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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 486 - (show annotations) (download)
Tue Jan 5 17:44:57 2010 UTC (4 years, 10 months ago) by ph10
File size: 180439 byte(s)
Fix HP compile problem for pcrecpp.cc.

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

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