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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 553 - (show annotations) (download)
Fri Oct 22 15:57:50 2010 UTC (3 years, 11 months ago) by ph10
File size: 191779 byte(s)
Change the way PCRE_PARTIAL_HARD handles \z, \Z, \b, \B, and $.

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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12