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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 598 - (show annotations)
Sat May 7 15:37:31 2011 UTC (4 years ago) by ph10
File size: 200100 byte(s)
Error occurred while calculating annotation data.
Pass back detailed info when UTF-8 check fails at runtime.
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 4. There were a number of related bugs in the code for matching backrefences
16 caselessly in UTF-8 mode when codes for the characters concerned were
17 different numbers of bytes. For example, U+023A and U+2C65 are an upper
18 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
19 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
20 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
21 2-byte code at the end of the subject (it thought there wasn't enough data
22 left).
23
24 5. Comprehensive information about what went wrong is now returned by
25 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
26 as the output vector has at least 2 elements. The offset of the start of
27 the failing character and a reason code are placed in the vector.
28
29 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
30 now returned is for the first byte of the failing character, instead of the
31 last byte inspected. This is an incompatible change, but I hope it is small
32 enough not to be a problem. It makes the returned offset consistent with
33 pcre_exec() and pcre_dfa_exec().
34
35 7. pcretest now gives a text phrase as well as the error number when
36 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
37 failure, the offset and reason code are output.
38
39
40
41 Version 8.12 15-Jan-2011
42 ------------------------
43
44 1. Fixed some typos in the markup of the man pages, and wrote a script that
45 checks for such things as part of the documentation building process.
46
47 2. On a big-endian 64-bit system, pcregrep did not correctly process the
48 --match-limit and --recursion-limit options (added for 8.11). In
49 particular, this made one of the standard tests fail. (The integer value
50 went into the wrong half of a long int.)
51
52 3. If the --colour option was given to pcregrep with -v (invert match), it
53 did strange things, either producing crazy output, or crashing. It should,
54 of course, ignore a request for colour when reporting lines that do not
55 match.
56
57 4. Another pcregrep bug caused similar problems if --colour was specified with
58 -M (multiline) and the pattern match finished with a line ending.
59
60 5. In pcregrep, when a pattern that ended with a literal newline sequence was
61 matched in multiline mode, the following line was shown as part of the
62 match. This seems wrong, so I have changed it.
63
64 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
65 the check for further matches in the same line (so they could be coloured)
66 to overrun the end of the current line. If another match was found, it was
67 incorrectly shown (and then shown again when found in the next line).
68
69 7. If pcregrep was compiled under Windows, there was a reference to the
70 function pcregrep_exit() before it was defined. I am assuming this was
71 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
72 reported by a user. I've moved the definition above the reference.
73
74
75 Version 8.11 10-Dec-2010
76 ------------------------
77
78 1. (*THEN) was not working properly if there were untried alternatives prior
79 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
80 backtracked to try for "b" instead of moving to the next alternative branch
81 at the same level (in this case, to look for "c"). The Perl documentation
82 is clear that when (*THEN) is backtracked onto, it goes to the "next
83 alternative in the innermost enclosing group".
84
85 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
86 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
87 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
88 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
89 (*THEN).
90
91 3. If \s appeared in a character class, it removed the VT character from
92 the class, even if it had been included by some previous item, for example
93 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
94 of \s, but is part of the POSIX "space" class.)
95
96 4. A partial match never returns an empty string (because you can always
97 match an empty string at the end of the subject); however the checking for
98 an empty string was starting at the "start of match" point. This has been
99 changed to the "earliest inspected character" point, because the returned
100 data for a partial match starts at this character. This means that, for
101 example, /(?<=abc)def/ gives a partial match for the subject "abc"
102 (previously it gave "no match").
103
104 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
105 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
106 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
107 has an implication that the given string is incomplete (because a partial
108 match is preferred over a full match). For this reason, these items now
109 give a partial match in this situation. [Aside: previously, the one case
110 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
111 match rather than a full match, which was wrong by the old rules, but is
112 now correct.]
113
114 6. There was a bug in the handling of #-introduced comments, recognized when
115 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
116 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
117 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
118 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
119 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
120 places in pcre_compile().
121
122 7. Related to (6) above, when pcre_compile() was skipping #-introduced
123 comments when looking ahead for named forward references to subpatterns,
124 the only newline sequence it recognized was NL. It now handles newlines
125 according to the set newline convention.
126
127 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
128 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
129 cater for a lack of strerror(). These oversights have been fixed.
130
131 9. Added --match-limit and --recursion-limit to pcregrep.
132
133 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
134
135 11. When the -o option was used, pcregrep was setting a return code of 1, even
136 when matches were found, and --line-buffered was not being honoured.
137
138 12. Added an optional parentheses number to the -o and --only-matching options
139 of pcregrep.
140
141 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
142 can match an empty string. The code to do it in pcretest and pcredemo
143 needed fixing:
144
145 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
146 only one byte after an empty string match just before CRLF (this case
147 just got forgotten; "any" and "anycrlf" were OK).
148
149 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
150 mode when an empty string match preceded an ASCII character followed by
151 a non-ASCII character. (The code for advancing by one character rather
152 than one byte was nonsense.)
153
154 (c) The pcredemo.c sample program did not have any code at all to handle
155 the cases when CRLF is a valid newline sequence.
156
157 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
158 as a starting offset was within the subject string. There is now a new
159 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
160 negative or greater than the length of the string. In order to test this,
161 pcretest is extended to allow the setting of negative starting offsets.
162
163 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
164 starting offset points to the beginning of a UTF-8 character was
165 unnecessarily clumsy. I tidied it up.
166
167 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
168 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
169
170 17. Nobody had reported that the --include_dir option, which was added in
171 release 7.7 should have been called --include-dir (hyphen, not underscore)
172 for compatibility with GNU grep. I have changed it to --include-dir, but
173 left --include_dir as an undocumented synonym, and the same for
174 --exclude-dir, though that is not available in GNU grep, at least as of
175 release 2.5.4.
176
177 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
178 characters from a string of bytes) have been redefined so as not to use
179 loops, in order to improve performance in some environments. At the same
180 time, I abstracted some of the common code into auxiliary macros to save
181 repetition (this should not affect the compiled code).
182
183 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
184 compile-time error is now given if \c is not followed by an ASCII
185 character, that is, a byte less than 128. (In EBCDIC mode, the code is
186 different, and any byte value is allowed.)
187
188 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
189 START_OPTIMIZE option, which is now allowed at compile time - but just
190 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
191 to pcregrep and other applications that have no direct access to PCRE
192 options. The new /Y option in pcretest sets this option when calling
193 pcre_compile().
194
195 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
196 back references. Groups containing recursive back references were forced to
197 be atomic by that change, but in the case of named groups, the amount of
198 memory required was incorrectly computed, leading to "Failed: internal
199 error: code overflow". This has been fixed.
200
201 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
202 pcretest.c, to avoid build problems in some Borland environments.
203
204
205 Version 8.10 25-Jun-2010
206 ------------------------
207
208 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
209 THEN.
210
211 2. (*ACCEPT) was not working when inside an atomic group.
212
213 3. Inside a character class, \B is treated as a literal by default, but
214 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
215 causes the error). The code is unchanged, but I tidied the documentation.
216
217 4. Inside a character class, PCRE always treated \R and \X as literals,
218 whereas Perl faults them if its -w option is set. I have changed PCRE so
219 that it faults them when PCRE_EXTRA is set.
220
221 5. Added support for \N, which always matches any character other than
222 newline. (It is the same as "." when PCRE_DOTALL is not set.)
223
224 6. When compiling pcregrep with newer versions of gcc which may have
225 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
226 declared with attribute warn_unused_result" were given. Just casting the
227 result to (void) does not stop the warnings; a more elaborate fudge is
228 needed. I've used a macro to implement this.
229
230 7. Minor change to pcretest.c to avoid a compiler warning.
231
232 8. Added four artifical Unicode properties to help with an option to make
233 \s etc use properties (see next item). The new properties are: Xan
234 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
235
236 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
237 use Unicode properties. (*UCP) at the start of a pattern can be used to set
238 this option. Modified pcretest to add /W to test this facility. Added
239 REG_UCP to make it available via the POSIX interface.
240
241 10. Added --line-buffered to pcregrep.
242
243 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
244 studied, and the match started with a letter with a code point greater than
245 127 whose first byte was different to the first byte of the other case of
246 the letter, the other case of this starting letter was not recognized
247 (#976).
248
249 12. If a pattern that was studied started with a repeated Unicode property
250 test, for example, \p{Nd}+, there was the theoretical possibility of
251 setting up an incorrect bitmap of starting bytes, but fortunately it could
252 not have actually happened in practice until change 8 above was made (it
253 added property types that matched character-matching opcodes).
254
255 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
256 possible starting bytes for non-anchored patterns.
257
258 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
259 \R, and also a number of cases that involve Unicode properties, both
260 explicit and implicit when PCRE_UCP is set.
261
262 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
263 input, it could crash or give wrong results if characters with values
264 greater than 0xc0 were present in the subject string. (Detail: it assumed
265 UTF-8 input when processing these items.)
266
267 16. Added a lot of (int) casts to avoid compiler warnings in systems where
268 size_t is 64-bit (#991).
269
270 17. Added a check for running out of memory when PCRE is compiled with
271 --disable-stack-for-recursion (#990).
272
273 18. If the last data line in a file for pcretest does not have a newline on
274 the end, a newline was missing in the output.
275
276 19. The default pcre_chartables.c file recognizes only ASCII characters (values
277 less than 128) in its various bitmaps. However, there is a facility for
278 generating tables according to the current locale when PCRE is compiled. It
279 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
280 characters, are recognized by isspace() and therefore were getting set in
281 these tables, and indeed these tables seem to approximate to ISO 8859. This
282 caused a problem in UTF-8 mode when pcre_study() was used to create a list
283 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
284 which of course cannot start UTF-8 characters. I have changed the code so
285 that only real ASCII characters (less than 128) and the correct starting
286 bytes for UTF-8 encodings are set for characters greater than 127 when in
287 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
288 altogether.)
289
290 20. Added the /T option to pcretest so as to be able to run tests with non-
291 standard character tables, thus making it possible to include the tests
292 used for 19 above in the standard set of tests.
293
294 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
295 reference to a subpattern the other side of a comment that contains an
296 opening parenthesis caused either an internal compiling error, or a
297 reference to the wrong subpattern.
298
299
300 Version 8.02 19-Mar-2010
301 ------------------------
302
303 1. The Unicode data tables have been updated to Unicode 5.2.0.
304
305 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
306 configured.
307
308 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
309 original author of that file, following a query about its status.
310
311 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
312 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
313
314 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
315 quantifier applied to a forward-referencing subroutine call, could compile
316 incorrect code or give the error "internal error: previously-checked
317 referenced subpattern not found".
318
319 6. Both MS Visual Studio and Symbian OS have problems with initializing
320 variables to point to external functions. For these systems, therefore,
321 pcre_malloc etc. are now initialized to local functions that call the
322 relevant global functions.
323
324 7. There were two entries missing in the vectors called coptable and poptable
325 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
326 I've fixed the data, and added a kludgy way of testing at compile time that
327 the lengths are correct (equal to the number of opcodes).
328
329 8. Following on from 7, I added a similar kludge to check the length of the
330 eint vector in pcreposix.c.
331
332 9. Error texts for pcre_compile() are held as one long string to avoid too
333 much relocation at load time. To find a text, the string is searched,
334 counting zeros. There was no check for running off the end of the string,
335 which could happen if a new error number was added without updating the
336 string.
337
338 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
339
340 11. \K was not working if it appeared in an atomic group or in a group that
341 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
342 \K is "not well defined" if used in an assertion. PCRE now accepts it if
343 the assertion is positive, but not if it is negative.
344
345 12. Change 11 fortuitously reduced the size of the stack frame used in the
346 "match()" function of pcre_exec.c by one pointer. Forthcoming
347 implementation of support for (*MARK) will need an extra pointer on the
348 stack; I have reserved it now, so that the stack frame size does not
349 decrease.
350
351 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
352 item in branch that calls a recursion is a subroutine call - as in the
353 second branch in the above example - was incorrectly given the compile-
354 time error "recursive call could loop indefinitely" because pcre_compile()
355 was not correctly checking the subroutine for matching a non-empty string.
356
357 14. The checks for overrunning compiling workspace could trigger after an
358 overrun had occurred. This is a "should never occur" error, but it can be
359 triggered by pathological patterns such as hundreds of nested parentheses.
360 The checks now trigger 100 bytes before the end of the workspace.
361
362 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
363
364
365 Version 8.01 19-Jan-2010
366 ------------------------
367
368 1. If a pattern contained a conditional subpattern with only one branch (in
369 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
370 computed the wrong minimum data length (which is of course zero for such
371 subpatterns). This could cause incorrect "no match" results.
372
373 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
374 the pattern is reset in the first branch, pcre_compile() failed with
375 "internal error: code overflow at offset...". This happened only when
376 the reset was to the original external option setting. (An optimization
377 abstracts leading options settings into an external setting, which was the
378 cause of this.)
379
380 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
381 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
382 assertion pattern did not match (meaning that the assertion was true), it
383 was incorrectly treated as false if the SKIP had been reached during the
384 matching. This also applied to assertions used as conditions.
385
386 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
387 assertion subpattern, including such a pattern used as a condition,
388 unpredictable results occurred, instead of the error return
389 PCRE_ERROR_DFA_UITEM.
390
391 5. The C++ GlobalReplace function was not working like Perl for the special
392 situation when an empty string is matched. It now does the fancy magic
393 stuff that is necessary.
394
395 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
396 removed. (These were left over from very, very early versions of PCRE.)
397
398 7. Some cosmetic changes to the code to make life easier when compiling it
399 as part of something else:
400
401 (a) Change DEBUG to PCRE_DEBUG.
402
403 (b) In pcre_compile(), rename the member of the "branch_chain" structure
404 called "current" as "current_branch", to prevent a collision with the
405 Linux macro when compiled as a kernel module.
406
407 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
408 prevent a collision with the Linux macro when compiled as a kernel
409 module.
410
411 8. In pcre_compile() there are some checks for integer overflows that used to
412 cast potentially large values to (double). This has been changed to that
413 when building, a check for int64_t is made, and if it is found, it is used
414 instead, thus avoiding the use of floating point arithmetic. (There is no
415 other use of FP in PCRE.) If int64_t is not found, the fallback is to
416 double.
417
418 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
419 2005 (difference between two addresses compared to an unsigned value).
420
421 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
422 custom one, because of the following reported problem in Windows:
423
424 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
425 under Win32.
426 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
427 therefore missing the function definition.
428 - The compiler thus generates a "C" signature for the test function.
429 - The linker fails to find the "C" function.
430 - PCRE fails to configure if asked to do so against libbz2.
431
432 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
433 messages were output:
434
435 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
436 rerunning libtoolize, to keep the correct libtool macros in-tree.
437 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
438
439 I have done both of these things.
440
441 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
442 most of the time, it *can* run out if it is given a pattern that contains a
443 runaway infinite recursion. I updated the discussion in the pcrestack man
444 page.
445
446 13. Now that we have gone to the x.xx style of version numbers, the minor
447 version may start with zero. Using 08 or 09 is a bad idea because users
448 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
449 interpreted as invalid octal numbers. I've updated the previous comment in
450 configure.ac, and also added a check that gives an error if 08 or 09 are
451 used.
452
453 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
454 causing partial matching to fail when the end of the subject matched \W
455 in a UTF-8 pattern where \W was quantified with a minimum of 3.
456
457 15. There were some discrepancies between the declarations in pcre_internal.h
458 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
459 their definitions. The declarations used "const uschar *" and the
460 definitions used USPTR. Even though USPTR is normally defined as "const
461 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
462 reported that: "This difference in casting confuses some C++ compilers, for
463 example, SunCC recognizes above declarations as different functions and
464 generates broken code for hbpcre." I have changed the declarations to use
465 USPTR.
466
467 16. GNU libtool is named differently on some systems. The autogen.sh script now
468 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
469 (FreeBSD).
470
471 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
472 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
473 comment: "Figure out how to create a longlong from a string: strtoll and
474 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
475 instance, but it only takes 2 args instead of 3!"
476
477 18. A subtle bug concerned with back references has been fixed by a change of
478 specification, with a corresponding code fix. A pattern such as
479 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
480 refers, was giving matches when it shouldn't. For example, xa=xaaa would
481 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
482 same bug. Such groups have to be quantified to be useful, or contained
483 inside another quantified group. (If there's no repetition, the reference
484 can never match.) The problem arises because, having left the group and
485 moved on to the rest of the pattern, a later failure that backtracks into
486 the group uses the captured value from the final iteration of the group
487 rather than the correct earlier one. I have fixed this in PCRE by forcing
488 any group that contains a reference to itself to be an atomic group; that
489 is, there cannot be any backtracking into it once it has completed. This is
490 similar to recursive and subroutine calls.
491
492
493 Version 8.00 19-Oct-09
494 ----------------------
495
496 1. The table for translating pcre_compile() error codes into POSIX error codes
497 was out-of-date, and there was no check on the pcre_compile() error code
498 being within the table. This could lead to an OK return being given in
499 error.
500
501 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
502 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
503 in a Windows environment.
504
505 3. The pcregrep --count option prints the count for each file even when it is
506 zero, as does GNU grep. However, pcregrep was also printing all files when
507 --files-with-matches was added. Now, when both options are given, it prints
508 counts only for those files that have at least one match. (GNU grep just
509 prints the file name in this circumstance, but including the count seems
510 more useful - otherwise, why use --count?) Also ensured that the
511 combination -clh just lists non-zero counts, with no names.
512
513 4. The long form of the pcregrep -F option was incorrectly implemented as
514 --fixed_strings instead of --fixed-strings. This is an incompatible change,
515 but it seems right to fix it, and I didn't think it was worth preserving
516 the old behaviour.
517
518 5. The command line items --regex=pattern and --regexp=pattern were not
519 recognized by pcregrep, which required --regex pattern or --regexp pattern
520 (with a space rather than an '='). The man page documented the '=' forms,
521 which are compatible with GNU grep; these now work.
522
523 6. No libpcreposix.pc file was created for pkg-config; there was just
524 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
525
526 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
527 when UCP support is not needed, by modifying the Python script that
528 generates it from Unicode data files. This should not matter if the module
529 is correctly used as a library, but I received one complaint about 50K of
530 unwanted data. My guess is that the person linked everything into his
531 program rather than using a library. Anyway, it does no harm.
532
533 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
534 was a minimum greater than 1 for a wide character in a possessive
535 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
536 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
537 character. Chaos in the form of incorrect output or a compiling loop could
538 result.
539
540 9. The restrictions on what a pattern can contain when partial matching is
541 requested for pcre_exec() have been removed. All patterns can now be
542 partially matched by this function. In addition, if there are at least two
543 slots in the offset vector, the offset of the earliest inspected character
544 for the match and the offset of the end of the subject are set in them when
545 PCRE_ERROR_PARTIAL is returned.
546
547 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
548 synonymous with PCRE_PARTIAL, for backwards compatibility, and
549 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
550 and may be more useful for multi-segment matching.
551
552 11. Partial matching with pcre_exec() is now more intuitive. A partial match
553 used to be given if ever the end of the subject was reached; now it is
554 given only if matching could not proceed because another character was
555 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
556 string "Z", which now yields "no match" instead of "partial match". In the
557 case of pcre_dfa_exec(), "no match" is given if every matching path for the
558 final character ended with (*FAIL).
559
560 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
561 if the pattern had a "must contain" character that was already found in the
562 earlier partial match, unless partial matching was again requested. For
563 example, with the pattern /dog.(body)?/, the "must contain" character is
564 "g". If the first part-match was for the string "dog", restarting with
565 "sbody" failed. This bug has been fixed.
566
567 13. The string returned by pcre_dfa_exec() after a partial match has been
568 changed so that it starts at the first inspected character rather than the
569 first character of the match. This makes a difference only if the pattern
570 starts with a lookbehind assertion or \b or \B (\K is not supported by
571 pcre_dfa_exec()). It's an incompatible change, but it makes the two
572 matching functions compatible, and I think it's the right thing to do.
573
574 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
575 so that the demonstration program is easily available in environments where
576 PCRE has not been installed from source.
577
578 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
579 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
580 library.
581
582 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
583 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
584 is not the first non-POSIX option to be added. Clearly some people find
585 these options useful.
586
587 17. If a caller to the POSIX matching function regexec() passes a non-zero
588 value for nmatch with a NULL value for pmatch, the value of
589 nmatch is forced to zero.
590
591 18. RunGrepTest did not have a test for the availability of the -u option of
592 the diff command, as RunTest does. It now checks in the same way as
593 RunTest, and also checks for the -b option.
594
595 19. If an odd number of negated classes containing just a single character
596 interposed, within parentheses, between a forward reference to a named
597 subpattern and the definition of the subpattern, compilation crashed with
598 an internal error, complaining that it could not find the referenced
599 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
600 [The bug was that it was starting one character too far in when skipping
601 over the character class, thus treating the ] as data rather than
602 terminating the class. This meant it could skip too much.]
603
604 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
605 /g option in pcretest when the pattern contains \K, which makes it possible
606 to have an empty string match not at the start, even when the pattern is
607 anchored. Updated pcretest and pcredemo to use this option.
608
609 21. If the maximum number of capturing subpatterns in a recursion was greater
610 than the maximum at the outer level, the higher number was returned, but
611 with unset values at the outer level. The correct (outer level) value is
612 now given.
613
614 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
615 PCRE did not set those parentheses (unlike Perl). I have now found a way to
616 make it do so. The string so far is captured, making this feature
617 compatible with Perl.
618
619 23. The tests have been re-organized, adding tests 11 and 12, to make it
620 possible to check the Perl 5.10 features against Perl 5.10.
621
622 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
623 pattern matches a fixed length string. PCRE did not allow this; now it
624 does. Neither allows recursion.
625
626 25. I finally figured out how to implement a request to provide the minimum
627 length of subject string that was needed in order to match a given pattern.
628 (It was back references and recursion that I had previously got hung up
629 on.) This code has now been added to pcre_study(); it finds a lower bound
630 to the length of subject needed. It is not necessarily the greatest lower
631 bound, but using it to avoid searching strings that are too short does give
632 some useful speed-ups. The value is available to calling programs via
633 pcre_fullinfo().
634
635 26. While implementing 25, I discovered to my embarrassment that pcretest had
636 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
637 study optimizations had never been tested with that matching function.
638 Oops. What is worse, even when it was passed study data, there was a bug in
639 pcre_dfa_exec() that meant it never actually used it. Double oops. There
640 were also very few tests of studied patterns with pcre_dfa_exec().
641
642 27. If (?| is used to create subpatterns with duplicate numbers, they are now
643 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
644 on the other side of the coin, they are no longer allowed to have different
645 names, because these cannot be distinguished in PCRE, and this has caused
646 confusion. (This is a difference from Perl.)
647
648 28. When duplicate subpattern names are present (necessarily with different
649 numbers, as required by 27 above), and a test is made by name in a
650 conditional pattern, either for a subpattern having been matched, or for
651 recursion in such a pattern, all the associated numbered subpatterns are
652 tested, and the overall condition is true if the condition is true for any
653 one of them. This is the way Perl works, and is also more like the way
654 testing by number works.
655
656
657 Version 7.9 11-Apr-09
658 ---------------------
659
660 1. When building with support for bzlib/zlib (pcregrep) and/or readline
661 (pcretest), all targets were linked against these libraries. This included
662 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
663 libraries. This caused unwanted dependencies to be created. This problem
664 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
665 pcretest is linked with readline.
666
667 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
668 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
669 moved outside it again, because FALSE and TRUE are already defined in AIX,
670 but BOOL is not.
671
672 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
673 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
674
675 4. The pcregrep documentation said spaces were inserted as well as colons (or
676 hyphens) following file names and line numbers when outputting matching
677 lines. This is not true; no spaces are inserted. I have also clarified the
678 wording for the --colour (or --color) option.
679
680 5. In pcregrep, when --colour was used with -o, the list of matching strings
681 was not coloured; this is different to GNU grep, so I have changed it to be
682 the same.
683
684 6. When --colo(u)r was used in pcregrep, only the first matching substring in
685 each matching line was coloured. Now it goes on to look for further matches
686 of any of the test patterns, which is the same behaviour as GNU grep.
687
688 7. A pattern that could match an empty string could cause pcregrep to loop; it
689 doesn't make sense to accept an empty string match in pcregrep, so I have
690 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
691 seems to be how GNU grep behaves.
692
693 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
694 start or after a newline", because the conditional assertion was not being
695 correctly handled. The rule now is that both the assertion and what follows
696 in the first alternative must satisfy the test.
697
698 9. If auto-callout was enabled in a pattern with a conditional group whose
699 condition was an assertion, PCRE could crash during matching, both with
700 pcre_exec() and pcre_dfa_exec().
701
702 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
703 used for matching.
704
705 11. Unicode property support in character classes was not working for
706 characters (bytes) greater than 127 when not in UTF-8 mode.
707
708 12. Added the -M command line option to pcretest.
709
710 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
711
712 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
713
714 16. Added comments and documentation about mis-use of no_arg in the C++
715 wrapper.
716
717 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
718 from Martin Jerabek that uses macro names for all relevant character and
719 string constants.
720
721 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
722 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
723 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
724 these, but not everybody uses configure.
725
726 19. A conditional group that had only one branch was not being correctly
727 recognized as an item that could match an empty string. This meant that an
728 enclosing group might also not be so recognized, causing infinite looping
729 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
730 with the subject "ab", where knowledge that the repeated group can match
731 nothing is needed in order to break the loop.
732
733 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
734 exec(), but without supplying a callout function, matching went wrong.
735
736 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
737 leak if the size of the offset vector was greater than 30. When the vector
738 is smaller, the saved offsets during recursion go onto a local stack
739 vector, but for larger vectors malloc() is used. It was failing to free
740 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
741 error, in fact).
742
743 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
744 heapframe that is used only when UTF-8 support is enabled. This caused no
745 problem, but was untidy.
746
747 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
748 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
749 included within another project.
750
751 24. Steven Van Ingelgem's patches to add more options to the CMake support,
752 slightly modified by me:
753
754 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
755 not building pcregrep.
756
757 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
758 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
759
760 25. Forward references, both numeric and by name, in patterns that made use of
761 duplicate group numbers, could behave incorrectly or give incorrect errors,
762 because when scanning forward to find the reference group, PCRE was not
763 taking into account the duplicate group numbers. A pattern such as
764 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
765
766 26. Changed a few more instances of "const unsigned char *" to USPTR, making
767 the feature of a custom pointer more persuasive (as requested by a user).
768
769 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
770 pcretest.c, inside #ifndefs, because it seems they are sometimes already
771 pre-defined.
772
773 28. Added support for (*UTF8) at the start of a pattern.
774
775 29. Arrange for flags added by the "release type" setting in CMake to be shown
776 in the configuration summary.
777
778
779 Version 7.8 05-Sep-08
780 ---------------------
781
782 1. Replaced UCP searching code with optimized version as implemented for Ad
783 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
784 stage table and inline lookup instead of a function, giving speed ups of 2
785 to 5 times on some simple patterns that I tested. Permission was given to
786 distribute the MultiStage2.py script that generates the tables (it's not in
787 the tarball, but is in the Subversion repository).
788
789 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
790 scripts.
791
792 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
793 a group with a zero qualifier. The result of the study could be incorrect,
794 or the function might crash, depending on the pattern.
795
796 4. Caseless matching was not working for non-ASCII characters in back
797 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
798 It now works when Unicode Property Support is available.
799
800 5. In pcretest, an escape such as \x{de} in the data was always generating
801 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
802 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
803 truncation.
804
805 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
806
807 7. Added two (int) casts to pcregrep when printing the difference of two
808 pointers, in case they are 64-bit values.
809
810 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
811 test 2 if it fails.
812
813 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
814 and a #define of that name to empty if it is not externally set. This is to
815 allow users of MSVC to set it if necessary.
816
817 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
818 the convenience functions in the pcre_get.c source file.
819
820 11. An option change at the start of a pattern that had top-level alternatives
821 could cause overwriting and/or a crash. This command provoked a crash in
822 some environments:
823
824 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
825
826 This potential security problem was recorded as CVE-2008-2371.
827
828 12. For a pattern where the match had to start at the beginning or immediately
829 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
830 pcre_dfa_exec() could read past the end of the passed subject if there was
831 no match. To help with detecting such bugs (e.g. with valgrind), I modified
832 pcretest so that it places the subject at the end of its malloc-ed buffer.
833
834 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
835 exec() might read past the end of the data buffer in UTF-8 mode.
836
837 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
838 the data contained the byte 0x85 as part of a UTF-8 character within its
839 first line. This applied both to normal and DFA matching.
840
841 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
842 /^[^d]*?$/8 failed to match "abc".
843
844 16. Added a missing copyright notice to pcrecpp_internal.h.
845
846 17. Make it more clear in the documentation that values returned from
847 pcre_exec() in ovector are byte offsets, not character counts.
848
849 18. Tidied a few places to stop certain compilers from issuing warnings.
850
851 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
852 supplied by Stefan Weber. I made a further small update for 7.8 because
853 there is a change of source arrangements: the pcre_searchfuncs.c module is
854 replaced by pcre_ucd.c.
855
856
857 Version 7.7 07-May-08
858 ---------------------
859
860 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
861 a string to a long long, pretend we don't even have a long long." This is
862 done by checking for the strtoq, strtoll, and _strtoi64 functions.
863
864 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
865 pre-7.6 versions, which defined a global no_arg variable instead of putting
866 it in the RE class. (See also #8 below.)
867
868 3. Remove a line of dead code, identified by coverity and reported by Nuno
869 Lopes.
870
871 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
872
873 (1) The include/exclude patterns were being applied to the whole pathnames
874 of files, instead of just to the final components.
875
876 (2) If there was more than one level of directory, the subdirectories were
877 skipped unless they satisfied the include/exclude conditions. This is
878 inconsistent with GNU grep (and could even be seen as contrary to the
879 pcregrep specification - which I improved to make it absolutely clear).
880 The action now is always to scan all levels of directory, and just
881 apply the include/exclude patterns to regular files.
882
883 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
884 --exclude_dir in the tests to avoid scanning .svn directories.
885
886 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
887 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
888 doesn't support NULs in patterns.
889
890 7. Added some missing "const"s to declarations of static tables in
891 pcre_compile.c and pcre_dfa_exec.c.
892
893 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
894 caused by fix #2 above. (Subsequently also a second patch to fix the
895 first patch. And a third patch - this was a messy problem.)
896
897 9. Applied Craig's patch to remove the use of push_back().
898
899 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
900 matching function regexec().
901
902 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
903 which, however, unlike Perl's \g{...}, are subroutine calls, not back
904 references. PCRE supports relative numbers with this syntax (I don't think
905 Oniguruma does).
906
907 12. Previously, a group with a zero repeat such as (...){0} was completely
908 omitted from the compiled regex. However, this means that if the group
909 was called as a subroutine from elsewhere in the pattern, things went wrong
910 (an internal error was given). Such groups are now left in the compiled
911 pattern, with a new opcode that causes them to be skipped at execution
912 time.
913
914 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
915 to the way PCRE behaves:
916
917 (a) A lone ] character is dis-allowed (Perl treats it as data).
918
919 (b) A back reference to an unmatched subpattern matches an empty string
920 (Perl fails the current match path).
921
922 (c) A data ] in a character class must be notated as \] because if the
923 first data character in a class is ], it defines an empty class. (In
924 Perl it is not possible to have an empty class.) The empty class []
925 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
926 The negative empty class [^] matches any one character, independently
927 of the DOTALL setting.
928
929 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
930 non-existent subpattern following a character class starting with ']' and
931 containing () gave an internal compiling error instead of "reference to
932 non-existent subpattern". Fortunately, when the pattern did exist, the
933 compiled code was correct. (When scanning forwards to check for the
934 existencd of the subpattern, it was treating the data ']' as terminating
935 the class, so got the count wrong. When actually compiling, the reference
936 was subsequently set up correctly.)
937
938 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
939 it was being rejected as not supported by pcre_dfa_exec(), even though
940 other assertions are supported. I have made pcre_dfa_exec() support
941 (*FAIL).
942
943 16. The implementation of 13c above involved the invention of a new opcode,
944 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
945 cannot be changed at match time, I realized I could make a small
946 improvement to matching performance by compiling OP_ALLANY instead of
947 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
948 on the OP_ANY path.
949
950 17. Compiling pcretest on Windows with readline support failed without the
951 following two fixes: (1) Make the unistd.h include conditional on
952 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
953
954 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
955 ncurses library to be included for pcretest when ReadLine support is
956 requested, but also to allow for it to be overridden. This patch came from
957 Daniel Bergström.
958
959 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
960 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
961 any errors with the current Unicode tables. Thanks to Peter Kankowski for
962 spotting this.
963
964
965 Version 7.6 28-Jan-08
966 ---------------------
967
968 1. A character class containing a very large number of characters with
969 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
970 overflow.
971
972 2. Patch to cut out the "long long" test in pcrecpp_unittest when
973 HAVE_LONG_LONG is not defined.
974
975 3. Applied Christian Ehrlicher's patch to update the CMake build files to
976 bring them up to date and include new features. This patch includes:
977
978 - Fixed PH's badly added libz and libbz2 support.
979 - Fixed a problem with static linking.
980 - Added pcredemo. [But later removed - see 7 below.]
981 - Fixed dftables problem and added an option.
982 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
983 HAVE_LONG_LONG.
984 - Added readline support for pcretest.
985 - Added an listing of the option settings after cmake has run.
986
987 4. A user submitted a patch to Makefile that makes it easy to create
988 "pcre.dll" under mingw when using Configure/Make. I added stuff to
989 Makefile.am that cause it to include this special target, without
990 affecting anything else. Note that the same mingw target plus all
991 the other distribution libraries and programs are now supported
992 when configuring with CMake (see 6 below) instead of with
993 Configure/Make.
994
995 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
996 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
997 exported in the Windows port". It has not yet been confirmed that the patch
998 solves the problem, but it does no harm.
999
1000 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1001 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1002 with CMake, and also correct the comment about stack recursion.
1003
1004 7. Remove the automatic building of pcredemo from the ./configure system and
1005 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1006 of a program that users should build themselves after PCRE is installed, so
1007 building it automatically is not really right. What is more, it gave
1008 trouble in some build environments.
1009
1010 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1011
1012
1013 Version 7.5 10-Jan-08
1014 ---------------------
1015
1016 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1017 values in parens when parsing an RE using the C++ wrapper."
1018
1019 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1020 Characters greater than 255 were excluded from the class instead of being
1021 included.
1022
1023 3. The same bug as (2) above applied to negated POSIX classes such as
1024 [:^space:].
1025
1026 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1027 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1028 I have changed it.
1029
1030 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1031 first named subpattern) and a construct such as (?&a) would reference the
1032 first named subpattern whose name started with "a" (in other words, the
1033 length check was missing). Both these problems are fixed. "Subpattern name
1034 expected" is now given for (?&) (a zero-length name), and this patch also
1035 makes it give the same error for \k'' (previously it complained that that
1036 was a reference to a non-existent subpattern).
1037
1038 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1039 this is right because (?- can be followed by option settings as well as by
1040 digits. I have, however, made the messages clearer.
1041
1042 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1043 than the number used in the conditional) now cause a compile-time error.
1044 This is actually not compatible with Perl, which accepts such patterns, but
1045 treats the conditional as always being FALSE (as PCRE used to), but it
1046 seems to me that giving a diagnostic is better.
1047
1048 8. Change "alphameric" to the more common word "alphanumeric" in comments
1049 and messages.
1050
1051 9. Fix two occurrences of "backslash" in comments that should have been
1052 "backspace".
1053
1054 10. Remove two redundant lines of code that can never be obeyed (their function
1055 was moved elsewhere).
1056
1057 11. The program that makes PCRE's Unicode character property table had a bug
1058 which caused it to generate incorrect table entries for sequences of
1059 characters that have the same character type, but are in different scripts.
1060 It amalgamated them into a single range, with the script of the first of
1061 them. In other words, some characters were in the wrong script. There were
1062 thirteen such cases, affecting characters in the following ranges:
1063
1064 U+002b0 - U+002c1
1065 U+0060c - U+0060d
1066 U+0061e - U+00612
1067 U+0064b - U+0065e
1068 U+0074d - U+0076d
1069 U+01800 - U+01805
1070 U+01d00 - U+01d77
1071 U+01d9b - U+01dbf
1072 U+0200b - U+0200f
1073 U+030fc - U+030fe
1074 U+03260 - U+0327f
1075 U+0fb46 - U+0fbb1
1076 U+10450 - U+1049d
1077
1078 12. The -o option (show only the matching part of a line) for pcregrep was not
1079 compatible with GNU grep in that, if there was more than one match in a
1080 line, it showed only the first of them. It now behaves in the same way as
1081 GNU grep.
1082
1083 13. If the -o and -v options were combined for pcregrep, it printed a blank
1084 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1085 does the same. The return code can be used to tell if there were any
1086 non-matching lines.
1087
1088 14. Added --file-offsets and --line-offsets to pcregrep.
1089
1090 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1091 infinitely looping recursion. The bug was that positive lookaheads were not
1092 being skipped when checking for a possible empty match (negative lookaheads
1093 and both kinds of lookbehind were skipped).
1094
1095 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1096 inclusion of <windows.h> to before rather than after the definition of
1097 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1098
1099 17. Specifying a possessive quantifier with a specific limit for a Unicode
1100 character property caused pcre_compile() to compile bad code, which led at
1101 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1102 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1103 caused the error; without that there was no problem.
1104
1105 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1106
1107 19. Added --enable-pcretest-libreadline.
1108
1109 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1110 RE::GlobalReplace(). As a result, the number of replacements returned was
1111 double what it should be. I removed one of the increments, but Craig sent a
1112 later patch that removed the other one (the right fix) and added unit tests
1113 that check the return values (which was not done before).
1114
1115 21. Several CMake things:
1116
1117 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1118 the names libpcre and libpcreposix, not just pcre and pcreposix.
1119
1120 (2) The above change means that pcretest and pcregrep are now correctly
1121 linked with the newly-built libraries, not previously installed ones.
1122
1123 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1124
1125 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1126 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1127 UTF-8 newline character). The key issue is that the pattern starts .*;
1128 this means that the match must be either at the beginning, or after a
1129 newline. The bug was in the code for advancing after a failed match and
1130 checking that the new position followed a newline. It was not taking
1131 account of UTF-8 characters correctly.
1132
1133 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1134 character classes. PCRE was not treating the sequence [:...:] as a
1135 character class unless the ... were all letters. Perl, however, seems to
1136 allow any characters between [: and :], though of course it rejects as
1137 unknown any "names" that contain non-letters, because all the known class
1138 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1139 for example, whereas PCRE did not - it did not recognize a POSIX character
1140 class. This seemed a bit dangerous, so the code has been changed to be
1141 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1142 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1143 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1144 Perl does, and where it didn't before.
1145
1146 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1147 Windows environments %n is disabled by default.
1148
1149
1150 Version 7.4 21-Sep-07
1151 ---------------------
1152
1153 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1154 means that a class such as [\s] counted as "explicit reference to CR or
1155 LF". That isn't really right - the whole point of the change was to try to
1156 help when there was an actual mention of one of the two characters. So now
1157 the change happens only if \r or \n (or a literal CR or LF) character is
1158 encountered.
1159
1160 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1161 of both had grown to the point where there were only 3 bits left.
1162 Fortunately, there was spare space in the data structure, and so I have
1163 moved the internal flags into a new 16-bit field to free up more option
1164 bits.
1165
1166 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1167 but did not set the internal JCHANGED flag - either of these is enough to
1168 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1169 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1170 start sets both bits.
1171
1172 4. Added options (at build time, compile time, exec time) to change \R from
1173 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1174
1175 5. doc/pcresyntax.html was missing from the distribution.
1176
1177 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1178 compatibility, even though it is no longer used.
1179
1180 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1181 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1182 windows.h file is present (where different names are used). [This was
1183 reversed later after testing - see 16 below.]
1184
1185 8. Changed all #include <config.h> to #include "config.h". There were also
1186 some further <pcre.h> cases that I changed to "pcre.h".
1187
1188 9. When pcregrep was used with the --colour option, it missed the line ending
1189 sequence off the lines that it output.
1190
1191 10. It was pointed out to me that arrays of string pointers cause lots of
1192 relocations when a shared library is dynamically loaded. A technique of
1193 using a single long string with a table of offsets can drastically reduce
1194 these. I have refactored PCRE in four places to do this. The result is
1195 dramatic:
1196
1197 Originally: 290
1198 After changing UCP table: 187
1199 After changing error message table: 43
1200 After changing table of "verbs" 36
1201 After changing table of Posix names 22
1202
1203 Thanks to the folks working on Gregex for glib for this insight.
1204
1205 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1206 unicode-properties was also set.
1207
1208 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1209
1210 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1211 checked only for CRLF.
1212
1213 14. Added casts to pcretest.c to avoid compiler warnings.
1214
1215 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1216
1217 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1218 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1219 entirely. This removes changes made in 7 above.
1220
1221 17. The CMake files have been updated, and there is now more information about
1222 building with CMake in the NON-UNIX-USE document.
1223
1224
1225 Version 7.3 28-Aug-07
1226 ---------------------
1227
1228 1. In the rejigging of the build system that eventually resulted in 7.1, the
1229 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1230 brackets there is not right, since it causes compilers to look for an
1231 installed pcre.h, not the version that is in the source that is being
1232 compiled (which of course may be different). I have changed it back to:
1233
1234 #include "pcre.h"
1235
1236 I have a vague recollection that the change was concerned with compiling in
1237 different directories, but in the new build system, that is taken care of
1238 by the VPATH setting the Makefile.
1239
1240 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1241 when the subject happened to end in the byte 0x85 (e.g. if the last
1242 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1243 characters but of course it shouldn't be taken as a newline when it is part
1244 of another character. The bug was that, for an unlimited repeat of . in
1245 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1246 characters when looking for a newline.
1247
1248 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1249
1250 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1251 in debug output.
1252
1253 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1254 long printing in the pcrecpp unittest when running under MinGW.
1255
1256 6. ESC_K was left out of the EBCDIC table.
1257
1258 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1259 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1260 limit also applies to "virtual nesting" when a pattern is recursive, and in
1261 this case 1000 isn't so big. I have been able to remove this limit at the
1262 expense of backing off one optimization in certain circumstances. Normally,
1263 when pcre_exec() would call its internal match() function recursively and
1264 immediately return the result unconditionally, it uses a "tail recursion"
1265 feature to save stack. However, when a subpattern that can match an empty
1266 string has an unlimited repetition quantifier, it no longer makes this
1267 optimization. That gives it a stack frame in which to save the data for
1268 checking that an empty string has been matched. Previously this was taken
1269 from the 1000-entry workspace that had been reserved. So now there is no
1270 explicit limit, but more stack is used.
1271
1272 8. Applied Daniel's patches to solve problems with the import/export magic
1273 syntax that is required for Windows, and which was going wrong for the
1274 pcreposix and pcrecpp parts of the library. These were overlooked when this
1275 problem was solved for the main library.
1276
1277 9. There were some crude static tests to avoid integer overflow when computing
1278 the size of patterns that contain repeated groups with explicit upper
1279 limits. As the maximum quantifier is 65535, the maximum group length was
1280 set at 30,000 so that the product of these two numbers did not overflow a
1281 32-bit integer. However, it turns out that people want to use groups that
1282 are longer than 30,000 bytes (though not repeat them that many times).
1283 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1284 made it possible to implement the integer overflow checks in a much more
1285 dynamic way, which I have now done. The artificial limitation on group
1286 length has been removed - we now have only the limit on the total length of
1287 the compiled pattern, which depends on the LINK_SIZE setting.
1288
1289 10. Fixed a bug in the documentation for get/copy named substring when
1290 duplicate names are permitted. If none of the named substrings are set, the
1291 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1292 empty string.
1293
1294 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1295 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1296 because the ] is interpreted as the first data character and the
1297 terminating ] is not found. PCRE has been made compatible with Perl in this
1298 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1299 cause memory overwriting.
1300
1301 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1302 string has been matched (to stop an infinite loop). It was not recognizing
1303 a conditional subpattern that could match an empty string if that
1304 subpattern was within another subpattern. For example, it looped when
1305 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1306 condition was not nested. This bug has been fixed.
1307
1308 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1309 past the start of the subject in the presence of bytes with the top bit
1310 set, for example "\x8aBCD".
1311
1312 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1313 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1314
1315 14. Optimized (?!) to (*FAIL).
1316
1317 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1318 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1319 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1320 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1321 does: it's just the validity check that is more restrictive.
1322
1323 16. Inserted checks for integer overflows during escape sequence (backslash)
1324 processing, and also fixed erroneous offset values for syntax errors during
1325 backslash processing.
1326
1327 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1328 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1329
1330 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1331 caused an overrun.
1332
1333 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1334 something other than just ASCII characters) inside a group that had an
1335 unlimited repeat caused a loop at compile time (while checking to see
1336 whether the group could match an empty string).
1337
1338 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1339 [\P{Any}] did so. (Error in the code for printing property names.)
1340
1341 21. An orphan \E inside a character class could cause a crash.
1342
1343 22. A repeated capturing bracket such as (A)? could cause a wild memory
1344 reference during compilation.
1345
1346 23. There are several functions in pcre_compile() that scan along a compiled
1347 expression for various reasons (e.g. to see if it's fixed length for look
1348 behind). There were bugs in these functions when a repeated \p or \P was
1349 present in the pattern. These operators have additional parameters compared
1350 with \d, etc, and these were not being taken into account when moving along
1351 the compiled data. Specifically:
1352
1353 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1354 length.
1355
1356 (b) An item such as \pL+ within a repeated group could cause crashes or
1357 loops.
1358
1359 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1360 "reference to non-existent subpattern" error.
1361
1362 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1363
1364 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1365 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1366
1367 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1368
1369 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1370 character were causing crashes (broken optimization).
1371
1372 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1373 \p or \P) caused a compile-time loop.
1374
1375 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1376 break. For example, the unstudied pattern [\r\n]A does not match the string
1377 "\r\nA" because change 7.0/46 below moves the current point on by two
1378 characters after failing to match at the start. However, the pattern \nA
1379 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1380 the same is true. There doesn't seem any very clean way out of this, but
1381 what I have chosen to do makes the common cases work: PCRE now takes note
1382 of whether there can be an explicit match for \r or \n anywhere in the
1383 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1384 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1385 pattern has explicit CR or LF references.
1386
1387 29. Added (*CR) etc for changing newline setting at start of pattern.
1388
1389
1390 Version 7.2 19-Jun-07
1391 ---------------------
1392
1393 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1394 which is apparently normally available under Windows.
1395
1396 2. Re-jig the pcregrep tests with different newline settings in an attempt
1397 to make them independent of the local environment's newline setting.
1398
1399 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1400
1401 4. Some of the "internals" tests were previously cut out when the link size
1402 was not 2, because the output contained actual offsets. The recent new
1403 "Z" feature of pcretest means that these can be cut out, making the tests
1404 usable with all link sizes.
1405
1406 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1407 stack recursion. This gives a massive performance boost under BSD, but just
1408 a small improvement under Linux. However, it saves one field in the frame
1409 in all cases.
1410
1411 6. Added more features from the forthcoming Perl 5.10:
1412
1413 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1414 recursion call. It refers to the nth most recently opened parentheses.
1415
1416 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1417 to be opened parentheses.
1418
1419 (c) Conditions that refer to capturing parentheses can be specified
1420 relatively, for example, (?(-2)... or (?(+3)...
1421
1422 (d) \K resets the start of the current match so that everything before
1423 is not part of it.
1424
1425 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1426
1427 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1428 reference syntax.
1429
1430 (g) (?| introduces a group in which the numbering of parentheses in each
1431 alternative starts with the same number.
1432
1433 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1434
1435 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1436 PCRE_INFO_JCHANGED.
1437
1438 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1439 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1440 for detecting groups that can match an empty string.
1441
1442 9. A pattern with a very large number of alternatives (more than several
1443 hundred) was running out of internal workspace during the pre-compile
1444 phase, where pcre_compile() figures out how much memory will be needed. A
1445 bit of new cunning has reduced the workspace needed for groups with
1446 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1447 workspace instead of running out of the 4096 that are available.
1448
1449 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1450
1451 11. Applied patch from Google to remove an optimization that didn't quite work.
1452 The report of the bug said:
1453
1454 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1455 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1456 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1457
1458 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1459 it matched the wrong number of bytes.
1460
1461
1462 Version 7.1 24-Apr-07
1463 ---------------------
1464
1465 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1466 that is more "standard", making use of automake and other Autotools. There
1467 is some re-arrangement of the files and adjustment of comments consequent
1468 on this.
1469
1470 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1471 for recursive directory scanning broke on some systems because the files
1472 are not scanned in any specific order and on different systems the order
1473 was different. A call to "sort" has been inserted into RunGrepTest for the
1474 approprate test as a short-term fix. In the longer term there may be an
1475 alternative.
1476
1477 3. I had an email from Eric Raymond about problems translating some of PCRE's
1478 man pages to HTML (despite the fact that I distribute HTML pages, some
1479 people do their own conversions for various reasons). The problems
1480 concerned the use of low-level troff macros .br and .in. I have therefore
1481 removed all such uses from the man pages (some were redundant, some could
1482 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1483 HTML has been updated to handle .nf/.fi and to complain if it encounters
1484 .br or .in.
1485
1486 4. Updated comments in configure.ac that get placed in config.h.in and also
1487 arranged for config.h to be included in the distribution, with the name
1488 config.h.generic, for the benefit of those who have to compile without
1489 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1490
1491 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1492 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1493 makevp.bat for the current PCRE, using the additional files
1494 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1495
1496 6. A Windows user reported a minor discrepancy with test 2, which turned out
1497 to be caused by a trailing space on an input line that had got lost in his
1498 copy. The trailing space was an accident, so I've just removed it.
1499
1500 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1501 that is needed.
1502
1503 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1504 as "const" (a) because they are and (b) because it helps the PHP
1505 maintainers who have recently made a script to detect big data structures
1506 in the php code that should be moved to the .rodata section. I remembered
1507 to update Builducptable as well, so it won't revert if ucptable.h is ever
1508 re-created.
1509
1510 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1511 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1512 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1513 support is not required. This saves 1.5-2K of code, which is important in
1514 some applications.
1515
1516 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1517 so as not to refer to the tables, even though these functions will never be
1518 called when UTF-8 support is disabled. Otherwise there are problems with a
1519 shared library.
1520
1521 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1522
1523 (a) It was defining its arguments as char * instead of void *.
1524
1525 (b) It was assuming that all moves were upwards in memory; this was true
1526 a long time ago when I wrote it, but is no longer the case.
1527
1528 The emulated memove() is provided for those environments that have neither
1529 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1530 is clearly not the case, as these two bugs were recently reported.
1531
1532 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1533 and Detrail to create the HTML documentation, the .txt form of the man
1534 pages, and it removes trailing spaces from listed files. It also creates
1535 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1536 case, it wraps all the #defines with #ifndefs. This script should be run
1537 before "make dist".
1538
1539 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1540 with Unicode property support.
1541
1542 (a) For a maximizing quantifier, if the two different cases of the
1543 character were of different lengths in their UTF-8 codings (there are
1544 some cases like this - I found 11), and the matching function had to
1545 back up over a mixture of the two cases, it incorrectly assumed they
1546 were both the same length.
1547
1548 (b) When PCRE was configured to use the heap rather than the stack for
1549 recursion during matching, it was not correctly preserving the data for
1550 the other case of a UTF-8 character when checking ahead for a match
1551 while processing a minimizing repeat. If the check also involved
1552 matching a wide character, but failed, corruption could cause an
1553 erroneous result when trying to check for a repeat of the original
1554 character.
1555
1556 13. Some tidying changes to the testing mechanism:
1557
1558 (a) The RunTest script now detects the internal link size and whether there
1559 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1560 values substituted by "configure". (The RunGrepTest script already did
1561 this for UTF-8.) The configure.ac script no longer substitutes the
1562 relevant variables.
1563
1564 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1565 with length and offset values. This means that the output is different
1566 for different internal link sizes. Test 2 is skipped for link sizes
1567 other than 2 because of this, bypassing the problem. Unfortunately,
1568 there was also a test in test 3 (the locale tests) that used /B and
1569 failed for link sizes other than 2. Rather than cut the whole test out,
1570 I have added a new /Z option to pcretest that replaces the length and
1571 offset values with spaces. This is now used to make test 3 independent
1572 of link size. (Test 2 will be tidied up later.)
1573
1574 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1575 segmentation fault instead of returning the appropriate error message.
1576
1577 15. In multiline mode when the newline sequence was set to "any", the pattern
1578 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1579 This doesn't seem right; it now treats the CRLF combination as the line
1580 ending, and so does not match in that case. It's only a pattern such as ^$
1581 that would hit this one: something like ^ABC$ would have failed after \r
1582 and then tried again after \r\n.
1583
1584 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1585 in an attempt to make files that differ only in their line terminators
1586 compare equal. This works on Linux.
1587
1588 17. Under certain error circumstances pcregrep might try to free random memory
1589 as it exited. This is now fixed, thanks to valgrind.
1590
1591 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1592 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1593 was because its rules for how to advance for /g after matching an empty
1594 string at the end of a line did not allow for this case. They now check for
1595 it specially.
1596
1597 20. pcretest is supposed to handle patterns and data of any length, by
1598 extending its buffers when necessary. It was getting this wrong when the
1599 buffer for a data line had to be extended.
1600
1601 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1602 CRLF as a newline sequence.
1603
1604 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1605 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1606 I have nevertheless tidied it up.
1607
1608 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1609
1610 24. Added a man page for pcre-config.
1611
1612
1613 Version 7.0 19-Dec-06
1614 ---------------------
1615
1616 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1617 moving to gcc 4.1.1.
1618
1619 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1620 sys/time.h, which is documented as needed for this function. It doesn't
1621 seem to matter on Linux, but it showed up on some releases of OS X.
1622
1623 3. It seems that there are systems where bytes whose values are greater than
1624 127 match isprint() in the "C" locale. The "C" locale should be the
1625 default when a C program starts up. In most systems, only ASCII printing
1626 characters match isprint(). This difference caused the output from pcretest
1627 to vary, making some of the tests fail. I have changed pcretest so that:
1628
1629 (a) When it is outputting text in the compiled version of a pattern, bytes
1630 other than 32-126 are always shown as hex escapes.
1631
1632 (b) When it is outputting text that is a matched part of a subject string,
1633 it does the same, unless a different locale has been set for the match
1634 (using the /L modifier). In this case, it uses isprint() to decide.
1635
1636 4. Fixed a major bug that caused incorrect computation of the amount of memory
1637 required for a compiled pattern when options that changed within the
1638 pattern affected the logic of the preliminary scan that determines the
1639 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1640 that the computed length was too small. The symptoms of this bug were
1641 either the PCRE error "internal error: code overflow" from pcre_compile(),
1642 or a glibc crash with a message such as "pcretest: free(): invalid next
1643 size (fast)". Examples of patterns that provoked this bug (shown in
1644 pcretest format) are:
1645
1646 /(?-x: )/x
1647 /(?x)(?-x: \s*#\s*)/
1648 /((?i)[\x{c0}])/8
1649 /(?i:[\x{c0}])/8
1650
1651 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1652 is now done differently.
1653
1654 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1655 wrapper classes; (b) implement a new function in the C++ scanner that is
1656 more efficient than the old way of doing things because it avoids levels of
1657 recursion in the regex matching; (c) add a paragraph to the documentation
1658 for the FullMatch() function.
1659
1660 6. The escape sequence \n was being treated as whatever was defined as
1661 "newline". Not only was this contrary to the documentation, which states
1662 that \n is character 10 (hex 0A), but it also went horribly wrong when
1663 "newline" was defined as CRLF. This has been fixed.
1664
1665 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1666 was being set to -1 for the "end of line" case (supposedly a value that no
1667 character can have). Though this value is never used (the check for end of
1668 line is "zero bytes in current character"), it caused compiler complaints.
1669 I've changed it to 0xffffffff.
1670
1671 8. In pcre_version.c, the version string was being built by a sequence of
1672 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1673 string (as it is for production releases) called a macro with an empty
1674 argument. The C standard says the result of this is undefined. The gcc
1675 compiler treats it as an empty string (which was what was wanted) but it is
1676 reported that Visual C gives an error. The source has been hacked around to
1677 avoid this problem.
1678
1679 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1680 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1681 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1682 of them did).
1683
1684 10. Originally, pcretest opened its input and output without "b"; then I was
1685 told that "b" was needed in some environments, so it was added for release
1686 5.0 to both the input and output. (It makes no difference on Unix-like
1687 systems.) Later I was told that it is wrong for the input on Windows. I've
1688 now abstracted the modes into two macros, to make it easier to fiddle with
1689 them, and removed "b" from the input mode under Windows.
1690
1691 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1692
1693 12. Added -help and --help to pcretest as an official way of being reminded
1694 of the options.
1695
1696 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1697 and pcrecpp.cc because they annoy compilers at high warning levels.
1698
1699 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1700
1701 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1702 scripts are not C programs :-) and which was not noticed because it works
1703 on Linux.
1704
1705 16. pcretest is supposed to handle any length of pattern and data line (as one
1706 line or as a continued sequence of lines) by extending its input buffer if
1707 necessary. This feature was broken for very long pattern lines, leading to
1708 a string of junk being passed to pcre_compile() if the pattern was longer
1709 than about 50K.
1710
1711 17. I have done a major re-factoring of the way pcre_compile() computes the
1712 amount of memory needed for a compiled pattern. Previously, there was code
1713 that made a preliminary scan of the pattern in order to do this. That was
1714 OK when PCRE was new, but as the facilities have expanded, it has become
1715 harder and harder to keep it in step with the real compile phase, and there
1716 have been a number of bugs (see for example, 4 above). I have now found a
1717 cunning way of running the real compile function in a "fake" mode that
1718 enables it to compute how much memory it would need, while actually only
1719 ever using a few hundred bytes of working memory and without too many
1720 tests of the mode. This should make future maintenance and development
1721 easier. A side effect of this work is that the limit of 200 on the nesting
1722 depth of parentheses has been removed (though this was never a serious
1723 limitation, I suspect). However, there is a downside: pcre_compile() now
1724 runs more slowly than before (30% or more, depending on the pattern). I
1725 hope this isn't a big issue. There is no effect on runtime performance.
1726
1727 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1728 newline (only possible for the last line of a file) and it was a
1729 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1730
1731 19. Added additional timing features to pcretest. (1) The -tm option now times
1732 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1733 separate command line item, by a number that specifies the number of
1734 repeats to use when timing. The default is 50000; this gives better
1735 precision, but takes uncomfortably long for very large patterns.
1736
1737 20. Extended pcre_study() to be more clever in cases where a branch of a
1738 subpattern has no definite first character. For example, (a*|b*)[cd] would
1739 previously give no result from pcre_study(). Now it recognizes that the
1740 first character must be a, b, c, or d.
1741
1742 21. There was an incorrect error "recursive call could loop indefinitely" if
1743 a subpattern (or the entire pattern) that was being tested for matching an
1744 empty string contained only one non-empty item after a nested subpattern.
1745 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1746 incorrectly, because the \d was being skipped in the check.
1747
1748 22. The pcretest program now has a new pattern option /B and a command line
1749 option -b, which is equivalent to adding /B to every pattern. This causes
1750 it to show the compiled bytecode, without the additional information that
1751 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1752 is the same as /B/I).
1753
1754 23. A new optimization is now able automatically to treat some sequences such
1755 as a*b as a*+b. More specifically, if something simple (such as a character
1756 or a simple class like \d) has an unlimited quantifier, and is followed by
1757 something that cannot possibly match the quantified thing, the quantifier
1758 is automatically "possessified".
1759
1760 24. A recursive reference to a subpattern whose number was greater than 39
1761 went wrong under certain circumstances in UTF-8 mode. This bug could also
1762 have affected the operation of pcre_study().
1763
1764 25. Realized that a little bit of performance could be had by replacing
1765 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1766
1767 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1768
1769 27. Possessive quantifiers such as a++ were previously implemented by turning
1770 them into atomic groups such as ($>a+). Now they have their own opcodes,
1771 which improves performance. This includes the automatically created ones
1772 from 23 above.
1773
1774 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1775 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1776 the first matched character to be a colon. This applied both to named and
1777 numbered groups.
1778
1779 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1780
1781 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1782 building PCRE on Windows easier, so I have included it in the distribution.
1783
1784 31. There is now a check in pcretest against a ridiculously large number being
1785 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1786 loop, the loop is abandoned.
1787
1788 32. Forward references to subpatterns in conditions such as (?(2)...) where
1789 subpattern 2 is defined later cause pcre_compile() to search forwards in
1790 the pattern for the relevant set of parentheses. This search went wrong
1791 when there were unescaped parentheses in a character class, parentheses
1792 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1793
1794 33. "Subroutine" calls and backreferences were previously restricted to
1795 referencing subpatterns earlier in the regex. This restriction has now
1796 been removed.
1797
1798 34. Added a number of extra features that are going to be in Perl 5.10. On the
1799 whole, these are just syntactic alternatives for features that PCRE had
1800 previously implemented using the Python syntax or my own invention. The
1801 other formats are all retained for compatibility.
1802
1803 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1804 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1805 also .NET compatible.
1806
1807 (b) A recursion or subroutine call to a named group can now be defined as
1808 (?&name) as well as (?P>name).
1809
1810 (c) A backreference to a named group can now be defined as \k<name> or
1811 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1812 5.10, are also .NET compatible.
1813
1814 (d) A conditional reference to a named group can now use the syntax
1815 (?(<name>) or (?('name') as well as (?(name).
1816
1817 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1818 groups (named and numbered) that are never evaluated inline, but can be
1819 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1820 is always false. There may be only one alternative in such a group.
1821
1822 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1823 as the simple (?(R). The condition is true only if the most recent
1824 recursion is that of the given number or name. It does not search out
1825 through the entire recursion stack.
1826
1827 (g) The escape \gN or \g{N} has been added, where N is a positive or
1828 negative number, specifying an absolute or relative reference.
1829
1830 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1831 some "unreachable code" warnings.
1832
1833 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1834 things, this adds five new scripts.
1835
1836 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1837 There were also incompatibilities regarding the handling of \Q..\E inside
1838 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1839 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1840
1841 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1842 matches an empty string, and forcibly breaks the loop. There were bugs in
1843 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1844 against aaaa the result was just "a" rather than "aaaa", for example. Two
1845 separate and independent bugs (that affected different cases) have been
1846 fixed.
1847
1848 39. Refactored the code to abolish the use of different opcodes for small
1849 capturing bracket numbers. This is a tidy that I avoided doing when I
1850 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1851 The new approach is not only tidier, it makes it possible to reduce the
1852 memory needed to fix the previous bug (38).
1853
1854 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1855 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1856 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1857 mode.
1858
1859 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1860 report.
1861
1862 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1863 copy construction and assignment in the C++ wrapper.
1864
1865 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1866 couple of bugs that could have given wrong results in the "--newline=crlf"
1867 case.
1868
1869 44. Added a number of casts and did some reorganization of signed/unsigned int
1870 variables following suggestions from Dair Grant. Also renamed the variable
1871 "this" as "item" because it is a C++ keyword.
1872
1873 45. Arranged for dftables to add
1874
1875 #include "pcre_internal.h"
1876
1877 to pcre_chartables.c because without it, gcc 4.x may remove the array
1878 definition from the final binary if PCRE is built into a static library and
1879 dead code stripping is activated.
1880
1881 46. For an unanchored pattern, if a match attempt fails at the start of a
1882 newline sequence, and the newline setting is CRLF or ANY, and the next two
1883 characters are CRLF, advance by two characters instead of one.
1884
1885
1886 Version 6.7 04-Jul-06
1887 ---------------------
1888
1889 1. In order to handle tests when input lines are enormously long, pcretest has
1890 been re-factored so that it automatically extends its buffers when
1891 necessary. The code is crude, but this _is_ just a test program. The
1892 default size has been increased from 32K to 50K.
1893
1894 2. The code in pcre_study() was using the value of the re argument before
1895 testing it for NULL. (Of course, in any sensible call of the function, it
1896 won't be NULL.)
1897
1898 3. The memmove() emulation function in pcre_internal.h, which is used on
1899 systems that lack both memmove() and bcopy() - that is, hardly ever -
1900 was missing a "static" storage class specifier.
1901
1902 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1903 containing an extended class (one that cannot be represented by a bitmap
1904 because it contains high-valued characters or Unicode property items, e.g.
1905 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1906 pattern, but PCRE should not loop if you do not (it no longer does).
1907 [Detail: two cases were found: (a) a repeated subpattern containing an
1908 extended class; (b) a recursive reference to a subpattern that followed a
1909 previous extended class. It wasn't skipping over the extended class
1910 correctly when UTF-8 mode was not set.]
1911
1912 5. A negated single-character class was not being recognized as fixed-length
1913 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1914 compile error "lookbehind assertion is not fixed length".
1915
1916 6. The RunPerlTest auxiliary script was showing an unexpected difference
1917 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1918 write a Perl script that can interpret lines of an input file either as
1919 byte characters or as UTF-8, which is what "perltest" was being required to
1920 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1921 can't do is switch easily at run time between having the "use utf8;" pragma
1922 or not. In the end, I fudged it by using the RunPerlTest script to insert
1923 "use utf8;" explicitly for the UTF-8 tests.
1924
1925 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1926 the end of the subject string, contrary to the documentation and to what
1927 Perl does. This was true of both matching functions. Now it matches only at
1928 the start of the subject and immediately after *internal* newlines.
1929
1930 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1931 a pointer to an int instead of a pointer to an unsigned long int. This
1932 caused problems on 64-bit systems.
1933
1934 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1935 instance of the 'standard' template library not being so standard".
1936
1937 10. There was no check on the number of named subpatterns nor the maximum
1938 length of a subpattern name. The product of these values is used to compute
1939 the size of the memory block for a compiled pattern. By supplying a very
1940 long subpattern name and a large number of named subpatterns, the size
1941 computation could be caused to overflow. This is now prevented by limiting
1942 the length of names to 32 characters, and the number of named subpatterns
1943 to 10,000.
1944
1945 11. Subpatterns that are repeated with specific counts have to be replicated in
1946 the compiled pattern. The size of memory for this was computed from the
1947 length of the subpattern and the repeat count. The latter is limited to
1948 65535, but there was no limit on the former, meaning that integer overflow
1949 could in principle occur. The compiled length of a repeated subpattern is
1950 now limited to 30,000 bytes in order to prevent this.
1951
1952 12. Added the optional facility to have named substrings with the same name.
1953
1954 13. Added the ability to use a named substring as a condition, using the
1955 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1956 are numbers (not recommended). Forward references are permitted.
1957
1958 14. Added forward references in named backreferences (if you see what I mean).
1959
1960 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1961 pattern could run off the end of the subject. For example, the pattern
1962 "(?s)(.{1,5})"8 did this with the subject "ab".
1963
1964 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1965 PCRE_CASELESS was set when matching characters that were quantified with ?
1966 or *.
1967
1968 17. A character class other than a single negated character that had a minimum
1969 but no maximum quantifier - for example [ab]{6,} - was not handled
1970 correctly by pce_dfa_exec(). It would match only one character.
1971
1972 18. A valid (though odd) pattern that looked like a POSIX character
1973 class but used an invalid character after [ (for example [[,abc,]]) caused
1974 pcre_compile() to give the error "Failed: internal error: code overflow" or
1975 in some cases to crash with a glibc free() error. This could even happen if
1976 the pattern terminated after [[ but there just happened to be a sequence of
1977 letters, a binary zero, and a closing ] in the memory that followed.
1978
1979 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1980 over the years. Originally (before any Unicode support), just the bottom 8
1981 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1982 output from "man perlunicode" includes this:
1983
1984 The regular expression compiler produces polymorphic opcodes. That
1985 is, the pattern adapts to the data and automatically switches to
1986 the Unicode character scheme when presented with Unicode data--or
1987 instead uses a traditional byte scheme when presented with byte
1988 data.
1989
1990 Sadly, a wide octal escape does not cause a switch, and in a string with
1991 no other multibyte characters, these octal escapes are treated as before.
1992 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1993 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1994 Unicode string.
1995
1996 I have not perpetrated such confusion in PCRE. Up till now, it took just
1997 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1998 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1999 translate to the appropriate multibyte character.
2000
2001 29. Applied some refactoring to reduce the number of warnings from Microsoft
2002 and Borland compilers. This has included removing the fudge introduced
2003 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2004 a warning about an unused variable.
2005
2006 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2007 characters since release 4.0, because Perl (from release 5.004) does not.
2008 [Or at least, is documented not to: some releases seem to be in conflict
2009 with the documentation.] However, when a pattern was studied with
2010 pcre_study() and all its branches started with \s, PCRE still included VT
2011 as a possible starting character. Of course, this did no harm; it just
2012 caused an unnecessary match attempt.
2013
2014 22. Removed a now-redundant internal flag bit that recorded the fact that case
2015 dependency changed within the pattern. This was once needed for "required
2016 byte" processing, but is no longer used. This recovers a now-scarce options
2017 bit. Also moved the least significant internal flag bit to the most-
2018 significant bit of the word, which was not previously used (hangover from
2019 the days when it was an int rather than a uint) to free up another bit for
2020 the future.
2021
2022 23. Added support for CRLF line endings as well as CR and LF. As well as the
2023 default being selectable at build time, it can now be changed at runtime
2024 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2025 specify that it is scanning data with non-default line endings.
2026
2027 24. Changed the definition of CXXLINK to make it agree with the definition of
2028 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2029
2030 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2031 recursions. This makes a big different to stack usage for some patterns.
2032
2033 26. If a subpattern containing a named recursion or subroutine reference such
2034 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2035 the space required for the compiled pattern went wrong and gave too small a
2036 value. Depending on the environment, this could lead to "Failed: internal
2037 error: code overflow at offset 49" or "glibc detected double free or
2038 corruption" errors.
2039
2040 27. Applied patches from Google (a) to support the new newline modes and (b) to
2041 advance over multibyte UTF-8 characters in GlobalReplace.
2042
2043 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2044 difference for some implementation of PCRE in some Windows version.
2045
2046 29. Added some extra testing facilities to pcretest:
2047
2048 \q<number> in a data line sets the "match limit" value
2049 \Q<number> in a data line sets the "match recursion limt" value
2050 -S <number> sets the stack size, where <number> is in megabytes
2051
2052 The -S option isn't available for Windows.
2053
2054
2055 Version 6.6 06-Feb-06
2056 ---------------------
2057
2058 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2059 in pcreposix.h. I have copied the definition from pcre.h.
2060
2061 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2062 because pcre.h is no longer a built file.
2063
2064 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2065 not normally included in the compiled code.
2066
2067
2068 Version 6.5 01-Feb-06
2069 ---------------------
2070
2071 1. When using the partial match feature with pcre_dfa_exec(), it was not
2072 anchoring the second and subsequent partial matches at the new starting
2073 point. This could lead to incorrect results. For example, with the pattern
2074 /1234/, partially matching against "123" and then "a4" gave a match.
2075
2076 2. Changes to pcregrep:
2077
2078 (a) All non-match returns from pcre_exec() were being treated as failures
2079 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2080 error message is output. Some extra information is given for the
2081 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2082 probably the only errors that are likely to be caused by users (by
2083 specifying a regex that has nested indefinite repeats, for instance).
2084 If there are more than 20 of these errors, pcregrep is abandoned.
2085
2086 (b) A binary zero was treated as data while matching, but terminated the
2087 output line if it was written out. This has been fixed: binary zeroes
2088 are now no different to any other data bytes.
2089
2090 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2091 used to set a locale for matching. The --locale=xxxx long option has
2092 been added (no short equivalent) to specify a locale explicitly on the
2093 pcregrep command, overriding the environment variables.
2094
2095 (d) When -B was used with -n, some line numbers in the output were one less
2096 than they should have been.
2097
2098 (e) Added the -o (--only-matching) option.
2099
2100 (f) If -A or -C was used with -c (count only), some lines of context were
2101 accidentally printed for the final match.
2102
2103 (g) Added the -H (--with-filename) option.
2104
2105 (h) The combination of options -rh failed to suppress file names for files
2106 that were found from directory arguments.
2107
2108 (i) Added the -D (--devices) and -d (--directories) options.
2109
2110 (j) Added the -F (--fixed-strings) option.
2111
2112 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2113
2114 (l) Added the --colo(u)r option.
2115
2116 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2117 is not present by default.
2118
2119 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2120 items such as (?R) or (?1), when the recursion could match a number of
2121 alternatives. If it matched one of the alternatives, but subsequently,
2122 outside the recursion, there was a failure, the code tried to back up into
2123 the recursion. However, because of the way PCRE is implemented, this is not
2124 possible, and the result was an incorrect result from the match.
2125
2126 In order to prevent this happening, the specification of recursion has
2127 been changed so that all such subpatterns are automatically treated as
2128 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2129
2130 4. I had overlooked the fact that, in some locales, there are characters for
2131 which isalpha() is true but neither isupper() nor islower() are true. In
2132 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2133 and ordfeminine) are like this. This affected the treatment of \w and \W
2134 when they appeared in character classes, but not when they appeared outside
2135 a character class. The bit map for "word" characters is now created
2136 separately from the results of isalnum() instead of just taking it from the
2137 upper, lower, and digit maps. (Plus the underscore character, of course.)
2138
2139 5. The above bug also affected the handling of POSIX character classes such as
2140 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2141 permanent tables. Instead, the bit maps for such a class were previously
2142 created as the appropriate unions of the upper, lower, and digit bitmaps.
2143 Now they are created by subtraction from the [[:word:]] class, which has
2144 its own bitmap.
2145
2146 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2147 It is created by subtracting the vertical space characters (\x09, \x0a,
2148 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2149 subtraction was done in the overall bitmap for a character class, meaning
2150 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2151 be recognized. This bug has been fixed.
2152
2153 7. Patches from the folks at Google:
2154
2155 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2156 real life, but is still worth protecting against".
2157
2158 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2159 regular expressions".
2160
2161 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2162 have it.
2163
2164 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2165 "configure" and the latter not, in order to fix a problem somebody had
2166 with compiling the Arg class on HP-UX.
2167
2168 (e) Improve the error-handling of the C++ wrapper a little bit.
2169
2170 (f) New tests for checking recursion limiting.
2171
2172 8. The pcre_memmove() function, which is used only if the environment does not
2173 have a standard memmove() function (and is therefore rarely compiled),
2174 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2175 returning a result (though PCRE never actually uses the result).
2176
2177 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2178 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2179 returned instead of calling malloc() with an overflowing number that would
2180 most likely cause subsequent chaos.
2181
2182 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2183
2184 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2185 with this option is matched, the nmatch and pmatch options of regexec() are
2186 ignored.
2187
2188 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2189 provided in case anyone wants to the the POSIX interface with UTF-8
2190 strings.
2191
2192 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2193 C++ linking (needed for some HP-UX environments).
2194
2195 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2196 (unused parameter) and in the pcre_printint() function (omitted "default"
2197 switch label when the default is to do nothing).
2198
2199 15. Added some code to make it possible, when PCRE is compiled as a C++
2200 library, to replace subject pointers for pcre_exec() with a smart pointer
2201 class, thus making it possible to process discontinuous strings.
2202
2203 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2204 much the same function. They were added by different people who were trying
2205 to make PCRE easy to compile on non-Unix systems. It has been suggested
2206 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2207 for compiling on Windows systems. I have therefore replaced it with
2208 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2209 defaults to "extern" for C or "extern C" for C++, which works fine on
2210 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2211 SCOPE with something explicit in config.h. In addition:
2212
2213 (a) pcreposix.h still had just "extern" instead of either of these macros;
2214 I have replaced it with PCRE_DATA_SCOPE.
2215
2216 (b) Functions such as _pcre_xclass(), which are internal to the library,
2217 but external in the C sense, all had PCRE_EXPORT in their definitions.
2218 This is apparently wrong for the Windows case, so I have removed it.
2219 (It makes no difference on Unix-like systems.)
2220
2221 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2222 of recursive calls to match(). This is different to MATCH_LIMIT because
2223 that limits the total number of calls to match(), not all of which increase
2224 the depth of recursion. Limiting the recursion depth limits the amount of
2225 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2226 when PCRE is compiled, and changed at run time. A patch from Google adds
2227 this functionality to the C++ interface.
2228
2229 18. Changes to the handling of Unicode character properties:
2230
2231 (a) Updated the table to Unicode 4.1.0.
2232
2233 (b) Recognize characters that are not in the table as "Cn" (undefined).
2234
2235 (c) I revised the way the table is implemented to a much improved format
2236 which includes recognition of ranges. It now supports the ranges that
2237 are defined in UnicodeData.txt, and it also amalgamates other
2238 characters into ranges. This has reduced the number of entries in the
2239 table from around 16,000 to around 3,000, thus reducing its size
2240 considerably. I realized I did not need to use a tree structure after
2241 all - a binary chop search is just as efficient. Having reduced the
2242 number of entries, I extended their size from 6 bytes to 8 bytes to
2243 allow for more data.
2244
2245 (d) Added support for Unicode script names via properties such as \p{Han}.
2246
2247 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2248 matching that character.
2249
2250 20. When matching a repeated Unicode property with a minimum greater than zero,
2251 (for example \pL{2,}), PCRE could look past the end of the subject if it
2252 reached it while seeking the minimum number of characters. This could
2253 happen only if some of the characters were more than one byte long, because
2254 there is a check for at least the minimum number of bytes.
2255
2256 21. Refactored the implementation of \p and \P so as to be more general, to
2257 allow for more different types of property in future. This has changed the
2258 compiled form incompatibly. Anybody with saved compiled patterns that use
2259 \p or \P will have to recompile them.
2260
2261 22. Added "Any" and "L&" to the supported property types.
2262
2263 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2264 but give a compile time error if the value is greater than 0xff.
2265
2266 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2267 accidentally not being installed or uninstalled.
2268
2269 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2270 made were to insert the current release number. This seemed silly, because
2271 it made things harder for people building PCRE on systems that don't run
2272 "configure". I have turned pcre.h into a distributed file, no longer built
2273 by "configure", with the version identification directly included. There is
2274 no longer a pcre.h.in file.
2275
2276 However, this change necessitated a change to the pcre-config script as
2277 well. It is built from pcre-config.in, and one of the substitutions was the
2278 release number. I have updated configure.ac so that ./configure now finds
2279 the release number by grepping pcre.h.
2280
2281 26. Added the ability to run the tests under valgrind.
2282
2283
2284 Version 6.4 05-Sep-05
2285 ---------------------
2286
2287 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2288 "--" to be printed when multiple files were scanned, even when none of the
2289 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2290 consider it to be a bug, and have restored the previous behaviour.
2291
2292 2. A couple of code tidies to get rid of compiler warnings.
2293
2294 3. The pcretest program used to cheat by referring to symbols in the library
2295 whose names begin with _pcre_. These are internal symbols that are not
2296 really supposed to be visible externally, and in some environments it is
2297 possible to suppress them. The cheating is now confined to including
2298 certain files from the library's source, which is a bit cleaner.
2299
2300 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2301 file's purpose clearer.
2302
2303 5. Reorganized pcre_ucp_findchar().
2304
2305
2306 Version 6.3 15-Aug-05
2307 ---------------------
2308
2309 1. The file libpcre.pc.in did not have general read permission in the tarball.
2310
2311 2. There were some problems when building without C++ support:
2312
2313 (a) If C++ support was not built, "make install" and "make test" still
2314 tried to test it.
2315
2316 (b) There were problems when the value of CXX was explicitly set. Some
2317 changes have been made to try to fix these, and ...
2318
2319 (c) --disable-cpp can now be used to explicitly disable C++ support.
2320
2321 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2322 backslash in a target when C++ was disabled. This confuses some
2323 versions of "make", apparently. Using an intermediate variable solves
2324 this. (Same for CPP_LOBJ.)
2325
2326 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2327 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2328 necessary on certain architectures.
2329
2330 4. Added a setting of -export-symbols-regex to the link command to remove
2331 those symbols that are exported in the C sense, but actually are local
2332 within the library, and not documented. Their names all begin with
2333 "_pcre_". This is not a perfect job, because (a) we have to except some
2334 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2335 available (and never for static libraries). I have made a note to try to
2336 find a way round (a) in the future.
2337
2338
2339 Version 6.2 01-Aug-05
2340 ---------------------
2341
2342 1. There was no test for integer overflow of quantifier values. A construction
2343 such as {1111111111111111} would give undefined results. What is worse, if
2344 a minimum quantifier for a parenthesized subpattern overflowed and became
2345 negative, the calculation of the memory size went wrong. This could have
2346 led to memory overwriting.
2347
2348 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2349
2350 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2351 operating environments where this matters.
2352
2353 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2354 PCRE options from within the C++ wrapper.
2355
2356 5. Named capturing subpatterns were not being correctly counted when a pattern
2357 was compiled. This caused two problems: (a) If there were more than 100
2358 such subpatterns, the calculation of the memory needed for the whole
2359 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2360 back references of the form \12, where the number was greater than 9, were
2361 not recognized as back references, even though there were sufficient
2362 previous subpatterns.
2363
2364 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2365 versions of gcc, e.g. 2.95.4.
2366
2367
2368 Version 6.1 21-Jun-05
2369 ---------------------
2370
2371 1. There was one reference to the variable "posix" in pcretest.c that was not
2372 surrounded by "#if !defined NOPOSIX".
2373
2374 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2375 the cross-check on the old pcre_info() function, for the benefit of the
2376 cut-down version of PCRE that is currently imported into Exim.
2377
2378 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2379 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2380 patterns that start (?i) but I don't think that matters. The use of (?i) is
2381 just an example; this all applies to the other options as well.
2382
2383 4. Since libtool seems to echo the compile commands it is issuing, the output
2384 from "make" can be reduced a bit by putting "@" in front of each libtool
2385 compile command.
2386
2387 5. Patch from the folks at Google for configure.in to be a bit more thorough
2388 in checking for a suitable C++ installation before trying to compile the
2389 C++ stuff. This should fix a reported problem when a compiler was present,
2390 but no suitable headers.
2391
2392 6. The man pages all had just "PCRE" as their title. I have changed them to
2393 be the relevant file name. I have also arranged that these names are
2394 retained in the file doc/pcre.txt, which is a concatenation in text format
2395 of all the man pages except the little individual ones for each function.
2396
2397 7. The NON-UNIX-USE file had not been updated for the different set of source
2398 files that come with release 6. I also added a few comments about the C++
2399 wrapper.
2400
2401
2402 Version 6.0 07-Jun-05
2403 ---------------------
2404
2405 1. Some minor internal re-organization to help with my DFA experiments.
2406
2407 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2408 didn't matter for the library itself when fully configured, but did matter
2409 when compiling without UCP support, or within Exim, where the ucp files are
2410 not imported.
2411
2412 3. Refactoring of the library code to split up the various functions into
2413 different source modules. The addition of the new DFA matching code (see
2414 below) to a single monolithic source would have made it really too
2415 unwieldy, quite apart from causing all the code to be include in a
2416 statically linked application, when only some functions are used. This is
2417 relevant even without the DFA addition now that patterns can be compiled in
2418 one application and matched in another.
2419
2420 The downside of splitting up is that there have to be some external
2421 functions and data tables that are used internally in different modules of
2422 the library but which are not part of the API. These have all had their
2423 names changed to start with "_pcre_" so that they are unlikely to clash
2424 with other external names.
2425
2426 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2427 a different (DFA) algorithm. Although it is slower than the original
2428 function, it does have some advantages for certain types of matching
2429 problem.
2430
2431 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2432 including restarting after a partial match.
2433
2434 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2435 defined when compiling for Windows was sent to me. I have put it into the
2436 code, though I have no means of testing or verifying it.
2437
2438 7. Added the pcre_refcount() auxiliary function.
2439
2440 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2441 match before or at the first newline in the subject string. In pcretest,
2442 the /f option on a pattern can be used to set this.
2443
2444 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2445 would behave wrongly. This has been present in PCRE since release 4.0.
2446
2447 10. A number of changes to the pcregrep command:
2448
2449 (a) Refactored how -x works; insert ^(...)$ instead of setting
2450 PCRE_ANCHORED and checking the length, in preparation for adding
2451 something similar for -w.
2452
2453 (b) Added the -w (match as a word) option.
2454
2455 (c) Refactored the way lines are read and buffered so as to have more
2456 than one at a time available.
2457
2458 (d) Implemented a pcregrep test script.
2459
2460 (e) Added the -M (multiline match) option. This allows patterns to match
2461 over several lines of the subject. The buffering ensures that at least
2462 8K, or the rest of the document (whichever is the shorter) is available
2463 for matching (and similarly the previous 8K for lookbehind assertions).
2464
2465 (f) Changed the --help output so that it now says
2466
2467 -w, --word-regex(p)
2468
2469 instead of two lines, one with "regex" and the other with "regexp"
2470 because that confused at least one person since the short forms are the
2471 same. (This required a bit of code, as the output is generated
2472 automatically from a table. It wasn't just a text change.)
2473
2474 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2475 option but starts with a hyphen. Could be a pattern or a path name
2476 starting with a hyphen, for instance.
2477
2478 (h) "-" can be given as a file name to represent stdin.
2479
2480 (i) When file names are being printed, "(standard input)" is used for
2481 the standard input, for compatibility with GNU grep. Previously
2482 "<stdin>" was used.
2483
2484 (j) The option --label=xxx can be used to supply a name to be used for
2485 stdin when file names are being printed. There is no short form.
2486
2487 (k) Re-factored the options decoding logic because we are going to add
2488 two more options that take data. Such options can now be given in four
2489 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2490
2491 (l) Added the -A, -B, and -C options for requesting that lines of context
2492 around matches be printed.
2493
2494 (m) Added the -L option to print the names of files that do not contain
2495 any matching lines, that is, the complement of -l.
2496
2497 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2498 continue to scan other files.
2499
2500 (o) The -s option was incorrectly implemented. For compatibility with other
2501 greps, it now suppresses the error message for a non-existent or non-
2502 accessible file (but not the return code). There is a new option called
2503 -q that suppresses the output of matching lines, which was what -s was
2504 previously doing.
2505
2506 (p) Added --include and --exclude options to specify files for inclusion
2507 and exclusion when recursing.
2508
2509 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2510 Hopefully, it now does.
2511
2512 12. Missing cast in pcre_study().
2513
2514 13. Added an "uninstall" target to the makefile.
2515
2516 14. Replaced "extern" in the function prototypes in Makefile.in with
2517 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2518 world, but is set differently for Windows.
2519
2520 15. Added a second compiling function called pcre_compile2(). The only
2521 difference is that it has an extra argument, which is a pointer to an
2522 integer error code. When there is a compile-time failure, this is set
2523 non-zero, in addition to the error test pointer being set to point to an
2524 error message. The new argument may be NULL if no error number is required
2525 (but then you may as well call pcre_compile(), which is now just a
2526 wrapper). This facility is provided because some applications need a
2527 numeric error indication, but it has also enabled me to tidy up the way
2528 compile-time errors are handled in the POSIX wrapper.
2529
2530 16. Added VPATH=.libs to the makefile; this should help when building with one
2531 prefix path and installing with another. (Or so I'm told by someone who
2532 knows more about this stuff than I do.)
2533
2534 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2535 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2536 match everything, including newlines. This is not POSIX-compatible, but
2537 somebody wanted the feature. From pcretest it can be activated by using
2538 both the P and the s flags.
2539
2540 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2541
2542 19. libpcre.pc was being incorrectly installed as executable.
2543
2544 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2545 it now also looks for '\r' so that it will work unmodified on Windows.
2546
2547 21. Added Google's contributed C++ wrapper to the distribution.
2548
2549 22. Added some untidy missing memory free() calls in pcretest, to keep
2550 Electric Fence happy when testing.
2551
2552
2553
2554 Version 5.0 13-Sep-04
2555 ---------------------
2556
2557 1. Internal change: literal characters are no longer packed up into items
2558 containing multiple characters in a single byte-string. Each character
2559 is now matched using a separate opcode. However, there may be more than one
2560 byte in the character in UTF-8 mode.
2561
2562 2. The pcre_callout_block structure has two new fields: pattern_position and
2563 next_item_length. These contain the offset in the pattern to the next match
2564 item, and its length, respectively.
2565
2566 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2567 insertion of callouts before each pattern item. Added the /C option to
2568 pcretest to make use of this.
2569
2570 4. On the advice of a Windows user, the lines
2571
2572 #if defined(_WIN32) || defined(WIN32)
2573 _setmode( _fileno( stdout ), 0x8000 );
2574 #endif /* defined(_WIN32) || defined(WIN32) */
2575
2576 have been added to the source of pcretest. This apparently does useful
2577 magic in relation to line terminators.
2578
2579 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2580 for the benefit of those environments where the "b" makes a difference.
2581
2582 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2583 to know about it. I have put a hack into configure.in that adds in code
2584 to set GCC=yes if CC=icc. This seems to end up at a point in the
2585 generated configure script that is early enough to affect the setting of
2586 compiler options, which is what is needed, but I have no means of testing
2587 whether it really works. (The user who reported this had patched the
2588 generated configure script, which of course I cannot do.)
2589
2590 LATER: After change 22 below (new libtool files), the configure script
2591 seems to know about icc (and also ecc). Therefore, I have commented out
2592 this hack in configure.in.
2593
2594 7. Added support for pkg-config (2 patches were sent in).
2595
2596 8. Negated POSIX character classes that used a combination of internal tables
2597 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2598 [[:^ascii]]. Typically, they would match almost any characters. The other
2599 POSIX classes were not broken in this way.
2600
2601 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2602 to find the match, as PCRE was deluded into thinking that the match had to
2603 start at the start point or following a newline. The same bug applied to
2604 patterns with negative forward assertions or any backward assertions
2605 preceding ".*" at the start, unless the pattern required a fixed first
2606 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2607
2608 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2609 starting at the last subject character, bytes beyond the end of the subject
2610 string were read.
2611
2612 11. Renamed the variable "class" as "classbits" to make life easier for C++
2613 users. (Previously there was a macro definition, but it apparently wasn't
2614 enough.)
2615
2616 12. Added the new field "tables" to the extra data so that tables can be passed
2617 in at exec time, or the internal tables can be re-selected. This allows
2618 a compiled regex to be saved and re-used at a later time by a different
2619 program that might have everything at different addresses.
2620
2621 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2622 -R library as well as a -L library.
2623
2624 14. The debugging options of pcretest (-d on the command line or D on a
2625 pattern) showed incorrect output for anything following an extended class
2626 that contained multibyte characters and which was followed by a quantifier.
2627
2628 15. Added optional support for general category Unicode character properties
2629 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2630 support. It adds about 90K to the size of the library. The meanings of the
2631 inbuilt class escapes such as \d and \s have NOT been changed.
2632
2633 16. Updated pcredemo.c to include calls to free() to release the memory for the
2634 compiled pattern.
2635
2636 17. The generated file chartables.c was being created in the source directory
2637 instead of in the building directory. This caused the build to fail if the
2638 source directory was different from the building directory, and was
2639 read-only.
2640
2641 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2642 file. No doubt somebody will tell me if they don't make sense... Also added
2643 Dan Mooney's comments about building on OpenVMS.
2644
2645 19. Added support for partial matching via the PCRE_PARTIAL option for
2646 pcre_exec() and the \P data escape in pcretest.
2647
2648 20. Extended pcretest with 3 new pattern features:
2649
2650 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2651 write the compiled pattern to the file whose name is "rest-of-line".
2652 This is a straight binary dump of the data, with the saved pointer to
2653 the character tables forced to be NULL. The study data, if any, is
2654 written too. After writing, pcretest reads a new pattern.
2655
2656 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2657 compiled pattern from the given file. There must not be any
2658 occurrences of "<" in the file name (pretty unlikely); if there are,
2659 pcretest will instead treat the initial "<" as a pattern delimiter.
2660 After reading in the pattern, pcretest goes on to read data lines as
2661 usual.
2662
2663 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2664 and 16-bit fields in a compiled pattern, to simulate a pattern that
2665 was compiled on a host of opposite endianness.
2666
2667 21. The pcre-exec() function can now cope with patterns that were compiled on
2668 hosts of opposite endianness, with this restriction:
2669
2670 As for any compiled expression that is saved and used later, the tables
2671 pointer field cannot be preserved; the extra_data field in the arguments
2672 to pcre_exec() should be used to pass in a tables address if a value
2673 other than the default internal tables were used at compile time.
2674
2675 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2676 now diagnosed as an error. Previously, most of the time, a negative number
2677 would have been treated as zero, but if in addition "ovector" was passed as
2678 NULL, a crash could occur.
2679
2680 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2681 new versions from the libtool 1.5 distribution (the last one is a copy of
2682 a file called libtool.m4). This seems to have fixed the need to patch
2683 "configure" to support Darwin 1.3 (which I used to do). However, I still
2684 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2685 workstation).
2686
2687 24. Changed the PCRE licence to be the more standard "BSD" licence.
2688
2689
2690 Version 4.5 01-Dec-03
2691 ---------------------
2692
2693 1. There has been some re-arrangement of the code for the match() function so
2694 that it can be compiled in a version that does not call itself recursively.
2695 Instead, it keeps those local variables that need separate instances for
2696 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2697 needs to "recurse". Keeping track of where control must go is done by means
2698 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2699 hide most of the details from the main code, and operates only if
2700 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2701 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2702 operating.
2703
2704 To make it easier for callers to provide specially tailored get/free
2705 functions for this usage, two new functions, pcre_stack_malloc, and
2706 pcre_stack_free, are used. They are always called in strict stacking order,
2707 and the size of block requested is always the same.
2708
2709 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2710 PCRE has been compiled to use the stack or the heap for recursion. The
2711 -C option of pcretest uses this to show which version is compiled.
2712
2713 A new data escape \S, is added to pcretest; it causes the amounts of store
2714 obtained and freed by both kinds of malloc/free at match time to be added
2715 to the output.
2716
2717 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2718 what's available on my current Linux desktop machine.
2719
2720 3. When matching a UTF-8 string, the test for a valid string at the start has
2721 been extended. If start_offset is not zero, PCRE now checks that it points
2722 to a byte that is the start of a UTF-8 character. If not, it returns
2723 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2724 this is necessary because there may be backward assertions in the pattern.
2725 When matching the same subject several times, it may save resources to use
2726 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2727
2728 4. The code for checking the validity of UTF-8 strings has been tightened so
2729 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2730 containing "overlong sequences".
2731
2732 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2733 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2734 should have been "&", but it just so happened that all the cases this let
2735 through by mistake were picked up later in the function.
2736
2737 6. I had used a variable called "isblank" - this is a C99 function, causing
2738 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2739
2740 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2741 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2742 so the version is shown; (c) stop "make test" echoing "./RunTest".
2743
2744 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2745
2746 9. The prototype for memmove() for systems that don't have it was using
2747 size_t, but the inclusion of the header that defines size_t was later. I've
2748 moved the #includes for the C headers earlier to avoid this.
2749
2750 10. Added some adjustments to the code to make it easier to compiler on certain
2751 special systems:
2752
2753 (a) Some "const" qualifiers were missing.
2754 (b) Added the macro EXPORT before all exported functions; by default this
2755 is defined to be empty.
2756 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2757 that it reads its output file name as an argument instead of writing
2758 to the standard output and assuming this can be redirected.
2759
2760 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2761 class containing characters with values greater than 255, PCRE compilation
2762 went into a loop.
2763
2764 12. A recursive reference to a subpattern that was within another subpattern
2765 that had a minimum quantifier of zero caused PCRE to crash. For example,
2766 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2767 recursion. If the recursively-called subpattern itself had a zero repeat,
2768 that was OK.
2769
2770 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2771 buffer into which it was copied (for escape processing) was still set at
2772 1024, so long lines caused crashes.
2773
2774 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2775 "internal error: code overflow...". This applied to any character class
2776 that was followed by a possessive quantifier.
2777
2778 15. Modified the Makefile to add libpcre.la as a prerequisite for
2779 libpcreposix.la because I was told this is needed for a parallel build to
2780 work.
2781
2782 16. If a pattern that contained .* following optional items at the start was
2783 studied, the wrong optimizing data was generated, leading to matching
2784 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2785 matching string must start with a or b or c. The correct conclusion for
2786 this pattern is that a match can start with any character.
2787
2788
2789 Version 4.4 13-Aug-03
2790 ---------------------
2791
2792 1. In UTF-8 mode, a character class containing characters with values between
2793 127 and 255 was not handled correctly if the compiled pattern was studied.
2794 In fixing this, I have also improved the studying algorithm for such
2795 classes (slightly).
2796
2797 2. Three internal functions had redundant arguments passed to them. Removal
2798 might give a very teeny performance improvement.
2799
2800 3. Documentation bug: the value of the capture_top field in a callout is *one
2801 more than* the number of the hightest numbered captured substring.
2802
2803 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2804 in incorrectly linking with a previously installed version. They now link
2805 explicitly with libpcre.la.
2806
2807 5. configure.in no longer needs to recognize Cygwin specially.
2808
2809 6. A problem in pcre.in for Windows platforms is fixed.
2810
2811 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2812 pcretest, it used to include the size of the study block as part of its
2813 output. Unfortunately, the structure contains a field that has a different
2814 size on different hardware architectures. This meant that the tests that
2815 showed this size failed. As the block is currently always of a fixed size,
2816 this information isn't actually particularly useful in pcretest output, so
2817 I have just removed it.
2818
2819 8. Three pre-processor statements accidentally did not start in column 1.
2820 Sadly, there are *still* compilers around that complain, even though
2821 standard C has not required this for well over a decade. Sigh.
2822
2823 9. In pcretest, the code for checking callouts passed small integers in the
2824 callout_data field, which is a void * field. However, some picky compilers
2825 complained about the casts involved for this on 64-bit systems. Now
2826 pcretest passes the address of the small integer instead, which should get
2827 rid of the warnings.
2828
2829 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2830 both compile and run time, and gives an error if an invalid UTF-8 sequence
2831 is found. There is a option for disabling this check in cases where the
2832 string is known to be correct and/or the maximum performance is wanted.
2833
2834 11. In response to a bug report, I changed one line in Makefile.in from
2835
2836 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2837 to
2838 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2839
2840 to look similar to other lines, but I have no way of telling whether this
2841 is the right thing to do, as I do not use Windows. No doubt I'll get told
2842 if it's wrong...
2843
2844
2845 Version 4.3 21-May-03
2846 ---------------------
2847
2848 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2849 Makefile.
2850
2851 2. Some refactoring to improve the quality of the code:
2852
2853 (i) The utf8_table... variables are now declared "const".
2854
2855 (ii) The code for \cx, which used the "case flipping" table to upper case
2856 lower case letters, now just substracts 32. This is ASCII-specific,
2857 but the whole concept of \cx is ASCII-specific, so it seems
2858 reasonable.
2859
2860 (iii) PCRE was using its character types table to recognize decimal and
2861 hexadecimal digits in the pattern. This is silly, because it handles
2862 only 0-9, a-f, and A-F, but the character types table is locale-
2863 specific, which means strange things might happen. A private
2864 table is now used for this - though it costs 256 bytes, a table is
2865 much faster than multiple explicit tests. Of course, the standard
2866 character types table is still used for matching digits in subject
2867 strings against \d.
2868
2869 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2870 ending in _t are). So I've renamed it as ESC_tee.
2871
2872 3. The first argument for regexec() in the POSIX wrapper should have been
2873 defined as "const".
2874
2875 4. Changed pcretest to use malloc() for its buffers so that they can be
2876 Electric Fenced for debugging.
2877
2878 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2879 to read one or more bytes before the start of the subject string. Often this
2880 had no effect on PCRE's behaviour, but in some circumstances it could
2881 provoke a segmentation fault.
2882
2883 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2884 to try to read one or more bytes before the start of the subject string.
2885
2886 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2887 UTF-8 support could misbehave in various ways if the subject string
2888 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2889 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2890 back over UTF-8 characters.)
2891
2892
2893 Version 4.2 14-Apr-03
2894 ---------------------
2895
2896 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2897
2898 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2899 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2900 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2901 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2902 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2903 and BUILD_EXEEXT
2904 Note: automatic setting of the BUILD variables is not yet working
2905 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2906 compile-time but not at link-time
2907 [LINK]: use for linking executables only
2908 make different versions for Windows and non-Windows
2909 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2910 libraries
2911 [LINK_FOR_BUILD]: new variable
2912 [OBJEXT]: use throughout
2913 [EXEEXT]: use throughout
2914 <winshared>: new target
2915 <wininstall>: new target
2916 <dftables.o>: use native compiler
2917 <dftables>: use native linker
2918 <install>: handle Windows platform correctly
2919 <clean>: ditto
2920 <check>: ditto
2921 copy DLL to top builddir before testing
2922
2923 As part of these changes, -no-undefined was removed again. This was reported
2924 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2925 in any case.
2926
2927 3. Some tidies to get rid of compiler warnings:
2928
2929 . In the match_data structure, match_limit was an unsigned long int, whereas
2930 match_call_count was an int. I've made them both unsigned long ints.
2931
2932 . In pcretest the fact that a const uschar * doesn't automatically cast to
2933 a void * provoked a warning.
2934
2935 . Turning on some more compiler warnings threw up some "shadow" variables
2936 and a few more missing casts.
2937
2938 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2939 option, a class that contained a single character with a value between 128
2940 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2941
2942 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2943 option, a class that contained several characters, but with at least one
2944 whose value was between 128 and 255 caused PCRE to crash.
2945
2946
2947 Version 4.1 12-Mar-03
2948 ---------------------
2949
2950 1. Compiling with gcc -pedantic found a couple of places where casts were
2951 needed, and a string in dftables.c that was longer than standard compilers are
2952 required to support.
2953
2954 2. Compiling with Sun's compiler found a few more places where the code could
2955 be tidied up in order to avoid warnings.
2956
2957 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2958 first of these names is deprecated in the latest Autoconf in favour of the name
2959 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2960 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2961 analogy I have changed it to CFLAGS_FOR_BUILD.
2962
2963 4. Added -no-undefined to the linking command in the Makefile, because this is
2964 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2965 linking step for the pcreposix library.
2966
2967 5. PCRE was failing to diagnose the case of two named groups with the same
2968 name.
2969
2970 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2971 literal character that is needed in the subject for a match, and scans along to
2972 ensure that it is present before embarking on the full matching process. This
2973 saves time in cases of nested unlimited repeats that are never going to match.
2974 Problem: the scan can take a lot of time if the subject is very long (e.g.
2975 megabytes), thus penalizing straightforward matches. It is now done only if the
2976 amount of subject to be scanned is less than 1000 bytes.
2977
2978 7. A lesser problem with the same optimization is that it was recording the
2979 first character of an anchored pattern as "needed", thus provoking a search
2980 right along the subject, even when the first match of the pattern was going to
2981 fail. The "needed" character is now not set for anchored patterns, unless it
2982 follows something in the pattern that is of non-fixed length. Thus, it still
2983 fulfils its original purpose of finding quick non-matches in cases of nested
2984 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2985
2986
2987 Version 4.0 17-Feb-03
2988 ---------------------
2989
2990 1. If a comment in an extended regex that started immediately after a meta-item
2991 extended to the end of string, PCRE compiled incorrect data. This could lead to
2992 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2993
2994 2. Moved to autoconf 2.53 and libtool 1.4.2.
2995
2996 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2997 the special perltest8 script is no longer needed - all the tests can be run
2998 from a single perltest script.
2999
3000 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3001 by \s. It has now been removed in PCRE. This means it isn't recognized as
3002 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3003 class [:space:] *does* include VT, thereby creating a mess.
3004
3005 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3006 space and tab.
3007
3008 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3009 its new features into the main test script, reducing the number of scripts.
3010
3011 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3012 were backward compatible, and made the (?i) apply to the whole pattern, as if
3013 /i were given. Now it behaves more logically, and applies the option setting
3014 only to what follows. PCRE has been changed to follow suit. However, if it
3015 finds options settings right at the start of the pattern, it extracts them into
3016 the global options, as before. Thus, they show up in the info data.
3017
3018 8. Added support for the \Q...\E escape sequence. Characters in between are
3019 treated as literals. This is slightly different from Perl in that $ and @ are
3020 also handled as literals inside the quotes. In Perl, they will cause variable
3021 interpolation. Note the following examples:
3022
3023 Pattern PCRE matches Perl matches
3024
3025 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3026 \Qabc\$xyz\E abc\$xyz abc\$xyz
3027 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3028
3029 For compatibility with Perl, \Q...\E sequences are recognized inside character
3030 classes as well as outside them.
3031
3032 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3033 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3034 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3035 signed/unsigned warnings.
3036
3037 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3038 option for pcretest, so I've replaced it by a simple function that does just
3039 that job.
3040
3041 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3042 "pcregrep -".
3043
3044 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3045 Java package. This provides some syntactic sugar for simple cases of what my
3046 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3047 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3048 item, you can use this simplified notation. Note that only makes sense with
3049 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3050 greediness, whatever the setting of the PCRE_UNGREEDY option.
3051
3052 13. A change of greediness default within a pattern was not taking effect at
3053 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3054 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3055 was abstracted outside.
3056
3057 14. PCRE now supports the \G assertion. It is true when the current matching
3058 position is at the start point of the match. This differs from \A when the
3059 starting offset is non-zero. Used with the /g option of pcretest (or similar
3060 code), it works in the same way as it does for Perl's /g option. If all
3061 alternatives of a regex begin with \G, the expression is anchored to the start
3062 match position, and the "anchored" flag is set in the compiled expression.
3063
3064 15. Some bugs concerning the handling of certain option changes within patterns
3065 have been fixed. These applied to options other than (?ims). For example,
3066 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3067 been the other way round. Some of this was related to change 7 above.
3068
3069 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3070 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3071 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3072 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3073
3074 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3075 mode. Unlike ".", it always matches newline, whatever the setting of
3076 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3077 assertions. Perl allows it, but it doesn't (in general) work because it can't
3078 calculate the length of the lookbehind. At least, that's the case for Perl
3079 5.8.0 - I've been told they are going to document that it doesn't work in
3080 future.
3081
3082 18. Added an error diagnosis for escapes that PCRE does not support: these are
3083 \L, \l, \N, \P, \p, \U, \u, and \X.
3084
3085 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3086 reading past the end of the pattern in cases such as /[abcd/.
3087
3088 20. PCRE was getting more memory than necessary for patterns with classes that
3089 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3090
3091 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3092 compiling PCRE for use with Virtual Pascal.
3093
3094 22. Small fix to the Makefile to make it work properly if the build is done
3095 outside the source tree.
3096
3097 23. Added a new extension: a condition to go with recursion. If a conditional
3098 subpattern starts with (?(R) the "true" branch is used if recursion has
3099 happened, whereas the "false" branch is used only at the top level.
3100
3101 24. When there was a very long string of literal characters (over 255 bytes
3102 without UTF support, over 250 bytes with UTF support), the computation of how
3103 much memory was required could be incorrect, leading to segfaults or other
3104 strange effects.
3105
3106 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3107 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3108 there was a subsequent back reference to those brackets. This meant that, for
3109 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3110 possible to check for precisely this case. All we can do is abandon the
3111 optimization if .* occurs inside capturing brackets when there are any back
3112 references whatsoever. (See below for a better fix that came later.)
3113
3114 26. The handling of the optimization for finding the first character of a
3115 non-anchored pattern, and for finding a character that is required later in the
3116 match were failing in some cases. This didn't break the matching; it just
3117 failed to optimize when it could. The way this is done has been re-implemented.
3118
3119 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3120
3121 28. Added a new feature that provides some of the functionality that Perl
3122 provides with (?{...}). The facility is termed a "callout". The way it is done
3123 in PCRE is for the caller to provide an optional function, by setting
3124 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3125 global variable. By default it is unset, which disables all calling out. To get
3126 the function called, the regex must include (?C) at appropriate points. This
3127 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3128 This provides a means of identifying different callout points. When PCRE
3129 reaches such a point in the regex, if pcre_callout has been set, the external
3130 function is called. It is provided with data in a structure called
3131 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3132 matching continues; if it returns a non-zero value, the match at the current
3133 point fails. However, backtracking will occur if possible. [This was changed
3134 later and other features added - see item 49 below.]
3135
3136 29. pcretest is upgraded to test the callout functionality. It provides a
3137 callout function that displays information. By default, it shows the start of
3138 the match and the current position in the text. There are some new data escapes
3139 to vary what happens:
3140
3141 \C+ in addition, show current contents of captured substrings
3142 \C- do not supply a callout function
3143 \C!n return 1 when callout number n is reached
3144 \C!n!m return 1 when callout number n is reached for the mth time
3145
3146 30. If pcregrep was called with the -l option and just a single file name, it
3147 output "<stdin>" if a match was found, instead of the file name.
3148
3149 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3150 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3151 pcre_exec(). This saves a malloc/free per call. The default value of
3152 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3153 when configuring.
3154
3155 32. The default maximum size of a compiled pattern is 64K. There have been a
3156 few cases of people hitting this limit. The code now uses macros to handle the
3157 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3158 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3159 configuring. Tests 2 and 5 work only with 2-byte links because they output
3160 debugging information about compiled patterns.
3161
3162 33. Internal code re-arrangements:
3163
3164 (a) Moved the debugging function for printing out a compiled regex into
3165 its own source file (printint.c) and used #include to pull it into
3166 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3167 separate copies.
3168
3169 (b) Defined the list of op-code names for debugging as a macro in
3170 internal.h so that it is next to the definition of the opcodes.
3171
3172 (c) Defined a table of op-code lengths for simpler skipping along compiled
3173 code. This is again a macro in internal.h so that it is next to the
3174 definition of the opcodes.
3175
3176 34. Added support for recursive calls to individual subpatterns, along the
3177 lines of Robin Houston's patch (but implemented somewhat differently).
3178
3179 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3180 allow it to read and process whole directories in Win32. This code was
3181 contributed by Lionel Fourquaux; it has not been tested by me.
3182
3183 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3184 used to name a group. Names consist of alphanumerics and underscores, and must
3185 be unique. Back references use the syntax (?P=name) and recursive calls use
3186 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3187 numbers. The function pcre_fullinfo() can be used after compilation to extract
3188 a name/number map. There are three relevant calls:
3189
3190 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3191 PCRE_INFO_NAMECOUNT yields the number of entries
3192 PCRE_INFO_NAMETABLE yields a pointer to the map.
3193
3194 The map is a vector of fixed-size entries. The size of each entry depends on
3195 the length of the longest name used. The first two bytes of each entry are the
3196 group number, most significant byte first. There follows the corresponding
3197 name, zero terminated. The names are in alphabetical order.
3198
3199 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3200 case instead of 255. Making it the same both with and without UTF-8 support
3201 means that the same test output works with both.
3202
3203 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3204 calling malloc() with a zero argument.
3205
3206 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3207 optimization. I've improved things by keeping a bitmap of backreferences with
3208 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3209 fact referenced, the optimization can be applied. It is unlikely that a
3210 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3211 the match to follow \n) will appear inside brackets with a number greater than
3212 31, but if it does, any back reference > 31 suppresses the optimization.
3213
3214 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3215 of disabling numbered capturing parentheses. Any opening parenthesis that is
3216 not followed by ? behaves as if it were followed by ?: but named parentheses
3217 can still be used for capturing (and they will acquire numbers in the usual
3218 way).
3219
3220 41. Redesigned the return codes from the match() function into yes/no/error so
3221 that errors can be passed back from deep inside the nested calls. A malloc
3222 failure while inside a recursive subpattern call now causes the
3223 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3224
3225 42. It is now possible to set a limit on the number of times the match()
3226 function is called in a call to pcre_exec(). This facility makes it possible to
3227 limit the amount of recursion and backtracking, though not in a directly
3228 obvious way, because the match() function is used in a number of different
3229 circumstances. The count starts from zero for each position in the subject
3230 string (for non-anchored patterns). The default limit is, for compatibility, a
3231 large number, namely 10 000 000. You can change this in two ways:
3232
3233 (a) When configuring PCRE before making, you can use --with-match-limit=n
3234 to set a default value for the compiled library.
3235
3236 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3237 a different value is set. See 45 below.
3238
3239 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3240
3241 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3242 of things that can be changed at compile time. The first argument specifies
3243 what is wanted and the second points to where the information is to be placed.
3244 The current list of available information is:
3245
3246 PCRE_CONFIG_UTF8
3247
3248 The output is an integer that is set to one if UTF-8 support is available;
3249 otherwise it is set to zero.
3250
3251 PCRE_CONFIG_NEWLINE
3252
3253 The output is an integer that it set to the value of the code that is used for
3254 newline. It is either LF (10) or CR (13).
3255
3256 PCRE_CONFIG_LINK_SIZE
3257
3258 The output is an integer that contains the number of bytes used for internal
3259 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3260
3261 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3262
3263 The output is an integer that contains the threshold above which the POSIX
3264 interface uses malloc() for output vectors. See item 31 above.
3265
3266 PCRE_CONFIG_MATCH_LIMIT
3267
3268 The output is an unsigned integer that contains the default limit of the number
3269 of match() calls in a pcre_exec() execution. See 42 above.
3270
3271 44. pcretest has been upgraded by the addition of the -C option. This causes it
3272 to extract all the available output from the new pcre_config() function, and to
3273 output it. The program then exits immediately.
3274
3275 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3276 order to support additional features. One way would have been to define
3277 pcre_exec2() (for example) with extra arguments, but this would not have been
3278 extensible, and would also have required all calls to the original function to
3279 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3280 is used for passing in "extra" data from pcre_study().
3281
3282 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3283 contains the following fields:
3284
3285 flags a bitmap indicating which of the following fields are set
3286 study_data opaque data from pcre_study()
3287 match_limit a way of specifying a limit on match() calls for a specific
3288 call to pcre_exec()
3289 callout_data data for callouts (see 49 below)
3290
3291 The flag bits are also defined in pcre.h, and are
3292
3293 PCRE_EXTRA_STUDY_DATA
3294 PCRE_EXTRA_MATCH_LIMIT
3295 PCRE_EXTRA_CALLOUT_DATA
3296
3297 The pcre_study() function now returns one of these new pcre_extra blocks, with
3298 the actual study data pointed to by the study_data field, and the
3299 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3300 before. That is, this change is entirely upwards-compatible and requires no
3301 change to existing code.
3302
3303 If you want to pass in additional data to pcre_exec(), you can either place it
3304 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3305 block.
3306
3307 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3308 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3309 times with different match limits, until it finds the minimum value needed for
3310 pcre_exec() to complete. The value is then output. This can be instructive; for
3311 most simple matches the number is quite small, but for pathological cases it
3312 gets very large very quickly.
3313
3314 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3315 returns the size of the data block pointed to by the study_data field in a
3316 pcre_extra block, that is, the value that was passed as the argument to
3317 pcre_malloc() when PCRE was getting memory in which to place the information
3318 created by pcre_study(). The fourth argument should point to a size_t variable.
3319 pcretest has been extended so that this information is shown after a successful
3320 pcre_study() call when information about the compiled regex is being displayed.
3321
3322 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3323 because what follows is always an absolute path. (Later: it turns out that this
3324 is more than cosmetic for MinGW, because it doesn't like empty path
3325 components.)
3326
3327 49. Some changes have been made to the callout feature (see 28 above):
3328
3329 (i) A callout function now has three choices for what it returns:
3330
3331 0 => success, carry on matching
3332 > 0 => failure at this point, but backtrack if possible
3333 < 0 => serious error, return this value from pcre_exec()
3334
3335 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3336 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3337 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3338 use by callout functions. It will never be used by PCRE itself.
3339
3340 (ii) The pcre_extra structure (see 45 above) has a void * field called
3341 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3342 pcre_callout_block structure has a field of the same name. The contents of
3343 the field passed in the pcre_extra structure are passed to the callout
3344 function in the corresponding field in the callout block. This makes it
3345 easier to use the same callout-containing regex from multiple threads. For
3346 testing, the pcretest program has a new data escape
3347
3348 \C*n pass the number n (may be negative) as callout_data
3349
3350 If the callout function in pcretest receives a non-zero value as
3351 callout_data, it returns that value.
3352
3353 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3354 there were some redundant $(CFLAGS) in commands that are now specified as
3355 $(LINK), which already includes $(CFLAGS).
3356
3357 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3358 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3359 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3360 one-byte characters throughout. Note that case-insensitive matching applies
3361 only to characters whose values are less than 256. PCRE doesn't support the
3362 notion of cases for higher-valued characters.
3363
3364 (i) A character class whose characters are all within 0-255 is handled as
3365 a bit map, and the map is inverted for negative classes. Previously, a
3366 character > 255 always failed to match such a class; however it should
3367 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3368
3369 (ii) A negated character class with a single character < 255 is coded as
3370 "not this character" (OP_NOT). This wasn't working properly when the test
3371 character was multibyte, either singly or repeated.
3372
3373 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3374 mode, for example: \x{100}{2,3}.
3375
3376 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3377 singly or repeated) now correctly test multibyte characters. However,
3378 PCRE doesn't recognize any characters with values greater than 255 as
3379 digits, spaces, or word characters. Such characters always match \D, \S,
3380 and \W, and never match \d, \s, or \w.
3381
3382 (v) Classes may now contain characters and character ranges with values
3383 greater than 255. For example: [ab\x{100}-\x{400}].
3384
3385 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3386 PCRE in UTF-8 mode.
3387
3388 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3389 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3390 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3391 value.)
3392
3393 53. The single man page has become too large. I have therefore split it up into
3394 a number of separate man pages. These also give rise to individual HTML pages;
3395 these are now put in a separate directory, and there is an index.html page that
3396 lists them all. Some hyperlinking between the pages has been installed.
3397
3398 54. Added convenience functions for handling named capturing parentheses.
3399
3400 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3401 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3402 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3403 are faulted.
3404
3405 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3406 calling configure. These values are used when compiling the dftables.c program
3407 which is run to generate the source of the default character tables. They
3408 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3409 you will need to set these values.
3410
3411 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3412
3413
3414 Version 3.9 02-Jan-02
3415 ---------------------
3416
3417 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3418
3419 2. If --disable-static was given, the building process failed when trying to
3420 build pcretest and pcregrep. (For some reason it was using libtool to compile
3421 them, which is not right, as they aren't part of the library.)
3422
3423
3424 Version 3.8 18-Dec-01
3425 ---------------------
3426
3427 1. The experimental UTF-8 code was completely screwed up. It was packing the
3428 bytes in the wrong order. How dumb can you get?
3429
3430
3431 Version 3.7 29-Oct-01
3432 ---------------------
3433
3434 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3435 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3436 this didn't happen under Solaris 8, where I normally test things.
3437
3438 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3439 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3440 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3441 this fix an hour or so after the initial 3.7 release.)
3442
3443
3444 Version 3.6 23-Oct-01
3445 ---------------------
3446
3447 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3448 offsets passed as NULL with zero offset count.
3449
3450 2. The config.guess and config.sub files had not been updated when I moved to
3451 the latest autoconf.
3452
3453
3454 Version 3.5 15-Aug-01
3455 ---------------------
3456
3457 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3458 had been forgotten.
3459
3460 2. By using declared but undefined structures, we can avoid using "void"
3461 definitions in pcre.h while keeping the internal definitions of the structures
3462 private.
3463
3464 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3465 user point of view, this means that both static and shared libraries are built
3466 by default, but this can be individually controlled. More of the work of
3467 handling this static/shared cases is now inside libtool instead of PCRE's make
3468 file.
3469
3470 4. The pcretest utility is now installed along with pcregrep because it is
3471 useful for users (to test regexs) and by doing this, it automatically gets
3472 relinked by libtool. The documentation has been turned into a man page, so
3473 there are now .1, .txt, and .html versions in /doc.
3474
3475 5. Upgrades to pcregrep:
3476 (i) Added long-form option names like gnu grep.
3477 (ii) Added --help to list all options with an explanatory phrase.
3478 (iii) Added -r, --recursive to recurse into sub-directories.
3479 (iv) Added -f, --file to read patterns from a file.
3480
3481 6. pcre_exec() was referring to its "code" argument before testing that
3482 argument for NULL (and giving an error if it was NULL).
3483
3484 7. Upgraded Makefile.in to allow for compiling in a different directory from
3485 the source directory.
3486
3487 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3488 options bits, the pointer it was passed was to an int instead of to an unsigned
3489 long int. This mattered only on 64-bit systems.
3490
3491 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3492 generated) instead of pcre.in, which it its source. Also made the same change
3493 in several of the .c files.
3494
3495 10. A new release of gcc defines printf() as a macro, which broke pcretest
3496 because it had an ifdef in the middle of a string argument for printf(). Fixed
3497 by using separate calls to printf().
3498
3499 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3500 script, to force use of CR or LF instead of \n in the source. On non-Unix
3501 systems, the value can be set in config.h.
3502
3503 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3504 absolute limit. Changed the text of the error message to make this clear, and
3505 likewise updated the man page.
3506
3507 13. The limit of 99 on the number of capturing subpatterns has been removed.
3508 The new limit is 65535, which I hope will not be a "real" limit.
3509
3510
3511 Version 3.4 22-Aug-00
3512 ---------------------
3513
3514 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3515
3516 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3517
3518
3519 Version 3.3 01-Aug-00
3520 ---------------------
3521
3522 1. If an octal character was given, but the value was greater than \377, it
3523 was not getting masked to the least significant bits, as documented. This could
3524 lead to crashes in some systems.
3525
3526 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3527 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3528
3529 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3530 These just pass their arguments on to (pcre_free)(), but they are provided
3531 because some uses of PCRE bind it to non-C systems that can call its functions,
3532 but cannot call free() or pcre_free() directly.
3533
3534 4. Add "make test" as a synonym for "make check". Corrected some comments in
3535 the Makefile.
3536
3537 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3538 Makefile.
3539
3540 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3541 command called pgrep for grepping around the active processes.
3542
3543 7. Added the beginnings of support for UTF-8 character strings.
3544
3545 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3546 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3547 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3548 out for the ar command.)
3549
3550
3551 Version 3.2 12-May-00
3552 ---------------------
3553
3554 This is purely a bug fixing release.
3555
3556 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3557 of ZA. This was just one example of several cases that could provoke this bug,
3558 which was introduced by change 9 of version 2.00. The code for breaking
3559 infinite loops after an iteration that matches an empty string was't working
3560 correctly.
3561
3562 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3563 when matched against abbab (for example). After matching an empty string, it
3564 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3565 caused it to match further down the string than it should.
3566
3567 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3568 was there because it doesn't seem to be needed, and it causes trouble on some
3569 systems, as it is not a Standard C header. It has been removed.
3570
3571 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3572 were reported on the Macintosh. The changes were from
3573
3574 while ((c = *(++ptr)) != 0 && c != '\n');
3575 to
3576 while ((c = *(++ptr)) != 0 && c != '\n') ;
3577
3578 Totally extraordinary, but if that's what it takes...
3579
3580 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3581 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3582 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3583 assumes the way PCRE uses memmove() (always moving upwards).
3584
3585 6. PCRE is being used in one environment where strchr() is not available. There
3586 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3587 faster code anyway.
3588
3589
3590 Version 3.1 09-Feb-00
3591 ---------------------
3592
3593 The only change in this release is the fixing of some bugs in Makefile.in for
3594 the "install" target:
3595
3596 (1) It was failing to install pcreposix.h.
3597
3598 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3599
3600
3601 Version 3.0 01-Feb-00
3602 ---------------------
3603
3604 1. Add support for the /+ modifier to perltest (to output $` like it does in
3605 pcretest).
3606
3607 2. Add support for the /g modifier to perltest.
3608
3609 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3610 matches null strings.
3611
3612 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3613 pattern. Perl treats empty patterns specially - it reuses the most recent
3614 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3615 effect.
3616
3617 5. The POSIX interface was broken in that it was just handing over the POSIX
3618 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3619 required a bigger vector, with some working space on the end. This means that
3620 the POSIX wrapper now has to get and free some memory, and copy the results.
3621
3622 6. Added some simple autoconf support, placing the test data and the
3623 documentation in separate directories, re-organizing some of the
3624 information files, and making it build pcre-config (a GNU standard). Also added
3625 libtool support for building PCRE as a shared library, which is now the
3626 default.
3627
3628 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3629 09 are not valid octal constants. Single digits will be used for minor values
3630 less than 10.
3631
3632 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3633 existing programs that set these in the POSIX interface can use PCRE without
3634 modification.
3635
3636 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3637 return all that pcre_info() returns, plus additional data. The pcre_info()
3638 function is retained for compatibility, but is considered to be obsolete.
3639
3640 10. Added experimental recursion feature (?R) to handle one common case that
3641 Perl 5.6 will be able to do with (?p{...}).
3642
3643 11. Added support for POSIX character classes like [:alpha:], which Perl is
3644 adopting.
3645
3646
3647 Version 2.08 31-Aug-99
3648 ----------------------
3649
3650 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3651 trying to match at the startoffset position, but instead was moving forward to
3652 the next newline as if a previous match had failed.
3653
3654 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3655 and could get into a loop if a null string was matched other than at the start
3656 of the subject.
3657
3658 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3659 be distinguished at compile time, and for completeness also added PCRE_DATE.
3660
3661 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3662 in GnuWin32 environments.
3663
3664
3665 Version 2.07 29-Jul-99
3666 ----------------------
3667
3668 1. The documentation is now supplied in plain text form and HTML as well as in
3669 the form of man page sources.
3670
3671 2. C++ compilers don't like assigning (void *) values to other pointer types.
3672 In particular this affects malloc(). Although there is no problem in Standard
3673 C, I've put in casts to keep C++ compilers happy.
3674
3675 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3676 should be (const char *).
3677
3678 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3679 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3680 However, I haven't made this a standard facility. The documentation doesn't
3681 mention it, and the Makefile doesn't support it.
3682
3683 5. The Makefile now contains an "install" target, with editable destinations at
3684 the top of the file. The pcretest program is not installed.
3685
3686 6. pgrep -V now gives the PCRE version number and date.
3687
3688 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3689 causing the entire string to be ignored, instead of just the last character.
3690
3691 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3692 non-matching string, it can take a very, very long time, even for strings of
3693 quite modest length, because of the nested recursion. PCRE now does better in
3694 some of these cases. It does this by remembering the last required literal
3695 character in the pattern, and pre-searching the subject to ensure it is present
3696 before running the real match. In other words, it applies a heuristic to detect
3697 some types of certain failure quickly, and in the above example, if presented
3698 with a string that has no trailing " it gives "no match" very quickly.
3699
3700 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3701 other alternatives are tried instead.
3702
3703
3704 Version 2.06 09-Jun-99
3705 ----------------------
3706
3707 1. Change pcretest's output for amount of store used to show just the code
3708 space, because the remainder (the data block) varies in size between 32-bit and
3709 64-bit systems.
3710
3711 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3712 start matching at. This allows lookbehinds to work when searching for multiple
3713 occurrences in a string.
3714
3715 3. Added additional options to pcretest for testing multiple occurrences:
3716
3717 /+ outputs the rest of the string that follows a match
3718 /g loops for multiple occurrences, using the new startoffset argument
3719 /G loops for multiple occurrences by passing an incremented pointer
3720
3721 4. PCRE wasn't doing the "first character" optimization for patterns starting
3722 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3723 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3724 the letter 'x'. On long subject strings, this gives a significant speed-up.
3725
3726
3727 Version 2.05 21-Apr-99
3728 ----------------------
3729
3730 1. Changed the type of magic_number from int to long int so that it works
3731 properly on 16-bit systems.
3732
3733 2. Fixed a bug which caused patterns starting with .* not to work correctly
3734 when the subject string contained newline characters. PCRE was assuming
3735 anchoring for such patterns in all cases, which is not correct because .* will
3736 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3737 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3738 must be retried after every newline in the subject.
3739
3740
3741 Version 2.04 18-Feb-99
3742 ----------------------
3743
3744 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3745 computation of the store needed to hold the pattern was incorrect (too large).
3746 If such patterns were nested a few deep, this could multiply and become a real
3747 problem.
3748
3749 2. Added /M option to pcretest to show the memory requirement of a specific
3750 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3751
3752 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3753 compiled in such a way that the backtracking after subsequent failure was
3754 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3755 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3756
3757
3758 Version 2.03 02-Feb-99
3759 ----------------------
3760
3761 1. Fixed typo and small mistake in man page.
3762
3763 2. Added 4th condition (GPL supersedes if conflict) and created separate
3764 LICENCE file containing the conditions.
3765
3766 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3767 Perl, that is the internal \ allows the delimiter to be included in the
3768 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3769 the final delimiter, add \ to the end of the pattern (to test the error).
3770
3771 4. Added the convenience functions for extracting substrings after a successful
3772 match. Updated pcretest to make it able to test these functions.
3773
3774
3775 Version 2.02 14-Jan-99
3776 ----------------------
3777
3778 1. Initialized the working variables associated with each extraction so that
3779 their saving and restoring doesn't refer to uninitialized store.
3780
3781 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3782 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3783 fix the problem.
3784
3785 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3786 calls, and wasn't printing the correct value for compiling calls. Increased the
3787 default value of LOOPREPEAT, and the number of significant figures in the
3788 times.
3789
3790 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3791
3792 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3793 a building problem on Windows NT with a FAT file system.
3794
3795
3796 Version 2.01 21-Oct-98
3797 ----------------------
3798
3799 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3800 to character tables built by pcre_maketables() in the current locale. If NULL
3801 is passed, the default tables are used.
3802
3803
3804 Version 2.00 24-Sep-98
3805 ----------------------
3806
3807 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3808 it any more.
3809
3810 2. Allow quantification of (?>) groups, and make it work correctly.
3811
3812 3. The first character computation wasn't working for (?>) groups.
3813
3814 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3815 end of the subject) and add 5.005's \z, which really does match only at the
3816 very end of the subject.
3817
3818 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3819
3820 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3821 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3822 localized options. All options to pcre_study() were also removed.
3823
3824 7. Add other new features from 5.005:
3825
3826 $(?<= positive lookbehind
3827 $(?<! negative lookbehind
3828 (?imsx-imsx) added the unsetting capability
3829 such a setting is global if at outer level; local otherwise
3830 (?imsx-imsx:) non-capturing groups with option setting
3831 (?(cond)re|re) conditional pattern matching
3832
3833 A backreference to itself in a repeated group matches the previous
3834 captured string.
3835
3836 8. General tidying up of studying (both automatic and via "study")
3837 consequential on the addition of new assertions.
3838
3839 9. As in 5.005, unlimited repeated groups that could match an empty substring
3840 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3841 runtime if any iteration does actually match an empty substring.
3842
3843 10. Include the RunTest script in the distribution.
3844
3845 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3846 discrepancies, some of which were old and were also with respect to 5.004. They
3847 have now been fixed.
3848
3849
3850 Version 1.09 28-Apr-98
3851 ----------------------
3852
3853 1. A negated single character class followed by a quantifier with a minimum
3854 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3855 program crashes, or just wrong answers. This did not apply to negated classes
3856 containing more than one character, or to minima other than one.
3857
3858
3859 Version 1.08 27-Mar-98
3860 ----------------------
3861
3862 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3863
3864 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3865 latter must appear before anything that relies on it in the pattern.
3866
3867
3868 Version 1.07 16-Feb-98
3869 ----------------------
3870
3871 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3872 repeat of a potentially empty string).
3873
3874
3875 Version 1.06 23-Jan-98
3876 ----------------------
3877
3878 1. Added Markus Oberhumer's little patches for C++.
3879
3880 2. Literal strings longer than 255 characters were broken.
3881
3882
3883 Version 1.05 23-Dec-97
3884 ----------------------
3885
3886 1. Negated character classes containing more than one character were failing if
3887 PCRE_CASELESS was set at run time.
3888
3889
3890 Version 1.04 19-Dec-97
3891 ----------------------
3892
3893 1. Corrected the man page, where some "const" qualifiers had been omitted.
3894
3895 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3896 input syntax.
3897
3898 3. Fixed memory leak which occurred when a regex with back references was
3899 matched with an offsets vector that wasn't big enough. The temporary memory
3900 that is used in this case wasn't being freed if the match failed.
3901
3902 4. Tidied pcretest to ensure it frees memory that it gets.
3903
3904 5. Temporary memory was being obtained in the case where the passed offsets
3905 vector was exactly big enough.
3906
3907 6. Corrected definition of offsetof() from change 5 below.
3908
3909 7. I had screwed up change 6 below and broken the rules for the use of
3910 setjmp(). Now fixed.
3911
3912
3913 Version 1.03 18-Dec-97
3914 ----------------------
3915
3916 1. A erroneous regex with a missing opening parenthesis was correctly
3917 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3918 on some systems.
3919
3920 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3921 it was reported that one broken compiler failed on the former because "code" is
3922 also an independent variable.
3923
3924 3. The erroneous regex a[]b caused an array overrun reference.
3925
3926 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3927 fail on data ending with that character. (It was going on too far, and checking
3928 the next character, typically a binary zero.) This was specific to the
3929 optimized code for single-character negative classes.
3930
3931 5. Added a contributed patch from the TIN world which does the following:
3932
3933 + Add an undef for memmove, in case the the system defines a macro for it.
3934
3935 + Add a definition of offsetof(), in case there isn't one. (I don't know
3936 the reason behind this - offsetof() is part of the ANSI standard - but
3937 it does no harm).
3938
3939 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3940 most of the places where whitespace preceded '#'. I have given up and
3941 allowed the remaining 2 cases to be at the margin.
3942
3943 + Rename some variables in pcre to eliminate shadowing. This seems very
3944 pedantic, but does no harm, of course.
3945
3946 6. Moved the call to setjmp() into its own function, to get rid of warnings
3947 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3948
3949 7. Constructs such as \d{8,} were compiling into the equivalent of
3950 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3951 outcome, but in this particular case used more store than had been allocated,
3952 which caused the bug to be discovered because it threw up an internal error.
3953
3954 8. The debugging code in both pcre and pcretest for outputting the compiled
3955 form of a regex was going wrong in the case of back references followed by
3956 curly-bracketed repeats.
3957
3958
3959 Version 1.02 12-Dec-97
3960 ----------------------
3961
3962 1. Typos in pcre.3 and comments in the source fixed.
3963
3964 2. Applied a contributed patch to get rid of places where it used to remove
3965 'const' from variables, and fixed some signed/unsigned and uninitialized
3966 variable warnings.
3967
3968 3. Added the "runtest" target to Makefile.
3969
3970 4. Set default compiler flag to -O2 rather than just -O.
3971
3972
3973 Version 1.01 19-Nov-97
3974 ----------------------
3975
3976 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3977 like /([ab]*)*/, that is, for classes with more than one character in them.
3978
3979 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3980 as /((?>a*))*/ (a PCRE_EXTRA facility).
3981
3982
3983 Version 1.00 18-Nov-97
3984 ----------------------
3985
3986 1. Added compile-time macros to support systems such as SunOS4 which don't have
3987 memmove() or strerror() but have other things that can be used instead.
3988
3989 2. Arranged that "make clean" removes the executables.
3990
3991
3992 Version 0.99 27-Oct-97
3993 ----------------------
3994
3995 1. Fixed bug in code for optimizing classes with only one character. It was
3996 initializing a 32-byte map regardless, which could cause it to run off the end
3997 of the memory it had got.
3998
3999 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4000
4001
4002 Version 0.98 22-Oct-97
4003 ----------------------
4004
4005 1. Fixed bug in code for handling temporary memory usage when there are more
4006 back references than supplied space in the ovector. This could cause segfaults.
4007
4008
4009 Version 0.97 21-Oct-97
4010 ----------------------
4011
4012 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4013
4014 2. Optimized negated single characters not to use a bit map.
4015
4016 3. Brought error texts together as macro definitions; clarified some of them;
4017 fixed one that was wrong - it said "range out of order" when it meant "invalid
4018 escape sequence".
4019
4020 4. Changed some char * arguments to const char *.
4021
4022 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4023
4024 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4025 pcretest.
4026
4027
4028 Version 0.96 16-Oct-97
4029 ----------------------
4030
4031 1. Added a simple "pgrep" utility to the distribution.
4032
4033 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4034 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4035 where "ddd" means "one or more decimal digits".
4036
4037 3. Fixed serious bug. If a pattern had a back reference, but the call to
4038 pcre_exec() didn't supply a large enough ovector to record the related
4039 identifying subpattern, the match always failed. PCRE now remembers the number
4040 of the largest back reference, and gets some temporary memory in which to save
4041 the offsets during matching if necessary, in order to ensure that
4042 backreferences always work.
4043
4044 4. Increased the compatibility with Perl in a number of ways:
4045
4046 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4047 to request this handling. The option can be set at compile or exec time.
4048
4049 (b) $ matches before a terminating newline by default; an option
4050 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4051 mode). The option can be set at compile or exec time.
4052
4053 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4054 the same as Perl's. If the decimal number it represents is less than 10
4055 or there aren't that many previous left capturing parentheses, an octal
4056 escape is read. Inside a character class, it's always an octal escape,
4057 even if it is a single digit.
4058
4059 (d) An escaped but undefined alphabetic character is taken as a literal,
4060 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4061 escapes.
4062
4063 (e) {0} is now permitted. (The previous item is removed from the compiled
4064 pattern).
4065
4066 5. Changed all the names of code files so that the basic parts are no longer
4067 than 10 characters, and abolished the teeny "globals.c" file.
4068
4069 6. Changed the handling of character classes; they are now done with a 32-byte
4070 bit map always.
4071
4072 7. Added the -d and /D options to pcretest to make it possible to look at the
4073 internals of compilation without having to recompile pcre.
4074
4075
4076 Version 0.95 23-Sep-97
4077 ----------------------
4078
4079 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4080 \x20 at the start of a run of normal characters. These were being treated as
4081 real characters, instead of the source characters being re-checked.
4082
4083
4084 Version 0.94 18-Sep-97
4085 ----------------------
4086
4087 1. The functions are now thread-safe, with the caveat that the global variables
4088 containing pointers to malloc() and free() or alternative functions are the
4089 same for all threads.
4090
4091 2. Get pcre_study() to generate a bitmap of initial characters for non-
4092 anchored patterns when this is possible, and use it if passed to pcre_exec().
4093
4094
4095 Version 0.93 15-Sep-97
4096 ----------------------
4097
4098 1. /(b)|(:+)/ was computing an incorrect first character.
4099
4100 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4101 but not actually doing anything yet.
4102
4103 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4104 as Perl does (e.g. [-az] or [az-]).
4105
4106 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4107 all possible positions.
4108
4109 5. Split up into different modules to avoid including unneeded functions in a
4110 compiled binary. However, compile and exec are still in one module. The "study"
4111 function is split off.
4112
4113 6. The character tables are now in a separate module whose source is generated
4114 by an auxiliary program - but can then be edited by hand if required. There are
4115 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4116 toupper() in the code.
4117
4118 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4119 make them global. Abolish the function for setting them, as the caller can now
4120 set them directly.
4121
4122
4123 Version 0.92 11-Sep-97
4124 ----------------------
4125
4126 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4127 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4128
4129 2. Caseless matching was not working in character classes if the characters in
4130 the pattern were in upper case.
4131
4132 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4133
4134 4. Make PCRE_ANCHORED public and accept as a compile option.
4135
4136 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4137 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4138 pass them.
4139
4140 6. Give an error if bad option bits passed at compile or run time.
4141
4142 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4143 pcretest to cause it to pass that flag.
4144
4145 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4146 options, and the first character, if set.
4147
4148 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4149
4150
4151 Version 0.91 10-Sep-97
4152 ----------------------
4153
4154 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4155 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4156
4157 2. PCRE was looping on encountering an indefinitely repeated back reference to
4158 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4159 Perl does - treats the match as successful.
4160
4161 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5