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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 471 - (show annotations) (download)
Fri Dec 11 15:11:55 2009 UTC (4 years, 7 months ago) by ph10
File size: 175088 byte(s)
Fix study bug with single-branch conditions, including (DEFINE).

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17 was out-of-date, and there was no check on the pcre_compile() error code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

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