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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 447 - (show annotations) (download)
Tue Sep 15 18:17:54 2009 UTC (5 years ago) by ph10
File size: 172757 byte(s)
Capture data when (*ACCEPT) is inside capturing parentheses.

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8 was out-of-date, and there was no check on the pcre_compile() error code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

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