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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 454 - (show annotations) (download)
Tue Sep 22 09:42:11 2009 UTC (4 years, 10 months ago) by ph10
File size: 173098 byte(s)
Allow fixed-length subroutine calls in lookbehinds.

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8 was out-of-date, and there was no check on the pcre_compile() error code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

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