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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 457 - (show annotations)
Sat Oct 3 16:24:08 2009 UTC (5 years, 9 months ago) by ph10
File size: 174499 byte(s)
Error occurred while calculating annotation data.
Allow duplicate names for same-numbered groups; forbid different names.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.00 ??-???-??
5 ----------------------
6
7 1. The table for translating pcre_compile() error codes into POSIX error codes
8 was out-of-date, and there was no check on the pcre_compile() error code
9 being within the table. This could lead to an OK return being given in
10 error.
11
12 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
13 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
14 in a Windows environment.
15
16 3. The pcregrep --count option prints the count for each file even when it is
17 zero, as does GNU grep. However, pcregrep was also printing all files when
18 --files-with-matches was added. Now, when both options are given, it prints
19 counts only for those files that have at least one match. (GNU grep just
20 prints the file name in this circumstance, but including the count seems
21 more useful - otherwise, why use --count?) Also ensured that the
22 combination -clh just lists non-zero counts, with no names.
23
24 4. The long form of the pcregrep -F option was incorrectly implemented as
25 --fixed_strings instead of --fixed-strings. This is an incompatible change,
26 but it seems right to fix it, and I didn't think it was worth preserving
27 the old behaviour.
28
29 5. The command line items --regex=pattern and --regexp=pattern were not
30 recognized by pcregrep, which required --regex pattern or --regexp pattern
31 (with a space rather than an '='). The man page documented the '=' forms,
32 which are compatible with GNU grep; these now work.
33
34 6. No libpcreposix.pc file was created for pkg-config; there was just
35 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
36
37 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
38 when UCP support is not needed, by modifying the Python script that
39 generates it from Unicode data files. This should not matter if the module
40 is correctly used as a library, but I received one complaint about 50K of
41 unwanted data. My guess is that the person linked everything into his
42 program rather than using a library. Anyway, it does no harm.
43
44 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
45 was a minimum greater than 1 for a wide character in a possessive
46 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
47 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
48 character. Chaos in the form of incorrect output or a compiling loop could
49 result.
50
51 9. The restrictions on what a pattern can contain when partial matching is
52 requested for pcre_exec() have been removed. All patterns can now be
53 partially matched by this function. In addition, if there are at least two
54 slots in the offset vector, the offset of the earliest inspected character
55 for the match and the offset of the end of the subject are set in them when
56 PCRE_ERROR_PARTIAL is returned.
57
58 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
59 synonymous with PCRE_PARTIAL, for backwards compatibility, and
60 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
61 and may be more useful for multi-segment matching, especially with
62 pcre_exec().
63
64 11. Partial matching with pcre_exec() is now more intuitive. A partial match
65 used to be given if ever the end of the subject was reached; now it is
66 given only if matching could not proceed because another character was
67 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
68 string "Z", which now yields "no match" instead of "partial match". In the
69 case of pcre_dfa_exec(), "no match" is given if every matching path for the
70 final character ended with (*FAIL).
71
72 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
73 if the pattern had a "must contain" character that was already found in the
74 earlier partial match, unless partial matching was again requested. For
75 example, with the pattern /dog.(body)?/, the "must contain" character is
76 "g". If the first part-match was for the string "dog", restarting with
77 "sbody" failed. This bug has been fixed.
78
79 13. The string returned by pcre_dfa_exec() after a partial match has been
80 changed so that it starts at the first inspected character rather than the
81 first character of the match. This makes a difference only if the pattern
82 starts with a lookbehind assertion or \b or \B (\K is not supported by
83 pcre_dfa_exec()). It's an incompatible change, but it makes the two
84 matching functions compatible, and I think it's the right thing to do.
85
86 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
87 so that the demonstration program is easily available in environments where
88 PCRE has not been installed from source.
89
90 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
91 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
92 library.
93
94 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
95 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
96 is not the first non-POSIX option to be added. Clearly some people find
97 these options useful.
98
99 17. If a caller to the POSIX matching function regexec() passes a non-zero
100 value for nmatch with a NULL value for pmatch, the value of
101 nmatch is forced to zero.
102
103 18. RunGrepTest did not have a test for the availability of the -u option of
104 the diff command, as RunTest does. It now checks in the same way as
105 RunTest, and also checks for the -b option.
106
107 19. If an odd number of negated classes containing just a single character
108 interposed, within parentheses, between a forward reference to a named
109 subpattern and the definition of the subpattern, compilation crashed with
110 an internal error, complaining that it could not find the referenced
111 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
112 [The bug was that it was starting one character too far in when skipping
113 over the character class, thus treating the ] as data rather than
114 terminating the class. This meant it could skip too much.]
115
116 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
117 /g option in pcretest when the pattern contains \K, which makes it possible
118 to have an empty string match not at the start, even when the pattern is
119 anchored. Updated pcretest and pcredemo to use this option.
120
121 21. If the maximum number of capturing subpatterns in a recursion was greater
122 than the maximum at the outer level, the higher number was returned, but
123 with unset values at the outer level. The correct (outer level) value is
124 now given.
125
126 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
127 PCRE did not set those parentheses (unlike Perl). I have now found a way to
128 make it do so. The string so far is captured, making this feature
129 compatible with Perl.
130
131 23. The tests have been re-organized, adding tests 11 and 12, to make it
132 possible to check the Perl 5.10 features against Perl 5.10.
133
134 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
135 pattern matches a fixed length string. PCRE did not allow this; now it
136 does. Neither allows recursion.
137
138 25. I finally figured out how to implement a request to provide the minimum
139 length of subject string that was needed in order to match a given pattern.
140 (It was back references and recursion that I had previously got hung up
141 on.) This code has now been added to pcre_study(); it finds a lower bound
142 to the length of subject needed. It is not necessarily the greatest lower
143 bound, but using it to avoid searching strings that are too short does give
144 some useful speed-ups. The value is available to calling programs via
145 pcre_fullinfo().
146
147 26. While implementing 25, I discovered to my embarrassment that pcretest had
148 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
149 study optimizations had never been tested with that matching function.
150 Oops. What is worse, even when it was passed study data, there was a bug in
151 pcre_dfa_exec() that meant it never actually used it. Double oops. There
152 were also very few tests of studied patterns with pcre_dfa_exec().
153
154 27. If (?| is used to create subpatterns with duplicate numbers, they are now
155 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
156 on the other side of the coin, they are no longer allowed to have different
157 names, because these cannot be distinguished in PCRE, and this has caused
158 confusion. (This is a difference from Perl.)
159
160
161 Version 7.9 11-Apr-09
162 ---------------------
163
164 1. When building with support for bzlib/zlib (pcregrep) and/or readline
165 (pcretest), all targets were linked against these libraries. This included
166 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
167 libraries. This caused unwanted dependencies to be created. This problem
168 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
169 pcretest is linked with readline.
170
171 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
172 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
173 moved outside it again, because FALSE and TRUE are already defined in AIX,
174 but BOOL is not.
175
176 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
177 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
178
179 4. The pcregrep documentation said spaces were inserted as well as colons (or
180 hyphens) following file names and line numbers when outputting matching
181 lines. This is not true; no spaces are inserted. I have also clarified the
182 wording for the --colour (or --color) option.
183
184 5. In pcregrep, when --colour was used with -o, the list of matching strings
185 was not coloured; this is different to GNU grep, so I have changed it to be
186 the same.
187
188 6. When --colo(u)r was used in pcregrep, only the first matching substring in
189 each matching line was coloured. Now it goes on to look for further matches
190 of any of the test patterns, which is the same behaviour as GNU grep.
191
192 7. A pattern that could match an empty string could cause pcregrep to loop; it
193 doesn't make sense to accept an empty string match in pcregrep, so I have
194 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
195 seems to be how GNU grep behaves.
196
197 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
198 start or after a newline", because the conditional assertion was not being
199 correctly handled. The rule now is that both the assertion and what follows
200 in the first alternative must satisfy the test.
201
202 9. If auto-callout was enabled in a pattern with a conditional group whose
203 condition was an assertion, PCRE could crash during matching, both with
204 pcre_exec() and pcre_dfa_exec().
205
206 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
207 used for matching.
208
209 11. Unicode property support in character classes was not working for
210 characters (bytes) greater than 127 when not in UTF-8 mode.
211
212 12. Added the -M command line option to pcretest.
213
214 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
215
216 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
217
218 16. Added comments and documentation about mis-use of no_arg in the C++
219 wrapper.
220
221 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
222 from Martin Jerabek that uses macro names for all relevant character and
223 string constants.
224
225 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
226 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
227 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
228 these, but not everybody uses configure.
229
230 19. A conditional group that had only one branch was not being correctly
231 recognized as an item that could match an empty string. This meant that an
232 enclosing group might also not be so recognized, causing infinite looping
233 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
234 with the subject "ab", where knowledge that the repeated group can match
235 nothing is needed in order to break the loop.
236
237 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
238 exec(), but without supplying a callout function, matching went wrong.
239
240 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
241 leak if the size of the offset vector was greater than 30. When the vector
242 is smaller, the saved offsets during recursion go onto a local stack
243 vector, but for larger vectors malloc() is used. It was failing to free
244 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
245 error, in fact).
246
247 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
248 heapframe that is used only when UTF-8 support is enabled. This caused no
249 problem, but was untidy.
250
251 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
252 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
253 included within another project.
254
255 24. Steven Van Ingelgem's patches to add more options to the CMake support,
256 slightly modified by me:
257
258 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
259 not building pcregrep.
260
261 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
262 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
263
264 25. Forward references, both numeric and by name, in patterns that made use of
265 duplicate group numbers, could behave incorrectly or give incorrect errors,
266 because when scanning forward to find the reference group, PCRE was not
267 taking into account the duplicate group numbers. A pattern such as
268 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
269
270 26. Changed a few more instances of "const unsigned char *" to USPTR, making
271 the feature of a custom pointer more persuasive (as requested by a user).
272
273 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
274 pcretest.c, inside #ifndefs, because it seems they are sometimes already
275 pre-defined.
276
277 28. Added support for (*UTF8) at the start of a pattern.
278
279 29. Arrange for flags added by the "release type" setting in CMake to be shown
280 in the configuration summary.
281
282
283 Version 7.8 05-Sep-08
284 ---------------------
285
286 1. Replaced UCP searching code with optimized version as implemented for Ad
287 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
288 stage table and inline lookup instead of a function, giving speed ups of 2
289 to 5 times on some simple patterns that I tested. Permission was given to
290 distribute the MultiStage2.py script that generates the tables (it's not in
291 the tarball, but is in the Subversion repository).
292
293 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
294 scripts.
295
296 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
297 a group with a zero qualifier. The result of the study could be incorrect,
298 or the function might crash, depending on the pattern.
299
300 4. Caseless matching was not working for non-ASCII characters in back
301 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
302 It now works when Unicode Property Support is available.
303
304 5. In pcretest, an escape such as \x{de} in the data was always generating
305 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
306 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
307 truncation.
308
309 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
310
311 7. Added two (int) casts to pcregrep when printing the difference of two
312 pointers, in case they are 64-bit values.
313
314 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
315 test 2 if it fails.
316
317 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
318 and a #define of that name to empty if it is not externally set. This is to
319 allow users of MSVC to set it if necessary.
320
321 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
322 the convenience functions in the pcre_get.c source file.
323
324 11. An option change at the start of a pattern that had top-level alternatives
325 could cause overwriting and/or a crash. This command provoked a crash in
326 some environments:
327
328 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
329
330 This potential security problem was recorded as CVE-2008-2371.
331
332 12. For a pattern where the match had to start at the beginning or immediately
333 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
334 pcre_dfa_exec() could read past the end of the passed subject if there was
335 no match. To help with detecting such bugs (e.g. with valgrind), I modified
336 pcretest so that it places the subject at the end of its malloc-ed buffer.
337
338 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
339 exec() might read past the end of the data buffer in UTF-8 mode.
340
341 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
342 the data contained the byte 0x85 as part of a UTF-8 character within its
343 first line. This applied both to normal and DFA matching.
344
345 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
346 /^[^d]*?$/8 failed to match "abc".
347
348 16. Added a missing copyright notice to pcrecpp_internal.h.
349
350 17. Make it more clear in the documentation that values returned from
351 pcre_exec() in ovector are byte offsets, not character counts.
352
353 18. Tidied a few places to stop certain compilers from issuing warnings.
354
355 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
356 supplied by Stefan Weber. I made a further small update for 7.8 because
357 there is a change of source arrangements: the pcre_searchfuncs.c module is
358 replaced by pcre_ucd.c.
359
360
361 Version 7.7 07-May-08
362 ---------------------
363
364 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
365 a string to a long long, pretend we don't even have a long long." This is
366 done by checking for the strtoq, strtoll, and _strtoi64 functions.
367
368 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
369 pre-7.6 versions, which defined a global no_arg variable instead of putting
370 it in the RE class. (See also #8 below.)
371
372 3. Remove a line of dead code, identified by coverity and reported by Nuno
373 Lopes.
374
375 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
376
377 (1) The include/exclude patterns were being applied to the whole pathnames
378 of files, instead of just to the final components.
379
380 (2) If there was more than one level of directory, the subdirectories were
381 skipped unless they satisfied the include/exclude conditions. This is
382 inconsistent with GNU grep (and could even be seen as contrary to the
383 pcregrep specification - which I improved to make it absolutely clear).
384 The action now is always to scan all levels of directory, and just
385 apply the include/exclude patterns to regular files.
386
387 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
388 --exclude_dir in the tests to avoid scanning .svn directories.
389
390 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
391 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
392 doesn't support NULs in patterns.
393
394 7. Added some missing "const"s to declarations of static tables in
395 pcre_compile.c and pcre_dfa_exec.c.
396
397 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
398 caused by fix #2 above. (Subsequently also a second patch to fix the
399 first patch. And a third patch - this was a messy problem.)
400
401 9. Applied Craig's patch to remove the use of push_back().
402
403 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
404 matching function regexec().
405
406 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
407 which, however, unlike Perl's \g{...}, are subroutine calls, not back
408 references. PCRE supports relative numbers with this syntax (I don't think
409 Oniguruma does).
410
411 12. Previously, a group with a zero repeat such as (...){0} was completely
412 omitted from the compiled regex. However, this means that if the group
413 was called as a subroutine from elsewhere in the pattern, things went wrong
414 (an internal error was given). Such groups are now left in the compiled
415 pattern, with a new opcode that causes them to be skipped at execution
416 time.
417
418 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
419 to the way PCRE behaves:
420
421 (a) A lone ] character is dis-allowed (Perl treats it as data).
422
423 (b) A back reference to an unmatched subpattern matches an empty string
424 (Perl fails the current match path).
425
426 (c) A data ] in a character class must be notated as \] because if the
427 first data character in a class is ], it defines an empty class. (In
428 Perl it is not possible to have an empty class.) The empty class []
429 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
430 The negative empty class [^] matches any one character, independently
431 of the DOTALL setting.
432
433 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
434 non-existent subpattern following a character class starting with ']' and
435 containing () gave an internal compiling error instead of "reference to
436 non-existent subpattern". Fortunately, when the pattern did exist, the
437 compiled code was correct. (When scanning forwards to check for the
438 existencd of the subpattern, it was treating the data ']' as terminating
439 the class, so got the count wrong. When actually compiling, the reference
440 was subsequently set up correctly.)
441
442 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
443 it was being rejected as not supported by pcre_dfa_exec(), even though
444 other assertions are supported. I have made pcre_dfa_exec() support
445 (*FAIL).
446
447 16. The implementation of 13c above involved the invention of a new opcode,
448 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
449 cannot be changed at match time, I realized I could make a small
450 improvement to matching performance by compiling OP_ALLANY instead of
451 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
452 on the OP_ANY path.
453
454 17. Compiling pcretest on Windows with readline support failed without the
455 following two fixes: (1) Make the unistd.h include conditional on
456 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
457
458 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
459 ncurses library to be included for pcretest when ReadLine support is
460 requested, but also to allow for it to be overridden. This patch came from
461 Daniel Bergström.
462
463 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
464 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
465 any errors with the current Unicode tables. Thanks to Peter Kankowski for
466 spotting this.
467
468
469 Version 7.6 28-Jan-08
470 ---------------------
471
472 1. A character class containing a very large number of characters with
473 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
474 overflow.
475
476 2. Patch to cut out the "long long" test in pcrecpp_unittest when
477 HAVE_LONG_LONG is not defined.
478
479 3. Applied Christian Ehrlicher's patch to update the CMake build files to
480 bring them up to date and include new features. This patch includes:
481
482 - Fixed PH's badly added libz and libbz2 support.
483 - Fixed a problem with static linking.
484 - Added pcredemo. [But later removed - see 7 below.]
485 - Fixed dftables problem and added an option.
486 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
487 HAVE_LONG_LONG.
488 - Added readline support for pcretest.
489 - Added an listing of the option settings after cmake has run.
490
491 4. A user submitted a patch to Makefile that makes it easy to create
492 "pcre.dll" under mingw when using Configure/Make. I added stuff to
493 Makefile.am that cause it to include this special target, without
494 affecting anything else. Note that the same mingw target plus all
495 the other distribution libraries and programs are now supported
496 when configuring with CMake (see 6 below) instead of with
497 Configure/Make.
498
499 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
500 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
501 exported in the Windows port". It has not yet been confirmed that the patch
502 solves the problem, but it does no harm.
503
504 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
505 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
506 with CMake, and also correct the comment about stack recursion.
507
508 7. Remove the automatic building of pcredemo from the ./configure system and
509 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
510 of a program that users should build themselves after PCRE is installed, so
511 building it automatically is not really right. What is more, it gave
512 trouble in some build environments.
513
514 8. Further tidies to CMakeLists.txt from Sheri and Christian.
515
516
517 Version 7.5 10-Jan-08
518 ---------------------
519
520 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
521 values in parens when parsing an RE using the C++ wrapper."
522
523 2. Negative specials like \S did not work in character classes in UTF-8 mode.
524 Characters greater than 255 were excluded from the class instead of being
525 included.
526
527 3. The same bug as (2) above applied to negated POSIX classes such as
528 [:^space:].
529
530 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
531 defined or documented. It seems to have been a typo for PCRE_STATIC, so
532 I have changed it.
533
534 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
535 first named subpattern) and a construct such as (?&a) would reference the
536 first named subpattern whose name started with "a" (in other words, the
537 length check was missing). Both these problems are fixed. "Subpattern name
538 expected" is now given for (?&) (a zero-length name), and this patch also
539 makes it give the same error for \k'' (previously it complained that that
540 was a reference to a non-existent subpattern).
541
542 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
543 this is right because (?- can be followed by option settings as well as by
544 digits. I have, however, made the messages clearer.
545
546 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
547 than the number used in the conditional) now cause a compile-time error.
548 This is actually not compatible with Perl, which accepts such patterns, but
549 treats the conditional as always being FALSE (as PCRE used to), but it
550 seems to me that giving a diagnostic is better.
551
552 8. Change "alphameric" to the more common word "alphanumeric" in comments
553 and messages.
554
555 9. Fix two occurrences of "backslash" in comments that should have been
556 "backspace".
557
558 10. Remove two redundant lines of code that can never be obeyed (their function
559 was moved elsewhere).
560
561 11. The program that makes PCRE's Unicode character property table had a bug
562 which caused it to generate incorrect table entries for sequences of
563 characters that have the same character type, but are in different scripts.
564 It amalgamated them into a single range, with the script of the first of
565 them. In other words, some characters were in the wrong script. There were
566 thirteen such cases, affecting characters in the following ranges:
567
568 U+002b0 - U+002c1
569 U+0060c - U+0060d
570 U+0061e - U+00612
571 U+0064b - U+0065e
572 U+0074d - U+0076d
573 U+01800 - U+01805
574 U+01d00 - U+01d77
575 U+01d9b - U+01dbf
576 U+0200b - U+0200f
577 U+030fc - U+030fe
578 U+03260 - U+0327f
579 U+0fb46 - U+0fbb1
580 U+10450 - U+1049d
581
582 12. The -o option (show only the matching part of a line) for pcregrep was not
583 compatible with GNU grep in that, if there was more than one match in a
584 line, it showed only the first of them. It now behaves in the same way as
585 GNU grep.
586
587 13. If the -o and -v options were combined for pcregrep, it printed a blank
588 line for every non-matching line. GNU grep prints nothing, and pcregrep now
589 does the same. The return code can be used to tell if there were any
590 non-matching lines.
591
592 14. Added --file-offsets and --line-offsets to pcregrep.
593
594 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
595 infinitely looping recursion. The bug was that positive lookaheads were not
596 being skipped when checking for a possible empty match (negative lookaheads
597 and both kinds of lookbehind were skipped).
598
599 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
600 inclusion of <windows.h> to before rather than after the definition of
601 INVALID_FILE_ATTRIBUTES (patch from David Byron).
602
603 17. Specifying a possessive quantifier with a specific limit for a Unicode
604 character property caused pcre_compile() to compile bad code, which led at
605 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
606 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
607 caused the error; without that there was no problem.
608
609 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
610
611 19. Added --enable-pcretest-libreadline.
612
613 20. In pcrecpp.cc, the variable 'count' was incremented twice in
614 RE::GlobalReplace(). As a result, the number of replacements returned was
615 double what it should be. I removed one of the increments, but Craig sent a
616 later patch that removed the other one (the right fix) and added unit tests
617 that check the return values (which was not done before).
618
619 21. Several CMake things:
620
621 (1) Arranged that, when cmake is used on Unix, the libraries end up with
622 the names libpcre and libpcreposix, not just pcre and pcreposix.
623
624 (2) The above change means that pcretest and pcregrep are now correctly
625 linked with the newly-built libraries, not previously installed ones.
626
627 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
628
629 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
630 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
631 UTF-8 newline character). The key issue is that the pattern starts .*;
632 this means that the match must be either at the beginning, or after a
633 newline. The bug was in the code for advancing after a failed match and
634 checking that the new position followed a newline. It was not taking
635 account of UTF-8 characters correctly.
636
637 23. PCRE was behaving differently from Perl in the way it recognized POSIX
638 character classes. PCRE was not treating the sequence [:...:] as a
639 character class unless the ... were all letters. Perl, however, seems to
640 allow any characters between [: and :], though of course it rejects as
641 unknown any "names" that contain non-letters, because all the known class
642 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
643 for example, whereas PCRE did not - it did not recognize a POSIX character
644 class. This seemed a bit dangerous, so the code has been changed to be
645 closer to Perl. The behaviour is not identical to Perl, because PCRE will
646 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
647 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
648 Perl does, and where it didn't before.
649
650 24. Rewrite so as to remove the single use of %n from pcregrep because in some
651 Windows environments %n is disabled by default.
652
653
654 Version 7.4 21-Sep-07
655 ---------------------
656
657 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
658 means that a class such as [\s] counted as "explicit reference to CR or
659 LF". That isn't really right - the whole point of the change was to try to
660 help when there was an actual mention of one of the two characters. So now
661 the change happens only if \r or \n (or a literal CR or LF) character is
662 encountered.
663
664 2. The 32-bit options word was also used for 6 internal flags, but the numbers
665 of both had grown to the point where there were only 3 bits left.
666 Fortunately, there was spare space in the data structure, and so I have
667 moved the internal flags into a new 16-bit field to free up more option
668 bits.
669
670 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
671 but did not set the internal JCHANGED flag - either of these is enough to
672 control the way the "get" function works - but the PCRE_INFO_JCHANGED
673 facility is supposed to tell if (?J) was ever used, so now (?J) at the
674 start sets both bits.
675
676 4. Added options (at build time, compile time, exec time) to change \R from
677 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
678
679 5. doc/pcresyntax.html was missing from the distribution.
680
681 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
682 compatibility, even though it is no longer used.
683
684 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
685 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
686 windows.h file is present (where different names are used). [This was
687 reversed later after testing - see 16 below.]
688
689 8. Changed all #include <config.h> to #include "config.h". There were also
690 some further <pcre.h> cases that I changed to "pcre.h".
691
692 9. When pcregrep was used with the --colour option, it missed the line ending
693 sequence off the lines that it output.
694
695 10. It was pointed out to me that arrays of string pointers cause lots of
696 relocations when a shared library is dynamically loaded. A technique of
697 using a single long string with a table of offsets can drastically reduce
698 these. I have refactored PCRE in four places to do this. The result is
699 dramatic:
700
701 Originally: 290
702 After changing UCP table: 187
703 After changing error message table: 43
704 After changing table of "verbs" 36
705 After changing table of Posix names 22
706
707 Thanks to the folks working on Gregex for glib for this insight.
708
709 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
710 unicode-properties was also set.
711
712 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
713
714 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
715 checked only for CRLF.
716
717 14. Added casts to pcretest.c to avoid compiler warnings.
718
719 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
720
721 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
722 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
723 entirely. This removes changes made in 7 above.
724
725 17. The CMake files have been updated, and there is now more information about
726 building with CMake in the NON-UNIX-USE document.
727
728
729 Version 7.3 28-Aug-07
730 ---------------------
731
732 1. In the rejigging of the build system that eventually resulted in 7.1, the
733 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
734 brackets there is not right, since it causes compilers to look for an
735 installed pcre.h, not the version that is in the source that is being
736 compiled (which of course may be different). I have changed it back to:
737
738 #include "pcre.h"
739
740 I have a vague recollection that the change was concerned with compiling in
741 different directories, but in the new build system, that is taken care of
742 by the VPATH setting the Makefile.
743
744 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
745 when the subject happened to end in the byte 0x85 (e.g. if the last
746 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
747 characters but of course it shouldn't be taken as a newline when it is part
748 of another character. The bug was that, for an unlimited repeat of . in
749 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
750 characters when looking for a newline.
751
752 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
753
754 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
755 in debug output.
756
757 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
758 long printing in the pcrecpp unittest when running under MinGW.
759
760 6. ESC_K was left out of the EBCDIC table.
761
762 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
763 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
764 limit also applies to "virtual nesting" when a pattern is recursive, and in
765 this case 1000 isn't so big. I have been able to remove this limit at the
766 expense of backing off one optimization in certain circumstances. Normally,
767 when pcre_exec() would call its internal match() function recursively and
768 immediately return the result unconditionally, it uses a "tail recursion"
769 feature to save stack. However, when a subpattern that can match an empty
770 string has an unlimited repetition quantifier, it no longer makes this
771 optimization. That gives it a stack frame in which to save the data for
772 checking that an empty string has been matched. Previously this was taken
773 from the 1000-entry workspace that had been reserved. So now there is no
774 explicit limit, but more stack is used.
775
776 8. Applied Daniel's patches to solve problems with the import/export magic
777 syntax that is required for Windows, and which was going wrong for the
778 pcreposix and pcrecpp parts of the library. These were overlooked when this
779 problem was solved for the main library.
780
781 9. There were some crude static tests to avoid integer overflow when computing
782 the size of patterns that contain repeated groups with explicit upper
783 limits. As the maximum quantifier is 65535, the maximum group length was
784 set at 30,000 so that the product of these two numbers did not overflow a
785 32-bit integer. However, it turns out that people want to use groups that
786 are longer than 30,000 bytes (though not repeat them that many times).
787 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
788 made it possible to implement the integer overflow checks in a much more
789 dynamic way, which I have now done. The artificial limitation on group
790 length has been removed - we now have only the limit on the total length of
791 the compiled pattern, which depends on the LINK_SIZE setting.
792
793 10. Fixed a bug in the documentation for get/copy named substring when
794 duplicate names are permitted. If none of the named substrings are set, the
795 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
796 empty string.
797
798 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
799 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
800 because the ] is interpreted as the first data character and the
801 terminating ] is not found. PCRE has been made compatible with Perl in this
802 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
803 cause memory overwriting.
804
805 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
806 string has been matched (to stop an infinite loop). It was not recognizing
807 a conditional subpattern that could match an empty string if that
808 subpattern was within another subpattern. For example, it looped when
809 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
810 condition was not nested. This bug has been fixed.
811
812 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
813 past the start of the subject in the presence of bytes with the top bit
814 set, for example "\x8aBCD".
815
816 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
817 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
818
819 14. Optimized (?!) to (*FAIL).
820
821 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
822 This restricts code points to be within the range 0 to 0x10FFFF, excluding
823 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
824 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
825 does: it's just the validity check that is more restrictive.
826
827 16. Inserted checks for integer overflows during escape sequence (backslash)
828 processing, and also fixed erroneous offset values for syntax errors during
829 backslash processing.
830
831 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
832 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
833
834 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
835 caused an overrun.
836
837 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
838 something other than just ASCII characters) inside a group that had an
839 unlimited repeat caused a loop at compile time (while checking to see
840 whether the group could match an empty string).
841
842 20. Debugging a pattern containing \p or \P could cause a crash. For example,
843 [\P{Any}] did so. (Error in the code for printing property names.)
844
845 21. An orphan \E inside a character class could cause a crash.
846
847 22. A repeated capturing bracket such as (A)? could cause a wild memory
848 reference during compilation.
849
850 23. There are several functions in pcre_compile() that scan along a compiled
851 expression for various reasons (e.g. to see if it's fixed length for look
852 behind). There were bugs in these functions when a repeated \p or \P was
853 present in the pattern. These operators have additional parameters compared
854 with \d, etc, and these were not being taken into account when moving along
855 the compiled data. Specifically:
856
857 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
858 length.
859
860 (b) An item such as \pL+ within a repeated group could cause crashes or
861 loops.
862
863 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
864 "reference to non-existent subpattern" error.
865
866 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
867
868 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
869 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
870
871 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
872
873 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
874 character were causing crashes (broken optimization).
875
876 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
877 \p or \P) caused a compile-time loop.
878
879 28. More problems have arisen in unanchored patterns when CRLF is a valid line
880 break. For example, the unstudied pattern [\r\n]A does not match the string
881 "\r\nA" because change 7.0/46 below moves the current point on by two
882 characters after failing to match at the start. However, the pattern \nA
883 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
884 the same is true. There doesn't seem any very clean way out of this, but
885 what I have chosen to do makes the common cases work: PCRE now takes note
886 of whether there can be an explicit match for \r or \n anywhere in the
887 pattern, and if so, 7.0/46 no longer applies. As part of this change,
888 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
889 pattern has explicit CR or LF references.
890
891 29. Added (*CR) etc for changing newline setting at start of pattern.
892
893
894 Version 7.2 19-Jun-07
895 ---------------------
896
897 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
898 which is apparently normally available under Windows.
899
900 2. Re-jig the pcregrep tests with different newline settings in an attempt
901 to make them independent of the local environment's newline setting.
902
903 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
904
905 4. Some of the "internals" tests were previously cut out when the link size
906 was not 2, because the output contained actual offsets. The recent new
907 "Z" feature of pcretest means that these can be cut out, making the tests
908 usable with all link sizes.
909
910 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
911 stack recursion. This gives a massive performance boost under BSD, but just
912 a small improvement under Linux. However, it saves one field in the frame
913 in all cases.
914
915 6. Added more features from the forthcoming Perl 5.10:
916
917 (a) (?-n) (where n is a string of digits) is a relative subroutine or
918 recursion call. It refers to the nth most recently opened parentheses.
919
920 (b) (?+n) is also a relative subroutine call; it refers to the nth next
921 to be opened parentheses.
922
923 (c) Conditions that refer to capturing parentheses can be specified
924 relatively, for example, (?(-2)... or (?(+3)...
925
926 (d) \K resets the start of the current match so that everything before
927 is not part of it.
928
929 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
930
931 (f) \g{name} is another synonym - part of Perl 5.10's unification of
932 reference syntax.
933
934 (g) (?| introduces a group in which the numbering of parentheses in each
935 alternative starts with the same number.
936
937 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
938
939 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
940 PCRE_INFO_JCHANGED.
941
942 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
943 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
944 for detecting groups that can match an empty string.
945
946 9. A pattern with a very large number of alternatives (more than several
947 hundred) was running out of internal workspace during the pre-compile
948 phase, where pcre_compile() figures out how much memory will be needed. A
949 bit of new cunning has reduced the workspace needed for groups with
950 alternatives. The 1000-alternative test pattern now uses 12 bytes of
951 workspace instead of running out of the 4096 that are available.
952
953 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
954
955 11. Applied patch from Google to remove an optimization that didn't quite work.
956 The report of the bug said:
957
958 pcrecpp::RE("a*").FullMatch("aaa") matches, while
959 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
960 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
961
962 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
963 it matched the wrong number of bytes.
964
965
966 Version 7.1 24-Apr-07
967 ---------------------
968
969 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
970 that is more "standard", making use of automake and other Autotools. There
971 is some re-arrangement of the files and adjustment of comments consequent
972 on this.
973
974 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
975 for recursive directory scanning broke on some systems because the files
976 are not scanned in any specific order and on different systems the order
977 was different. A call to "sort" has been inserted into RunGrepTest for the
978 approprate test as a short-term fix. In the longer term there may be an
979 alternative.
980
981 3. I had an email from Eric Raymond about problems translating some of PCRE's
982 man pages to HTML (despite the fact that I distribute HTML pages, some
983 people do their own conversions for various reasons). The problems
984 concerned the use of low-level troff macros .br and .in. I have therefore
985 removed all such uses from the man pages (some were redundant, some could
986 be replaced by .nf/.fi pairs). The 132html script that I use to generate
987 HTML has been updated to handle .nf/.fi and to complain if it encounters
988 .br or .in.
989
990 4. Updated comments in configure.ac that get placed in config.h.in and also
991 arranged for config.h to be included in the distribution, with the name
992 config.h.generic, for the benefit of those who have to compile without
993 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
994
995 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
996 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
997 makevp.bat for the current PCRE, using the additional files
998 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
999
1000 6. A Windows user reported a minor discrepancy with test 2, which turned out
1001 to be caused by a trailing space on an input line that had got lost in his
1002 copy. The trailing space was an accident, so I've just removed it.
1003
1004 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1005 that is needed.
1006
1007 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1008 as "const" (a) because they are and (b) because it helps the PHP
1009 maintainers who have recently made a script to detect big data structures
1010 in the php code that should be moved to the .rodata section. I remembered
1011 to update Builducptable as well, so it won't revert if ucptable.h is ever
1012 re-created.
1013
1014 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1015 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1016 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1017 support is not required. This saves 1.5-2K of code, which is important in
1018 some applications.
1019
1020 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1021 so as not to refer to the tables, even though these functions will never be
1022 called when UTF-8 support is disabled. Otherwise there are problems with a
1023 shared library.
1024
1025 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1026
1027 (a) It was defining its arguments as char * instead of void *.
1028
1029 (b) It was assuming that all moves were upwards in memory; this was true
1030 a long time ago when I wrote it, but is no longer the case.
1031
1032 The emulated memove() is provided for those environments that have neither
1033 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1034 is clearly not the case, as these two bugs were recently reported.
1035
1036 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1037 and Detrail to create the HTML documentation, the .txt form of the man
1038 pages, and it removes trailing spaces from listed files. It also creates
1039 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1040 case, it wraps all the #defines with #ifndefs. This script should be run
1041 before "make dist".
1042
1043 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1044 with Unicode property support.
1045
1046 (a) For a maximizing quantifier, if the two different cases of the
1047 character were of different lengths in their UTF-8 codings (there are
1048 some cases like this - I found 11), and the matching function had to
1049 back up over a mixture of the two cases, it incorrectly assumed they
1050 were both the same length.
1051
1052 (b) When PCRE was configured to use the heap rather than the stack for
1053 recursion during matching, it was not correctly preserving the data for
1054 the other case of a UTF-8 character when checking ahead for a match
1055 while processing a minimizing repeat. If the check also involved
1056 matching a wide character, but failed, corruption could cause an
1057 erroneous result when trying to check for a repeat of the original
1058 character.
1059
1060 13. Some tidying changes to the testing mechanism:
1061
1062 (a) The RunTest script now detects the internal link size and whether there
1063 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1064 values substituted by "configure". (The RunGrepTest script already did
1065 this for UTF-8.) The configure.ac script no longer substitutes the
1066 relevant variables.
1067
1068 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1069 with length and offset values. This means that the output is different
1070 for different internal link sizes. Test 2 is skipped for link sizes
1071 other than 2 because of this, bypassing the problem. Unfortunately,
1072 there was also a test in test 3 (the locale tests) that used /B and
1073 failed for link sizes other than 2. Rather than cut the whole test out,
1074 I have added a new /Z option to pcretest that replaces the length and
1075 offset values with spaces. This is now used to make test 3 independent
1076 of link size. (Test 2 will be tidied up later.)
1077
1078 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1079 segmentation fault instead of returning the appropriate error message.
1080
1081 15. In multiline mode when the newline sequence was set to "any", the pattern
1082 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1083 This doesn't seem right; it now treats the CRLF combination as the line
1084 ending, and so does not match in that case. It's only a pattern such as ^$
1085 that would hit this one: something like ^ABC$ would have failed after \r
1086 and then tried again after \r\n.
1087
1088 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1089 in an attempt to make files that differ only in their line terminators
1090 compare equal. This works on Linux.
1091
1092 17. Under certain error circumstances pcregrep might try to free random memory
1093 as it exited. This is now fixed, thanks to valgrind.
1094
1095 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1096 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1097 was because its rules for how to advance for /g after matching an empty
1098 string at the end of a line did not allow for this case. They now check for
1099 it specially.
1100
1101 20. pcretest is supposed to handle patterns and data of any length, by
1102 extending its buffers when necessary. It was getting this wrong when the
1103 buffer for a data line had to be extended.
1104
1105 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1106 CRLF as a newline sequence.
1107
1108 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1109 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1110 I have nevertheless tidied it up.
1111
1112 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1113
1114 24. Added a man page for pcre-config.
1115
1116
1117 Version 7.0 19-Dec-06
1118 ---------------------
1119
1120 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1121 moving to gcc 4.1.1.
1122
1123 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1124 sys/time.h, which is documented as needed for this function. It doesn't
1125 seem to matter on Linux, but it showed up on some releases of OS X.
1126
1127 3. It seems that there are systems where bytes whose values are greater than
1128 127 match isprint() in the "C" locale. The "C" locale should be the
1129 default when a C program starts up. In most systems, only ASCII printing
1130 characters match isprint(). This difference caused the output from pcretest
1131 to vary, making some of the tests fail. I have changed pcretest so that:
1132
1133 (a) When it is outputting text in the compiled version of a pattern, bytes
1134 other than 32-126 are always shown as hex escapes.
1135
1136 (b) When it is outputting text that is a matched part of a subject string,
1137 it does the same, unless a different locale has been set for the match
1138 (using the /L modifier). In this case, it uses isprint() to decide.
1139
1140 4. Fixed a major bug that caused incorrect computation of the amount of memory
1141 required for a compiled pattern when options that changed within the
1142 pattern affected the logic of the preliminary scan that determines the
1143 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1144 that the computed length was too small. The symptoms of this bug were
1145 either the PCRE error "internal error: code overflow" from pcre_compile(),
1146 or a glibc crash with a message such as "pcretest: free(): invalid next
1147 size (fast)". Examples of patterns that provoked this bug (shown in
1148 pcretest format) are:
1149
1150 /(?-x: )/x
1151 /(?x)(?-x: \s*#\s*)/
1152 /((?i)[\x{c0}])/8
1153 /(?i:[\x{c0}])/8
1154
1155 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1156 is now done differently.
1157
1158 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1159 wrapper classes; (b) implement a new function in the C++ scanner that is
1160 more efficient than the old way of doing things because it avoids levels of
1161 recursion in the regex matching; (c) add a paragraph to the documentation
1162 for the FullMatch() function.
1163
1164 6. The escape sequence \n was being treated as whatever was defined as
1165 "newline". Not only was this contrary to the documentation, which states
1166 that \n is character 10 (hex 0A), but it also went horribly wrong when
1167 "newline" was defined as CRLF. This has been fixed.
1168
1169 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1170 was being set to -1 for the "end of line" case (supposedly a value that no
1171 character can have). Though this value is never used (the check for end of
1172 line is "zero bytes in current character"), it caused compiler complaints.
1173 I've changed it to 0xffffffff.
1174
1175 8. In pcre_version.c, the version string was being built by a sequence of
1176 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1177 string (as it is for production releases) called a macro with an empty
1178 argument. The C standard says the result of this is undefined. The gcc
1179 compiler treats it as an empty string (which was what was wanted) but it is
1180 reported that Visual C gives an error. The source has been hacked around to
1181 avoid this problem.
1182
1183 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1184 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1185 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1186 of them did).
1187
1188 10. Originally, pcretest opened its input and output without "b"; then I was
1189 told that "b" was needed in some environments, so it was added for release
1190 5.0 to both the input and output. (It makes no difference on Unix-like
1191 systems.) Later I was told that it is wrong for the input on Windows. I've
1192 now abstracted the modes into two macros, to make it easier to fiddle with
1193 them, and removed "b" from the input mode under Windows.
1194
1195 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1196
1197 12. Added -help and --help to pcretest as an official way of being reminded
1198 of the options.
1199
1200 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1201 and pcrecpp.cc because they annoy compilers at high warning levels.
1202
1203 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1204
1205 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1206 scripts are not C programs :-) and which was not noticed because it works
1207 on Linux.
1208
1209 16. pcretest is supposed to handle any length of pattern and data line (as one
1210 line or as a continued sequence of lines) by extending its input buffer if
1211 necessary. This feature was broken for very long pattern lines, leading to
1212 a string of junk being passed to pcre_compile() if the pattern was longer
1213 than about 50K.
1214
1215 17. I have done a major re-factoring of the way pcre_compile() computes the
1216 amount of memory needed for a compiled pattern. Previously, there was code
1217 that made a preliminary scan of the pattern in order to do this. That was
1218 OK when PCRE was new, but as the facilities have expanded, it has become
1219 harder and harder to keep it in step with the real compile phase, and there
1220 have been a number of bugs (see for example, 4 above). I have now found a
1221 cunning way of running the real compile function in a "fake" mode that
1222 enables it to compute how much memory it would need, while actually only
1223 ever using a few hundred bytes of working memory and without too many
1224 tests of the mode. This should make future maintenance and development
1225 easier. A side effect of this work is that the limit of 200 on the nesting
1226 depth of parentheses has been removed (though this was never a serious
1227 limitation, I suspect). However, there is a downside: pcre_compile() now
1228 runs more slowly than before (30% or more, depending on the pattern). I
1229 hope this isn't a big issue. There is no effect on runtime performance.
1230
1231 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1232 newline (only possible for the last line of a file) and it was a
1233 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1234
1235 19. Added additional timing features to pcretest. (1) The -tm option now times
1236 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1237 separate command line item, by a number that specifies the number of
1238 repeats to use when timing. The default is 50000; this gives better
1239 precision, but takes uncomfortably long for very large patterns.
1240
1241 20. Extended pcre_study() to be more clever in cases where a branch of a
1242 subpattern has no definite first character. For example, (a*|b*)[cd] would
1243 previously give no result from pcre_study(). Now it recognizes that the
1244 first character must be a, b, c, or d.
1245
1246 21. There was an incorrect error "recursive call could loop indefinitely" if
1247 a subpattern (or the entire pattern) that was being tested for matching an
1248 empty string contained only one non-empty item after a nested subpattern.
1249 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1250 incorrectly, because the \d was being skipped in the check.
1251
1252 22. The pcretest program now has a new pattern option /B and a command line
1253 option -b, which is equivalent to adding /B to every pattern. This causes
1254 it to show the compiled bytecode, without the additional information that
1255 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1256 is the same as /B/I).
1257
1258 23. A new optimization is now able automatically to treat some sequences such
1259 as a*b as a*+b. More specifically, if something simple (such as a character
1260 or a simple class like \d) has an unlimited quantifier, and is followed by
1261 something that cannot possibly match the quantified thing, the quantifier
1262 is automatically "possessified".
1263
1264 24. A recursive reference to a subpattern whose number was greater than 39
1265 went wrong under certain circumstances in UTF-8 mode. This bug could also
1266 have affected the operation of pcre_study().
1267
1268 25. Realized that a little bit of performance could be had by replacing
1269 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1270
1271 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1272
1273 27. Possessive quantifiers such as a++ were previously implemented by turning
1274 them into atomic groups such as ($>a+). Now they have their own opcodes,
1275 which improves performance. This includes the automatically created ones
1276 from 23 above.
1277
1278 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1279 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1280 the first matched character to be a colon. This applied both to named and
1281 numbered groups.
1282
1283 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1284
1285 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1286 building PCRE on Windows easier, so I have included it in the distribution.
1287
1288 31. There is now a check in pcretest against a ridiculously large number being
1289 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1290 loop, the loop is abandoned.
1291
1292 32. Forward references to subpatterns in conditions such as (?(2)...) where
1293 subpattern 2 is defined later cause pcre_compile() to search forwards in
1294 the pattern for the relevant set of parentheses. This search went wrong
1295 when there were unescaped parentheses in a character class, parentheses
1296 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1297
1298 33. "Subroutine" calls and backreferences were previously restricted to
1299 referencing subpatterns earlier in the regex. This restriction has now
1300 been removed.
1301
1302 34. Added a number of extra features that are going to be in Perl 5.10. On the
1303 whole, these are just syntactic alternatives for features that PCRE had
1304 previously implemented using the Python syntax or my own invention. The
1305 other formats are all retained for compatibility.
1306
1307 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1308 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1309 also .NET compatible.
1310
1311 (b) A recursion or subroutine call to a named group can now be defined as
1312 (?&name) as well as (?P>name).
1313
1314 (c) A backreference to a named group can now be defined as \k<name> or
1315 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1316 5.10, are also .NET compatible.
1317
1318 (d) A conditional reference to a named group can now use the syntax
1319 (?(<name>) or (?('name') as well as (?(name).
1320
1321 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1322 groups (named and numbered) that are never evaluated inline, but can be
1323 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1324 is always false. There may be only one alternative in such a group.
1325
1326 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1327 as the simple (?(R). The condition is true only if the most recent
1328 recursion is that of the given number or name. It does not search out
1329 through the entire recursion stack.
1330
1331 (g) The escape \gN or \g{N} has been added, where N is a positive or
1332 negative number, specifying an absolute or relative reference.
1333
1334 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1335 some "unreachable code" warnings.
1336
1337 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1338 things, this adds five new scripts.
1339
1340 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1341 There were also incompatibilities regarding the handling of \Q..\E inside
1342 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1343 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1344
1345 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1346 matches an empty string, and forcibly breaks the loop. There were bugs in
1347 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1348 against aaaa the result was just "a" rather than "aaaa", for example. Two
1349 separate and independent bugs (that affected different cases) have been
1350 fixed.
1351
1352 39. Refactored the code to abolish the use of different opcodes for small
1353 capturing bracket numbers. This is a tidy that I avoided doing when I
1354 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1355 The new approach is not only tidier, it makes it possible to reduce the
1356 memory needed to fix the previous bug (38).
1357
1358 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1359 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1360 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1361 mode.
1362
1363 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1364 report.
1365
1366 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1367 copy construction and assignment in the C++ wrapper.
1368
1369 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1370 couple of bugs that could have given wrong results in the "--newline=crlf"
1371 case.
1372
1373 44. Added a number of casts and did some reorganization of signed/unsigned int
1374 variables following suggestions from Dair Grant. Also renamed the variable
1375 "this" as "item" because it is a C++ keyword.
1376
1377 45. Arranged for dftables to add
1378
1379 #include "pcre_internal.h"
1380
1381 to pcre_chartables.c because without it, gcc 4.x may remove the array
1382 definition from the final binary if PCRE is built into a static library and
1383 dead code stripping is activated.
1384
1385 46. For an unanchored pattern, if a match attempt fails at the start of a
1386 newline sequence, and the newline setting is CRLF or ANY, and the next two
1387 characters are CRLF, advance by two characters instead of one.
1388
1389
1390 Version 6.7 04-Jul-06
1391 ---------------------
1392
1393 1. In order to handle tests when input lines are enormously long, pcretest has
1394 been re-factored so that it automatically extends its buffers when
1395 necessary. The code is crude, but this _is_ just a test program. The
1396 default size has been increased from 32K to 50K.
1397
1398 2. The code in pcre_study() was using the value of the re argument before
1399 testing it for NULL. (Of course, in any sensible call of the function, it
1400 won't be NULL.)
1401
1402 3. The memmove() emulation function in pcre_internal.h, which is used on
1403 systems that lack both memmove() and bcopy() - that is, hardly ever -
1404 was missing a "static" storage class specifier.
1405
1406 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1407 containing an extended class (one that cannot be represented by a bitmap
1408 because it contains high-valued characters or Unicode property items, e.g.
1409 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1410 pattern, but PCRE should not loop if you do not (it no longer does).
1411 [Detail: two cases were found: (a) a repeated subpattern containing an
1412 extended class; (b) a recursive reference to a subpattern that followed a
1413 previous extended class. It wasn't skipping over the extended class
1414 correctly when UTF-8 mode was not set.]
1415
1416 5. A negated single-character class was not being recognized as fixed-length
1417 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1418 compile error "lookbehind assertion is not fixed length".
1419
1420 6. The RunPerlTest auxiliary script was showing an unexpected difference
1421 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1422 write a Perl script that can interpret lines of an input file either as
1423 byte characters or as UTF-8, which is what "perltest" was being required to
1424 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1425 can't do is switch easily at run time between having the "use utf8;" pragma
1426 or not. In the end, I fudged it by using the RunPerlTest script to insert
1427 "use utf8;" explicitly for the UTF-8 tests.
1428
1429 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1430 the end of the subject string, contrary to the documentation and to what
1431 Perl does. This was true of both matching functions. Now it matches only at
1432 the start of the subject and immediately after *internal* newlines.
1433
1434 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1435 a pointer to an int instead of a pointer to an unsigned long int. This
1436 caused problems on 64-bit systems.
1437
1438 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1439 instance of the 'standard' template library not being so standard".
1440
1441 10. There was no check on the number of named subpatterns nor the maximum
1442 length of a subpattern name. The product of these values is used to compute
1443 the size of the memory block for a compiled pattern. By supplying a very
1444 long subpattern name and a large number of named subpatterns, the size
1445 computation could be caused to overflow. This is now prevented by limiting
1446 the length of names to 32 characters, and the number of named subpatterns
1447 to 10,000.
1448
1449 11. Subpatterns that are repeated with specific counts have to be replicated in
1450 the compiled pattern. The size of memory for this was computed from the
1451 length of the subpattern and the repeat count. The latter is limited to
1452 65535, but there was no limit on the former, meaning that integer overflow
1453 could in principle occur. The compiled length of a repeated subpattern is
1454 now limited to 30,000 bytes in order to prevent this.
1455
1456 12. Added the optional facility to have named substrings with the same name.
1457
1458 13. Added the ability to use a named substring as a condition, using the
1459 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1460 are numbers (not recommended). Forward references are permitted.
1461
1462 14. Added forward references in named backreferences (if you see what I mean).
1463
1464 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1465 pattern could run off the end of the subject. For example, the pattern
1466 "(?s)(.{1,5})"8 did this with the subject "ab".
1467
1468 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1469 PCRE_CASELESS was set when matching characters that were quantified with ?
1470 or *.
1471
1472 17. A character class other than a single negated character that had a minimum
1473 but no maximum quantifier - for example [ab]{6,} - was not handled
1474 correctly by pce_dfa_exec(). It would match only one character.
1475
1476 18. A valid (though odd) pattern that looked like a POSIX character
1477 class but used an invalid character after [ (for example [[,abc,]]) caused
1478 pcre_compile() to give the error "Failed: internal error: code overflow" or
1479 in some cases to crash with a glibc free() error. This could even happen if
1480 the pattern terminated after [[ but there just happened to be a sequence of
1481 letters, a binary zero, and a closing ] in the memory that followed.
1482
1483 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1484 over the years. Originally (before any Unicode support), just the bottom 8
1485 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1486 output from "man perlunicode" includes this:
1487
1488 The regular expression compiler produces polymorphic opcodes. That
1489 is, the pattern adapts to the data and automatically switches to
1490 the Unicode character scheme when presented with Unicode data--or
1491 instead uses a traditional byte scheme when presented with byte
1492 data.
1493
1494 Sadly, a wide octal escape does not cause a switch, and in a string with
1495 no other multibyte characters, these octal escapes are treated as before.
1496 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1497 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1498 Unicode string.
1499
1500 I have not perpetrated such confusion in PCRE. Up till now, it took just
1501 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1502 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1503 translate to the appropriate multibyte character.
1504
1505 29. Applied some refactoring to reduce the number of warnings from Microsoft
1506 and Borland compilers. This has included removing the fudge introduced
1507 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1508 a warning about an unused variable.
1509
1510 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1511 characters since release 4.0, because Perl (from release 5.004) does not.
1512 [Or at least, is documented not to: some releases seem to be in conflict
1513 with the documentation.] However, when a pattern was studied with
1514 pcre_study() and all its branches started with \s, PCRE still included VT
1515 as a possible starting character. Of course, this did no harm; it just
1516 caused an unnecessary match attempt.
1517
1518 22. Removed a now-redundant internal flag bit that recorded the fact that case
1519 dependency changed within the pattern. This was once needed for "required
1520 byte" processing, but is no longer used. This recovers a now-scarce options
1521 bit. Also moved the least significant internal flag bit to the most-
1522 significant bit of the word, which was not previously used (hangover from
1523 the days when it was an int rather than a uint) to free up another bit for
1524 the future.
1525
1526 23. Added support for CRLF line endings as well as CR and LF. As well as the
1527 default being selectable at build time, it can now be changed at runtime
1528 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1529 specify that it is scanning data with non-default line endings.
1530
1531 24. Changed the definition of CXXLINK to make it agree with the definition of
1532 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1533
1534 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1535 recursions. This makes a big different to stack usage for some patterns.
1536
1537 26. If a subpattern containing a named recursion or subroutine reference such
1538 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1539 the space required for the compiled pattern went wrong and gave too small a
1540 value. Depending on the environment, this could lead to "Failed: internal
1541 error: code overflow at offset 49" or "glibc detected double free or
1542 corruption" errors.
1543
1544 27. Applied patches from Google (a) to support the new newline modes and (b) to
1545 advance over multibyte UTF-8 characters in GlobalReplace.
1546
1547 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1548 difference for some implementation of PCRE in some Windows version.
1549
1550 29. Added some extra testing facilities to pcretest:
1551
1552 \q<number> in a data line sets the "match limit" value
1553 \Q<number> in a data line sets the "match recursion limt" value
1554 -S <number> sets the stack size, where <number> is in megabytes
1555
1556 The -S option isn't available for Windows.
1557
1558
1559 Version 6.6 06-Feb-06
1560 ---------------------
1561
1562 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1563 in pcreposix.h. I have copied the definition from pcre.h.
1564
1565 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1566 because pcre.h is no longer a built file.
1567
1568 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1569 not normally included in the compiled code.
1570
1571
1572 Version 6.5 01-Feb-06
1573 ---------------------
1574
1575 1. When using the partial match feature with pcre_dfa_exec(), it was not
1576 anchoring the second and subsequent partial matches at the new starting
1577 point. This could lead to incorrect results. For example, with the pattern
1578 /1234/, partially matching against "123" and then "a4" gave a match.
1579
1580 2. Changes to pcregrep:
1581
1582 (a) All non-match returns from pcre_exec() were being treated as failures
1583 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1584 error message is output. Some extra information is given for the
1585 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1586 probably the only errors that are likely to be caused by users (by
1587 specifying a regex that has nested indefinite repeats, for instance).
1588 If there are more than 20 of these errors, pcregrep is abandoned.
1589
1590 (b) A binary zero was treated as data while matching, but terminated the
1591 output line if it was written out. This has been fixed: binary zeroes
1592 are now no different to any other data bytes.
1593
1594 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1595 used to set a locale for matching. The --locale=xxxx long option has
1596 been added (no short equivalent) to specify a locale explicitly on the
1597 pcregrep command, overriding the environment variables.
1598
1599 (d) When -B was used with -n, some line numbers in the output were one less
1600 than they should have been.
1601
1602 (e) Added the -o (--only-matching) option.
1603
1604 (f) If -A or -C was used with -c (count only), some lines of context were
1605 accidentally printed for the final match.
1606
1607 (g) Added the -H (--with-filename) option.
1608
1609 (h) The combination of options -rh failed to suppress file names for files
1610 that were found from directory arguments.
1611
1612 (i) Added the -D (--devices) and -d (--directories) options.
1613
1614 (j) Added the -F (--fixed-strings) option.
1615
1616 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1617
1618 (l) Added the --colo(u)r option.
1619
1620 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1621 is not present by default.
1622
1623 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1624 items such as (?R) or (?1), when the recursion could match a number of
1625 alternatives. If it matched one of the alternatives, but subsequently,
1626 outside the recursion, there was a failure, the code tried to back up into
1627 the recursion. However, because of the way PCRE is implemented, this is not
1628 possible, and the result was an incorrect result from the match.
1629
1630 In order to prevent this happening, the specification of recursion has
1631 been changed so that all such subpatterns are automatically treated as
1632 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1633
1634 4. I had overlooked the fact that, in some locales, there are characters for
1635 which isalpha() is true but neither isupper() nor islower() are true. In
1636 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1637 and ordfeminine) are like this. This affected the treatment of \w and \W
1638 when they appeared in character classes, but not when they appeared outside
1639 a character class. The bit map for "word" characters is now created
1640 separately from the results of isalnum() instead of just taking it from the
1641 upper, lower, and digit maps. (Plus the underscore character, of course.)
1642
1643 5. The above bug also affected the handling of POSIX character classes such as
1644 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1645 permanent tables. Instead, the bit maps for such a class were previously
1646 created as the appropriate unions of the upper, lower, and digit bitmaps.
1647 Now they are created by subtraction from the [[:word:]] class, which has
1648 its own bitmap.
1649
1650 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1651 It is created by subtracting the vertical space characters (\x09, \x0a,
1652 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1653 subtraction was done in the overall bitmap for a character class, meaning
1654 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1655 be recognized. This bug has been fixed.
1656
1657 7. Patches from the folks at Google:
1658
1659 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1660 real life, but is still worth protecting against".
1661
1662 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1663 regular expressions".
1664
1665 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1666 have it.
1667
1668 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1669 "configure" and the latter not, in order to fix a problem somebody had
1670 with compiling the Arg class on HP-UX.
1671
1672 (e) Improve the error-handling of the C++ wrapper a little bit.
1673
1674 (f) New tests for checking recursion limiting.
1675
1676 8. The pcre_memmove() function, which is used only if the environment does not
1677 have a standard memmove() function (and is therefore rarely compiled),
1678 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1679 returning a result (though PCRE never actually uses the result).
1680
1681 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1682 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1683 returned instead of calling malloc() with an overflowing number that would
1684 most likely cause subsequent chaos.
1685
1686 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1687
1688 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1689 with this option is matched, the nmatch and pmatch options of regexec() are
1690 ignored.
1691
1692 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1693 provided in case anyone wants to the the POSIX interface with UTF-8
1694 strings.
1695
1696 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1697 C++ linking (needed for some HP-UX environments).
1698
1699 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1700 (unused parameter) and in the pcre_printint() function (omitted "default"
1701 switch label when the default is to do nothing).
1702
1703 15. Added some code to make it possible, when PCRE is compiled as a C++
1704 library, to replace subject pointers for pcre_exec() with a smart pointer
1705 class, thus making it possible to process discontinuous strings.
1706
1707 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1708 much the same function. They were added by different people who were trying
1709 to make PCRE easy to compile on non-Unix systems. It has been suggested
1710 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1711 for compiling on Windows systems. I have therefore replaced it with
1712 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1713 defaults to "extern" for C or "extern C" for C++, which works fine on
1714 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1715 SCOPE with something explicit in config.h. In addition:
1716
1717 (a) pcreposix.h still had just "extern" instead of either of these macros;
1718 I have replaced it with PCRE_DATA_SCOPE.
1719
1720 (b) Functions such as _pcre_xclass(), which are internal to the library,
1721 but external in the C sense, all had PCRE_EXPORT in their definitions.
1722 This is apparently wrong for the Windows case, so I have removed it.
1723 (It makes no difference on Unix-like systems.)
1724
1725 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1726 of recursive calls to match(). This is different to MATCH_LIMIT because
1727 that limits the total number of calls to match(), not all of which increase
1728 the depth of recursion. Limiting the recursion depth limits the amount of
1729 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1730 when PCRE is compiled, and changed at run time. A patch from Google adds
1731 this functionality to the C++ interface.
1732
1733 18. Changes to the handling of Unicode character properties:
1734
1735 (a) Updated the table to Unicode 4.1.0.
1736
1737 (b) Recognize characters that are not in the table as "Cn" (undefined).
1738
1739 (c) I revised the way the table is implemented to a much improved format
1740 which includes recognition of ranges. It now supports the ranges that
1741 are defined in UnicodeData.txt, and it also amalgamates other
1742 characters into ranges. This has reduced the number of entries in the
1743 table from around 16,000 to around 3,000, thus reducing its size
1744 considerably. I realized I did not need to use a tree structure after
1745 all - a binary chop search is just as efficient. Having reduced the
1746 number of entries, I extended their size from 6 bytes to 8 bytes to
1747 allow for more data.
1748
1749 (d) Added support for Unicode script names via properties such as \p{Han}.
1750
1751 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1752 matching that character.
1753
1754 20. When matching a repeated Unicode property with a minimum greater than zero,
1755 (for example \pL{2,}), PCRE could look past the end of the subject if it
1756 reached it while seeking the minimum number of characters. This could
1757 happen only if some of the characters were more than one byte long, because
1758 there is a check for at least the minimum number of bytes.
1759
1760 21. Refactored the implementation of \p and \P so as to be more general, to
1761 allow for more different types of property in future. This has changed the
1762 compiled form incompatibly. Anybody with saved compiled patterns that use
1763 \p or \P will have to recompile them.
1764
1765 22. Added "Any" and "L&" to the supported property types.
1766
1767 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1768 but give a compile time error if the value is greater than 0xff.
1769
1770 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1771 accidentally not being installed or uninstalled.
1772
1773 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1774 made were to insert the current release number. This seemed silly, because
1775 it made things harder for people building PCRE on systems that don't run
1776 "configure". I have turned pcre.h into a distributed file, no longer built
1777 by "configure", with the version identification directly included. There is
1778 no longer a pcre.h.in file.
1779
1780 However, this change necessitated a change to the pcre-config script as
1781 well. It is built from pcre-config.in, and one of the substitutions was the
1782 release number. I have updated configure.ac so that ./configure now finds
1783 the release number by grepping pcre.h.
1784
1785 26. Added the ability to run the tests under valgrind.
1786
1787
1788 Version 6.4 05-Sep-05
1789 ---------------------
1790
1791 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1792 "--" to be printed when multiple files were scanned, even when none of the
1793 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1794 consider it to be a bug, and have restored the previous behaviour.
1795
1796 2. A couple of code tidies to get rid of compiler warnings.
1797
1798 3. The pcretest program used to cheat by referring to symbols in the library
1799 whose names begin with _pcre_. These are internal symbols that are not
1800 really supposed to be visible externally, and in some environments it is
1801 possible to suppress them. The cheating is now confined to including
1802 certain files from the library's source, which is a bit cleaner.
1803
1804 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1805 file's purpose clearer.
1806
1807 5. Reorganized pcre_ucp_findchar().
1808
1809
1810 Version 6.3 15-Aug-05
1811 ---------------------
1812
1813 1. The file libpcre.pc.in did not have general read permission in the tarball.
1814
1815 2. There were some problems when building without C++ support:
1816
1817 (a) If C++ support was not built, "make install" and "make test" still
1818 tried to test it.
1819
1820 (b) There were problems when the value of CXX was explicitly set. Some
1821 changes have been made to try to fix these, and ...
1822
1823 (c) --disable-cpp can now be used to explicitly disable C++ support.
1824
1825 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1826 backslash in a target when C++ was disabled. This confuses some
1827 versions of "make", apparently. Using an intermediate variable solves
1828 this. (Same for CPP_LOBJ.)
1829
1830 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1831 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1832 necessary on certain architectures.
1833
1834 4. Added a setting of -export-symbols-regex to the link command to remove
1835 those symbols that are exported in the C sense, but actually are local
1836 within the library, and not documented. Their names all begin with
1837 "_pcre_". This is not a perfect job, because (a) we have to except some
1838 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1839 available (and never for static libraries). I have made a note to try to
1840 find a way round (a) in the future.
1841
1842
1843 Version 6.2 01-Aug-05
1844 ---------------------
1845
1846 1. There was no test for integer overflow of quantifier values. A construction
1847 such as {1111111111111111} would give undefined results. What is worse, if
1848 a minimum quantifier for a parenthesized subpattern overflowed and became
1849 negative, the calculation of the memory size went wrong. This could have
1850 led to memory overwriting.
1851
1852 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1853
1854 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1855 operating environments where this matters.
1856
1857 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1858 PCRE options from within the C++ wrapper.
1859
1860 5. Named capturing subpatterns were not being correctly counted when a pattern
1861 was compiled. This caused two problems: (a) If there were more than 100
1862 such subpatterns, the calculation of the memory needed for the whole
1863 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1864 back references of the form \12, where the number was greater than 9, were
1865 not recognized as back references, even though there were sufficient
1866 previous subpatterns.
1867
1868 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1869 versions of gcc, e.g. 2.95.4.
1870
1871
1872 Version 6.1 21-Jun-05
1873 ---------------------
1874
1875 1. There was one reference to the variable "posix" in pcretest.c that was not
1876 surrounded by "#if !defined NOPOSIX".
1877
1878 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1879 the cross-check on the old pcre_info() function, for the benefit of the
1880 cut-down version of PCRE that is currently imported into Exim.
1881
1882 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1883 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1884 patterns that start (?i) but I don't think that matters. The use of (?i) is
1885 just an example; this all applies to the other options as well.
1886
1887 4. Since libtool seems to echo the compile commands it is issuing, the output
1888 from "make" can be reduced a bit by putting "@" in front of each libtool
1889 compile command.
1890
1891 5. Patch from the folks at Google for configure.in to be a bit more thorough
1892 in checking for a suitable C++ installation before trying to compile the
1893 C++ stuff. This should fix a reported problem when a compiler was present,
1894 but no suitable headers.
1895
1896 6. The man pages all had just "PCRE" as their title. I have changed them to
1897 be the relevant file name. I have also arranged that these names are
1898 retained in the file doc/pcre.txt, which is a concatenation in text format
1899 of all the man pages except the little individual ones for each function.
1900
1901 7. The NON-UNIX-USE file had not been updated for the different set of source
1902 files that come with release 6. I also added a few comments about the C++
1903 wrapper.
1904
1905
1906 Version 6.0 07-Jun-05
1907 ---------------------
1908
1909 1. Some minor internal re-organization to help with my DFA experiments.
1910
1911 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1912 didn't matter for the library itself when fully configured, but did matter
1913 when compiling without UCP support, or within Exim, where the ucp files are
1914 not imported.
1915
1916 3. Refactoring of the library code to split up the various functions into
1917 different source modules. The addition of the new DFA matching code (see
1918 below) to a single monolithic source would have made it really too
1919 unwieldy, quite apart from causing all the code to be include in a
1920 statically linked application, when only some functions are used. This is
1921 relevant even without the DFA addition now that patterns can be compiled in
1922 one application and matched in another.
1923
1924 The downside of splitting up is that there have to be some external
1925 functions and data tables that are used internally in different modules of
1926 the library but which are not part of the API. These have all had their
1927 names changed to start with "_pcre_" so that they are unlikely to clash
1928 with other external names.
1929
1930 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1931 a different (DFA) algorithm. Although it is slower than the original
1932 function, it does have some advantages for certain types of matching
1933 problem.
1934
1935 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1936 including restarting after a partial match.
1937
1938 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1939 defined when compiling for Windows was sent to me. I have put it into the
1940 code, though I have no means of testing or verifying it.
1941
1942 7. Added the pcre_refcount() auxiliary function.
1943
1944 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1945 match before or at the first newline in the subject string. In pcretest,
1946 the /f option on a pattern can be used to set this.
1947
1948 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1949 would behave wrongly. This has been present in PCRE since release 4.0.
1950
1951 10. A number of changes to the pcregrep command:
1952
1953 (a) Refactored how -x works; insert ^(...)$ instead of setting
1954 PCRE_ANCHORED and checking the length, in preparation for adding
1955 something similar for -w.
1956
1957 (b) Added the -w (match as a word) option.
1958
1959 (c) Refactored the way lines are read and buffered so as to have more
1960 than one at a time available.
1961
1962 (d) Implemented a pcregrep test script.
1963
1964 (e) Added the -M (multiline match) option. This allows patterns to match
1965 over several lines of the subject. The buffering ensures that at least
1966 8K, or the rest of the document (whichever is the shorter) is available
1967 for matching (and similarly the previous 8K for lookbehind assertions).
1968
1969 (f) Changed the --help output so that it now says
1970
1971 -w, --word-regex(p)
1972
1973 instead of two lines, one with "regex" and the other with "regexp"
1974 because that confused at least one person since the short forms are the
1975 same. (This required a bit of code, as the output is generated
1976 automatically from a table. It wasn't just a text change.)
1977
1978 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1979 option but starts with a hyphen. Could be a pattern or a path name
1980 starting with a hyphen, for instance.
1981
1982 (h) "-" can be given as a file name to represent stdin.
1983
1984 (i) When file names are being printed, "(standard input)" is used for
1985 the standard input, for compatibility with GNU grep. Previously
1986 "<stdin>" was used.
1987
1988 (j) The option --label=xxx can be used to supply a name to be used for
1989 stdin when file names are being printed. There is no short form.
1990
1991 (k) Re-factored the options decoding logic because we are going to add
1992 two more options that take data. Such options can now be given in four
1993 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1994
1995 (l) Added the -A, -B, and -C options for requesting that lines of context
1996 around matches be printed.
1997
1998 (m) Added the -L option to print the names of files that do not contain
1999 any matching lines, that is, the complement of -l.
2000
2001 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2002 continue to scan other files.
2003
2004 (o) The -s option was incorrectly implemented. For compatibility with other
2005 greps, it now suppresses the error message for a non-existent or non-
2006 accessible file (but not the return code). There is a new option called
2007 -q that suppresses the output of matching lines, which was what -s was
2008 previously doing.
2009
2010 (p) Added --include and --exclude options to specify files for inclusion
2011 and exclusion when recursing.
2012
2013 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2014 Hopefully, it now does.
2015
2016 12. Missing cast in pcre_study().
2017
2018 13. Added an "uninstall" target to the makefile.
2019
2020 14. Replaced "extern" in the function prototypes in Makefile.in with
2021 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2022 world, but is set differently for Windows.
2023
2024 15. Added a second compiling function called pcre_compile2(). The only
2025 difference is that it has an extra argument, which is a pointer to an
2026 integer error code. When there is a compile-time failure, this is set
2027 non-zero, in addition to the error test pointer being set to point to an
2028 error message. The new argument may be NULL if no error number is required
2029 (but then you may as well call pcre_compile(), which is now just a
2030 wrapper). This facility is provided because some applications need a
2031 numeric error indication, but it has also enabled me to tidy up the way
2032 compile-time errors are handled in the POSIX wrapper.
2033
2034 16. Added VPATH=.libs to the makefile; this should help when building with one
2035 prefix path and installing with another. (Or so I'm told by someone who
2036 knows more about this stuff than I do.)
2037
2038 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2039 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2040 match everything, including newlines. This is not POSIX-compatible, but
2041 somebody wanted the feature. From pcretest it can be activated by using
2042 both the P and the s flags.
2043
2044 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2045
2046 19. libpcre.pc was being incorrectly installed as executable.
2047
2048 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2049 it now also looks for '\r' so that it will work unmodified on Windows.
2050
2051 21. Added Google's contributed C++ wrapper to the distribution.
2052
2053 22. Added some untidy missing memory free() calls in pcretest, to keep
2054 Electric Fence happy when testing.
2055
2056
2057
2058 Version 5.0 13-Sep-04
2059 ---------------------
2060
2061 1. Internal change: literal characters are no longer packed up into items
2062 containing multiple characters in a single byte-string. Each character
2063 is now matched using a separate opcode. However, there may be more than one
2064 byte in the character in UTF-8 mode.
2065
2066 2. The pcre_callout_block structure has two new fields: pattern_position and
2067 next_item_length. These contain the offset in the pattern to the next match
2068 item, and its length, respectively.
2069
2070 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2071 insertion of callouts before each pattern item. Added the /C option to
2072 pcretest to make use of this.
2073
2074 4. On the advice of a Windows user, the lines
2075
2076 #if defined(_WIN32) || defined(WIN32)
2077 _setmode( _fileno( stdout ), 0x8000 );
2078 #endif /* defined(_WIN32) || defined(WIN32) */
2079
2080 have been added to the source of pcretest. This apparently does useful
2081 magic in relation to line terminators.
2082
2083 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2084 for the benefit of those environments where the "b" makes a difference.
2085
2086 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2087 to know about it. I have put a hack into configure.in that adds in code
2088 to set GCC=yes if CC=icc. This seems to end up at a point in the
2089 generated configure script that is early enough to affect the setting of
2090 compiler options, which is what is needed, but I have no means of testing
2091 whether it really works. (The user who reported this had patched the
2092 generated configure script, which of course I cannot do.)
2093
2094 LATER: After change 22 below (new libtool files), the configure script
2095 seems to know about icc (and also ecc). Therefore, I have commented out
2096 this hack in configure.in.
2097
2098 7. Added support for pkg-config (2 patches were sent in).
2099
2100 8. Negated POSIX character classes that used a combination of internal tables
2101 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2102 [[:^ascii]]. Typically, they would match almost any characters. The other
2103 POSIX classes were not broken in this way.
2104
2105 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2106 to find the match, as PCRE was deluded into thinking that the match had to
2107 start at the start point or following a newline. The same bug applied to
2108 patterns with negative forward assertions or any backward assertions
2109 preceding ".*" at the start, unless the pattern required a fixed first
2110 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2111
2112 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2113 starting at the last subject character, bytes beyond the end of the subject
2114 string were read.
2115
2116 11. Renamed the variable "class" as "classbits" to make life easier for C++
2117 users. (Previously there was a macro definition, but it apparently wasn't
2118 enough.)
2119
2120 12. Added the new field "tables" to the extra data so that tables can be passed
2121 in at exec time, or the internal tables can be re-selected. This allows
2122 a compiled regex to be saved and re-used at a later time by a different
2123 program that might have everything at different addresses.
2124
2125 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2126 -R library as well as a -L library.
2127
2128 14. The debugging options of pcretest (-d on the command line or D on a
2129 pattern) showed incorrect output for anything following an extended class
2130 that contained multibyte characters and which was followed by a quantifier.
2131
2132 15. Added optional support for general category Unicode character properties
2133 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2134 support. It adds about 90K to the size of the library. The meanings of the
2135 inbuilt class escapes such as \d and \s have NOT been changed.
2136
2137 16. Updated pcredemo.c to include calls to free() to release the memory for the
2138 compiled pattern.
2139
2140 17. The generated file chartables.c was being created in the source directory
2141 instead of in the building directory. This caused the build to fail if the
2142 source directory was different from the building directory, and was
2143 read-only.
2144
2145 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2146 file. No doubt somebody will tell me if they don't make sense... Also added
2147 Dan Mooney's comments about building on OpenVMS.
2148
2149 19. Added support for partial matching via the PCRE_PARTIAL option for
2150 pcre_exec() and the \P data escape in pcretest.
2151
2152 20. Extended pcretest with 3 new pattern features:
2153
2154 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2155 write the compiled pattern to the file whose name is "rest-of-line".
2156 This is a straight binary dump of the data, with the saved pointer to
2157 the character tables forced to be NULL. The study data, if any, is
2158 written too. After writing, pcretest reads a new pattern.
2159
2160 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2161 compiled pattern from the given file. There must not be any
2162 occurrences of "<" in the file name (pretty unlikely); if there are,
2163 pcretest will instead treat the initial "<" as a pattern delimiter.
2164 After reading in the pattern, pcretest goes on to read data lines as
2165 usual.
2166
2167 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2168 and 16-bit fields in a compiled pattern, to simulate a pattern that
2169 was compiled on a host of opposite endianness.
2170
2171 21. The pcre-exec() function can now cope with patterns that were compiled on
2172 hosts of opposite endianness, with this restriction:
2173
2174 As for any compiled expression that is saved and used later, the tables
2175 pointer field cannot be preserved; the extra_data field in the arguments
2176 to pcre_exec() should be used to pass in a tables address if a value
2177 other than the default internal tables were used at compile time.
2178
2179 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2180 now diagnosed as an error. Previously, most of the time, a negative number
2181 would have been treated as zero, but if in addition "ovector" was passed as
2182 NULL, a crash could occur.
2183
2184 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2185 new versions from the libtool 1.5 distribution (the last one is a copy of
2186 a file called libtool.m4). This seems to have fixed the need to patch
2187 "configure" to support Darwin 1.3 (which I used to do). However, I still
2188 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2189 workstation).
2190
2191 24. Changed the PCRE licence to be the more standard "BSD" licence.
2192
2193
2194 Version 4.5 01-Dec-03
2195 ---------------------
2196
2197 1. There has been some re-arrangement of the code for the match() function so
2198 that it can be compiled in a version that does not call itself recursively.
2199 Instead, it keeps those local variables that need separate instances for
2200 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2201 needs to "recurse". Keeping track of where control must go is done by means
2202 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2203 hide most of the details from the main code, and operates only if
2204 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2205 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2206 operating.
2207
2208 To make it easier for callers to provide specially tailored get/free
2209 functions for this usage, two new functions, pcre_stack_malloc, and
2210 pcre_stack_free, are used. They are always called in strict stacking order,
2211 and the size of block requested is always the same.
2212
2213 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2214 PCRE has been compiled to use the stack or the heap for recursion. The
2215 -C option of pcretest uses this to show which version is compiled.
2216
2217 A new data escape \S, is added to pcretest; it causes the amounts of store
2218 obtained and freed by both kinds of malloc/free at match time to be added
2219 to the output.
2220
2221 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2222 what's available on my current Linux desktop machine.
2223
2224 3. When matching a UTF-8 string, the test for a valid string at the start has
2225 been extended. If start_offset is not zero, PCRE now checks that it points
2226 to a byte that is the start of a UTF-8 character. If not, it returns
2227 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2228 this is necessary because there may be backward assertions in the pattern.
2229 When matching the same subject several times, it may save resources to use
2230 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2231
2232 4. The code for checking the validity of UTF-8 strings has been tightened so
2233 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2234 containing "overlong sequences".
2235
2236 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2237 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2238 should have been "&", but it just so happened that all the cases this let
2239 through by mistake were picked up later in the function.
2240
2241 6. I had used a variable called "isblank" - this is a C99 function, causing
2242 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2243
2244 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2245 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2246 so the version is shown; (c) stop "make test" echoing "./RunTest".
2247
2248 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2249
2250 9. The prototype for memmove() for systems that don't have it was using
2251 size_t, but the inclusion of the header that defines size_t was later. I've
2252 moved the #includes for the C headers earlier to avoid this.
2253
2254 10. Added some adjustments to the code to make it easier to compiler on certain
2255 special systems:
2256
2257 (a) Some "const" qualifiers were missing.
2258 (b) Added the macro EXPORT before all exported functions; by default this
2259 is defined to be empty.
2260 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2261 that it reads its output file name as an argument instead of writing
2262 to the standard output and assuming this can be redirected.
2263
2264 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2265 class containing characters with values greater than 255, PCRE compilation
2266 went into a loop.
2267
2268 12. A recursive reference to a subpattern that was within another subpattern
2269 that had a minimum quantifier of zero caused PCRE to crash. For example,
2270 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2271 recursion. If the recursively-called subpattern itself had a zero repeat,
2272 that was OK.
2273
2274 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2275 buffer into which it was copied (for escape processing) was still set at
2276 1024, so long lines caused crashes.
2277
2278 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2279 "internal error: code overflow...". This applied to any character class
2280 that was followed by a possessive quantifier.
2281
2282 15. Modified the Makefile to add libpcre.la as a prerequisite for
2283 libpcreposix.la because I was told this is needed for a parallel build to
2284 work.
2285
2286 16. If a pattern that contained .* following optional items at the start was
2287 studied, the wrong optimizing data was generated, leading to matching
2288 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2289 matching string must start with a or b or c. The correct conclusion for
2290 this pattern is that a match can start with any character.
2291
2292
2293 Version 4.4 13-Aug-03
2294 ---------------------
2295
2296 1. In UTF-8 mode, a character class containing characters with values between
2297 127 and 255 was not handled correctly if the compiled pattern was studied.
2298 In fixing this, I have also improved the studying algorithm for such
2299 classes (slightly).
2300
2301 2. Three internal functions had redundant arguments passed to them. Removal
2302 might give a very teeny performance improvement.
2303
2304 3. Documentation bug: the value of the capture_top field in a callout is *one
2305 more than* the number of the hightest numbered captured substring.
2306
2307 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2308 in incorrectly linking with a previously installed version. They now link
2309 explicitly with libpcre.la.
2310
2311 5. configure.in no longer needs to recognize Cygwin specially.
2312
2313 6. A problem in pcre.in for Windows platforms is fixed.
2314
2315 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2316 pcretest, it used to include the size of the study block as part of its
2317 output. Unfortunately, the structure contains a field that has a different
2318 size on different hardware architectures. This meant that the tests that
2319 showed this size failed. As the block is currently always of a fixed size,
2320 this information isn't actually particularly useful in pcretest output, so
2321 I have just removed it.
2322
2323 8. Three pre-processor statements accidentally did not start in column 1.
2324 Sadly, there are *still* compilers around that complain, even though
2325 standard C has not required this for well over a decade. Sigh.
2326
2327 9. In pcretest, the code for checking callouts passed small integers in the
2328 callout_data field, which is a void * field. However, some picky compilers
2329 complained about the casts involved for this on 64-bit systems. Now
2330 pcretest passes the address of the small integer instead, which should get
2331 rid of the warnings.
2332
2333 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2334 both compile and run time, and gives an error if an invalid UTF-8 sequence
2335 is found. There is a option for disabling this check in cases where the
2336 string is known to be correct and/or the maximum performance is wanted.
2337
2338 11. In response to a bug report, I changed one line in Makefile.in from
2339
2340 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2341 to
2342 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2343
2344 to look similar to other lines, but I have no way of telling whether this
2345 is the right thing to do, as I do not use Windows. No doubt I'll get told
2346 if it's wrong...
2347
2348
2349 Version 4.3 21-May-03
2350 ---------------------
2351
2352 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2353 Makefile.
2354
2355 2. Some refactoring to improve the quality of the code:
2356
2357 (i) The utf8_table... variables are now declared "const".
2358
2359 (ii) The code for \cx, which used the "case flipping" table to upper case
2360 lower case letters, now just substracts 32. This is ASCII-specific,
2361 but the whole concept of \cx is ASCII-specific, so it seems
2362 reasonable.
2363
2364 (iii) PCRE was using its character types table to recognize decimal and
2365 hexadecimal digits in the pattern. This is silly, because it handles
2366 only 0-9, a-f, and A-F, but the character types table is locale-
2367 specific, which means strange things might happen. A private
2368 table is now used for this - though it costs 256 bytes, a table is
2369 much faster than multiple explicit tests. Of course, the standard
2370 character types table is still used for matching digits in subject
2371 strings against \d.
2372
2373 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2374 ending in _t are). So I've renamed it as ESC_tee.
2375
2376 3. The first argument for regexec() in the POSIX wrapper should have been
2377 defined as "const".
2378
2379 4. Changed pcretest to use malloc() for its buffers so that they can be
2380 Electric Fenced for debugging.
2381
2382 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2383 to read one or more bytes before the start of the subject string. Often this
2384 had no effect on PCRE's behaviour, but in some circumstances it could
2385 provoke a segmentation fault.
2386
2387 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2388 to try to read one or more bytes before the start of the subject string.
2389
2390 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2391 UTF-8 support could misbehave in various ways if the subject string
2392 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2393 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2394 back over UTF-8 characters.)
2395
2396
2397 Version 4.2 14-Apr-03
2398 ---------------------
2399
2400 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2401
2402 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2403 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2404 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2405 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2406 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2407 and BUILD_EXEEXT
2408 Note: automatic setting of the BUILD variables is not yet working
2409 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2410 compile-time but not at link-time
2411 [LINK]: use for linking executables only
2412 make different versions for Windows and non-Windows
2413 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2414 libraries
2415 [LINK_FOR_BUILD]: new variable
2416 [OBJEXT]: use throughout
2417 [EXEEXT]: use throughout
2418 <winshared>: new target
2419 <wininstall>: new target
2420 <dftables.o>: use native compiler
2421 <dftables>: use native linker
2422 <install>: handle Windows platform correctly
2423 <clean>: ditto
2424 <check>: ditto
2425 copy DLL to top builddir before testing
2426
2427 As part of these changes, -no-undefined was removed again. This was reported
2428 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2429 in any case.
2430
2431 3. Some tidies to get rid of compiler warnings:
2432
2433 . In the match_data structure, match_limit was an unsigned long int, whereas
2434 match_call_count was an int. I've made them both unsigned long ints.
2435
2436 . In pcretest the fact that a const uschar * doesn't automatically cast to
2437 a void * provoked a warning.
2438
2439 . Turning on some more compiler warnings threw up some "shadow" variables
2440 and a few more missing casts.
2441
2442 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2443 option, a class that contained a single character with a value between 128
2444 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2445
2446 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2447 option, a class that contained several characters, but with at least one
2448 whose value was between 128 and 255 caused PCRE to crash.
2449
2450
2451 Version 4.1 12-Mar-03
2452 ---------------------
2453
2454 1. Compiling with gcc -pedantic found a couple of places where casts were
2455 needed, and a string in dftables.c that was longer than standard compilers are
2456 required to support.
2457
2458 2. Compiling with Sun's compiler found a few more places where the code could
2459 be tidied up in order to avoid warnings.
2460
2461 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2462 first of these names is deprecated in the latest Autoconf in favour of the name
2463 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2464 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2465 analogy I have changed it to CFLAGS_FOR_BUILD.
2466
2467 4. Added -no-undefined to the linking command in the Makefile, because this is
2468 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2469 linking step for the pcreposix library.
2470
2471 5. PCRE was failing to diagnose the case of two named groups with the same
2472 name.
2473
2474 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2475 literal character that is needed in the subject for a match, and scans along to
2476 ensure that it is present before embarking on the full matching process. This
2477 saves time in cases of nested unlimited repeats that are never going to match.
2478 Problem: the scan can take a lot of time if the subject is very long (e.g.
2479 megabytes), thus penalizing straightforward matches. It is now done only if the
2480 amount of subject to be scanned is less than 1000 bytes.
2481
2482 7. A lesser problem with the same optimization is that it was recording the
2483 first character of an anchored pattern as "needed", thus provoking a search
2484 right along the subject, even when the first match of the pattern was going to
2485 fail. The "needed" character is now not set for anchored patterns, unless it
2486 follows something in the pattern that is of non-fixed length. Thus, it still
2487 fulfils its original purpose of finding quick non-matches in cases of nested
2488 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2489
2490
2491 Version 4.0 17-Feb-03
2492 ---------------------
2493
2494 1. If a comment in an extended regex that started immediately after a meta-item
2495 extended to the end of string, PCRE compiled incorrect data. This could lead to
2496 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2497
2498 2. Moved to autoconf 2.53 and libtool 1.4.2.
2499
2500 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2501 the special perltest8 script is no longer needed - all the tests can be run
2502 from a single perltest script.
2503
2504 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2505 by \s. It has now been removed in PCRE. This means it isn't recognized as
2506 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2507 class [:space:] *does* include VT, thereby creating a mess.
2508
2509 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2510 space and tab.
2511
2512 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2513 its new features into the main test script, reducing the number of scripts.
2514
2515 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2516 were backward compatible, and made the (?i) apply to the whole pattern, as if
2517 /i were given. Now it behaves more logically, and applies the option setting
2518 only to what follows. PCRE has been changed to follow suit. However, if it
2519 finds options settings right at the start of the pattern, it extracts them into
2520 the global options, as before. Thus, they show up in the info data.
2521
2522 8. Added support for the \Q...\E escape sequence. Characters in between are
2523 treated as literals. This is slightly different from Perl in that $ and @ are
2524 also handled as literals inside the quotes. In Perl, they will cause variable
2525 interpolation. Note the following examples:
2526
2527 Pattern PCRE matches Perl matches
2528
2529 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2530 \Qabc\$xyz\E abc\$xyz abc\$xyz
2531 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2532
2533 For compatibility with Perl, \Q...\E sequences are recognized inside character
2534 classes as well as outside them.
2535
2536 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2537 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2538 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2539 signed/unsigned warnings.
2540
2541 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2542 option for pcretest, so I've replaced it by a simple function that does just
2543 that job.
2544
2545 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2546 "pcregrep -".
2547
2548 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2549 Java package. This provides some syntactic sugar for simple cases of what my
2550 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2551 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2552 item, you can use this simplified notation. Note that only makes sense with
2553 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2554 greediness, whatever the setting of the PCRE_UNGREEDY option.
2555
2556 13. A change of greediness default within a pattern was not taking effect at
2557 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2558 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2559 was abstracted outside.
2560
2561 14. PCRE now supports the \G assertion. It is true when the current matching
2562 position is at the start point of the match. This differs from \A when the
2563 starting offset is non-zero. Used with the /g option of pcretest (or similar
2564 code), it works in the same way as it does for Perl's /g option. If all
2565 alternatives of a regex begin with \G, the expression is anchored to the start
2566 match position, and the "anchored" flag is set in the compiled expression.
2567
2568 15. Some bugs concerning the handling of certain option changes within patterns
2569 have been fixed. These applied to options other than (?ims). For example,
2570 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2571 been the other way round. Some of this was related to change 7 above.
2572
2573 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2574 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2575 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2576 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2577
2578 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2579 mode. Unlike ".", it always matches newline, whatever the setting of
2580 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2581 assertions. Perl allows it, but it doesn't (in general) work because it can't
2582 calculate the length of the lookbehind. At least, that's the case for Perl
2583 5.8.0 - I've been told they are going to document that it doesn't work in
2584 future.
2585
2586 18. Added an error diagnosis for escapes that PCRE does not support: these are
2587 \L, \l, \N, \P, \p, \U, \u, and \X.
2588
2589 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2590 reading past the end of the pattern in cases such as /[abcd/.
2591
2592 20. PCRE was getting more memory than necessary for patterns with classes that
2593 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2594
2595 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2596 compiling PCRE for use with Virtual Pascal.
2597
2598 22. Small fix to the Makefile to make it work properly if the build is done
2599 outside the source tree.
2600
2601 23. Added a new extension: a condition to go with recursion. If a conditional
2602 subpattern starts with (?(R) the "true" branch is used if recursion has
2603 happened, whereas the "false" branch is used only at the top level.
2604
2605 24. When there was a very long string of literal characters (over 255 bytes
2606 without UTF support, over 250 bytes with UTF support), the computation of how
2607 much memory was required could be incorrect, leading to segfaults or other
2608 strange effects.
2609
2610 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2611 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2612 there was a subsequent back reference to those brackets. This meant that, for
2613 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2614 possible to check for precisely this case. All we can do is abandon the
2615 optimization if .* occurs inside capturing brackets when there are any back
2616 references whatsoever. (See below for a better fix that came later.)
2617
2618 26. The handling of the optimization for finding the first character of a
2619 non-anchored pattern, and for finding a character that is required later in the
2620 match were failing in some cases. This didn't break the matching; it just
2621 failed to optimize when it could. The way this is done has been re-implemented.
2622
2623 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2624
2625 28. Added a new feature that provides some of the functionality that Perl
2626 provides with (?{...}). The facility is termed a "callout". The way it is done
2627 in PCRE is for the caller to provide an optional function, by setting
2628 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2629 global variable. By default it is unset, which disables all calling out. To get
2630 the function called, the regex must include (?C) at appropriate points. This
2631 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2632 This provides a means of identifying different callout points. When PCRE
2633 reaches such a point in the regex, if pcre_callout has been set, the external
2634 function is called. It is provided with data in a structure called
2635 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2636 matching continues; if it returns a non-zero value, the match at the current
2637 point fails. However, backtracking will occur if possible. [This was changed
2638 later and other features added - see item 49 below.]
2639
2640 29. pcretest is upgraded to test the callout functionality. It provides a
2641 callout function that displays information. By default, it shows the start of
2642 the match and the current position in the text. There are some new data escapes
2643 to vary what happens:
2644
2645 \C+ in addition, show current contents of captured substrings
2646 \C- do not supply a callout function
2647 \C!n return 1 when callout number n is reached
2648 \C!n!m return 1 when callout number n is reached for the mth time
2649
2650 30. If pcregrep was called with the -l option and just a single file name, it
2651 output "<stdin>" if a match was found, instead of the file name.
2652
2653 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2654 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2655 pcre_exec(). This saves a malloc/free per call. The default value of
2656 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2657 when configuring.
2658
2659 32. The default maximum size of a compiled pattern is 64K. There have been a
2660 few cases of people hitting this limit. The code now uses macros to handle the
2661 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2662 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2663 configuring. Tests 2 and 5 work only with 2-byte links because they output
2664 debugging information about compiled patterns.
2665
2666 33. Internal code re-arrangements:
2667
2668 (a) Moved the debugging function for printing out a compiled regex into
2669 its own source file (printint.c) and used #include to pull it into
2670 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2671 separate copies.
2672
2673 (b) Defined the list of op-code names for debugging as a macro in
2674 internal.h so that it is next to the definition of the opcodes.
2675
2676 (c) Defined a table of op-code lengths for simpler skipping along compiled
2677 code. This is again a macro in internal.h so that it is next to the
2678 definition of the opcodes.
2679
2680 34. Added support for recursive calls to individual subpatterns, along the
2681 lines of Robin Houston's patch (but implemented somewhat differently).
2682
2683 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2684 allow it to read and process whole directories in Win32. This code was
2685 contributed by Lionel Fourquaux; it has not been tested by me.
2686
2687 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2688 used to name a group. Names consist of alphanumerics and underscores, and must
2689 be unique. Back references use the syntax (?P=name) and recursive calls use
2690 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2691 numbers. The function pcre_fullinfo() can be used after compilation to extract
2692 a name/number map. There are three relevant calls:
2693
2694 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2695 PCRE_INFO_NAMECOUNT yields the number of entries
2696 PCRE_INFO_NAMETABLE yields a pointer to the map.
2697
2698 The map is a vector of fixed-size entries. The size of each entry depends on
2699 the length of the longest name used. The first two bytes of each entry are the
2700 group number, most significant byte first. There follows the corresponding
2701 name, zero terminated. The names are in alphabetical order.
2702
2703 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2704 case instead of 255. Making it the same both with and without UTF-8 support
2705 means that the same test output works with both.
2706
2707 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2708 calling malloc() with a zero argument.
2709
2710 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2711 optimization. I've improved things by keeping a bitmap of backreferences with
2712 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2713 fact referenced, the optimization can be applied. It is unlikely that a
2714 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2715 the match to follow \n) will appear inside brackets with a number greater than
2716 31, but if it does, any back reference > 31 suppresses the optimization.
2717
2718 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2719 of disabling numbered capturing parentheses. Any opening parenthesis that is
2720 not followed by ? behaves as if it were followed by ?: but named parentheses
2721 can still be used for capturing (and they will acquire numbers in the usual
2722 way).
2723
2724 41. Redesigned the return codes from the match() function into yes/no/error so
2725 that errors can be passed back from deep inside the nested calls. A malloc
2726 failure while inside a recursive subpattern call now causes the
2727 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2728
2729 42. It is now possible to set a limit on the number of times the match()
2730 function is called in a call to pcre_exec(). This facility makes it possible to
2731 limit the amount of recursion and backtracking, though not in a directly
2732 obvious way, because the match() function is used in a number of different
2733 circumstances. The count starts from zero for each position in the subject
2734 string (for non-anchored patterns). The default limit is, for compatibility, a
2735 large number, namely 10 000 000. You can change this in two ways:
2736
2737 (a) When configuring PCRE before making, you can use --with-match-limit=n
2738 to set a default value for the compiled library.
2739
2740 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2741 a different value is set. See 45 below.
2742
2743 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2744
2745 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2746 of things that can be changed at compile time. The first argument specifies
2747 what is wanted and the second points to where the information is to be placed.
2748 The current list of available information is:
2749
2750 PCRE_CONFIG_UTF8
2751
2752 The output is an integer that is set to one if UTF-8 support is available;
2753 otherwise it is set to zero.
2754
2755 PCRE_CONFIG_NEWLINE
2756
2757 The output is an integer that it set to the value of the code that is used for
2758 newline. It is either LF (10) or CR (13).
2759
2760 PCRE_CONFIG_LINK_SIZE
2761
2762 The output is an integer that contains the number of bytes used for internal
2763 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2764
2765 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2766
2767 The output is an integer that contains the threshold above which the POSIX
2768 interface uses malloc() for output vectors. See item 31 above.
2769
2770 PCRE_CONFIG_MATCH_LIMIT
2771
2772 The output is an unsigned integer that contains the default limit of the number
2773 of match() calls in a pcre_exec() execution. See 42 above.
2774
2775 44. pcretest has been upgraded by the addition of the -C option. This causes it
2776 to extract all the available output from the new pcre_config() function, and to
2777 output it. The program then exits immediately.
2778
2779 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2780 order to support additional features. One way would have been to define
2781 pcre_exec2() (for example) with extra arguments, but this would not have been
2782 extensible, and would also have required all calls to the original function to
2783 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2784 is used for passing in "extra" data from pcre_study().
2785
2786 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2787 contains the following fields:
2788
2789 flags a bitmap indicating which of the following fields are set
2790 study_data opaque data from pcre_study()
2791 match_limit a way of specifying a limit on match() calls for a specific
2792 call to pcre_exec()
2793 callout_data data for callouts (see 49 below)
2794
2795 The flag bits are also defined in pcre.h, and are
2796
2797 PCRE_EXTRA_STUDY_DATA
2798 PCRE_EXTRA_MATCH_LIMIT
2799 PCRE_EXTRA_CALLOUT_DATA
2800
2801 The pcre_study() function now returns one of these new pcre_extra blocks, with
2802 the actual study data pointed to by the study_data field, and the
2803 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2804 before. That is, this change is entirely upwards-compatible and requires no
2805 change to existing code.
2806
2807 If you want to pass in additional data to pcre_exec(), you can either place it
2808 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2809 block.
2810
2811 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2812 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2813 times with different match limits, until it finds the minimum value needed for
2814 pcre_exec() to complete. The value is then output. This can be instructive; for
2815 most simple matches the number is quite small, but for pathological cases it
2816 gets very large very quickly.
2817
2818 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2819 returns the size of the data block pointed to by the study_data field in a
2820 pcre_extra block, that is, the value that was passed as the argument to
2821 pcre_malloc() when PCRE was getting memory in which to place the information
2822 created by pcre_study(). The fourth argument should point to a size_t variable.
2823 pcretest has been extended so that this information is shown after a successful
2824 pcre_study() call when information about the compiled regex is being displayed.
2825
2826 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2827 because what follows is always an absolute path. (Later: it turns out that this
2828 is more than cosmetic for MinGW, because it doesn't like empty path
2829 components.)
2830
2831 49. Some changes have been made to the callout feature (see 28 above):
2832
2833 (i) A callout function now has three choices for what it returns:
2834
2835 0 => success, carry on matching
2836 > 0 => failure at this point, but backtrack if possible
2837 < 0 => serious error, return this value from pcre_exec()
2838
2839 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2840 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2841 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2842 use by callout functions. It will never be used by PCRE itself.
2843
2844 (ii) The pcre_extra structure (see 45 above) has a void * field called
2845 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2846 pcre_callout_block structure has a field of the same name. The contents of
2847 the field passed in the pcre_extra structure are passed to the callout
2848 function in the corresponding field in the callout block. This makes it
2849 easier to use the same callout-containing regex from multiple threads. For
2850 testing, the pcretest program has a new data escape
2851
2852 \C*n pass the number n (may be negative) as callout_data
2853
2854 If the callout function in pcretest receives a non-zero value as
2855 callout_data, it returns that value.
2856
2857 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2858 there were some redundant $(CFLAGS) in commands that are now specified as
2859 $(LINK), which already includes $(CFLAGS).
2860
2861 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2862 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2863 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2864 one-byte characters throughout. Note that case-insensitive matching applies
2865 only to characters whose values are less than 256. PCRE doesn't support the
2866 notion of cases for higher-valued characters.
2867
2868 (i) A character class whose characters are all within 0-255 is handled as
2869 a bit map, and the map is inverted for negative classes. Previously, a
2870 character > 255 always failed to match such a class; however it should
2871 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2872
2873 (ii) A negated character class with a single character < 255 is coded as
2874 "not this character" (OP_NOT). This wasn't working properly when the test
2875 character was multibyte, either singly or repeated.
2876
2877 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2878 mode, for example: \x{100}{2,3}.
2879
2880 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2881 singly or repeated) now correctly test multibyte characters. However,
2882 PCRE doesn't recognize any characters with values greater than 255 as
2883 digits, spaces, or word characters. Such characters always match \D, \S,
2884 and \W, and never match \d, \s, or \w.
2885
2886 (v) Classes may now contain characters and character ranges with values
2887 greater than 255. For example: [ab\x{100}-\x{400}].
2888
2889 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2890 PCRE in UTF-8 mode.
2891
2892 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2893 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2894 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2895 value.)
2896
2897 53. The single man page has become too large. I have therefore split it up into
2898 a number of separate man pages. These also give rise to individual HTML pages;
2899 these are now put in a separate directory, and there is an index.html page that
2900 lists them all. Some hyperlinking between the pages has been installed.
2901
2902 54. Added convenience functions for handling named capturing parentheses.
2903
2904 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2905 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2906 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2907 are faulted.
2908
2909 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2910 calling configure. These values are used when compiling the dftables.c program
2911 which is run to generate the source of the default character tables. They
2912 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2913 you will need to set these values.
2914
2915 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2916
2917
2918 Version 3.9 02-Jan-02
2919 ---------------------
2920
2921 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2922
2923 2. If --disable-static was given, the building process failed when trying to
2924 build pcretest and pcregrep. (For some reason it was using libtool to compile
2925 them, which is not right, as they aren't part of the library.)
2926
2927
2928 Version 3.8 18-Dec-01
2929 ---------------------
2930
2931 1. The experimental UTF-8 code was completely screwed up. It was packing the
2932 bytes in the wrong order. How dumb can you get?
2933
2934
2935 Version 3.7 29-Oct-01
2936 ---------------------
2937
2938 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2939 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2940 this didn't happen under Solaris 8, where I normally test things.
2941
2942 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2943 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2944 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2945 this fix an hour or so after the initial 3.7 release.)
2946
2947
2948 Version 3.6 23-Oct-01
2949 ---------------------
2950
2951 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2952 offsets passed as NULL with zero offset count.
2953
2954 2. The config.guess and config.sub files had not been updated when I moved to
2955 the latest autoconf.
2956
2957
2958 Version 3.5 15-Aug-01
2959 ---------------------
2960
2961 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2962 had been forgotten.
2963
2964 2. By using declared but undefined structures, we can avoid using "void"
2965 definitions in pcre.h while keeping the internal definitions of the structures
2966 private.
2967
2968 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2969 user point of view, this means that both static and shared libraries are built
2970 by default, but this can be individually controlled. More of the work of
2971 handling this static/shared cases is now inside libtool instead of PCRE's make
2972 file.
2973
2974 4. The pcretest utility is now installed along with pcregrep because it is
2975 useful for users (to test regexs) and by doing this, it automatically gets
2976 relinked by libtool. The documentation has been turned into a man page, so
2977 there are now .1, .txt, and .html versions in /doc.
2978
2979 5. Upgrades to pcregrep:
2980 (i) Added long-form option names like gnu grep.
2981 (ii) Added --help to list all options with an explanatory phrase.
2982 (iii) Added -r, --recursive to recurse into sub-directories.
2983 (iv) Added -f, --file to read patterns from a file.
2984
2985 6. pcre_exec() was referring to its "code" argument before testing that
2986 argument for NULL (and giving an error if it was NULL).
2987
2988 7. Upgraded Makefile.in to allow for compiling in a different directory from
2989 the source directory.
2990
2991 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2992 options bits, the pointer it was passed was to an int instead of to an unsigned
2993 long int. This mattered only on 64-bit systems.
2994
2995 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2996 generated) instead of pcre.in, which it its source. Also made the same change
2997 in several of the .c files.
2998
2999 10. A new release of gcc defines printf() as a macro, which broke pcretest
3000 because it had an ifdef in the middle of a string argument for printf(). Fixed
3001 by using separate calls to printf().
3002
3003 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3004 script, to force use of CR or LF instead of \n in the source. On non-Unix
3005 systems, the value can be set in config.h.
3006
3007 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3008 absolute limit. Changed the text of the error message to make this clear, and
3009 likewise updated the man page.
3010
3011 13. The limit of 99 on the number of capturing subpatterns has been removed.
3012 The new limit is 65535, which I hope will not be a "real" limit.
3013
3014
3015 Version 3.4 22-Aug-00
3016 ---------------------
3017
3018 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3019
3020 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3021
3022
3023 Version 3.3 01-Aug-00
3024 ---------------------
3025
3026 1. If an octal character was given, but the value was greater than \377, it
3027 was not getting masked to the least significant bits, as documented. This could
3028 lead to crashes in some systems.
3029
3030 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3031 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3032
3033 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3034 These just pass their arguments on to (pcre_free)(), but they are provided
3035 because some uses of PCRE bind it to non-C systems that can call its functions,
3036 but cannot call free() or pcre_free() directly.
3037
3038 4. Add "make test" as a synonym for "make check". Corrected some comments in
3039 the Makefile.
3040
3041 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3042 Makefile.
3043
3044 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3045 command called pgrep for grepping around the active processes.
3046
3047 7. Added the beginnings of support for UTF-8 character strings.
3048
3049 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3050 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3051 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3052 out for the ar command.)
3053
3054
3055 Version 3.2 12-May-00
3056 ---------------------
3057
3058 This is purely a bug fixing release.
3059
3060 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3061 of ZA. This was just one example of several cases that could provoke this bug,
3062 which was introduced by change 9 of version 2.00. The code for breaking
3063 infinite loops after an iteration that matches an empty string was't working
3064 correctly.
3065
3066 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3067 when matched against abbab (for example). After matching an empty string, it
3068 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3069 caused it to match further down the string than it should.
3070
3071 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3072 was there because it doesn't seem to be needed, and it causes trouble on some
3073 systems, as it is not a Standard C header. It has been removed.
3074
3075 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3076 were reported on the Macintosh. The changes were from
3077
3078 while ((c = *(++ptr)) != 0 && c != '\n');
3079 to
3080 while ((c = *(++ptr)) != 0 && c != '\n') ;
3081
3082 Totally extraordinary, but if that's what it takes...
3083
3084 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3085 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3086 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3087 assumes the way PCRE uses memmove() (always moving upwards).
3088
3089 6. PCRE is being used in one environment where strchr() is not available. There
3090 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3091 faster code anyway.
3092
3093
3094 Version 3.1 09-Feb-00
3095 ---------------------
3096
3097 The only change in this release is the fixing of some bugs in Makefile.in for
3098 the "install" target:
3099
3100 (1) It was failing to install pcreposix.h.
3101
3102 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3103
3104
3105 Version 3.0 01-Feb-00
3106 ---------------------
3107
3108 1. Add support for the /+ modifier to perltest (to output $` like it does in
3109 pcretest).
3110
3111 2. Add support for the /g modifier to perltest.
3112
3113 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3114 matches null strings.
3115
3116 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3117 pattern. Perl treats empty patterns specially - it reuses the most recent
3118 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3119 effect.
3120
3121 5. The POSIX interface was broken in that it was just handing over the POSIX
3122 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3123 required a bigger vector, with some working space on the end. This means that
3124 the POSIX wrapper now has to get and free some memory, and copy the results.
3125
3126 6. Added some simple autoconf support, placing the test data and the
3127 documentation in separate directories, re-organizing some of the
3128 information files, and making it build pcre-config (a GNU standard). Also added
3129 libtool support for building PCRE as a shared library, which is now the
3130 default.
3131
3132 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3133 09 are not valid octal constants. Single digits will be used for minor values
3134 less than 10.
3135
3136 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3137 existing programs that set these in the POSIX interface can use PCRE without
3138 modification.
3139
3140 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3141 return all that pcre_info() returns, plus additional data. The pcre_info()
3142 function is retained for compatibility, but is considered to be obsolete.
3143
3144 10. Added experimental recursion feature (?R) to handle one common case that
3145 Perl 5.6 will be able to do with (?p{...}).
3146
3147 11. Added support for POSIX character classes like [:alpha:], which Perl is
3148 adopting.
3149
3150
3151 Version 2.08 31-Aug-99
3152 ----------------------
3153
3154 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3155 trying to match at the startoffset position, but instead was moving forward to
3156 the next newline as if a previous match had failed.
3157
3158 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3159 and could get into a loop if a null string was matched other than at the start
3160 of the subject.
3161
3162 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3163 be distinguished at compile time, and for completeness also added PCRE_DATE.
3164
3165 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3166 in GnuWin32 environments.
3167
3168
3169 Version 2.07 29-Jul-99
3170 ----------------------
3171
3172 1. The documentation is now supplied in plain text form and HTML as well as in
3173 the form of man page sources.
3174
3175 2. C++ compilers don't like assigning (void *) values to other pointer types.
3176 In particular this affects malloc(). Although there is no problem in Standard
3177 C, I've put in casts to keep C++ compilers happy.
3178
3179 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3180 should be (const char *).
3181
3182 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3183 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3184 However, I haven't made this a standard facility. The documentation doesn't
3185 mention it, and the Makefile doesn't support it.
3186
3187 5. The Makefile now contains an "install" target, with editable destinations at
3188 the top of the file. The pcretest program is not installed.
3189
3190 6. pgrep -V now gives the PCRE version number and date.
3191
3192 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3193 causing the entire string to be ignored, instead of just the last character.
3194
3195 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3196 non-matching string, it can take a very, very long time, even for strings of
3197 quite modest length, because of the nested recursion. PCRE now does better in
3198 some of these cases. It does this by remembering the last required literal
3199 character in the pattern, and pre-searching the subject to ensure it is present
3200 before running the real match. In other words, it applies a heuristic to detect
3201 some types of certain failure quickly, and in the above example, if presented
3202 with a string that has no trailing " it gives "no match" very quickly.
3203
3204 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3205 other alternatives are tried instead.
3206
3207
3208 Version 2.06 09-Jun-99
3209 ----------------------
3210
3211 1. Change pcretest's output for amount of store used to show just the code
3212 space, because the remainder (the data block) varies in size between 32-bit and
3213 64-bit systems.
3214
3215 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3216 start matching at. This allows lookbehinds to work when searching for multiple
3217 occurrences in a string.
3218
3219 3. Added additional options to pcretest for testing multiple occurrences:
3220
3221 /+ outputs the rest of the string that follows a match
3222 /g loops for multiple occurrences, using the new startoffset argument
3223 /G loops for multiple occurrences by passing an incremented pointer
3224
3225 4. PCRE wasn't doing the "first character" optimization for patterns starting
3226 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3227 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3228 the letter 'x'. On long subject strings, this gives a significant speed-up.
3229
3230
3231 Version 2.05 21-Apr-99
3232 ----------------------
3233
3234 1. Changed the type of magic_number from int to long int so that it works
3235 properly on 16-bit systems.
3236
3237 2. Fixed a bug which caused patterns starting with .* not to work correctly
3238 when the subject string contained newline characters. PCRE was assuming
3239 anchoring for such patterns in all cases, which is not correct because .* will
3240 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3241 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3242 must be retried after every newline in the subject.
3243
3244
3245 Version 2.04 18-Feb-99
3246 ----------------------
3247
3248 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3249 computation of the store needed to hold the pattern was incorrect (too large).
3250 If such patterns were nested a few deep, this could multiply and become a real
3251 problem.
3252
3253 2. Added /M option to pcretest to show the memory requirement of a specific
3254 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3255
3256 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3257 compiled in such a way that the backtracking after subsequent failure was
3258 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3259 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3260
3261
3262 Version 2.03 02-Feb-99
3263 ----------------------
3264
3265 1. Fixed typo and small mistake in man page.
3266
3267 2. Added 4th condition (GPL supersedes if conflict) and created separate
3268 LICENCE file containing the conditions.
3269
3270 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3271 Perl, that is the internal \ allows the delimiter to be included in the
3272 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3273 the final delimiter, add \ to the end of the pattern (to test the error).
3274
3275 4. Added the convenience functions for extracting substrings after a successful
3276 match. Updated pcretest to make it able to test these functions.
3277
3278
3279 Version 2.02 14-Jan-99
3280 ----------------------
3281
3282 1. Initialized the working variables associated with each extraction so that
3283 their saving and restoring doesn't refer to uninitialized store.
3284
3285 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3286 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3287 fix the problem.
3288
3289 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3290 calls, and wasn't printing the correct value for compiling calls. Increased the
3291 default value of LOOPREPEAT, and the number of significant figures in the
3292 times.
3293
3294 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3295
3296 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3297 a building problem on Windows NT with a FAT file system.
3298
3299
3300 Version 2.01 21-Oct-98
3301 ----------------------
3302
3303 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3304 to character tables built by pcre_maketables() in the current locale. If NULL
3305 is passed, the default tables are used.
3306
3307
3308 Version 2.00 24-Sep-98
3309 ----------------------
3310
3311 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3312 it any more.
3313
3314 2. Allow quantification of (?>) groups, and make it work correctly.
3315
3316 3. The first character computation wasn't working for (?>) groups.
3317
3318 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3319 end of the subject) and add 5.005's \z, which really does match only at the
3320 very end of the subject.
3321
3322 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3323
3324 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3325 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3326 localized options. All options to pcre_study() were also removed.
3327
3328 7. Add other new features from 5.005:
3329
3330 $(?<= positive lookbehind
3331 $(?<! negative lookbehind
3332 (?imsx-imsx) added the unsetting capability
3333 such a setting is global if at outer level; local otherwise
3334 (?imsx-imsx:) non-capturing groups with option setting
3335 (?(cond)re|re) conditional pattern matching
3336
3337 A backreference to itself in a repeated group matches the previous
3338 captured string.
3339
3340 8. General tidying up of studying (both automatic and via "study")
3341 consequential on the addition of new assertions.
3342
3343 9. As in 5.005, unlimited repeated groups that could match an empty substring
3344 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3345 runtime if any iteration does actually match an empty substring.
3346
3347 10. Include the RunTest script in the distribution.
3348
3349 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3350 discrepancies, some of which were old and were also with respect to 5.004. They
3351 have now been fixed.
3352
3353
3354 Version 1.09 28-Apr-98
3355 ----------------------
3356
3357 1. A negated single character class followed by a quantifier with a minimum
3358 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3359 program crashes, or just wrong answers. This did not apply to negated classes
3360 containing more than one character, or to minima other than one.
3361
3362
3363 Version 1.08 27-Mar-98
3364 ----------------------
3365
3366 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3367
3368 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3369 latter must appear before anything that relies on it in the pattern.
3370
3371
3372 Version 1.07 16-Feb-98
3373 ----------------------
3374
3375 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3376 repeat of a potentially empty string).
3377
3378
3379 Version 1.06 23-Jan-98
3380 ----------------------
3381
3382 1. Added Markus Oberhumer's little patches for C++.
3383
3384 2. Literal strings longer than 255 characters were broken.
3385
3386
3387 Version 1.05 23-Dec-97
3388 ----------------------
3389
3390 1. Negated character classes containing more than one character were failing if
3391 PCRE_CASELESS was set at run time.
3392
3393
3394 Version 1.04 19-Dec-97
3395 ----------------------
3396
3397 1. Corrected the man page, where some "const" qualifiers had been omitted.
3398
3399 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3400 input syntax.
3401
3402 3. Fixed memory leak which occurred when a regex with back references was
3403 matched with an offsets vector that wasn't big enough. The temporary memory
3404 that is used in this case wasn't being freed if the match failed.
3405
3406 4. Tidied pcretest to ensure it frees memory that it gets.
3407
3408 5. Temporary memory was being obtained in the case where the passed offsets
3409 vector was exactly big enough.
3410
3411 6. Corrected definition of offsetof() from change 5 below.
3412
3413 7. I had screwed up change 6 below and broken the rules for the use of
3414 setjmp(). Now fixed.
3415
3416
3417 Version 1.03 18-Dec-97
3418 ----------------------
3419
3420 1. A erroneous regex with a missing opening parenthesis was correctly
3421 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3422 on some systems.
3423
3424 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3425 it was reported that one broken compiler failed on the former because "code" is
3426 also an independent variable.
3427
3428 3. The erroneous regex a[]b caused an array overrun reference.
3429
3430 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3431 fail on data ending with that character. (It was going on too far, and checking
3432 the next character, typically a binary zero.) This was specific to the
3433 optimized code for single-character negative classes.
3434
3435 5. Added a contributed patch from the TIN world which does the following:
3436
3437 + Add an undef for memmove, in case the the system defines a macro for it.
3438
3439 + Add a definition of offsetof(), in case there isn't one. (I don't know
3440 the reason behind this - offsetof() is part of the ANSI standard - but
3441 it does no harm).
3442
3443 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3444 most of the places where whitespace preceded '#'. I have given up and
3445 allowed the remaining 2 cases to be at the margin.
3446
3447 + Rename some variables in pcre to eliminate shadowing. This seems very
3448 pedantic, but does no harm, of course.
3449
3450 6. Moved the call to setjmp() into its own function, to get rid of warnings
3451 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3452
3453 7. Constructs such as \d{8,} were compiling into the equivalent of
3454 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3455 outcome, but in this particular case used more store than had been allocated,
3456 which caused the bug to be discovered because it threw up an internal error.
3457
3458 8. The debugging code in both pcre and pcretest for outputting the compiled
3459 form of a regex was going wrong in the case of back references followed by
3460 curly-bracketed repeats.
3461
3462
3463 Version 1.02 12-Dec-97
3464 ----------------------
3465
3466 1. Typos in pcre.3 and comments in the source fixed.
3467
3468 2. Applied a contributed patch to get rid of places where it used to remove
3469 'const' from variables, and fixed some signed/unsigned and uninitialized
3470 variable warnings.
3471
3472 3. Added the "runtest" target to Makefile.
3473
3474 4. Set default compiler flag to -O2 rather than just -O.
3475
3476
3477 Version 1.01 19-Nov-97
3478 ----------------------
3479
3480 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3481 like /([ab]*)*/, that is, for classes with more than one character in them.
3482
3483 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3484 as /((?>a*))*/ (a PCRE_EXTRA facility).
3485
3486
3487 Version 1.00 18-Nov-97
3488 ----------------------
3489
3490 1. Added compile-time macros to support systems such as SunOS4 which don't have
3491 memmove() or strerror() but have other things that can be used instead.
3492
3493 2. Arranged that "make clean" removes the executables.
3494
3495
3496 Version 0.99 27-Oct-97
3497 ----------------------
3498
3499 1. Fixed bug in code for optimizing classes with only one character. It was
3500 initializing a 32-byte map regardless, which could cause it to run off the end
3501 of the memory it had got.
3502
3503 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3504
3505
3506 Version 0.98 22-Oct-97
3507 ----------------------
3508
3509 1. Fixed bug in code for handling temporary memory usage when there are more
3510 back references than supplied space in the ovector. This could cause segfaults.
3511
3512
3513 Version 0.97 21-Oct-97
3514 ----------------------
3515
3516 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3517
3518 2. Optimized negated single characters not to use a bit map.
3519
3520 3. Brought error texts together as macro definitions; clarified some of them;
3521 fixed one that was wrong - it said "range out of order" when it meant "invalid
3522 escape sequence".
3523
3524 4. Changed some char * arguments to const char *.
3525
3526 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3527
3528 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3529 pcretest.
3530
3531
3532 Version 0.96 16-Oct-97
3533 ----------------------
3534
3535 1. Added a simple "pgrep" utility to the distribution.
3536
3537 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3538 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3539 where "ddd" means "one or more decimal digits".
3540
3541 3. Fixed serious bug. If a pattern had a back reference, but the call to
3542 pcre_exec() didn't supply a large enough ovector to record the related
3543 identifying subpattern, the match always failed. PCRE now remembers the number
3544 of the largest back reference, and gets some temporary memory in which to save
3545 the offsets during matching if necessary, in order to ensure that
3546 backreferences always work.
3547
3548 4. Increased the compatibility with Perl in a number of ways:
3549
3550 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3551 to request this handling. The option can be set at compile or exec time.
3552
3553 (b) $ matches before a terminating newline by default; an option
3554 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3555 mode). The option can be set at compile or exec time.
3556
3557 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3558 the same as Perl's. If the decimal number it represents is less than 10
3559 or there aren't that many previous left capturing parentheses, an octal
3560 escape is read. Inside a character class, it's always an octal escape,
3561 even if it is a single digit.
3562
3563 (d) An escaped but undefined alphabetic character is taken as a literal,
3564 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3565 escapes.
3566
3567 (e) {0} is now permitted. (The previous item is removed from the compiled
3568 pattern).
3569
3570 5. Changed all the names of code files so that the basic parts are no longer
3571 than 10 characters, and abolished the teeny "globals.c" file.
3572
3573 6. Changed the handling of character classes; they are now done with a 32-byte
3574 bit map always.
3575
3576 7. Added the -d and /D options to pcretest to make it possible to look at the
3577 internals of compilation without having to recompile pcre.
3578
3579
3580 Version 0.95 23-Sep-97
3581 ----------------------
3582
3583 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3584 \x20 at the start of a run of normal characters. These were being treated as
3585 real characters, instead of the source characters being re-checked.
3586
3587
3588 Version 0.94 18-Sep-97
3589 ----------------------
3590
3591 1. The functions are now thread-safe, with the caveat that the global variables
3592 containing pointers to malloc() and free() or alternative functions are the
3593 same for all threads.
3594
3595 2. Get pcre_study() to generate a bitmap of initial characters for non-
3596 anchored patterns when this is possible, and use it if passed to pcre_exec().
3597
3598
3599 Version 0.93 15-Sep-97
3600 ----------------------
3601
3602 1. /(b)|(:+)/ was computing an incorrect first character.
3603
3604 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3605 but not actually doing anything yet.
3606
3607 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3608 as Perl does (e.g. [-az] or [az-]).
3609
3610 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3611 all possible positions.
3612
3613 5. Split up into different modules to avoid including unneeded functions in a
3614 compiled binary. However, compile and exec are still in one module. The "study"
3615 function is split off.
3616
3617 6. The character tables are now in a separate module whose source is generated
3618 by an auxiliary program - but can then be edited by hand if required. There are
3619 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3620 toupper() in the code.
3621
3622 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3623 make them global. Abolish the function for setting them, as the caller can now
3624 set them directly.
3625
3626
3627 Version 0.92 11-Sep-97
3628 ----------------------
3629
3630 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3631 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3632
3633 2. Caseless matching was not working in character classes if the characters in
3634 the pattern were in upper case.
3635
3636 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3637
3638 4. Make PCRE_ANCHORED public and accept as a compile option.
3639
3640 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3641 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3642 pass them.
3643
3644 6. Give an error if bad option bits passed at compile or run time.
3645
3646 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3647 pcretest to cause it to pass that flag.
3648
3649 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3650 options, and the first character, if set.
3651
3652 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3653
3654
3655 Version 0.91 10-Sep-97
3656 ----------------------
3657
3658 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3659 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3660
3661 2. PCRE was looping on encountering an indefinitely repeated back reference to
3662 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3663 Perl does - treats the match as successful.
3664
3665 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5