amount of recursion and backtracking, though not in a directly
3203 obvious way, because the match() function is used in a number of different
3204 circumstances. The count starts from zero for each position in the subject
3205 string (for non-anchored patterns). The default limit is, for compatibility, a
3206 large number, namely 10 000 000. You can change this in two ways:
3207
3208 (a) When configuring PCRE before making, you can use --with-match-limit=n
3209 to set a default value for the compiled library.
3210
3211 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3212 a different value is set. See 45 below.
3213
3214 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3215
3216 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3217 of things that can be changed at compile time. The first argument specifies
3218 what is wanted and the second points to where the information is to be placed.
3219 The current list of available information is:
3220
3221 PCRE_CONFIG_UTF8
3222
3223 The output is an integer that is set to one if UTF-8 support is available;
3224 otherwise it is set to zero.
3225
3226 PCRE_CONFIG_NEWLINE
3227
3228 The output is an integer that it set to the value of the code that is used for
3229 newline. It is either LF (10) or CR (13).
3230
3231 PCRE_CONFIG_LINK_SIZE
3232
3233 The output is an integer that contains the number of bytes used for internal
3234 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3235
3236 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3237
3238 The output is an integer that contains the threshold above which the POSIX
3239 interface uses malloc() for output vectors. See item 31 above.
3240
3241 PCRE_CONFIG_MATCH_LIMIT
3242
3243 The output is an unsigned integer that contains the default limit of the number
3244 of match() calls in a pcre_exec() execution. See 42 above.
3245
3246 44. pcretest has been upgraded by the addition of the -C option. This causes it
3247 to extract all the available output from the new pcre_config() function, and to
3248 output it. The program then exits immediately.
3249
3250 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3251 order to support additional features. One way would have been to define
3252 pcre_exec2() (for example) with extra arguments, but this would not have been
3253 extensible, and would also have required all calls to the original function to
3254 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3255 is used for passing in "extra" data from pcre_study().
3256
3257 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3258 contains the following fields:
3259
3260 flags a bitmap indicating which of the following fields are set
3261 study_data opaque data from pcre_study()
3262 match_limit a way of specifying a limit on match() calls for a specific
3263 call to pcre_exec()
3264 callout_data data for callouts (see 49 below)
3265
3266 The flag bits are also defined in pcre.h, and are
3267
3268 PCRE_EXTRA_STUDY_DATA
3269 PCRE_EXTRA_MATCH_LIMIT
3270 PCRE_EXTRA_CALLOUT_DATA
3271
3272 The pcre_study() function now returns one of these new pcre_extra blocks, with
3273 the actual study data pointed to by the study_data field, and the
3274 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3275 before. That is, this change is entirely upwards-compatible and requires no
3276 change to existing code.
3277
3278 If you want to pass in additional data to pcre_exec(), you can either place it
3279 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3280 block.
3281
3282 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3283 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3284 times with different match limits, until it finds the minimum value needed for
3285 pcre_exec() to complete. The value is then output. This can be instructive; for
3286 most simple matches the number is quite small, but for pathological cases it
3287 gets very large very quickly.
3288
3289 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3290 returns the size of the data block pointed to by the study_data field in a
3291 pcre_extra block, that is, the value that was passed as the argument to
3292 pcre_malloc() when PCRE was getting memory in which to place the information
3293 created by pcre_study(). The fourth argument should point to a size_t variable.
3294 pcretest has been extended so that this information is shown after a successful
3295 pcre_study() call when information about the compiled regex is being displayed.
3296
3297 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3298 because what follows is always an absolute path. (Later: it turns out that this
3299 is more than cosmetic for MinGW, because it doesn't like empty path
3300 components.)
3301
3302 49. Some changes have been made to the callout feature (see 28 above):
3303
3304 (i) A callout function now has three choices for what it returns:
3305
3306 0 => success, carry on matching
3307 > 0 => failure at this point, but backtrack if possible
3308 < 0 => serious error, return this value from pcre_exec()
3309
3310 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3311 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3312 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3313 use by callout functions. It will never be used by PCRE itself.
3314
3315 (ii) The pcre_extra structure (see 45 above) has a void * field called
3316 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3317 pcre_callout_block structure has a field of the same name. The contents of
3318 the field passed in the pcre_extra structure are passed to the callout
3319 function in the corresponding field in the callout block. This makes it
3320 easier to use the same callout-containing regex from multiple threads. For
3321 testing, the pcretest program has a new data escape
3322
3323 \C*n pass the number n (may be negative) as callout_data
3324
3325 If the callout function in pcretest receives a non-zero value as
3326 callout_data, it returns that value.
3327
3328 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3329 there were some redundant $(CFLAGS) in commands that are now specified as
3330 $(LINK), which already includes $(CFLAGS).
3331
3332 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3333 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3334 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3335 one-byte characters throughout. Note that case-insensitive matching applies
3336 only to characters whose values are less than 256. PCRE doesn't support the
3337 notion of cases for higher-valued characters.
3338
3339 (i) A character class whose characters are all within 0-255 is handled as
3340 a bit map, and the map is inverted for negative classes. Previously, a
3341 character > 255 always failed to match such a class; however it should
3342 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3343
3344 (ii) A negated character class with a single character < 255 is coded as
3345 "not this character" (OP_NOT). This wasn't working properly when the test
3346 character was multibyte, either singly or repeated.
3347
3348 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3349 mode, for example: \x{100}{2,3}.
3350
3351 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3352 singly or repeated) now correctly test multibyte characters. However,
3353 PCRE doesn't recognize any characters with values greater than 255 as
3354 digits, spaces, or word characters. Such characters always match \D, \S,
3355 and \W, and never match \d, \s, or \w.
3356
3357 (v) Classes may now contain characters and character ranges with values
3358 greater than 255. For example: [ab\x{100}-\x{400}].
3359
3360 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3361 PCRE in UTF-8 mode.
3362
3363 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3364 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3365 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3366 value.)
3367
3368 53. The single man page has become too large. I have therefore split it up into
3369 a number of separate man pages. These also give rise to individual HTML pages;
3370 these are now put in a separate directory, and there is an index.html page that
3371 lists them all. Some hyperlinking between the pages has been installed.
3372
3373 54. Added convenience functions for handling named capturing parentheses.
3374
3375 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3376 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3377 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3378 are faulted.
3379
3380 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3381 calling configure. These values are used when compiling the dftables.c program
3382 which is run to generate the source of the default character tables. They
3383 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3384 you will need to set these values.
3385
3386 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3387
3388
3389 Version 3.9 02-Jan-02
3390 ---------------------
3391
3392 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3393
3394 2. If --disable-static was given, the building process failed when trying to
3395 build pcretest and pcregrep. (For some reason it was using libtool to compile
3396 them, which is not right, as they aren't part of the library.)
3397
3398
3399 Version 3.8 18-Dec-01
3400 ---------------------
3401
3402 1. The experimental UTF-8 code was completely screwed up. It was packing the
3403 bytes in the wrong order. How dumb can you get?
3404
3405
3406 Version 3.7 29-Oct-01
3407 ---------------------
3408
3409 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3410 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3411 this didn't happen under Solaris 8, where I normally test things.
3412
3413 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3414 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3415 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3416 this fix an hour or so after the initial 3.7 release.)
3417
3418
3419 Version 3.6 23-Oct-01
3420 ---------------------
3421
3422 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3423 offsets passed as NULL with zero offset count.
3424
3425 2. The config.guess and config.sub files had not been updated when I moved to
3426 the latest autoconf.
3427
3428
3429 Version 3.5 15-Aug-01
3430 ---------------------
3431
3432 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3433 had been forgotten.
3434
3435 2. By using declared but undefined structures, we can avoid using "void"
3436 definitions in pcre.h while keeping the internal definitions of the structures
3437 private.
3438
3439 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3440 user point of view, this means that both static and shared libraries are built
3441 by default, but this can be individually controlled. More of the work of
3442 handling this static/shared cases is now inside libtool instead of PCRE's make
3443 file.
3444
3445 4. The pcretest utility is now installed along with pcregrep because it is
3446 useful for users (to test regexs) and by doing this, it automatically gets
3447 relinked by libtool. The documentation has been turned into a man page, so
3448 there are now .1, .txt, and .html versions in /doc.
3449
3450 5. Upgrades to pcregrep:
3451 (i) Added long-form option names like gnu grep.
3452 (ii) Added --help to list all options with an explanatory phrase.
3453 (iii) Added -r, --recursive to recurse into sub-directories.
3454 (iv) Added -f, --file to read patterns from a file.
3455
3456 6. pcre_exec() was referring to its "code" argument before testing that
3457 argument for NULL (and giving an error if it was NULL).
3458
3459 7. Upgraded Makefile.in to allow for compiling in a different directory from
3460 the source directory.
3461
3462 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3463 options bits, the pointer it was passed was to an int instead of to an unsigned
3464 long int. This mattered only on 64-bit systems.
3465
3466 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3467 generated) instead of pcre.in, which it its source. Also made the same change
3468 in several of the .c files.
3469
3470 10. A new release of gcc defines printf() as a macro, which broke pcretest
3471 because it had an ifdef in the middle of a string argument for printf(). Fixed
3472 by using separate calls to printf().
3473
3474 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3475 script, to force use of CR or LF instead of \n in the source. On non-Unix
3476 systems, the value can be set in config.h.
3477
3478 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3479 absolute limit. Changed the text of the error message to make this clear, and
3480 likewise updated the man page.
3481
3482 13. The limit of 99 on the number of capturing subpatterns has been removed.
3483 The new limit is 65535, which I hope will not be a "real" limit.
3484
3485
3486 Version 3.4 22-Aug-00
3487 ---------------------
3488
3489 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3490
3491 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3492
3493
3494 Version 3.3 01-Aug-00
3495 ---------------------
3496
3497 1. If an octal character was given, but the value was greater than \377, it
3498 was not getting masked to the least significant bits, as documented. This could
3499 lead to crashes in some systems.
3500
3501 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3502 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3503
3504 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3505 These just pass their arguments on to (pcre_free)(), but they are provided
3506 because some uses of PCRE bind it to non-C systems that can call its functions,
3507 but cannot call free() or pcre_free() directly.
3508
3509 4. Add "make test" as a synonym for "make check". Corrected some comments in
3510 the Makefile.
3511
3512 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3513 Makefile.
3514
3515 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3516 command called pgrep for grepping around the active processes.
3517
3518 7. Added the beginnings of support for UTF-8 character strings.
3519
3520 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3521 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3522 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3523 out for the ar command.)
3524
3525
3526 Version 3.2 12-May-00
3527 ---------------------
3528
3529 This is purely a bug fixing release.
3530
3531 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3532 of ZA. This was just one example of several cases that could provoke this bug,
3533 which was introduced by change 9 of version 2.00. The code for breaking
3534 infinite loops after an iteration that matches an empty string was't working
3535 correctly.
3536
3537 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3538 when matched against abbab (for example). After matching an empty string, it
3539 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3540 caused it to match further down the string than it should.
3541
3542 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3543 was there because it doesn't seem to be needed, and it causes trouble on some
3544 systems, as it is not a Standard C header. It has been removed.
3545
3546 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3547 were reported on the Macintosh. The changes were from
3548
3549 while ((c = *(++ptr)) != 0 && c != '\n');
3550 to
3551 while ((c = *(++ptr)) != 0 && c != '\n') ;
3552
3553 Totally extraordinary, but if that's what it takes...
3554
3555 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3556 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3557 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3558 assumes the way PCRE uses memmove() (always moving upwards).
3559
3560 6. PCRE is being used in one environment where strchr() is not available. There
3561 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3562 faster code anyway.
3563
3564
3565 Version 3.1 09-Feb-00
3566 ---------------------
3567
3568 The only change in this release is the fixing of some bugs in Makefile.in for
3569 the "install" target:
3570
3571 (1) It was failing to install pcreposix.h.
3572
3573 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3574
3575
3576 Version 3.0 01-Feb-00
3577 ---------------------
3578
3579 1. Add support for the /+ modifier to perltest (to output $` like it does in
3580 pcretest).
3581
3582 2. Add support for the /g modifier to perltest.
3583
3584 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3585 matches null strings.
3586
3587 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3588 pattern. Perl treats empty patterns specially - it reuses the most recent
3589 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3590 effect.
3591
3592 5. The POSIX interface was broken in that it was just handing over the POSIX
3593 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3594 required a bigger vector, with some working space on the end. This means that
3595 the POSIX wrapper now has to get and free some memory, and copy the results.
3596
3597 6. Added some simple autoconf support, placing the test data and the
3598 documentation in separate directories, re-organizing some of the
3599 information files, and making it build pcre-config (a GNU standard). Also added
3600 libtool support for building PCRE as a shared library, which is now the
3601 default.
3602
3603 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3604 09 are not valid octal constants. Single digits will be used for minor values
3605 less than 10.
3606
3607 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3608 existing programs that set these in the POSIX interface can use PCRE without
3609 modification.
3610
3611 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3612 return all that pcre_info() returns, plus additional data. The pcre_info()
3613 function is retained for compatibility, but is considered to be obsolete.
3614
3615 10. Added experimental recursion feature (?R) to handle one common case that
3616 Perl 5.6 will be able to do with (?p{...}).
3617
3618 11. Added support for POSIX character classes like [:alpha:], which Perl is
3619 adopting.
3620
3621
3622 Version 2.08 31-Aug-99
3623 ----------------------
3624
3625 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3626 trying to match at the startoffset position, but instead was moving forward to
3627 the next newline as if a previous match had failed.
3628
3629 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3630 and could get into a loop if a null string was matched other than at the start
3631 of the subject.
3632
3633 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3634 be distinguished at compile time, and for completeness also added PCRE_DATE.
3635
3636 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3637 in GnuWin32 environments.
3638
3639
3640 Version 2.07 29-Jul-99
3641 ----------------------
3642
3643 1. The documentation is now supplied in plain text form and HTML as well as in
3644 the form of man page sources.
3645
3646 2. C++ compilers don't like assigning (void *) values to other pointer types.
3647 In particular this affects malloc(). Although there is no problem in Standard
3648 C, I've put in casts to keep C++ compilers happy.
3649
3650 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3651 should be (const char *).
3652
3653 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3654 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3655 However, I haven't made this a standard facility. The documentation doesn't
3656 mention it, and the Makefile doesn't support it.
3657
3658 5. The Makefile now contains an "install" target, with editable destinations at
3659 the top of the file. The pcretest program is not installed.
3660
3661 6. pgrep -V now gives the PCRE version number and date.
3662
3663 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3664 causing the entire string to be ignored, instead of just the last character.
3665
3666 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3667 non-matching string, it can take a very, very long time, even for strings of
3668 quite modest length, because of the nested recursion. PCRE now does better in
3669 some of these cases. It does this by remembering the last required literal
3670 character in the pattern, and pre-searching the subject to ensure it is present
3671 before running the real match. In other words, it applies a heuristic to detect
3672 some types of certain failure quickly, and in the above example, if presented
3673 with a string that has no trailing " it gives "no match" very quickly.
3674
3675 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3676 other alternatives are tried instead.
3677
3678
3679 Version 2.06 09-Jun-99
3680 ----------------------
3681
3682 1. Change pcretest's output for amount of store used to show just the code
3683 space, because the remainder (the data block) varies in size between 32-bit and
3684 64-bit systems.
3685
3686 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3687 start matching at. This allows lookbehinds to work when searching for multiple
3688 occurrences in a string.
3689
3690 3. Added additional options to pcretest for testing multiple occurrences:
3691
3692 /+ outputs the rest of the string that follows a match
3693 /g loops for multiple occurrences, using the new startoffset argument
3694 /G loops for multiple occurrences by passing an incremented pointer
3695
3696 4. PCRE wasn't doing the "first character" optimization for patterns starting
3697 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3698 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3699 the letter 'x'. On long subject strings, this gives a significant speed-up.
3700
3701
3702 Version 2.05 21-Apr-99
3703 ----------------------
3704
3705 1. Changed the type of magic_number from int to long int so that it works
3706 properly on 16-bit systems.
3707
3708 2. Fixed a bug which caused patterns starting with .* not to work correctly
3709 when the subject string contained newline characters. PCRE was assuming
3710 anchoring for such patterns in all cases, which is not correct because .* will
3711 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3712 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3713 must be retried after every newline in the subject.
3714
3715
3716 Version 2.04 18-Feb-99
3717 ----------------------
3718
3719 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3720 computation of the store needed to hold the pattern was incorrect (too large).
3721 If such patterns were nested a few deep, this could multiply and become a real
3722 problem.
3723
3724 2. Added /M option to pcretest to show the memory requirement of a specific
3725 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3726
3727 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3728 compiled in such a way that the backtracking after subsequent failure was
3729 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3730 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3731
3732
3733 Version 2.03 02-Feb-99
3734 ----------------------
3735
3736 1. Fixed typo and small mistake in man page.
3737
3738 2. Added 4th condition (GPL supersedes if conflict) and created separate
3739 LICENCE file containing the conditions.
3740
3741 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3742 Perl, that is the internal \ allows the delimiter to be included in the
3743 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3744 the final delimiter, add \ to the end of the pattern (to test the error).
3745
3746 4. Added the convenience functions for extracting substrings after a successful
3747 match. Updated pcretest to make it able to test these functions.
3748
3749
3750 Version 2.02 14-Jan-99
3751 ----------------------
3752
3753 1. Initialized the working variables associated with each extraction so that
3754 their saving and restoring doesn't refer to uninitialized store.
3755
3756 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3757 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3758 fix the problem.
3759
3760 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3761 calls, and wasn't printing the correct value for compiling calls. Increased the
3762 default value of LOOPREPEAT, and the number of significant figures in the
3763 times.
3764
3765 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3766
3767 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3768 a building problem on Windows NT with a FAT file system.
3769
3770
3771 Version 2.01 21-Oct-98
3772 ----------------------
3773
3774 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3775 to character tables built by pcre_maketables() in the current locale. If NULL
3776 is passed, the default tables are used.
3777
3778
3779 Version 2.00 24-Sep-98
3780 ----------------------
3781
3782 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3783 it any more.
3784
3785 2. Allow quantification of (?>) groups, and make it work correctly.
3786
3787 3. The first character computation wasn't working for (?>) groups.
3788
3789 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3790 end of the subject) and add 5.005's \z, which really does match only at the
3791 very end of the subject.
3792
3793 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3794
3795 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3796 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3797 localized options. All options to pcre_study() were also removed.
3798
3799 7. Add other new features from 5.005:
3800
3801 $(?<= positive lookbehind
3802 $(?<! negative lookbehind
3803 (?imsx-imsx) added the unsetting capability
3804 such a setting is global if at outer level; local otherwise
3805 (?imsx-imsx:) non-capturing groups with option setting
3806 (?(cond)re|re) conditional pattern matching
3807
3808 A backreference to itself in a repeated group matches the previous
3809 captured string.
3810
3811 8. General tidying up of studying (both automatic and via "study")
3812 consequential on the addition of new assertions.
3813
3814 9. As in 5.005, unlimited repeated groups that could match an empty substring
3815 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3816 runtime if any iteration does actually match an empty substring.
3817
3818 10. Include the RunTest script in the distribution.
3819
3820 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3821 discrepancies, some of which were old and were also with respect to 5.004. They
3822 have now been fixed.
3823
3824
3825 Version 1.09 28-Apr-98
3826 ----------------------
3827
3828 1. A negated single character class followed by a quantifier with a minimum
3829 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3830 program crashes, or just wrong answers. This did not apply to negated classes
3831 containing more than one character, or to minima other than one.
3832
3833
3834 Version 1.08 27-Mar-98
3835 ----------------------
3836
3837 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3838
3839 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3840 latter must appear before anything that relies on it in the pattern.
3841
3842
3843 Version 1.07 16-Feb-98
3844 ----------------------
3845
3846 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3847 repeat of a potentially empty string).
3848
3849
3850 Version 1.06 23-Jan-98
3851 ----------------------
3852
3853 1. Added Markus Oberhumer's little patches for C++.
3854
3855 2. Literal strings longer than 255 characters were broken.
3856
3857
3858 Version 1.05 23-Dec-97
3859 ----------------------
3860
3861 1. Negated character classes containing more than one character were failing if
3862 PCRE_CASELESS was set at run time.
3863
3864
3865 Version 1.04 19-Dec-97
3866 ----------------------
3867
3868 1. Corrected the man page, where some "const" qualifiers had been omitted.
3869
3870 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3871 input syntax.
3872
3873 3. Fixed memory leak which occurred when a regex with back references was
3874 matched with an offsets vector that wasn't big enough. The temporary memory
3875 that is used in this case wasn't being freed if the match failed.
3876
3877 4. Tidied pcretest to ensure it frees memory that it gets.
3878
3879 5. Temporary memory was being obtained in the case where the passed offsets
3880 vector was exactly big enough.
3881
3882 6. Corrected definition of offsetof() from change 5 below.
3883
3884 7. I had screwed up change 6 below and broken the rules for the use of
3885 setjmp(). Now fixed.
3886
3887
3888 Version 1.03 18-Dec-97
3889 ----------------------
3890
3891 1. A erroneous regex with a missing opening parenthesis was correctly
3892 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3893 on some systems.
3894
3895 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3896 it was reported that one broken compiler failed on the former because "code" is
3897 also an independent variable.
3898
3899 3. The erroneous regex a[]b caused an array overrun reference.
3900
3901 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3902 fail on data ending with that character. (It was going on too far, and checking
3903 the next character, typically a binary zero.) This was specific to the
3904 optimized code for single-character negative classes.
3905
3906 5. Added a contributed patch from the TIN world which does the following:
3907
3908 + Add an undef for memmove, in case the the system defines a macro for it.
3909
3910 + Add a definition of offsetof(), in case there isn't one. (I don't know
3911 the reason behind this - offsetof() is part of the ANSI standard - but
3912 it does no harm).
3913
3914 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3915 most of the places where whitespace preceded '#'. I have given up and
3916 allowed the remaining 2 cases to be at the margin.
3917
3918 + Rename some variables in pcre to eliminate shadowing. This seems very
3919 pedantic, but does no harm, of course.
3920
3921 6. Moved the call to setjmp() into its own function, to get rid of warnings
3922 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3923
3924 7. Constructs such as \d{8,} were compiling into the equivalent of
3925 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3926 outcome, but in this particular case used more store than had been allocated,
3927 which caused the bug to be discovered because it threw up an internal error.
3928
3929 8. The debugging code in both pcre and pcretest for outputting the compiled
3930 form of a regex was going wrong in the case of back references followed by
3931 curly-bracketed repeats.
3932
3933
3934 Version 1.02 12-Dec-97
3935 ----------------------
3936
3937 1. Typos in pcre.3 and comments in the source fixed.
3938
3939 2. Applied a contributed patch to get rid of places where it used to remove
3940 'const' from variables, and fixed some signed/unsigned and uninitialized
3941 variable warnings.
3942
3943 3. Added the "runtest" target to Makefile.
3944
3945 4. Set default compiler flag to -O2 rather than just -O.
3946
3947
3948 Version 1.01 19-Nov-97
3949 ----------------------
3950
3951 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3952 like /([ab]*)*/, that is, for classes with more than one character in them.
3953
3954 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3955 as /((?>a*))*/ (a PCRE_EXTRA facility).
3956
3957
3958 Version 1.00 18-Nov-97
3959 ----------------------
3960
3961 1. Added compile-time macros to support systems such as SunOS4 which don't have
3962 memmove() or strerror() but have other things that can be used instead.
3963
3964 2. Arranged that "make clean" removes the executables.
3965
3966
3967 Version 0.99 27-Oct-97
3968 ----------------------
3969
3970 1. Fixed bug in code for optimizing classes with only one character. It was
3971 initializing a 32-byte map regardless, which could cause it to run off the end
3972 of the memory it had got.
3973
3974 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3975
3976
3977 Version 0.98 22-Oct-97
3978 ----------------------
3979
3980 1. Fixed bug in code for handling temporary memory usage when there are more
3981 back references than supplied space in the ovector. This could cause segfaults.
3982
3983
3984 Version 0.97 21-Oct-97
3985 ----------------------
3986
3987 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3988
3989 2. Optimized negated single characters not to use a bit map.
3990
3991 3. Brought error texts together as macro definitions; clarified some of them;
3992 fixed one that was wrong - it said "range out of order" when it meant "invalid
3993 escape sequence".
3994
3995 4. Changed some char * arguments to const char *.
3996
3997 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3998
3999 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4000 pcretest.
4001
4002
4003 Version 0.96 16-Oct-97
4004 ----------------------
4005
4006 1. Added a simple "pgrep" utility to the distribution.
4007
4008 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4009 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4010 where "ddd" means "one or more decimal digits".
4011
4012 3. Fixed serious bug. If a pattern had a back reference, but the call to
4013 pcre_exec() didn't supply a large enough ovector to record the related
4014 identifying subpattern, the match always failed. PCRE now remembers the number
4015 of the largest back reference, and gets some temporary memory in which to save
4016 the offsets during matching if necessary, in order to ensure that
4017 backreferences always work.
4018
4019 4. Increased the compatibility with Perl in a number of ways:
4020
4021 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4022 to request this handling. The option can be set at compile or exec time.
4023
4024 (b) $ matches before a terminating newline by default; an option
4025 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4026 mode). The option can be set at compile or exec time.
4027
4028 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4029 the same as Perl's. If the decimal number it represents is less than 10
4030 or there aren't that many previous left capturing parentheses, an octal
4031 escape is read. Inside a character class, it's always an octal escape,
4032 even if it is a single digit.
4033
4034 (d) An escaped but undefined alphabetic character is taken as a literal,
4035 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4036 escapes.
4037
4038 (e) {0} is now permitted. (The previous item is removed from the compiled
4039 pattern).
4040
4041 5. Changed all the names of code files so that the basic parts are no longer
4042 than 10 characters, and abolished the teeny "globals.c" file.
4043
4044 6. Changed the handling of character classes; they are now done with a 32-byte
4045 bit map always.
4046
4047 7. Added the -d and /D options to pcretest to make it possible to look at the
4048 internals of compilation without having to recompile pcre.
4049
4050
4051 Version 0.95 23-Sep-97
4052 ----------------------
4053
4054 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4055 \x20 at the start of a run of normal characters. These were being treated as
4056 real characters, instead of the source characters being re-checked.
4057
4058
4059 Version 0.94 18-Sep-97
4060 ----------------------
4061
4062 1. The functions are now thread-safe, with the caveat that the global variables
4063 containing pointers to malloc() and free() or alternative functions are the
4064 same for all threads.
4065
4066 2. Get pcre_study() to generate a bitmap of initial characters for non-
4067 anchored patterns when this is possible, and use it if passed to pcre_exec().
4068
4069
4070 Version 0.93 15-Sep-97
4071 ----------------------
4072
4073 1. /(b)|(:+)/ was computing an incorrect first character.
4074
4075 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4076 but not actually doing anything yet.
4077
4078 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4079 as Perl does (e.g. [-az] or [az-]).
4080
4081 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4082 all possible positions.
4083
4084 5. Split up into different modules to avoid including unneeded functions in a
4085 compiled binary. However, compile and exec are still in one module. The "study"
4086 function is split off.
4087
4088 6. The character tables are now in a separate module whose source is generated
4089 by an auxiliary program - but can then be edited by hand if required. There are
4090 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4091 toupper() in the code.
4092
4093 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4094 make them global. Abolish the function for setting them, as the caller can now
4095 set them directly.
4096
4097
4098 Version 0.92 11-Sep-97
4099 ----------------------
4100
4101 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4102 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4103
4104 2. Caseless matching was not working in character classes if the characters in
4105 the pattern were in upper case.
4106
4107 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4108
4109 4. Make PCRE_ANCHORED public and accept as a compile option.
4110
4111 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4112 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4113 pass them.
4114
4115 6. Give an error if bad option bits passed at compile or run time.
4116
4117 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4118 pcretest to cause it to pass that flag.
4119
4120 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4121 options, and the first character, if set.
4122
4123 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4124
4125
4126 Version 0.91 10-Sep-97
4127 ----------------------
4128
4129 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4130 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4131
4132 2. PCRE was looping on encountering an indefinitely repeated back reference to
4133 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4134 Perl does - treats the match as successful.
4135
4136 ****

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