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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 551 - (show annotations) (download)
Sun Oct 10 17:33:07 2010 UTC (3 years, 6 months ago) by ph10
File size: 190344 byte(s)
Make (*COMMIT) override (*THEN) and similar.

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

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