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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 442 - (show annotations) (download)
Fri Sep 11 10:21:02 2009 UTC (4 years, 7 months ago) by ph10
File size: 172245 byte(s)
Added PCRE_NOTEMPTY_ATSTART to fix /g bug when \K is present.

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8 was out-of-date, and there was no check on the pcre_compile() error code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

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