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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 474 - (show annotations) (download)
Sat Jan 2 16:30:46 2010 UTC (4 years, 3 months ago) by ph10
File size: 176290 byte(s)
Fix C++ wrapper GlobalReplace function for empty matches.

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39 was out-of-date, and there was no check on the pcre_compile() error code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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12