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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 472 - (show annotations) (download)
Fri Dec 11 16:42:50 2009 UTC (4 years, 7 months ago) by ph10
File size: 175497 byte(s)
Fix options set and reset at top level bug.

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24 was out-of-date, and there was no check on the pcre_compile() error code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

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