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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 479 - (show annotations) (download)
Sun Jan 3 16:31:39 2010 UTC (4 years, 11 months ago) by ph10
File size: 178440 byte(s)
Add AC_CONFIG_MACRO_DIR([m4]) to configure.ac and -I m4 to ACLOCAL_AMFLAGS in 
Makefile.am, as suggested by libtoolize.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.01 11-Dec-09
5 ----------------------
6
7 1. If a pattern contained a conditional subpattern with only one branch (in
8 particular, this includes all (DEFINE) patterns), a call to pcre_study()
9 computed the wrong minimum data length (which is of course zero for such
10 subpatterns).
11
12 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
13 the pattern is reset in the first branch, pcre_compile() failed with
14 "internal error: code overflow at offset...". This happened only when
15 the reset was to the original external option setting. (An optimization
16 abstracts leading options settings into an external setting, which was the
17 cause of this.)
18
19 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
20 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
21 assertion pattern did not match (meaning that the assertion was true), it
22 was incorrectly treated as false if the SKIP had been reached during the
23 matching. This also applied to assertions used as conditions.
24
25 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
26 assertion subpattern, including such a pattern used as a condition,
27 unpredictable results occurred, instead of the error return
28 PCRE_ERROR_DFA_UITEM.
29
30 5. The C++ GlobalReplace function was not working like Perl for the special
31 situation when an empty string is matched. It now does the fancy magic
32 stuff that is necessary.
33
34 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
35 removed. (These were left over from very, very early versions of PCRE.)
36
37 7. Some cosmetic changes to the code to make life easier when compiling it
38 as part of something else:
39
40 (a) Change DEBUG to PCRE_DEBUG.
41
42 (b) In pcre_compile(), rename the member of the "branch_chain" structure
43 called "current" as "current_branch", to prevent a collision with the
44 Linux macro when compiled as a kernel module.
45
46 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
47 prevent a collision with the Linux macro when compiled as a kernel
48 module.
49
50 8. In pcre_compile() there are some checks for integer overflows that used to
51 cast potentially large values to (double). This has been changed to that
52 when building, a check for int64_t is made, and if it is found, it is used
53 instead, thus avoiding the use of floating point arithmetic. (There is no
54 other use of FP in PCRE.) If int64_t is not found, the fallback is to
55 double.
56
57 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
58 2005 (difference between two addresses compared to an unsigned value).
59
60 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
61 custom one, because of the following reported problem in Windows:
62
63 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
64 under Win32.
65 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
66 therefore missing the function definition.
67 - The compiler thus generates a "C" signature for the test function.
68 - The linker fails to find the "C" function.
69 - PCRE fails to configure if asked to do so against libbz2.
70
71 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
72 messages were output:
73
74 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
75 rerunning libtoolize, to keep the correct libtool macros in-tree.
76 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
77
78 I have done both of these things.
79
80
81
82 Version 8.00 19-Oct-09
83 ----------------------
84
85 1. The table for translating pcre_compile() error codes into POSIX error codes
86 was out-of-date, and there was no check on the pcre_compile() error code
87 being within the table. This could lead to an OK return being given in
88 error.
89
90 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
91 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
92 in a Windows environment.
93
94 3. The pcregrep --count option prints the count for each file even when it is
95 zero, as does GNU grep. However, pcregrep was also printing all files when
96 --files-with-matches was added. Now, when both options are given, it prints
97 counts only for those files that have at least one match. (GNU grep just
98 prints the file name in this circumstance, but including the count seems
99 more useful - otherwise, why use --count?) Also ensured that the
100 combination -clh just lists non-zero counts, with no names.
101
102 4. The long form of the pcregrep -F option was incorrectly implemented as
103 --fixed_strings instead of --fixed-strings. This is an incompatible change,
104 but it seems right to fix it, and I didn't think it was worth preserving
105 the old behaviour.
106
107 5. The command line items --regex=pattern and --regexp=pattern were not
108 recognized by pcregrep, which required --regex pattern or --regexp pattern
109 (with a space rather than an '='). The man page documented the '=' forms,
110 which are compatible with GNU grep; these now work.
111
112 6. No libpcreposix.pc file was created for pkg-config; there was just
113 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
114
115 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
116 when UCP support is not needed, by modifying the Python script that
117 generates it from Unicode data files. This should not matter if the module
118 is correctly used as a library, but I received one complaint about 50K of
119 unwanted data. My guess is that the person linked everything into his
120 program rather than using a library. Anyway, it does no harm.
121
122 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
123 was a minimum greater than 1 for a wide character in a possessive
124 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
125 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
126 character. Chaos in the form of incorrect output or a compiling loop could
127 result.
128
129 9. The restrictions on what a pattern can contain when partial matching is
130 requested for pcre_exec() have been removed. All patterns can now be
131 partially matched by this function. In addition, if there are at least two
132 slots in the offset vector, the offset of the earliest inspected character
133 for the match and the offset of the end of the subject are set in them when
134 PCRE_ERROR_PARTIAL is returned.
135
136 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
137 synonymous with PCRE_PARTIAL, for backwards compatibility, and
138 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
139 and may be more useful for multi-segment matching.
140
141 11. Partial matching with pcre_exec() is now more intuitive. A partial match
142 used to be given if ever the end of the subject was reached; now it is
143 given only if matching could not proceed because another character was
144 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
145 string "Z", which now yields "no match" instead of "partial match". In the
146 case of pcre_dfa_exec(), "no match" is given if every matching path for the
147 final character ended with (*FAIL).
148
149 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
150 if the pattern had a "must contain" character that was already found in the
151 earlier partial match, unless partial matching was again requested. For
152 example, with the pattern /dog.(body)?/, the "must contain" character is
153 "g". If the first part-match was for the string "dog", restarting with
154 "sbody" failed. This bug has been fixed.
155
156 13. The string returned by pcre_dfa_exec() after a partial match has been
157 changed so that it starts at the first inspected character rather than the
158 first character of the match. This makes a difference only if the pattern
159 starts with a lookbehind assertion or \b or \B (\K is not supported by
160 pcre_dfa_exec()). It's an incompatible change, but it makes the two
161 matching functions compatible, and I think it's the right thing to do.
162
163 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
164 so that the demonstration program is easily available in environments where
165 PCRE has not been installed from source.
166
167 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
168 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
169 library.
170
171 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
172 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
173 is not the first non-POSIX option to be added. Clearly some people find
174 these options useful.
175
176 17. If a caller to the POSIX matching function regexec() passes a non-zero
177 value for nmatch with a NULL value for pmatch, the value of
178 nmatch is forced to zero.
179
180 18. RunGrepTest did not have a test for the availability of the -u option of
181 the diff command, as RunTest does. It now checks in the same way as
182 RunTest, and also checks for the -b option.
183
184 19. If an odd number of negated classes containing just a single character
185 interposed, within parentheses, between a forward reference to a named
186 subpattern and the definition of the subpattern, compilation crashed with
187 an internal error, complaining that it could not find the referenced
188 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
189 [The bug was that it was starting one character too far in when skipping
190 over the character class, thus treating the ] as data rather than
191 terminating the class. This meant it could skip too much.]
192
193 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
194 /g option in pcretest when the pattern contains \K, which makes it possible
195 to have an empty string match not at the start, even when the pattern is
196 anchored. Updated pcretest and pcredemo to use this option.
197
198 21. If the maximum number of capturing subpatterns in a recursion was greater
199 than the maximum at the outer level, the higher number was returned, but
200 with unset values at the outer level. The correct (outer level) value is
201 now given.
202
203 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
204 PCRE did not set those parentheses (unlike Perl). I have now found a way to
205 make it do so. The string so far is captured, making this feature
206 compatible with Perl.
207
208 23. The tests have been re-organized, adding tests 11 and 12, to make it
209 possible to check the Perl 5.10 features against Perl 5.10.
210
211 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
212 pattern matches a fixed length string. PCRE did not allow this; now it
213 does. Neither allows recursion.
214
215 25. I finally figured out how to implement a request to provide the minimum
216 length of subject string that was needed in order to match a given pattern.
217 (It was back references and recursion that I had previously got hung up
218 on.) This code has now been added to pcre_study(); it finds a lower bound
219 to the length of subject needed. It is not necessarily the greatest lower
220 bound, but using it to avoid searching strings that are too short does give
221 some useful speed-ups. The value is available to calling programs via
222 pcre_fullinfo().
223
224 26. While implementing 25, I discovered to my embarrassment that pcretest had
225 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
226 study optimizations had never been tested with that matching function.
227 Oops. What is worse, even when it was passed study data, there was a bug in
228 pcre_dfa_exec() that meant it never actually used it. Double oops. There
229 were also very few tests of studied patterns with pcre_dfa_exec().
230
231 27. If (?| is used to create subpatterns with duplicate numbers, they are now
232 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
233 on the other side of the coin, they are no longer allowed to have different
234 names, because these cannot be distinguished in PCRE, and this has caused
235 confusion. (This is a difference from Perl.)
236
237 28. When duplicate subpattern names are present (necessarily with different
238 numbers, as required by 27 above), and a test is made by name in a
239 conditional pattern, either for a subpattern having been matched, or for
240 recursion in such a pattern, all the associated numbered subpatterns are
241 tested, and the overall condition is true if the condition is true for any
242 one of them. This is the way Perl works, and is also more like the way
243 testing by number works.
244
245
246 Version 7.9 11-Apr-09
247 ---------------------
248
249 1. When building with support for bzlib/zlib (pcregrep) and/or readline
250 (pcretest), all targets were linked against these libraries. This included
251 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
252 libraries. This caused unwanted dependencies to be created. This problem
253 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
254 pcretest is linked with readline.
255
256 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
257 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
258 moved outside it again, because FALSE and TRUE are already defined in AIX,
259 but BOOL is not.
260
261 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
262 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
263
264 4. The pcregrep documentation said spaces were inserted as well as colons (or
265 hyphens) following file names and line numbers when outputting matching
266 lines. This is not true; no spaces are inserted. I have also clarified the
267 wording for the --colour (or --color) option.
268
269 5. In pcregrep, when --colour was used with -o, the list of matching strings
270 was not coloured; this is different to GNU grep, so I have changed it to be
271 the same.
272
273 6. When --colo(u)r was used in pcregrep, only the first matching substring in
274 each matching line was coloured. Now it goes on to look for further matches
275 of any of the test patterns, which is the same behaviour as GNU grep.
276
277 7. A pattern that could match an empty string could cause pcregrep to loop; it
278 doesn't make sense to accept an empty string match in pcregrep, so I have
279 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
280 seems to be how GNU grep behaves.
281
282 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
283 start or after a newline", because the conditional assertion was not being
284 correctly handled. The rule now is that both the assertion and what follows
285 in the first alternative must satisfy the test.
286
287 9. If auto-callout was enabled in a pattern with a conditional group whose
288 condition was an assertion, PCRE could crash during matching, both with
289 pcre_exec() and pcre_dfa_exec().
290
291 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
292 used for matching.
293
294 11. Unicode property support in character classes was not working for
295 characters (bytes) greater than 127 when not in UTF-8 mode.
296
297 12. Added the -M command line option to pcretest.
298
299 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
300
301 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
302
303 16. Added comments and documentation about mis-use of no_arg in the C++
304 wrapper.
305
306 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
307 from Martin Jerabek that uses macro names for all relevant character and
308 string constants.
309
310 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
311 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
312 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
313 these, but not everybody uses configure.
314
315 19. A conditional group that had only one branch was not being correctly
316 recognized as an item that could match an empty string. This meant that an
317 enclosing group might also not be so recognized, causing infinite looping
318 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
319 with the subject "ab", where knowledge that the repeated group can match
320 nothing is needed in order to break the loop.
321
322 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
323 exec(), but without supplying a callout function, matching went wrong.
324
325 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
326 leak if the size of the offset vector was greater than 30. When the vector
327 is smaller, the saved offsets during recursion go onto a local stack
328 vector, but for larger vectors malloc() is used. It was failing to free
329 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
330 error, in fact).
331
332 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
333 heapframe that is used only when UTF-8 support is enabled. This caused no
334 problem, but was untidy.
335
336 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
337 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
338 included within another project.
339
340 24. Steven Van Ingelgem's patches to add more options to the CMake support,
341 slightly modified by me:
342
343 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
344 not building pcregrep.
345
346 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
347 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
348
349 25. Forward references, both numeric and by name, in patterns that made use of
350 duplicate group numbers, could behave incorrectly or give incorrect errors,
351 because when scanning forward to find the reference group, PCRE was not
352 taking into account the duplicate group numbers. A pattern such as
353 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
354
355 26. Changed a few more instances of "const unsigned char *" to USPTR, making
356 the feature of a custom pointer more persuasive (as requested by a user).
357
358 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
359 pcretest.c, inside #ifndefs, because it seems they are sometimes already
360 pre-defined.
361
362 28. Added support for (*UTF8) at the start of a pattern.
363
364 29. Arrange for flags added by the "release type" setting in CMake to be shown
365 in the configuration summary.
366
367
368 Version 7.8 05-Sep-08
369 ---------------------
370
371 1. Replaced UCP searching code with optimized version as implemented for Ad
372 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
373 stage table and inline lookup instead of a function, giving speed ups of 2
374 to 5 times on some simple patterns that I tested. Permission was given to
375 distribute the MultiStage2.py script that generates the tables (it's not in
376 the tarball, but is in the Subversion repository).
377
378 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
379 scripts.
380
381 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
382 a group with a zero qualifier. The result of the study could be incorrect,
383 or the function might crash, depending on the pattern.
384
385 4. Caseless matching was not working for non-ASCII characters in back
386 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
387 It now works when Unicode Property Support is available.
388
389 5. In pcretest, an escape such as \x{de} in the data was always generating
390 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
391 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
392 truncation.
393
394 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
395
396 7. Added two (int) casts to pcregrep when printing the difference of two
397 pointers, in case they are 64-bit values.
398
399 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
400 test 2 if it fails.
401
402 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
403 and a #define of that name to empty if it is not externally set. This is to
404 allow users of MSVC to set it if necessary.
405
406 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
407 the convenience functions in the pcre_get.c source file.
408
409 11. An option change at the start of a pattern that had top-level alternatives
410 could cause overwriting and/or a crash. This command provoked a crash in
411 some environments:
412
413 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
414
415 This potential security problem was recorded as CVE-2008-2371.
416
417 12. For a pattern where the match had to start at the beginning or immediately
418 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
419 pcre_dfa_exec() could read past the end of the passed subject if there was
420 no match. To help with detecting such bugs (e.g. with valgrind), I modified
421 pcretest so that it places the subject at the end of its malloc-ed buffer.
422
423 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
424 exec() might read past the end of the data buffer in UTF-8 mode.
425
426 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
427 the data contained the byte 0x85 as part of a UTF-8 character within its
428 first line. This applied both to normal and DFA matching.
429
430 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
431 /^[^d]*?$/8 failed to match "abc".
432
433 16. Added a missing copyright notice to pcrecpp_internal.h.
434
435 17. Make it more clear in the documentation that values returned from
436 pcre_exec() in ovector are byte offsets, not character counts.
437
438 18. Tidied a few places to stop certain compilers from issuing warnings.
439
440 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
441 supplied by Stefan Weber. I made a further small update for 7.8 because
442 there is a change of source arrangements: the pcre_searchfuncs.c module is
443 replaced by pcre_ucd.c.
444
445
446 Version 7.7 07-May-08
447 ---------------------
448
449 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
450 a string to a long long, pretend we don't even have a long long." This is
451 done by checking for the strtoq, strtoll, and _strtoi64 functions.
452
453 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
454 pre-7.6 versions, which defined a global no_arg variable instead of putting
455 it in the RE class. (See also #8 below.)
456
457 3. Remove a line of dead code, identified by coverity and reported by Nuno
458 Lopes.
459
460 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
461
462 (1) The include/exclude patterns were being applied to the whole pathnames
463 of files, instead of just to the final components.
464
465 (2) If there was more than one level of directory, the subdirectories were
466 skipped unless they satisfied the include/exclude conditions. This is
467 inconsistent with GNU grep (and could even be seen as contrary to the
468 pcregrep specification - which I improved to make it absolutely clear).
469 The action now is always to scan all levels of directory, and just
470 apply the include/exclude patterns to regular files.
471
472 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
473 --exclude_dir in the tests to avoid scanning .svn directories.
474
475 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
476 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
477 doesn't support NULs in patterns.
478
479 7. Added some missing "const"s to declarations of static tables in
480 pcre_compile.c and pcre_dfa_exec.c.
481
482 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
483 caused by fix #2 above. (Subsequently also a second patch to fix the
484 first patch. And a third patch - this was a messy problem.)
485
486 9. Applied Craig's patch to remove the use of push_back().
487
488 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
489 matching function regexec().
490
491 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
492 which, however, unlike Perl's \g{...}, are subroutine calls, not back
493 references. PCRE supports relative numbers with this syntax (I don't think
494 Oniguruma does).
495
496 12. Previously, a group with a zero repeat such as (...){0} was completely
497 omitted from the compiled regex. However, this means that if the group
498 was called as a subroutine from elsewhere in the pattern, things went wrong
499 (an internal error was given). Such groups are now left in the compiled
500 pattern, with a new opcode that causes them to be skipped at execution
501 time.
502
503 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
504 to the way PCRE behaves:
505
506 (a) A lone ] character is dis-allowed (Perl treats it as data).
507
508 (b) A back reference to an unmatched subpattern matches an empty string
509 (Perl fails the current match path).
510
511 (c) A data ] in a character class must be notated as \] because if the
512 first data character in a class is ], it defines an empty class. (In
513 Perl it is not possible to have an empty class.) The empty class []
514 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
515 The negative empty class [^] matches any one character, independently
516 of the DOTALL setting.
517
518 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
519 non-existent subpattern following a character class starting with ']' and
520 containing () gave an internal compiling error instead of "reference to
521 non-existent subpattern". Fortunately, when the pattern did exist, the
522 compiled code was correct. (When scanning forwards to check for the
523 existencd of the subpattern, it was treating the data ']' as terminating
524 the class, so got the count wrong. When actually compiling, the reference
525 was subsequently set up correctly.)
526
527 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
528 it was being rejected as not supported by pcre_dfa_exec(), even though
529 other assertions are supported. I have made pcre_dfa_exec() support
530 (*FAIL).
531
532 16. The implementation of 13c above involved the invention of a new opcode,
533 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
534 cannot be changed at match time, I realized I could make a small
535 improvement to matching performance by compiling OP_ALLANY instead of
536 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
537 on the OP_ANY path.
538
539 17. Compiling pcretest on Windows with readline support failed without the
540 following two fixes: (1) Make the unistd.h include conditional on
541 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
542
543 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
544 ncurses library to be included for pcretest when ReadLine support is
545 requested, but also to allow for it to be overridden. This patch came from
546 Daniel Bergström.
547
548 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
549 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
550 any errors with the current Unicode tables. Thanks to Peter Kankowski for
551 spotting this.
552
553
554 Version 7.6 28-Jan-08
555 ---------------------
556
557 1. A character class containing a very large number of characters with
558 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
559 overflow.
560
561 2. Patch to cut out the "long long" test in pcrecpp_unittest when
562 HAVE_LONG_LONG is not defined.
563
564 3. Applied Christian Ehrlicher's patch to update the CMake build files to
565 bring them up to date and include new features. This patch includes:
566
567 - Fixed PH's badly added libz and libbz2 support.
568 - Fixed a problem with static linking.
569 - Added pcredemo. [But later removed - see 7 below.]
570 - Fixed dftables problem and added an option.
571 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
572 HAVE_LONG_LONG.
573 - Added readline support for pcretest.
574 - Added an listing of the option settings after cmake has run.
575
576 4. A user submitted a patch to Makefile that makes it easy to create
577 "pcre.dll" under mingw when using Configure/Make. I added stuff to
578 Makefile.am that cause it to include this special target, without
579 affecting anything else. Note that the same mingw target plus all
580 the other distribution libraries and programs are now supported
581 when configuring with CMake (see 6 below) instead of with
582 Configure/Make.
583
584 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
585 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
586 exported in the Windows port". It has not yet been confirmed that the patch
587 solves the problem, but it does no harm.
588
589 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
590 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
591 with CMake, and also correct the comment about stack recursion.
592
593 7. Remove the automatic building of pcredemo from the ./configure system and
594 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
595 of a program that users should build themselves after PCRE is installed, so
596 building it automatically is not really right. What is more, it gave
597 trouble in some build environments.
598
599 8. Further tidies to CMakeLists.txt from Sheri and Christian.
600
601
602 Version 7.5 10-Jan-08
603 ---------------------
604
605 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
606 values in parens when parsing an RE using the C++ wrapper."
607
608 2. Negative specials like \S did not work in character classes in UTF-8 mode.
609 Characters greater than 255 were excluded from the class instead of being
610 included.
611
612 3. The same bug as (2) above applied to negated POSIX classes such as
613 [:^space:].
614
615 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
616 defined or documented. It seems to have been a typo for PCRE_STATIC, so
617 I have changed it.
618
619 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
620 first named subpattern) and a construct such as (?&a) would reference the
621 first named subpattern whose name started with "a" (in other words, the
622 length check was missing). Both these problems are fixed. "Subpattern name
623 expected" is now given for (?&) (a zero-length name), and this patch also
624 makes it give the same error for \k'' (previously it complained that that
625 was a reference to a non-existent subpattern).
626
627 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
628 this is right because (?- can be followed by option settings as well as by
629 digits. I have, however, made the messages clearer.
630
631 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
632 than the number used in the conditional) now cause a compile-time error.
633 This is actually not compatible with Perl, which accepts such patterns, but
634 treats the conditional as always being FALSE (as PCRE used to), but it
635 seems to me that giving a diagnostic is better.
636
637 8. Change "alphameric" to the more common word "alphanumeric" in comments
638 and messages.
639
640 9. Fix two occurrences of "backslash" in comments that should have been
641 "backspace".
642
643 10. Remove two redundant lines of code that can never be obeyed (their function
644 was moved elsewhere).
645
646 11. The program that makes PCRE's Unicode character property table had a bug
647 which caused it to generate incorrect table entries for sequences of
648 characters that have the same character type, but are in different scripts.
649 It amalgamated them into a single range, with the script of the first of
650 them. In other words, some characters were in the wrong script. There were
651 thirteen such cases, affecting characters in the following ranges:
652
653 U+002b0 - U+002c1
654 U+0060c - U+0060d
655 U+0061e - U+00612
656 U+0064b - U+0065e
657 U+0074d - U+0076d
658 U+01800 - U+01805
659 U+01d00 - U+01d77
660 U+01d9b - U+01dbf
661 U+0200b - U+0200f
662 U+030fc - U+030fe
663 U+03260 - U+0327f
664 U+0fb46 - U+0fbb1
665 U+10450 - U+1049d
666
667 12. The -o option (show only the matching part of a line) for pcregrep was not
668 compatible with GNU grep in that, if there was more than one match in a
669 line, it showed only the first of them. It now behaves in the same way as
670 GNU grep.
671
672 13. If the -o and -v options were combined for pcregrep, it printed a blank
673 line for every non-matching line. GNU grep prints nothing, and pcregrep now
674 does the same. The return code can be used to tell if there were any
675 non-matching lines.
676
677 14. Added --file-offsets and --line-offsets to pcregrep.
678
679 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
680 infinitely looping recursion. The bug was that positive lookaheads were not
681 being skipped when checking for a possible empty match (negative lookaheads
682 and both kinds of lookbehind were skipped).
683
684 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
685 inclusion of <windows.h> to before rather than after the definition of
686 INVALID_FILE_ATTRIBUTES (patch from David Byron).
687
688 17. Specifying a possessive quantifier with a specific limit for a Unicode
689 character property caused pcre_compile() to compile bad code, which led at
690 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
691 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
692 caused the error; without that there was no problem.
693
694 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
695
696 19. Added --enable-pcretest-libreadline.
697
698 20. In pcrecpp.cc, the variable 'count' was incremented twice in
699 RE::GlobalReplace(). As a result, the number of replacements returned was
700 double what it should be. I removed one of the increments, but Craig sent a
701 later patch that removed the other one (the right fix) and added unit tests
702 that check the return values (which was not done before).
703
704 21. Several CMake things:
705
706 (1) Arranged that, when cmake is used on Unix, the libraries end up with
707 the names libpcre and libpcreposix, not just pcre and pcreposix.
708
709 (2) The above change means that pcretest and pcregrep are now correctly
710 linked with the newly-built libraries, not previously installed ones.
711
712 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
713
714 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
715 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
716 UTF-8 newline character). The key issue is that the pattern starts .*;
717 this means that the match must be either at the beginning, or after a
718 newline. The bug was in the code for advancing after a failed match and
719 checking that the new position followed a newline. It was not taking
720 account of UTF-8 characters correctly.
721
722 23. PCRE was behaving differently from Perl in the way it recognized POSIX
723 character classes. PCRE was not treating the sequence [:...:] as a
724 character class unless the ... were all letters. Perl, however, seems to
725 allow any characters between [: and :], though of course it rejects as
726 unknown any "names" that contain non-letters, because all the known class
727 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
728 for example, whereas PCRE did not - it did not recognize a POSIX character
729 class. This seemed a bit dangerous, so the code has been changed to be
730 closer to Perl. The behaviour is not identical to Perl, because PCRE will
731 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
732 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
733 Perl does, and where it didn't before.
734
735 24. Rewrite so as to remove the single use of %n from pcregrep because in some
736 Windows environments %n is disabled by default.
737
738
739 Version 7.4 21-Sep-07
740 ---------------------
741
742 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
743 means that a class such as [\s] counted as "explicit reference to CR or
744 LF". That isn't really right - the whole point of the change was to try to
745 help when there was an actual mention of one of the two characters. So now
746 the change happens only if \r or \n (or a literal CR or LF) character is
747 encountered.
748
749 2. The 32-bit options word was also used for 6 internal flags, but the numbers
750 of both had grown to the point where there were only 3 bits left.
751 Fortunately, there was spare space in the data structure, and so I have
752 moved the internal flags into a new 16-bit field to free up more option
753 bits.
754
755 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
756 but did not set the internal JCHANGED flag - either of these is enough to
757 control the way the "get" function works - but the PCRE_INFO_JCHANGED
758 facility is supposed to tell if (?J) was ever used, so now (?J) at the
759 start sets both bits.
760
761 4. Added options (at build time, compile time, exec time) to change \R from
762 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
763
764 5. doc/pcresyntax.html was missing from the distribution.
765
766 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
767 compatibility, even though it is no longer used.
768
769 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
770 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
771 windows.h file is present (where different names are used). [This was
772 reversed later after testing - see 16 below.]
773
774 8. Changed all #include <config.h> to #include "config.h". There were also
775 some further <pcre.h> cases that I changed to "pcre.h".
776
777 9. When pcregrep was used with the --colour option, it missed the line ending
778 sequence off the lines that it output.
779
780 10. It was pointed out to me that arrays of string pointers cause lots of
781 relocations when a shared library is dynamically loaded. A technique of
782 using a single long string with a table of offsets can drastically reduce
783 these. I have refactored PCRE in four places to do this. The result is
784 dramatic:
785
786 Originally: 290
787 After changing UCP table: 187
788 After changing error message table: 43
789 After changing table of "verbs" 36
790 After changing table of Posix names 22
791
792 Thanks to the folks working on Gregex for glib for this insight.
793
794 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
795 unicode-properties was also set.
796
797 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
798
799 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
800 checked only for CRLF.
801
802 14. Added casts to pcretest.c to avoid compiler warnings.
803
804 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
805
806 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
807 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
808 entirely. This removes changes made in 7 above.
809
810 17. The CMake files have been updated, and there is now more information about
811 building with CMake in the NON-UNIX-USE document.
812
813
814 Version 7.3 28-Aug-07
815 ---------------------
816
817 1. In the rejigging of the build system that eventually resulted in 7.1, the
818 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
819 brackets there is not right, since it causes compilers to look for an
820 installed pcre.h, not the version that is in the source that is being
821 compiled (which of course may be different). I have changed it back to:
822
823 #include "pcre.h"
824
825 I have a vague recollection that the change was concerned with compiling in
826 different directories, but in the new build system, that is taken care of
827 by the VPATH setting the Makefile.
828
829 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
830 when the subject happened to end in the byte 0x85 (e.g. if the last
831 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
832 characters but of course it shouldn't be taken as a newline when it is part
833 of another character. The bug was that, for an unlimited repeat of . in
834 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
835 characters when looking for a newline.
836
837 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
838
839 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
840 in debug output.
841
842 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
843 long printing in the pcrecpp unittest when running under MinGW.
844
845 6. ESC_K was left out of the EBCDIC table.
846
847 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
848 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
849 limit also applies to "virtual nesting" when a pattern is recursive, and in
850 this case 1000 isn't so big. I have been able to remove this limit at the
851 expense of backing off one optimization in certain circumstances. Normally,
852 when pcre_exec() would call its internal match() function recursively and
853 immediately return the result unconditionally, it uses a "tail recursion"
854 feature to save stack. However, when a subpattern that can match an empty
855 string has an unlimited repetition quantifier, it no longer makes this
856 optimization. That gives it a stack frame in which to save the data for
857 checking that an empty string has been matched. Previously this was taken
858 from the 1000-entry workspace that had been reserved. So now there is no
859 explicit limit, but more stack is used.
860
861 8. Applied Daniel's patches to solve problems with the import/export magic
862 syntax that is required for Windows, and which was going wrong for the
863 pcreposix and pcrecpp parts of the library. These were overlooked when this
864 problem was solved for the main library.
865
866 9. There were some crude static tests to avoid integer overflow when computing
867 the size of patterns that contain repeated groups with explicit upper
868 limits. As the maximum quantifier is 65535, the maximum group length was
869 set at 30,000 so that the product of these two numbers did not overflow a
870 32-bit integer. However, it turns out that people want to use groups that
871 are longer than 30,000 bytes (though not repeat them that many times).
872 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
873 made it possible to implement the integer overflow checks in a much more
874 dynamic way, which I have now done. The artificial limitation on group
875 length has been removed - we now have only the limit on the total length of
876 the compiled pattern, which depends on the LINK_SIZE setting.
877
878 10. Fixed a bug in the documentation for get/copy named substring when
879 duplicate names are permitted. If none of the named substrings are set, the
880 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
881 empty string.
882
883 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
884 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
885 because the ] is interpreted as the first data character and the
886 terminating ] is not found. PCRE has been made compatible with Perl in this
887 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
888 cause memory overwriting.
889
890 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
891 string has been matched (to stop an infinite loop). It was not recognizing
892 a conditional subpattern that could match an empty string if that
893 subpattern was within another subpattern. For example, it looped when
894 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
895 condition was not nested. This bug has been fixed.
896
897 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
898 past the start of the subject in the presence of bytes with the top bit
899 set, for example "\x8aBCD".
900
901 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
902 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
903
904 14. Optimized (?!) to (*FAIL).
905
906 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
907 This restricts code points to be within the range 0 to 0x10FFFF, excluding
908 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
909 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
910 does: it's just the validity check that is more restrictive.
911
912 16. Inserted checks for integer overflows during escape sequence (backslash)
913 processing, and also fixed erroneous offset values for syntax errors during
914 backslash processing.
915
916 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
917 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
918
919 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
920 caused an overrun.
921
922 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
923 something other than just ASCII characters) inside a group that had an
924 unlimited repeat caused a loop at compile time (while checking to see
925 whether the group could match an empty string).
926
927 20. Debugging a pattern containing \p or \P could cause a crash. For example,
928 [\P{Any}] did so. (Error in the code for printing property names.)
929
930 21. An orphan \E inside a character class could cause a crash.
931
932 22. A repeated capturing bracket such as (A)? could cause a wild memory
933 reference during compilation.
934
935 23. There are several functions in pcre_compile() that scan along a compiled
936 expression for various reasons (e.g. to see if it's fixed length for look
937 behind). There were bugs in these functions when a repeated \p or \P was
938 present in the pattern. These operators have additional parameters compared
939 with \d, etc, and these were not being taken into account when moving along
940 the compiled data. Specifically:
941
942 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
943 length.
944
945 (b) An item such as \pL+ within a repeated group could cause crashes or
946 loops.
947
948 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
949 "reference to non-existent subpattern" error.
950
951 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
952
953 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
954 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
955
956 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
957
958 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
959 character were causing crashes (broken optimization).
960
961 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
962 \p or \P) caused a compile-time loop.
963
964 28. More problems have arisen in unanchored patterns when CRLF is a valid line
965 break. For example, the unstudied pattern [\r\n]A does not match the string
966 "\r\nA" because change 7.0/46 below moves the current point on by two
967 characters after failing to match at the start. However, the pattern \nA
968 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
969 the same is true. There doesn't seem any very clean way out of this, but
970 what I have chosen to do makes the common cases work: PCRE now takes note
971 of whether there can be an explicit match for \r or \n anywhere in the
972 pattern, and if so, 7.0/46 no longer applies. As part of this change,
973 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
974 pattern has explicit CR or LF references.
975
976 29. Added (*CR) etc for changing newline setting at start of pattern.
977
978
979 Version 7.2 19-Jun-07
980 ---------------------
981
982 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
983 which is apparently normally available under Windows.
984
985 2. Re-jig the pcregrep tests with different newline settings in an attempt
986 to make them independent of the local environment's newline setting.
987
988 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
989
990 4. Some of the "internals" tests were previously cut out when the link size
991 was not 2, because the output contained actual offsets. The recent new
992 "Z" feature of pcretest means that these can be cut out, making the tests
993 usable with all link sizes.
994
995 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
996 stack recursion. This gives a massive performance boost under BSD, but just
997 a small improvement under Linux. However, it saves one field in the frame
998 in all cases.
999
1000 6. Added more features from the forthcoming Perl 5.10:
1001
1002 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1003 recursion call. It refers to the nth most recently opened parentheses.
1004
1005 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1006 to be opened parentheses.
1007
1008 (c) Conditions that refer to capturing parentheses can be specified
1009 relatively, for example, (?(-2)... or (?(+3)...
1010
1011 (d) \K resets the start of the current match so that everything before
1012 is not part of it.
1013
1014 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1015
1016 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1017 reference syntax.
1018
1019 (g) (?| introduces a group in which the numbering of parentheses in each
1020 alternative starts with the same number.
1021
1022 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1023
1024 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1025 PCRE_INFO_JCHANGED.
1026
1027 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1028 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1029 for detecting groups that can match an empty string.
1030
1031 9. A pattern with a very large number of alternatives (more than several
1032 hundred) was running out of internal workspace during the pre-compile
1033 phase, where pcre_compile() figures out how much memory will be needed. A
1034 bit of new cunning has reduced the workspace needed for groups with
1035 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1036 workspace instead of running out of the 4096 that are available.
1037
1038 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1039
1040 11. Applied patch from Google to remove an optimization that didn't quite work.
1041 The report of the bug said:
1042
1043 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1044 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1045 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1046
1047 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1048 it matched the wrong number of bytes.
1049
1050
1051 Version 7.1 24-Apr-07
1052 ---------------------
1053
1054 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1055 that is more "standard", making use of automake and other Autotools. There
1056 is some re-arrangement of the files and adjustment of comments consequent
1057 on this.
1058
1059 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1060 for recursive directory scanning broke on some systems because the files
1061 are not scanned in any specific order and on different systems the order
1062 was different. A call to "sort" has been inserted into RunGrepTest for the
1063 approprate test as a short-term fix. In the longer term there may be an
1064 alternative.
1065
1066 3. I had an email from Eric Raymond about problems translating some of PCRE's
1067 man pages to HTML (despite the fact that I distribute HTML pages, some
1068 people do their own conversions for various reasons). The problems
1069 concerned the use of low-level troff macros .br and .in. I have therefore
1070 removed all such uses from the man pages (some were redundant, some could
1071 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1072 HTML has been updated to handle .nf/.fi and to complain if it encounters
1073 .br or .in.
1074
1075 4. Updated comments in configure.ac that get placed in config.h.in and also
1076 arranged for config.h to be included in the distribution, with the name
1077 config.h.generic, for the benefit of those who have to compile without
1078 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1079
1080 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1081 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1082 makevp.bat for the current PCRE, using the additional files
1083 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1084
1085 6. A Windows user reported a minor discrepancy with test 2, which turned out
1086 to be caused by a trailing space on an input line that had got lost in his
1087 copy. The trailing space was an accident, so I've just removed it.
1088
1089 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1090 that is needed.
1091
1092 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1093 as "const" (a) because they are and (b) because it helps the PHP
1094 maintainers who have recently made a script to detect big data structures
1095 in the php code that should be moved to the .rodata section. I remembered
1096 to update Builducptable as well, so it won't revert if ucptable.h is ever
1097 re-created.
1098
1099 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1100 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1101 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1102 support is not required. This saves 1.5-2K of code, which is important in
1103 some applications.
1104
1105 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1106 so as not to refer to the tables, even though these functions will never be
1107 called when UTF-8 support is disabled. Otherwise there are problems with a
1108 shared library.
1109
1110 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1111
1112 (a) It was defining its arguments as char * instead of void *.
1113
1114 (b) It was assuming that all moves were upwards in memory; this was true
1115 a long time ago when I wrote it, but is no longer the case.
1116
1117 The emulated memove() is provided for those environments that have neither
1118 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1119 is clearly not the case, as these two bugs were recently reported.
1120
1121 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1122 and Detrail to create the HTML documentation, the .txt form of the man
1123 pages, and it removes trailing spaces from listed files. It also creates
1124 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1125 case, it wraps all the #defines with #ifndefs. This script should be run
1126 before "make dist".
1127
1128 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1129 with Unicode property support.
1130
1131 (a) For a maximizing quantifier, if the two different cases of the
1132 character were of different lengths in their UTF-8 codings (there are
1133 some cases like this - I found 11), and the matching function had to
1134 back up over a mixture of the two cases, it incorrectly assumed they
1135 were both the same length.
1136
1137 (b) When PCRE was configured to use the heap rather than the stack for
1138 recursion during matching, it was not correctly preserving the data for
1139 the other case of a UTF-8 character when checking ahead for a match
1140 while processing a minimizing repeat. If the check also involved
1141 matching a wide character, but failed, corruption could cause an
1142 erroneous result when trying to check for a repeat of the original
1143 character.
1144
1145 13. Some tidying changes to the testing mechanism:
1146
1147 (a) The RunTest script now detects the internal link size and whether there
1148 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1149 values substituted by "configure". (The RunGrepTest script already did
1150 this for UTF-8.) The configure.ac script no longer substitutes the
1151 relevant variables.
1152
1153 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1154 with length and offset values. This means that the output is different
1155 for different internal link sizes. Test 2 is skipped for link sizes
1156 other than 2 because of this, bypassing the problem. Unfortunately,
1157 there was also a test in test 3 (the locale tests) that used /B and
1158 failed for link sizes other than 2. Rather than cut the whole test out,
1159 I have added a new /Z option to pcretest that replaces the length and
1160 offset values with spaces. This is now used to make test 3 independent
1161 of link size. (Test 2 will be tidied up later.)
1162
1163 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1164 segmentation fault instead of returning the appropriate error message.
1165
1166 15. In multiline mode when the newline sequence was set to "any", the pattern
1167 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1168 This doesn't seem right; it now treats the CRLF combination as the line
1169 ending, and so does not match in that case. It's only a pattern such as ^$
1170 that would hit this one: something like ^ABC$ would have failed after \r
1171 and then tried again after \r\n.
1172
1173 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1174 in an attempt to make files that differ only in their line terminators
1175 compare equal. This works on Linux.
1176
1177 17. Under certain error circumstances pcregrep might try to free random memory
1178 as it exited. This is now fixed, thanks to valgrind.
1179
1180 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1181 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1182 was because its rules for how to advance for /g after matching an empty
1183 string at the end of a line did not allow for this case. They now check for
1184 it specially.
1185
1186 20. pcretest is supposed to handle patterns and data of any length, by
1187 extending its buffers when necessary. It was getting this wrong when the
1188 buffer for a data line had to be extended.
1189
1190 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1191 CRLF as a newline sequence.
1192
1193 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1194 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1195 I have nevertheless tidied it up.
1196
1197 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1198
1199 24. Added a man page for pcre-config.
1200
1201
1202 Version 7.0 19-Dec-06
1203 ---------------------
1204
1205 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1206 moving to gcc 4.1.1.
1207
1208 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1209 sys/time.h, which is documented as needed for this function. It doesn't
1210 seem to matter on Linux, but it showed up on some releases of OS X.
1211
1212 3. It seems that there are systems where bytes whose values are greater than
1213 127 match isprint() in the "C" locale. The "C" locale should be the
1214 default when a C program starts up. In most systems, only ASCII printing
1215 characters match isprint(). This difference caused the output from pcretest
1216 to vary, making some of the tests fail. I have changed pcretest so that:
1217
1218 (a) When it is outputting text in the compiled version of a pattern, bytes
1219 other than 32-126 are always shown as hex escapes.
1220
1221 (b) When it is outputting text that is a matched part of a subject string,
1222 it does the same, unless a different locale has been set for the match
1223 (using the /L modifier). In this case, it uses isprint() to decide.
1224
1225 4. Fixed a major bug that caused incorrect computation of the amount of memory
1226 required for a compiled pattern when options that changed within the
1227 pattern affected the logic of the preliminary scan that determines the
1228 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1229 that the computed length was too small. The symptoms of this bug were
1230 either the PCRE error "internal error: code overflow" from pcre_compile(),
1231 or a glibc crash with a message such as "pcretest: free(): invalid next
1232 size (fast)". Examples of patterns that provoked this bug (shown in
1233 pcretest format) are:
1234
1235 /(?-x: )/x
1236 /(?x)(?-x: \s*#\s*)/
1237 /((?i)[\x{c0}])/8
1238 /(?i:[\x{c0}])/8
1239
1240 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1241 is now done differently.
1242
1243 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1244 wrapper classes; (b) implement a new function in the C++ scanner that is
1245 more efficient than the old way of doing things because it avoids levels of
1246 recursion in the regex matching; (c) add a paragraph to the documentation
1247 for the FullMatch() function.
1248
1249 6. The escape sequence \n was being treated as whatever was defined as
1250 "newline". Not only was this contrary to the documentation, which states
1251 that \n is character 10 (hex 0A), but it also went horribly wrong when
1252 "newline" was defined as CRLF. This has been fixed.
1253
1254 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1255 was being set to -1 for the "end of line" case (supposedly a value that no
1256 character can have). Though this value is never used (the check for end of
1257 line is "zero bytes in current character"), it caused compiler complaints.
1258 I've changed it to 0xffffffff.
1259
1260 8. In pcre_version.c, the version string was being built by a sequence of
1261 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1262 string (as it is for production releases) called a macro with an empty
1263 argument. The C standard says the result of this is undefined. The gcc
1264 compiler treats it as an empty string (which was what was wanted) but it is
1265 reported that Visual C gives an error. The source has been hacked around to
1266 avoid this problem.
1267
1268 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1269 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1270 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1271 of them did).
1272
1273 10. Originally, pcretest opened its input and output without "b"; then I was
1274 told that "b" was needed in some environments, so it was added for release
1275 5.0 to both the input and output. (It makes no difference on Unix-like
1276 systems.) Later I was told that it is wrong for the input on Windows. I've
1277 now abstracted the modes into two macros, to make it easier to fiddle with
1278 them, and removed "b" from the input mode under Windows.
1279
1280 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1281
1282 12. Added -help and --help to pcretest as an official way of being reminded
1283 of the options.
1284
1285 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1286 and pcrecpp.cc because they annoy compilers at high warning levels.
1287
1288 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1289
1290 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1291 scripts are not C programs :-) and which was not noticed because it works
1292 on Linux.
1293
1294 16. pcretest is supposed to handle any length of pattern and data line (as one
1295 line or as a continued sequence of lines) by extending its input buffer if
1296 necessary. This feature was broken for very long pattern lines, leading to
1297 a string of junk being passed to pcre_compile() if the pattern was longer
1298 than about 50K.
1299
1300 17. I have done a major re-factoring of the way pcre_compile() computes the
1301 amount of memory needed for a compiled pattern. Previously, there was code
1302 that made a preliminary scan of the pattern in order to do this. That was
1303 OK when PCRE was new, but as the facilities have expanded, it has become
1304 harder and harder to keep it in step with the real compile phase, and there
1305 have been a number of bugs (see for example, 4 above). I have now found a
1306 cunning way of running the real compile function in a "fake" mode that
1307 enables it to compute how much memory it would need, while actually only
1308 ever using a few hundred bytes of working memory and without too many
1309 tests of the mode. This should make future maintenance and development
1310 easier. A side effect of this work is that the limit of 200 on the nesting
1311 depth of parentheses has been removed (though this was never a serious
1312 limitation, I suspect). However, there is a downside: pcre_compile() now
1313 runs more slowly than before (30% or more, depending on the pattern). I
1314 hope this isn't a big issue. There is no effect on runtime performance.
1315
1316 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1317 newline (only possible for the last line of a file) and it was a
1318 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1319
1320 19. Added additional timing features to pcretest. (1) The -tm option now times
1321 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1322 separate command line item, by a number that specifies the number of
1323 repeats to use when timing. The default is 50000; this gives better
1324 precision, but takes uncomfortably long for very large patterns.
1325
1326 20. Extended pcre_study() to be more clever in cases where a branch of a
1327 subpattern has no definite first character. For example, (a*|b*)[cd] would
1328 previously give no result from pcre_study(). Now it recognizes that the
1329 first character must be a, b, c, or d.
1330
1331 21. There was an incorrect error "recursive call could loop indefinitely" if
1332 a subpattern (or the entire pattern) that was being tested for matching an
1333 empty string contained only one non-empty item after a nested subpattern.
1334 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1335 incorrectly, because the \d was being skipped in the check.
1336
1337 22. The pcretest program now has a new pattern option /B and a command line
1338 option -b, which is equivalent to adding /B to every pattern. This causes
1339 it to show the compiled bytecode, without the additional information that
1340 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1341 is the same as /B/I).
1342
1343 23. A new optimization is now able automatically to treat some sequences such
1344 as a*b as a*+b. More specifically, if something simple (such as a character
1345 or a simple class like \d) has an unlimited quantifier, and is followed by
1346 something that cannot possibly match the quantified thing, the quantifier
1347 is automatically "possessified".
1348
1349 24. A recursive reference to a subpattern whose number was greater than 39
1350 went wrong under certain circumstances in UTF-8 mode. This bug could also
1351 have affected the operation of pcre_study().
1352
1353 25. Realized that a little bit of performance could be had by replacing
1354 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1355
1356 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1357
1358 27. Possessive quantifiers such as a++ were previously implemented by turning
1359 them into atomic groups such as ($>a+). Now they have their own opcodes,
1360 which improves performance. This includes the automatically created ones
1361 from 23 above.
1362
1363 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1364 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1365 the first matched character to be a colon. This applied both to named and
1366 numbered groups.
1367
1368 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1369
1370 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1371 building PCRE on Windows easier, so I have included it in the distribution.
1372
1373 31. There is now a check in pcretest against a ridiculously large number being
1374 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1375 loop, the loop is abandoned.
1376
1377 32. Forward references to subpatterns in conditions such as (?(2)...) where
1378 subpattern 2 is defined later cause pcre_compile() to search forwards in
1379 the pattern for the relevant set of parentheses. This search went wrong
1380 when there were unescaped parentheses in a character class, parentheses
1381 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1382
1383 33. "Subroutine" calls and backreferences were previously restricted to
1384 referencing subpatterns earlier in the regex. This restriction has now
1385 been removed.
1386
1387 34. Added a number of extra features that are going to be in Perl 5.10. On the
1388 whole, these are just syntactic alternatives for features that PCRE had
1389 previously implemented using the Python syntax or my own invention. The
1390 other formats are all retained for compatibility.
1391
1392 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1393 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1394 also .NET compatible.
1395
1396 (b) A recursion or subroutine call to a named group can now be defined as
1397 (?&name) as well as (?P>name).
1398
1399 (c) A backreference to a named group can now be defined as \k<name> or
1400 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1401 5.10, are also .NET compatible.
1402
1403 (d) A conditional reference to a named group can now use the syntax
1404 (?(<name>) or (?('name') as well as (?(name).
1405
1406 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1407 groups (named and numbered) that are never evaluated inline, but can be
1408 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1409 is always false. There may be only one alternative in such a group.
1410
1411 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1412 as the simple (?(R). The condition is true only if the most recent
1413 recursion is that of the given number or name. It does not search out
1414 through the entire recursion stack.
1415
1416 (g) The escape \gN or \g{N} has been added, where N is a positive or
1417 negative number, specifying an absolute or relative reference.
1418
1419 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1420 some "unreachable code" warnings.
1421
1422 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1423 things, this adds five new scripts.
1424
1425 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1426 There were also incompatibilities regarding the handling of \Q..\E inside
1427 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1428 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1429
1430 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1431 matches an empty string, and forcibly breaks the loop. There were bugs in
1432 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1433 against aaaa the result was just "a" rather than "aaaa", for example. Two
1434 separate and independent bugs (that affected different cases) have been
1435 fixed.
1436
1437 39. Refactored the code to abolish the use of different opcodes for small
1438 capturing bracket numbers. This is a tidy that I avoided doing when I
1439 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1440 The new approach is not only tidier, it makes it possible to reduce the
1441 memory needed to fix the previous bug (38).
1442
1443 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1444 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1445 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1446 mode.
1447
1448 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1449 report.
1450
1451 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1452 copy construction and assignment in the C++ wrapper.
1453
1454 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1455 couple of bugs that could have given wrong results in the "--newline=crlf"
1456 case.
1457
1458 44. Added a number of casts and did some reorganization of signed/unsigned int
1459 variables following suggestions from Dair Grant. Also renamed the variable
1460 "this" as "item" because it is a C++ keyword.
1461
1462 45. Arranged for dftables to add
1463
1464 #include "pcre_internal.h"
1465
1466 to pcre_chartables.c because without it, gcc 4.x may remove the array
1467 definition from the final binary if PCRE is built into a static library and
1468 dead code stripping is activated.
1469
1470 46. For an unanchored pattern, if a match attempt fails at the start of a
1471 newline sequence, and the newline setting is CRLF or ANY, and the next two
1472 characters are CRLF, advance by two characters instead of one.
1473
1474
1475 Version 6.7 04-Jul-06
1476 ---------------------
1477
1478 1. In order to handle tests when input lines are enormously long, pcretest has
1479 been re-factored so that it automatically extends its buffers when
1480 necessary. The code is crude, but this _is_ just a test program. The
1481 default size has been increased from 32K to 50K.
1482
1483 2. The code in pcre_study() was using the value of the re argument before
1484 testing it for NULL. (Of course, in any sensible call of the function, it
1485 won't be NULL.)
1486
1487 3. The memmove() emulation function in pcre_internal.h, which is used on
1488 systems that lack both memmove() and bcopy() - that is, hardly ever -
1489 was missing a "static" storage class specifier.
1490
1491 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1492 containing an extended class (one that cannot be represented by a bitmap
1493 because it contains high-valued characters or Unicode property items, e.g.
1494 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1495 pattern, but PCRE should not loop if you do not (it no longer does).
1496 [Detail: two cases were found: (a) a repeated subpattern containing an
1497 extended class; (b) a recursive reference to a subpattern that followed a
1498 previous extended class. It wasn't skipping over the extended class
1499 correctly when UTF-8 mode was not set.]
1500
1501 5. A negated single-character class was not being recognized as fixed-length
1502 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1503 compile error "lookbehind assertion is not fixed length".
1504
1505 6. The RunPerlTest auxiliary script was showing an unexpected difference
1506 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1507 write a Perl script that can interpret lines of an input file either as
1508 byte characters or as UTF-8, which is what "perltest" was being required to
1509 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1510 can't do is switch easily at run time between having the "use utf8;" pragma
1511 or not. In the end, I fudged it by using the RunPerlTest script to insert
1512 "use utf8;" explicitly for the UTF-8 tests.
1513
1514 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1515 the end of the subject string, contrary to the documentation and to what
1516 Perl does. This was true of both matching functions. Now it matches only at
1517 the start of the subject and immediately after *internal* newlines.
1518
1519 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1520 a pointer to an int instead of a pointer to an unsigned long int. This
1521 caused problems on 64-bit systems.
1522
1523 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1524 instance of the 'standard' template library not being so standard".
1525
1526 10. There was no check on the number of named subpatterns nor the maximum
1527 length of a subpattern name. The product of these values is used to compute
1528 the size of the memory block for a compiled pattern. By supplying a very
1529 long subpattern name and a large number of named subpatterns, the size
1530 computation could be caused to overflow. This is now prevented by limiting
1531 the length of names to 32 characters, and the number of named subpatterns
1532 to 10,000.
1533
1534 11. Subpatterns that are repeated with specific counts have to be replicated in
1535 the compiled pattern. The size of memory for this was computed from the
1536 length of the subpattern and the repeat count. The latter is limited to
1537 65535, but there was no limit on the former, meaning that integer overflow
1538 could in principle occur. The compiled length of a repeated subpattern is
1539 now limited to 30,000 bytes in order to prevent this.
1540
1541 12. Added the optional facility to have named substrings with the same name.
1542
1543 13. Added the ability to use a named substring as a condition, using the
1544 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1545 are numbers (not recommended). Forward references are permitted.
1546
1547 14. Added forward references in named backreferences (if you see what I mean).
1548
1549 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1550 pattern could run off the end of the subject. For example, the pattern
1551 "(?s)(.{1,5})"8 did this with the subject "ab".
1552
1553 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1554 PCRE_CASELESS was set when matching characters that were quantified with ?
1555 or *.
1556
1557 17. A character class other than a single negated character that had a minimum
1558 but no maximum quantifier - for example [ab]{6,} - was not handled
1559 correctly by pce_dfa_exec(). It would match only one character.
1560
1561 18. A valid (though odd) pattern that looked like a POSIX character
1562 class but used an invalid character after [ (for example [[,abc,]]) caused
1563 pcre_compile() to give the error "Failed: internal error: code overflow" or
1564 in some cases to crash with a glibc free() error. This could even happen if
1565 the pattern terminated after [[ but there just happened to be a sequence of
1566 letters, a binary zero, and a closing ] in the memory that followed.
1567
1568 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1569 over the years. Originally (before any Unicode support), just the bottom 8
1570 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1571 output from "man perlunicode" includes this:
1572
1573 The regular expression compiler produces polymorphic opcodes. That
1574 is, the pattern adapts to the data and automatically switches to
1575 the Unicode character scheme when presented with Unicode data--or
1576 instead uses a traditional byte scheme when presented with byte
1577 data.
1578
1579 Sadly, a wide octal escape does not cause a switch, and in a string with
1580 no other multibyte characters, these octal escapes are treated as before.
1581 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1582 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1583 Unicode string.
1584
1585 I have not perpetrated such confusion in PCRE. Up till now, it took just
1586 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1587 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1588 translate to the appropriate multibyte character.
1589
1590 29. Applied some refactoring to reduce the number of warnings from Microsoft
1591 and Borland compilers. This has included removing the fudge introduced
1592 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1593 a warning about an unused variable.
1594
1595 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1596 characters since release 4.0, because Perl (from release 5.004) does not.
1597 [Or at least, is documented not to: some releases seem to be in conflict
1598 with the documentation.] However, when a pattern was studied with
1599 pcre_study() and all its branches started with \s, PCRE still included VT
1600 as a possible starting character. Of course, this did no harm; it just
1601 caused an unnecessary match attempt.
1602
1603 22. Removed a now-redundant internal flag bit that recorded the fact that case
1604 dependency changed within the pattern. This was once needed for "required
1605 byte" processing, but is no longer used. This recovers a now-scarce options
1606 bit. Also moved the least significant internal flag bit to the most-
1607 significant bit of the word, which was not previously used (hangover from
1608 the days when it was an int rather than a uint) to free up another bit for
1609 the future.
1610
1611 23. Added support for CRLF line endings as well as CR and LF. As well as the
1612 default being selectable at build time, it can now be changed at runtime
1613 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1614 specify that it is scanning data with non-default line endings.
1615
1616 24. Changed the definition of CXXLINK to make it agree with the definition of
1617 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1618
1619 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1620 recursions. This makes a big different to stack usage for some patterns.
1621
1622 26. If a subpattern containing a named recursion or subroutine reference such
1623 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1624 the space required for the compiled pattern went wrong and gave too small a
1625 value. Depending on the environment, this could lead to "Failed: internal
1626 error: code overflow at offset 49" or "glibc detected double free or
1627 corruption" errors.
1628
1629 27. Applied patches from Google (a) to support the new newline modes and (b) to
1630 advance over multibyte UTF-8 characters in GlobalReplace.
1631
1632 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1633 difference for some implementation of PCRE in some Windows version.
1634
1635 29. Added some extra testing facilities to pcretest:
1636
1637 \q<number> in a data line sets the "match limit" value
1638 \Q<number> in a data line sets the "match recursion limt" value
1639 -S <number> sets the stack size, where <number> is in megabytes
1640
1641 The -S option isn't available for Windows.
1642
1643
1644 Version 6.6 06-Feb-06
1645 ---------------------
1646
1647 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1648 in pcreposix.h. I have copied the definition from pcre.h.
1649
1650 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1651 because pcre.h is no longer a built file.
1652
1653 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1654 not normally included in the compiled code.
1655
1656
1657 Version 6.5 01-Feb-06
1658 ---------------------
1659
1660 1. When using the partial match feature with pcre_dfa_exec(), it was not
1661 anchoring the second and subsequent partial matches at the new starting
1662 point. This could lead to incorrect results. For example, with the pattern
1663 /1234/, partially matching against "123" and then "a4" gave a match.
1664
1665 2. Changes to pcregrep:
1666
1667 (a) All non-match returns from pcre_exec() were being treated as failures
1668 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1669 error message is output. Some extra information is given for the
1670 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1671 probably the only errors that are likely to be caused by users (by
1672 specifying a regex that has nested indefinite repeats, for instance).
1673 If there are more than 20 of these errors, pcregrep is abandoned.
1674
1675 (b) A binary zero was treated as data while matching, but terminated the
1676 output line if it was written out. This has been fixed: binary zeroes
1677 are now no different to any other data bytes.
1678
1679 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1680 used to set a locale for matching. The --locale=xxxx long option has
1681 been added (no short equivalent) to specify a locale explicitly on the
1682 pcregrep command, overriding the environment variables.
1683
1684 (d) When -B was used with -n, some line numbers in the output were one less
1685 than they should have been.
1686
1687 (e) Added the -o (--only-matching) option.
1688
1689 (f) If -A or -C was used with -c (count only), some lines of context were
1690 accidentally printed for the final match.
1691
1692 (g) Added the -H (--with-filename) option.
1693
1694 (h) The combination of options -rh failed to suppress file names for files
1695 that were found from directory arguments.
1696
1697 (i) Added the -D (--devices) and -d (--directories) options.
1698
1699 (j) Added the -F (--fixed-strings) option.
1700
1701 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1702
1703 (l) Added the --colo(u)r option.
1704
1705 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1706 is not present by default.
1707
1708 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1709 items such as (?R) or (?1), when the recursion could match a number of
1710 alternatives. If it matched one of the alternatives, but subsequently,
1711 outside the recursion, there was a failure, the code tried to back up into
1712 the recursion. However, because of the way PCRE is implemented, this is not
1713 possible, and the result was an incorrect result from the match.
1714
1715 In order to prevent this happening, the specification of recursion has
1716 been changed so that all such subpatterns are automatically treated as
1717 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1718
1719 4. I had overlooked the fact that, in some locales, there are characters for
1720 which isalpha() is true but neither isupper() nor islower() are true. In
1721 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1722 and ordfeminine) are like this. This affected the treatment of \w and \W
1723 when they appeared in character classes, but not when they appeared outside
1724 a character class. The bit map for "word" characters is now created
1725 separately from the results of isalnum() instead of just taking it from the
1726 upper, lower, and digit maps. (Plus the underscore character, of course.)
1727
1728 5. The above bug also affected the handling of POSIX character classes such as
1729 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1730 permanent tables. Instead, the bit maps for such a class were previously
1731 created as the appropriate unions of the upper, lower, and digit bitmaps.
1732 Now they are created by subtraction from the [[:word:]] class, which has
1733 its own bitmap.
1734
1735 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1736 It is created by subtracting the vertical space characters (\x09, \x0a,
1737 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1738 subtraction was done in the overall bitmap for a character class, meaning
1739 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1740 be recognized. This bug has been fixed.
1741
1742 7. Patches from the folks at Google:
1743
1744 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1745 real life, but is still worth protecting against".
1746
1747 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1748 regular expressions".
1749
1750 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1751 have it.
1752
1753 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1754 "configure" and the latter not, in order to fix a problem somebody had
1755 with compiling the Arg class on HP-UX.
1756
1757 (e) Improve the error-handling of the C++ wrapper a little bit.
1758
1759 (f) New tests for checking recursion limiting.
1760
1761 8. The pcre_memmove() function, which is used only if the environment does not
1762 have a standard memmove() function (and is therefore rarely compiled),
1763 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1764 returning a result (though PCRE never actually uses the result).
1765
1766 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1767 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1768 returned instead of calling malloc() with an overflowing number that would
1769 most likely cause subsequent chaos.
1770
1771 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1772
1773 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1774 with this option is matched, the nmatch and pmatch options of regexec() are
1775 ignored.
1776
1777 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1778 provided in case anyone wants to the the POSIX interface with UTF-8
1779 strings.
1780
1781 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1782 C++ linking (needed for some HP-UX environments).
1783
1784 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1785 (unused parameter) and in the pcre_printint() function (omitted "default"
1786 switch label when the default is to do nothing).
1787
1788 15. Added some code to make it possible, when PCRE is compiled as a C++
1789 library, to replace subject pointers for pcre_exec() with a smart pointer
1790 class, thus making it possible to process discontinuous strings.
1791
1792 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1793 much the same function. They were added by different people who were trying
1794 to make PCRE easy to compile on non-Unix systems. It has been suggested
1795 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1796 for compiling on Windows systems. I have therefore replaced it with
1797 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1798 defaults to "extern" for C or "extern C" for C++, which works fine on
1799 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1800 SCOPE with something explicit in config.h. In addition:
1801
1802 (a) pcreposix.h still had just "extern" instead of either of these macros;
1803 I have replaced it with PCRE_DATA_SCOPE.
1804
1805 (b) Functions such as _pcre_xclass(), which are internal to the library,
1806 but external in the C sense, all had PCRE_EXPORT in their definitions.
1807 This is apparently wrong for the Windows case, so I have removed it.
1808 (It makes no difference on Unix-like systems.)
1809
1810 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1811 of recursive calls to match(). This is different to MATCH_LIMIT because
1812 that limits the total number of calls to match(), not all of which increase
1813 the depth of recursion. Limiting the recursion depth limits the amount of
1814 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1815 when PCRE is compiled, and changed at run time. A patch from Google adds
1816 this functionality to the C++ interface.
1817
1818 18. Changes to the handling of Unicode character properties:
1819
1820 (a) Updated the table to Unicode 4.1.0.
1821
1822 (b) Recognize characters that are not in the table as "Cn" (undefined).
1823
1824 (c) I revised the way the table is implemented to a much improved format
1825 which includes recognition of ranges. It now supports the ranges that
1826 are defined in UnicodeData.txt, and it also amalgamates other
1827 characters into ranges. This has reduced the number of entries in the
1828 table from around 16,000 to around 3,000, thus reducing its size
1829 considerably. I realized I did not need to use a tree structure after
1830 all - a binary chop search is just as efficient. Having reduced the
1831 number of entries, I extended their size from 6 bytes to 8 bytes to
1832 allow for more data.
1833
1834 (d) Added support for Unicode script names via properties such as \p{Han}.
1835
1836 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1837 matching that character.
1838
1839 20. When matching a repeated Unicode property with a minimum greater than zero,
1840 (for example \pL{2,}), PCRE could look past the end of the subject if it
1841 reached it while seeking the minimum number of characters. This could
1842 happen only if some of the characters were more than one byte long, because
1843 there is a check for at least the minimum number of bytes.
1844
1845 21. Refactored the implementation of \p and \P so as to be more general, to
1846 allow for more different types of property in future. This has changed the
1847 compiled form incompatibly. Anybody with saved compiled patterns that use
1848 \p or \P will have to recompile them.
1849
1850 22. Added "Any" and "L&" to the supported property types.
1851
1852 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1853 but give a compile time error if the value is greater than 0xff.
1854
1855 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1856 accidentally not being installed or uninstalled.
1857
1858 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1859 made were to insert the current release number. This seemed silly, because
1860 it made things harder for people building PCRE on systems that don't run
1861 "configure". I have turned pcre.h into a distributed file, no longer built
1862 by "configure", with the version identification directly included. There is
1863 no longer a pcre.h.in file.
1864
1865 However, this change necessitated a change to the pcre-config script as
1866 well. It is built from pcre-config.in, and one of the substitutions was the
1867 release number. I have updated configure.ac so that ./configure now finds
1868 the release number by grepping pcre.h.
1869
1870 26. Added the ability to run the tests under valgrind.
1871
1872
1873 Version 6.4 05-Sep-05
1874 ---------------------
1875
1876 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1877 "--" to be printed when multiple files were scanned, even when none of the
1878 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1879 consider it to be a bug, and have restored the previous behaviour.
1880
1881 2. A couple of code tidies to get rid of compiler warnings.
1882
1883 3. The pcretest program used to cheat by referring to symbols in the library
1884 whose names begin with _pcre_. These are internal symbols that are not
1885 really supposed to be visible externally, and in some environments it is
1886 possible to suppress them. The cheating is now confined to including
1887 certain files from the library's source, which is a bit cleaner.
1888
1889 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1890 file's purpose clearer.
1891
1892 5. Reorganized pcre_ucp_findchar().
1893
1894
1895 Version 6.3 15-Aug-05
1896 ---------------------
1897
1898 1. The file libpcre.pc.in did not have general read permission in the tarball.
1899
1900 2. There were some problems when building without C++ support:
1901
1902 (a) If C++ support was not built, "make install" and "make test" still
1903 tried to test it.
1904
1905 (b) There were problems when the value of CXX was explicitly set. Some
1906 changes have been made to try to fix these, and ...
1907
1908 (c) --disable-cpp can now be used to explicitly disable C++ support.
1909
1910 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1911 backslash in a target when C++ was disabled. This confuses some
1912 versions of "make", apparently. Using an intermediate variable solves
1913 this. (Same for CPP_LOBJ.)
1914
1915 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1916 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1917 necessary on certain architectures.
1918
1919 4. Added a setting of -export-symbols-regex to the link command to remove
1920 those symbols that are exported in the C sense, but actually are local
1921 within the library, and not documented. Their names all begin with
1922 "_pcre_". This is not a perfect job, because (a) we have to except some
1923 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1924 available (and never for static libraries). I have made a note to try to
1925 find a way round (a) in the future.
1926
1927
1928 Version 6.2 01-Aug-05
1929 ---------------------
1930
1931 1. There was no test for integer overflow of quantifier values. A construction
1932 such as {1111111111111111} would give undefined results. What is worse, if
1933 a minimum quantifier for a parenthesized subpattern overflowed and became
1934 negative, the calculation of the memory size went wrong. This could have
1935 led to memory overwriting.
1936
1937 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1938
1939 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1940 operating environments where this matters.
1941
1942 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1943 PCRE options from within the C++ wrapper.
1944
1945 5. Named capturing subpatterns were not being correctly counted when a pattern
1946 was compiled. This caused two problems: (a) If there were more than 100
1947 such subpatterns, the calculation of the memory needed for the whole
1948 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1949 back references of the form \12, where the number was greater than 9, were
1950 not recognized as back references, even though there were sufficient
1951 previous subpatterns.
1952
1953 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1954 versions of gcc, e.g. 2.95.4.
1955
1956
1957 Version 6.1 21-Jun-05
1958 ---------------------
1959
1960 1. There was one reference to the variable "posix" in pcretest.c that was not
1961 surrounded by "#if !defined NOPOSIX".
1962
1963 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1964 the cross-check on the old pcre_info() function, for the benefit of the
1965 cut-down version of PCRE that is currently imported into Exim.
1966
1967 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1968 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1969 patterns that start (?i) but I don't think that matters. The use of (?i) is
1970 just an example; this all applies to the other options as well.
1971
1972 4. Since libtool seems to echo the compile commands it is issuing, the output
1973 from "make" can be reduced a bit by putting "@" in front of each libtool
1974 compile command.
1975
1976 5. Patch from the folks at Google for configure.in to be a bit more thorough
1977 in checking for a suitable C++ installation before trying to compile the
1978 C++ stuff. This should fix a reported problem when a compiler was present,
1979 but no suitable headers.
1980
1981 6. The man pages all had just "PCRE" as their title. I have changed them to
1982 be the relevant file name. I have also arranged that these names are
1983 retained in the file doc/pcre.txt, which is a concatenation in text format
1984 of all the man pages except the little individual ones for each function.
1985
1986 7. The NON-UNIX-USE file had not been updated for the different set of source
1987 files that come with release 6. I also added a few comments about the C++
1988 wrapper.
1989
1990
1991 Version 6.0 07-Jun-05
1992 ---------------------
1993
1994 1. Some minor internal re-organization to help with my DFA experiments.
1995
1996 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1997 didn't matter for the library itself when fully configured, but did matter
1998 when compiling without UCP support, or within Exim, where the ucp files are
1999 not imported.
2000
2001 3. Refactoring of the library code to split up the various functions into
2002 different source modules. The addition of the new DFA matching code (see
2003 below) to a single monolithic source would have made it really too
2004 unwieldy, quite apart from causing all the code to be include in a
2005 statically linked application, when only some functions are used. This is
2006 relevant even without the DFA addition now that patterns can be compiled in
2007 one application and matched in another.
2008
2009 The downside of splitting up is that there have to be some external
2010 functions and data tables that are used internally in different modules of
2011 the library but which are not part of the API. These have all had their
2012 names changed to start with "_pcre_" so that they are unlikely to clash
2013 with other external names.
2014
2015 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2016 a different (DFA) algorithm. Although it is slower than the original
2017 function, it does have some advantages for certain types of matching
2018 problem.
2019
2020 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2021 including restarting after a partial match.
2022
2023 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2024 defined when compiling for Windows was sent to me. I have put it into the
2025 code, though I have no means of testing or verifying it.
2026
2027 7. Added the pcre_refcount() auxiliary function.
2028
2029 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2030 match before or at the first newline in the subject string. In pcretest,
2031 the /f option on a pattern can be used to set this.
2032
2033 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2034 would behave wrongly. This has been present in PCRE since release 4.0.
2035
2036 10. A number of changes to the pcregrep command:
2037
2038 (a) Refactored how -x works; insert ^(...)$ instead of setting
2039 PCRE_ANCHORED and checking the length, in preparation for adding
2040 something similar for -w.
2041
2042 (b) Added the -w (match as a word) option.
2043
2044 (c) Refactored the way lines are read and buffered so as to have more
2045 than one at a time available.
2046
2047 (d) Implemented a pcregrep test script.
2048
2049 (e) Added the -M (multiline match) option. This allows patterns to match
2050 over several lines of the subject. The buffering ensures that at least
2051 8K, or the rest of the document (whichever is the shorter) is available
2052 for matching (and similarly the previous 8K for lookbehind assertions).
2053
2054 (f) Changed the --help output so that it now says
2055
2056 -w, --word-regex(p)
2057
2058 instead of two lines, one with "regex" and the other with "regexp"
2059 because that confused at least one person since the short forms are the
2060 same. (This required a bit of code, as the output is generated
2061 automatically from a table. It wasn't just a text change.)
2062
2063 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2064 option but starts with a hyphen. Could be a pattern or a path name
2065 starting with a hyphen, for instance.
2066
2067 (h) "-" can be given as a file name to represent stdin.
2068
2069 (i) When file names are being printed, "(standard input)" is used for
2070 the standard input, for compatibility with GNU grep. Previously
2071 "<stdin>" was used.
2072
2073 (j) The option --label=xxx can be used to supply a name to be used for
2074 stdin when file names are being printed. There is no short form.
2075
2076 (k) Re-factored the options decoding logic because we are going to add
2077 two more options that take data. Such options can now be given in four
2078 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2079
2080 (l) Added the -A, -B, and -C options for requesting that lines of context
2081 around matches be printed.
2082
2083 (m) Added the -L option to print the names of files that do not contain
2084 any matching lines, that is, the complement of -l.
2085
2086 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2087 continue to scan other files.
2088
2089 (o) The -s option was incorrectly implemented. For compatibility with other
2090 greps, it now suppresses the error message for a non-existent or non-
2091 accessible file (but not the return code). There is a new option called
2092 -q that suppresses the output of matching lines, which was what -s was
2093 previously doing.
2094
2095 (p) Added --include and --exclude options to specify files for inclusion
2096 and exclusion when recursing.
2097
2098 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2099 Hopefully, it now does.
2100
2101 12. Missing cast in pcre_study().
2102
2103 13. Added an "uninstall" target to the makefile.
2104
2105 14. Replaced "extern" in the function prototypes in Makefile.in with
2106 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2107 world, but is set differently for Windows.
2108
2109 15. Added a second compiling function called pcre_compile2(). The only
2110 difference is that it has an extra argument, which is a pointer to an
2111 integer error code. When there is a compile-time failure, this is set
2112 non-zero, in addition to the error test pointer being set to point to an
2113 error message. The new argument may be NULL if no error number is required
2114 (but then you may as well call pcre_compile(), which is now just a
2115 wrapper). This facility is provided because some applications need a
2116 numeric error indication, but it has also enabled me to tidy up the way
2117 compile-time errors are handled in the POSIX wrapper.
2118
2119 16. Added VPATH=.libs to the makefile; this should help when building with one
2120 prefix path and installing with another. (Or so I'm told by someone who
2121 knows more about this stuff than I do.)
2122
2123 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2124 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2125 match everything, including newlines. This is not POSIX-compatible, but
2126 somebody wanted the feature. From pcretest it can be activated by using
2127 both the P and the s flags.
2128
2129 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2130
2131 19. libpcre.pc was being incorrectly installed as executable.
2132
2133 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2134 it now also looks for '\r' so that it will work unmodified on Windows.
2135
2136 21. Added Google's contributed C++ wrapper to the distribution.
2137
2138 22. Added some untidy missing memory free() calls in pcretest, to keep
2139 Electric Fence happy when testing.
2140
2141
2142
2143 Version 5.0 13-Sep-04
2144 ---------------------
2145
2146 1. Internal change: literal characters are no longer packed up into items
2147 containing multiple characters in a single byte-string. Each character
2148 is now matched using a separate opcode. However, there may be more than one
2149 byte in the character in UTF-8 mode.
2150
2151 2. The pcre_callout_block structure has two new fields: pattern_position and
2152 next_item_length. These contain the offset in the pattern to the next match
2153 item, and its length, respectively.
2154
2155 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2156 insertion of callouts before each pattern item. Added the /C option to
2157 pcretest to make use of this.
2158
2159 4. On the advice of a Windows user, the lines
2160
2161 #if defined(_WIN32) || defined(WIN32)
2162 _setmode( _fileno( stdout ), 0x8000 );
2163 #endif /* defined(_WIN32) || defined(WIN32) */
2164
2165 have been added to the source of pcretest. This apparently does useful
2166 magic in relation to line terminators.
2167
2168 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2169 for the benefit of those environments where the "b" makes a difference.
2170
2171 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2172 to know about it. I have put a hack into configure.in that adds in code
2173 to set GCC=yes if CC=icc. This seems to end up at a point in the
2174 generated configure script that is early enough to affect the setting of
2175 compiler options, which is what is needed, but I have no means of testing
2176 whether it really works. (The user who reported this had patched the
2177 generated configure script, which of course I cannot do.)
2178
2179 LATER: After change 22 below (new libtool files), the configure script
2180 seems to know about icc (and also ecc). Therefore, I have commented out
2181 this hack in configure.in.
2182
2183 7. Added support for pkg-config (2 patches were sent in).
2184
2185 8. Negated POSIX character classes that used a combination of internal tables
2186 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2187 [[:^ascii]]. Typically, they would match almost any characters. The other
2188 POSIX classes were not broken in this way.
2189
2190 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2191 to find the match, as PCRE was deluded into thinking that the match had to
2192 start at the start point or following a newline. The same bug applied to
2193 patterns with negative forward assertions or any backward assertions
2194 preceding ".*" at the start, unless the pattern required a fixed first
2195 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2196
2197 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2198 starting at the last subject character, bytes beyond the end of the subject
2199 string were read.
2200
2201 11. Renamed the variable "class" as "classbits" to make life easier for C++
2202 users. (Previously there was a macro definition, but it apparently wasn't
2203 enough.)
2204
2205 12. Added the new field "tables" to the extra data so that tables can be passed
2206 in at exec time, or the internal tables can be re-selected. This allows
2207 a compiled regex to be saved and re-used at a later time by a different
2208 program that might have everything at different addresses.
2209
2210 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2211 -R library as well as a -L library.
2212
2213 14. The debugging options of pcretest (-d on the command line or D on a
2214 pattern) showed incorrect output for anything following an extended class
2215 that contained multibyte characters and which was followed by a quantifier.
2216
2217 15. Added optional support for general category Unicode character properties
2218 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2219 support. It adds about 90K to the size of the library. The meanings of the
2220 inbuilt class escapes such as \d and \s have NOT been changed.
2221
2222 16. Updated pcredemo.c to include calls to free() to release the memory for the
2223 compiled pattern.
2224
2225 17. The generated file chartables.c was being created in the source directory
2226 instead of in the building directory. This caused the build to fail if the
2227 source directory was different from the building directory, and was
2228 read-only.
2229
2230 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2231 file. No doubt somebody will tell me if they don't make sense... Also added
2232 Dan Mooney's comments about building on OpenVMS.
2233
2234 19. Added support for partial matching via the PCRE_PARTIAL option for
2235 pcre_exec() and the \P data escape in pcretest.
2236
2237 20. Extended pcretest with 3 new pattern features:
2238
2239 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2240 write the compiled pattern to the file whose name is "rest-of-line".
2241 This is a straight binary dump of the data, with the saved pointer to
2242 the character tables forced to be NULL. The study data, if any, is
2243 written too. After writing, pcretest reads a new pattern.
2244
2245 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2246 compiled pattern from the given file. There must not be any
2247 occurrences of "<" in the file name (pretty unlikely); if there are,
2248 pcretest will instead treat the initial "<" as a pattern delimiter.
2249 After reading in the pattern, pcretest goes on to read data lines as
2250 usual.
2251
2252 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2253 and 16-bit fields in a compiled pattern, to simulate a pattern that
2254 was compiled on a host of opposite endianness.
2255
2256 21. The pcre-exec() function can now cope with patterns that were compiled on
2257 hosts of opposite endianness, with this restriction:
2258
2259 As for any compiled expression that is saved and used later, the tables
2260 pointer field cannot be preserved; the extra_data field in the arguments
2261 to pcre_exec() should be used to pass in a tables address if a value
2262 other than the default internal tables were used at compile time.
2263
2264 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2265 now diagnosed as an error. Previously, most of the time, a negative number
2266 would have been treated as zero, but if in addition "ovector" was passed as
2267 NULL, a crash could occur.
2268
2269 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2270 new versions from the libtool 1.5 distribution (the last one is a copy of
2271 a file called libtool.m4). This seems to have fixed the need to patch
2272 "configure" to support Darwin 1.3 (which I used to do). However, I still
2273 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2274 workstation).
2275
2276 24. Changed the PCRE licence to be the more standard "BSD" licence.
2277
2278
2279 Version 4.5 01-Dec-03
2280 ---------------------
2281
2282 1. There has been some re-arrangement of the code for the match() function so
2283 that it can be compiled in a version that does not call itself recursively.
2284 Instead, it keeps those local variables that need separate instances for
2285 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2286 needs to "recurse". Keeping track of where control must go is done by means
2287 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2288 hide most of the details from the main code, and operates only if
2289 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2290 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2291 operating.
2292
2293 To make it easier for callers to provide specially tailored get/free
2294 functions for this usage, two new functions, pcre_stack_malloc, and
2295 pcre_stack_free, are used. They are always called in strict stacking order,
2296 and the size of block requested is always the same.
2297
2298 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2299 PCRE has been compiled to use the stack or the heap for recursion. The
2300 -C option of pcretest uses this to show which version is compiled.
2301
2302 A new data escape \S, is added to pcretest; it causes the amounts of store
2303 obtained and freed by both kinds of malloc/free at match time to be added
2304 to the output.
2305
2306 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2307 what's available on my current Linux desktop machine.
2308
2309 3. When matching a UTF-8 string, the test for a valid string at the start has
2310 been extended. If start_offset is not zero, PCRE now checks that it points
2311 to a byte that is the start of a UTF-8 character. If not, it returns
2312 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2313 this is necessary because there may be backward assertions in the pattern.
2314 When matching the same subject several times, it may save resources to use
2315 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2316
2317 4. The code for checking the validity of UTF-8 strings has been tightened so
2318 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2319 containing "overlong sequences".
2320
2321 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2322 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2323 should have been "&", but it just so happened that all the cases this let
2324 through by mistake were picked up later in the function.
2325
2326 6. I had used a variable called "isblank" - this is a C99 function, causing
2327 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2328
2329 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2330 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2331 so the version is shown; (c) stop "make test" echoing "./RunTest".
2332
2333 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2334
2335 9. The prototype for memmove() for systems that don't have it was using
2336 size_t, but the inclusion of the header that defines size_t was later. I've
2337 moved the #includes for the C headers earlier to avoid this.
2338
2339 10. Added some adjustments to the code to make it easier to compiler on certain
2340 special systems:
2341
2342 (a) Some "const" qualifiers were missing.
2343 (b) Added the macro EXPORT before all exported functions; by default this
2344 is defined to be empty.
2345 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2346 that it reads its output file name as an argument instead of writing
2347 to the standard output and assuming this can be redirected.
2348
2349 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2350 class containing characters with values greater than 255, PCRE compilation
2351 went into a loop.
2352
2353 12. A recursive reference to a subpattern that was within another subpattern
2354 that had a minimum quantifier of zero caused PCRE to crash. For example,
2355 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2356 recursion. If the recursively-called subpattern itself had a zero repeat,
2357 that was OK.
2358
2359 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2360 buffer into which it was copied (for escape processing) was still set at
2361 1024, so long lines caused crashes.
2362
2363 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2364 "internal error: code overflow...". This applied to any character class
2365 that was followed by a possessive quantifier.
2366
2367 15. Modified the Makefile to add libpcre.la as a prerequisite for
2368 libpcreposix.la because I was told this is needed for a parallel build to
2369 work.
2370
2371 16. If a pattern that contained .* following optional items at the start was
2372 studied, the wrong optimizing data was generated, leading to matching
2373 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2374 matching string must start with a or b or c. The correct conclusion for
2375 this pattern is that a match can start with any character.
2376
2377
2378 Version 4.4 13-Aug-03
2379 ---------------------
2380
2381 1. In UTF-8 mode, a character class containing characters with values between
2382 127 and 255 was not handled correctly if the compiled pattern was studied.
2383 In fixing this, I have also improved the studying algorithm for such
2384 classes (slightly).
2385
2386 2. Three internal functions had redundant arguments passed to them. Removal
2387 might give a very teeny performance improvement.
2388
2389 3. Documentation bug: the value of the capture_top field in a callout is *one
2390 more than* the number of the hightest numbered captured substring.
2391
2392 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2393 in incorrectly linking with a previously installed version. They now link
2394 explicitly with libpcre.la.
2395
2396 5. configure.in no longer needs to recognize Cygwin specially.
2397
2398 6. A problem in pcre.in for Windows platforms is fixed.
2399
2400 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2401 pcretest, it used to include the size of the study block as part of its
2402 output. Unfortunately, the structure contains a field that has a different
2403 size on different hardware architectures. This meant that the tests that
2404 showed this size failed. As the block is currently always of a fixed size,
2405 this information isn't actually particularly useful in pcretest output, so
2406 I have just removed it.
2407
2408 8. Three pre-processor statements accidentally did not start in column 1.
2409 Sadly, there are *still* compilers around that complain, even though
2410 standard C has not required this for well over a decade. Sigh.
2411
2412 9. In pcretest, the code for checking callouts passed small integers in the
2413 callout_data field, which is a void * field. However, some picky compilers
2414 complained about the casts involved for this on 64-bit systems. Now
2415 pcretest passes the address of the small integer instead, which should get
2416 rid of the warnings.
2417
2418 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2419 both compile and run time, and gives an error if an invalid UTF-8 sequence
2420 is found. There is a option for disabling this check in cases where the
2421 string is known to be correct and/or the maximum performance is wanted.
2422
2423 11. In response to a bug report, I changed one line in Makefile.in from
2424
2425 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2426 to
2427 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2428
2429 to look similar to other lines, but I have no way of telling whether this
2430 is the right thing to do, as I do not use Windows. No doubt I'll get told
2431 if it's wrong...
2432
2433
2434 Version 4.3 21-May-03
2435 ---------------------
2436
2437 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2438 Makefile.
2439
2440 2. Some refactoring to improve the quality of the code:
2441
2442 (i) The utf8_table... variables are now declared "const".
2443
2444 (ii) The code for \cx, which used the "case flipping" table to upper case
2445 lower case letters, now just substracts 32. This is ASCII-specific,
2446 but the whole concept of \cx is ASCII-specific, so it seems
2447 reasonable.
2448
2449 (iii) PCRE was using its character types table to recognize decimal and
2450 hexadecimal digits in the pattern. This is silly, because it handles
2451 only 0-9, a-f, and A-F, but the character types table is locale-
2452 specific, which means strange things might happen. A private
2453 table is now used for this - though it costs 256 bytes, a table is
2454 much faster than multiple explicit tests. Of course, the standard
2455 character types table is still used for matching digits in subject
2456 strings against \d.
2457
2458 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2459 ending in _t are). So I've renamed it as ESC_tee.
2460
2461 3. The first argument for regexec() in the POSIX wrapper should have been
2462 defined as "const".
2463
2464 4. Changed pcretest to use malloc() for its buffers so that they can be
2465 Electric Fenced for debugging.
2466
2467 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2468 to read one or more bytes before the start of the subject string. Often this
2469 had no effect on PCRE's behaviour, but in some circumstances it could
2470 provoke a segmentation fault.
2471
2472 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2473 to try to read one or more bytes before the start of the subject string.
2474
2475 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2476 UTF-8 support could misbehave in various ways if the subject string
2477 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2478 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2479 back over UTF-8 characters.)
2480
2481
2482 Version 4.2 14-Apr-03
2483 ---------------------
2484
2485 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2486
2487 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2488 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2489 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2490 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2491 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2492 and BUILD_EXEEXT
2493 Note: automatic setting of the BUILD variables is not yet working
2494 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2495 compile-time but not at link-time
2496 [LINK]: use for linking executables only
2497 make different versions for Windows and non-Windows
2498 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2499 libraries
2500 [LINK_FOR_BUILD]: new variable
2501 [OBJEXT]: use throughout
2502 [EXEEXT]: use throughout
2503 <winshared>: new target
2504 <wininstall>: new target
2505 <dftables.o>: use native compiler
2506 <dftables>: use native linker
2507 <install>: handle Windows platform correctly
2508 <clean>: ditto
2509 <check>: ditto
2510 copy DLL to top builddir before testing
2511
2512 As part of these changes, -no-undefined was removed again. This was reported
2513 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2514 in any case.
2515
2516 3. Some tidies to get rid of compiler warnings:
2517
2518 . In the match_data structure, match_limit was an unsigned long int, whereas
2519 match_call_count was an int. I've made them both unsigned long ints.
2520
2521 . In pcretest the fact that a const uschar * doesn't automatically cast to
2522 a void * provoked a warning.
2523
2524 . Turning on some more compiler warnings threw up some "shadow" variables
2525 and a few more missing casts.
2526
2527 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2528 option, a class that contained a single character with a value between 128
2529 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2530
2531 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2532 option, a class that contained several characters, but with at least one
2533 whose value was between 128 and 255 caused PCRE to crash.
2534
2535
2536 Version 4.1 12-Mar-03
2537 ---------------------
2538
2539 1. Compiling with gcc -pedantic found a couple of places where casts were
2540 needed, and a string in dftables.c that was longer than standard compilers are
2541 required to support.
2542
2543 2. Compiling with Sun's compiler found a few more places where the code could
2544 be tidied up in order to avoid warnings.
2545
2546 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2547 first of these names is deprecated in the latest Autoconf in favour of the name
2548 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2549 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2550 analogy I have changed it to CFLAGS_FOR_BUILD.
2551
2552 4. Added -no-undefined to the linking command in the Makefile, because this is
2553 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2554 linking step for the pcreposix library.
2555
2556 5. PCRE was failing to diagnose the case of two named groups with the same
2557 name.
2558
2559 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2560 literal character that is needed in the subject for a match, and scans along to
2561 ensure that it is present before embarking on the full matching process. This
2562 saves time in cases of nested unlimited repeats that are never going to match.
2563 Problem: the scan can take a lot of time if the subject is very long (e.g.
2564 megabytes), thus penalizing straightforward matches. It is now done only if the
2565 amount of subject to be scanned is less than 1000 bytes.
2566
2567 7. A lesser problem with the same optimization is that it was recording the
2568 first character of an anchored pattern as "needed", thus provoking a search
2569 right along the subject, even when the first match of the pattern was going to
2570 fail. The "needed" character is now not set for anchored patterns, unless it
2571 follows something in the pattern that is of non-fixed length. Thus, it still
2572 fulfils its original purpose of finding quick non-matches in cases of nested
2573 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2574
2575
2576 Version 4.0 17-Feb-03
2577 ---------------------
2578
2579 1. If a comment in an extended regex that started immediately after a meta-item
2580 extended to the end of string, PCRE compiled incorrect data. This could lead to
2581 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2582
2583 2. Moved to autoconf 2.53 and libtool 1.4.2.
2584
2585 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2586 the special perltest8 script is no longer needed - all the tests can be run
2587 from a single perltest script.
2588
2589 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2590 by \s. It has now been removed in PCRE. This means it isn't recognized as
2591 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2592 class [:space:] *does* include VT, thereby creating a mess.
2593
2594 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2595 space and tab.
2596
2597 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2598 its new features into the main test script, reducing the number of scripts.
2599
2600 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2601 were backward compatible, and made the (?i) apply to the whole pattern, as if
2602 /i were given. Now it behaves more logically, and applies the option setting
2603 only to what follows. PCRE has been changed to follow suit. However, if it
2604 finds options settings right at the start of the pattern, it extracts them into
2605 the global options, as before. Thus, they show up in the info data.
2606
2607 8. Added support for the \Q...\E escape sequence. Characters in between are
2608 treated as literals. This is slightly different from Perl in that $ and @ are
2609 also handled as literals inside the quotes. In Perl, they will cause variable
2610 interpolation. Note the following examples:
2611
2612 Pattern PCRE matches Perl matches
2613
2614 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2615 \Qabc\$xyz\E abc\$xyz abc\$xyz
2616 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2617
2618 For compatibility with Perl, \Q...\E sequences are recognized inside character
2619 classes as well as outside them.
2620
2621 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2622 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2623 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2624 signed/unsigned warnings.
2625
2626 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2627 option for pcretest, so I've replaced it by a simple function that does just
2628 that job.
2629
2630 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2631 "pcregrep -".
2632
2633 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2634 Java package. This provides some syntactic sugar for simple cases of what my
2635 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2636 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2637 item, you can use this simplified notation. Note that only makes sense with
2638 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2639 greediness, whatever the setting of the PCRE_UNGREEDY option.
2640
2641 13. A change of greediness default within a pattern was not taking effect at
2642 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2643 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2644 was abstracted outside.
2645
2646 14. PCRE now supports the \G assertion. It is true when the current matching
2647 position is at the start point of the match. This differs from \A when the
2648 starting offset is non-zero. Used with the /g option of pcretest (or similar
2649 code), it works in the same way as it does for Perl's /g option. If all
2650 alternatives of a regex begin with \G, the expression is anchored to the start
2651 match position, and the "anchored" flag is set in the compiled expression.
2652
2653 15. Some bugs concerning the handling of certain option changes within patterns
2654 have been fixed. These applied to options other than (?ims). For example,
2655 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2656 been the other way round. Some of this was related to change 7 above.
2657
2658 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2659 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2660 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2661 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2662
2663 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2664 mode. Unlike ".", it always matches newline, whatever the setting of
2665 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2666 assertions. Perl allows it, but it doesn't (in general) work because it can't
2667 calculate the length of the lookbehind. At least, that's the case for Perl
2668 5.8.0 - I've been told they are going to document that it doesn't work in
2669 future.
2670
2671 18. Added an error diagnosis for escapes that PCRE does not support: these are
2672 \L, \l, \N, \P, \p, \U, \u, and \X.
2673
2674 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2675 reading past the end of the pattern in cases such as /[abcd/.
2676
2677 20. PCRE was getting more memory than necessary for patterns with classes that
2678 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2679
2680 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2681 compiling PCRE for use with Virtual Pascal.
2682
2683 22. Small fix to the Makefile to make it work properly if the build is done
2684 outside the source tree.
2685
2686 23. Added a new extension: a condition to go with recursion. If a conditional
2687 subpattern starts with (?(R) the "true" branch is used if recursion has
2688 happened, whereas the "false" branch is used only at the top level.
2689
2690 24. When there was a very long string of literal characters (over 255 bytes
2691 without UTF support, over 250 bytes with UTF support), the computation of how
2692 much memory was required could be incorrect, leading to segfaults or other
2693 strange effects.
2694
2695 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2696 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2697 there was a subsequent back reference to those brackets. This meant that, for
2698 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2699 possible to check for precisely this case. All we can do is abandon the
2700 optimization if .* occurs inside capturing brackets when there are any back
2701 references whatsoever. (See below for a better fix that came later.)
2702
2703 26. The handling of the optimization for finding the first character of a
2704 non-anchored pattern, and for finding a character that is required later in the
2705 match were failing in some cases. This didn't break the matching; it just
2706 failed to optimize when it could. The way this is done has been re-implemented.
2707
2708 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2709
2710 28. Added a new feature that provides some of the functionality that Perl
2711 provides with (?{...}). The facility is termed a "callout". The way it is done
2712 in PCRE is for the caller to provide an optional function, by setting
2713 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2714 global variable. By default it is unset, which disables all calling out. To get
2715 the function called, the regex must include (?C) at appropriate points. This
2716 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2717 This provides a means of identifying different callout points. When PCRE
2718 reaches such a point in the regex, if pcre_callout has been set, the external
2719 function is called. It is provided with data in a structure called
2720 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2721 matching continues; if it returns a non-zero value, the match at the current
2722 point fails. However, backtracking will occur if possible. [This was changed
2723 later and other features added - see item 49 below.]
2724
2725 29. pcretest is upgraded to test the callout functionality. It provides a
2726 callout function that displays information. By default, it shows the start of
2727 the match and the current position in the text. There are some new data escapes
2728 to vary what happens:
2729
2730 \C+ in addition, show current contents of captured substrings
2731 \C- do not supply a callout function
2732 \C!n return 1 when callout number n is reached
2733 \C!n!m return 1 when callout number n is reached for the mth time
2734
2735 30. If pcregrep was called with the -l option and just a single file name, it
2736 output "<stdin>" if a match was found, instead of the file name.
2737
2738 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2739 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2740 pcre_exec(). This saves a malloc/free per call. The default value of
2741 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2742 when configuring.
2743
2744 32. The default maximum size of a compiled pattern is 64K. There have been a
2745 few cases of people hitting this limit. The code now uses macros to handle the
2746 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2747 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2748 configuring. Tests 2 and 5 work only with 2-byte links because they output
2749 debugging information about compiled patterns.
2750
2751 33. Internal code re-arrangements:
2752
2753 (a) Moved the debugging function for printing out a compiled regex into
2754 its own source file (printint.c) and used #include to pull it into
2755 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2756 separate copies.
2757
2758 (b) Defined the list of op-code names for debugging as a macro in
2759 internal.h so that it is next to the definition of the opcodes.
2760
2761 (c) Defined a table of op-code lengths for simpler skipping along compiled
2762 code. This is again a macro in internal.h so that it is next to the
2763 definition of the opcodes.
2764
2765 34. Added support for recursive calls to individual subpatterns, along the
2766 lines of Robin Houston's patch (but implemented somewhat differently).
2767
2768 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2769 allow it to read and process whole directories in Win32. This code was
2770 contributed by Lionel Fourquaux; it has not been tested by me.
2771
2772 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2773 used to name a group. Names consist of alphanumerics and underscores, and must
2774 be unique. Back references use the syntax (?P=name) and recursive calls use
2775 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2776 numbers. The function pcre_fullinfo() can be used after compilation to extract
2777 a name/number map. There are three relevant calls:
2778
2779 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2780 PCRE_INFO_NAMECOUNT yields the number of entries
2781 PCRE_INFO_NAMETABLE yields a pointer to the map.
2782
2783 The map is a vector of fixed-size entries. The size of each entry depends on
2784 the length of the longest name used. The first two bytes of each entry are the
2785 group number, most significant byte first. There follows the corresponding
2786 name, zero terminated. The names are in alphabetical order.
2787
2788 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2789 case instead of 255. Making it the same both with and without UTF-8 support
2790 means that the same test output works with both.
2791
2792 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2793 calling malloc() with a zero argument.
2794
2795 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2796 optimization. I've improved things by keeping a bitmap of backreferences with
2797 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2798 fact referenced, the optimization can be applied. It is unlikely that a
2799 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2800 the match to follow \n) will appear inside brackets with a number greater than
2801 31, but if it does, any back reference > 31 suppresses the optimization.
2802
2803 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2804 of disabling numbered capturing parentheses. Any opening parenthesis that is
2805 not followed by ? behaves as if it were followed by ?: but named parentheses
2806 can still be used for capturing (and they will acquire numbers in the usual
2807 way).
2808
2809 41. Redesigned the return codes from the match() function into yes/no/error so
2810 that errors can be passed back from deep inside the nested calls. A malloc
2811 failure while inside a recursive subpattern call now causes the
2812 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2813
2814 42. It is now possible to set a limit on the number of times the match()
2815 function is called in a call to pcre_exec(). This facility makes it possible to
2816 limit the amount of recursion and backtracking, though not in a directly
2817 obvious way, because the match() function is used in a number of different
2818 circumstances. The count starts from zero for each position in the subject
2819 string (for non-anchored patterns). The default limit is, for compatibility, a
2820 large number, namely 10 000 000. You can change this in two ways:
2821
2822 (a) When configuring PCRE before making, you can use --with-match-limit=n
2823 to set a default value for the compiled library.
2824
2825 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2826 a different value is set. See 45 below.
2827
2828 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2829
2830 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2831 of things that can be changed at compile time. The first argument specifies
2832 what is wanted and the second points to where the information is to be placed.
2833 The current list of available information is:
2834
2835 PCRE_CONFIG_UTF8
2836
2837 The output is an integer that is set to one if UTF-8 support is available;
2838 otherwise it is set to zero.
2839
2840 PCRE_CONFIG_NEWLINE
2841
2842 The output is an integer that it set to the value of the code that is used for
2843 newline. It is either LF (10) or CR (13).
2844
2845 PCRE_CONFIG_LINK_SIZE
2846
2847 The output is an integer that contains the number of bytes used for internal
2848 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2849
2850 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2851
2852 The output is an integer that contains the threshold above which the POSIX
2853 interface uses malloc() for output vectors. See item 31 above.
2854
2855 PCRE_CONFIG_MATCH_LIMIT
2856
2857 The output is an unsigned integer that contains the default limit of the number
2858 of match() calls in a pcre_exec() execution. See 42 above.
2859
2860 44. pcretest has been upgraded by the addition of the -C option. This causes it
2861 to extract all the available output from the new pcre_config() function, and to
2862 output it. The program then exits immediately.
2863
2864 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2865 order to support additional features. One way would have been to define
2866 pcre_exec2() (for example) with extra arguments, but this would not have been
2867 extensible, and would also have required all calls to the original function to
2868 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2869 is used for passing in "extra" data from pcre_study().
2870
2871 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2872 contains the following fields:
2873
2874 flags a bitmap indicating which of the following fields are set
2875 study_data opaque data from pcre_study()
2876 match_limit a way of specifying a limit on match() calls for a specific
2877 call to pcre_exec()
2878 callout_data data for callouts (see 49 below)
2879
2880 The flag bits are also defined in pcre.h, and are
2881
2882 PCRE_EXTRA_STUDY_DATA
2883 PCRE_EXTRA_MATCH_LIMIT
2884 PCRE_EXTRA_CALLOUT_DATA
2885
2886 The pcre_study() function now returns one of these new pcre_extra blocks, with
2887 the actual study data pointed to by the study_data field, and the
2888 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2889 before. That is, this change is entirely upwards-compatible and requires no
2890 change to existing code.
2891
2892 If you want to pass in additional data to pcre_exec(), you can either place it
2893 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2894 block.
2895
2896 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2897 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2898 times with different match limits, until it finds the minimum value needed for
2899 pcre_exec() to complete. The value is then output. This can be instructive; for
2900 most simple matches the number is quite small, but for pathological cases it
2901 gets very large very quickly.
2902
2903 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2904 returns the size of the data block pointed to by the study_data field in a
2905 pcre_extra block, that is, the value that was passed as the argument to
2906 pcre_malloc() when PCRE was getting memory in which to place the information
2907 created by pcre_study(). The fourth argument should point to a size_t variable.
2908 pcretest has been extended so that this information is shown after a successful
2909 pcre_study() call when information about the compiled regex is being displayed.
2910
2911 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2912 because what follows is always an absolute path. (Later: it turns out that this
2913 is more than cosmetic for MinGW, because it doesn't like empty path
2914 components.)
2915
2916 49. Some changes have been made to the callout feature (see 28 above):
2917
2918 (i) A callout function now has three choices for what it returns:
2919
2920 0 => success, carry on matching
2921 > 0 => failure at this point, but backtrack if possible
2922 < 0 => serious error, return this value from pcre_exec()
2923
2924 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2925 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2926 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2927 use by callout functions. It will never be used by PCRE itself.
2928
2929 (ii) The pcre_extra structure (see 45 above) has a void * field called
2930 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2931 pcre_callout_block structure has a field of the same name. The contents of
2932 the field passed in the pcre_extra structure are passed to the callout
2933 function in the corresponding field in the callout block. This makes it
2934 easier to use the same callout-containing regex from multiple threads. For
2935 testing, the pcretest program has a new data escape
2936
2937 \C*n pass the number n (may be negative) as callout_data
2938
2939 If the callout function in pcretest receives a non-zero value as
2940 callout_data, it returns that value.
2941
2942 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2943 there were some redundant $(CFLAGS) in commands that are now specified as
2944 $(LINK), which already includes $(CFLAGS).
2945
2946 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2947 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2948 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2949 one-byte characters throughout. Note that case-insensitive matching applies
2950 only to characters whose values are less than 256. PCRE doesn't support the
2951 notion of cases for higher-valued characters.
2952
2953 (i) A character class whose characters are all within 0-255 is handled as
2954 a bit map, and the map is inverted for negative classes. Previously, a
2955 character > 255 always failed to match such a class; however it should
2956 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2957
2958 (ii) A negated character class with a single character < 255 is coded as
2959 "not this character" (OP_NOT). This wasn't working properly when the test
2960 character was multibyte, either singly or repeated.
2961
2962 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2963 mode, for example: \x{100}{2,3}.
2964
2965 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2966 singly or repeated) now correctly test multibyte characters. However,
2967 PCRE doesn't recognize any characters with values greater than 255 as
2968 digits, spaces, or word characters. Such characters always match \D, \S,
2969 and \W, and never match \d, \s, or \w.
2970
2971 (v) Classes may now contain characters and character ranges with values
2972 greater than 255. For example: [ab\x{100}-\x{400}].
2973
2974 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2975 PCRE in UTF-8 mode.
2976
2977 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2978 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2979 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2980 value.)
2981
2982 53. The single man page has become too large. I have therefore split it up into
2983 a number of separate man pages. These also give rise to individual HTML pages;
2984 these are now put in a separate directory, and there is an index.html page that
2985 lists them all. Some hyperlinking between the pages has been installed.
2986
2987 54. Added convenience functions for handling named capturing parentheses.
2988
2989 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2990 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2991 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2992 are faulted.
2993
2994 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2995 calling configure. These values are used when compiling the dftables.c program
2996 which is run to generate the source of the default character tables. They
2997 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2998 you will need to set these values.
2999
3000 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3001
3002
3003 Version 3.9 02-Jan-02
3004 ---------------------
3005
3006 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3007
3008 2. If --disable-static was given, the building process failed when trying to
3009 build pcretest and pcregrep. (For some reason it was using libtool to compile
3010 them, which is not right, as they aren't part of the library.)
3011
3012
3013 Version 3.8 18-Dec-01
3014 ---------------------
3015
3016 1. The experimental UTF-8 code was completely screwed up. It was packing the
3017 bytes in the wrong order. How dumb can you get?
3018
3019
3020 Version 3.7 29-Oct-01
3021 ---------------------
3022
3023 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3024 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3025 this didn't happen under Solaris 8, where I normally test things.
3026
3027 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3028 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3029 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3030 this fix an hour or so after the initial 3.7 release.)
3031
3032
3033 Version 3.6 23-Oct-01
3034 ---------------------
3035
3036 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3037 offsets passed as NULL with zero offset count.
3038
3039 2. The config.guess and config.sub files had not been updated when I moved to
3040 the latest autoconf.
3041
3042
3043 Version 3.5 15-Aug-01
3044 ---------------------
3045
3046 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3047 had been forgotten.
3048
3049 2. By using declared but undefined structures, we can avoid using "void"
3050 definitions in pcre.h while keeping the internal definitions of the structures
3051 private.
3052
3053 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3054 user point of view, this means that both static and shared libraries are built
3055 by default, but this can be individually controlled. More of the work of
3056 handling this static/shared cases is now inside libtool instead of PCRE's make
3057 file.
3058
3059 4. The pcretest utility is now installed along with pcregrep because it is
3060 useful for users (to test regexs) and by doing this, it automatically gets
3061 relinked by libtool. The documentation has been turned into a man page, so
3062 there are now .1, .txt, and .html versions in /doc.
3063
3064 5. Upgrades to pcregrep:
3065 (i) Added long-form option names like gnu grep.
3066 (ii) Added --help to list all options with an explanatory phrase.
3067 (iii) Added -r, --recursive to recurse into sub-directories.
3068 (iv) Added -f, --file to read patterns from a file.
3069
3070 6. pcre_exec() was referring to its "code" argument before testing that
3071 argument for NULL (and giving an error if it was NULL).
3072
3073 7. Upgraded Makefile.in to allow for compiling in a different directory from
3074 the source directory.
3075
3076 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3077 options bits, the pointer it was passed was to an int instead of to an unsigned
3078 long int. This mattered only on 64-bit systems.
3079
3080 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3081 generated) instead of pcre.in, which it its source. Also made the same change
3082 in several of the .c files.
3083
3084 10. A new release of gcc defines printf() as a macro, which broke pcretest
3085 because it had an ifdef in the middle of a string argument for printf(). Fixed
3086 by using separate calls to printf().
3087
3088 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3089 script, to force use of CR or LF instead of \n in the source. On non-Unix
3090 systems, the value can be set in config.h.
3091
3092 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3093 absolute limit. Changed the text of the error message to make this clear, and
3094 likewise updated the man page.
3095
3096 13. The limit of 99 on the number of capturing subpatterns has been removed.
3097 The new limit is 65535, which I hope will not be a "real" limit.
3098
3099
3100 Version 3.4 22-Aug-00
3101 ---------------------
3102
3103 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3104
3105 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3106
3107
3108 Version 3.3 01-Aug-00
3109 ---------------------
3110
3111 1. If an octal character was given, but the value was greater than \377, it
3112 was not getting masked to the least significant bits, as documented. This could
3113 lead to crashes in some systems.
3114
3115 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3116 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3117
3118 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3119 These just pass their arguments on to (pcre_free)(), but they are provided
3120 because some uses of PCRE bind it to non-C systems that can call its functions,
3121 but cannot call free() or pcre_free() directly.
3122
3123 4. Add "make test" as a synonym for "make check". Corrected some comments in
3124 the Makefile.
3125
3126 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3127 Makefile.
3128
3129 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3130 command called pgrep for grepping around the active processes.
3131
3132 7. Added the beginnings of support for UTF-8 character strings.
3133
3134 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3135 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3136 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3137 out for the ar command.)
3138
3139
3140 Version 3.2 12-May-00
3141 ---------------------
3142
3143 This is purely a bug fixing release.
3144
3145 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3146 of ZA. This was just one example of several cases that could provoke this bug,
3147 which was introduced by change 9 of version 2.00. The code for breaking
3148 infinite loops after an iteration that matches an empty string was't working
3149 correctly.
3150
3151 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3152 when matched against abbab (for example). After matching an empty string, it
3153 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3154 caused it to match further down the string than it should.
3155
3156 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3157 was there because it doesn't seem to be needed, and it causes trouble on some
3158 systems, as it is not a Standard C header. It has been removed.
3159
3160 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3161 were reported on the Macintosh. The changes were from
3162
3163 while ((c = *(++ptr)) != 0 && c != '\n');
3164 to
3165 while ((c = *(++ptr)) != 0 && c != '\n') ;
3166
3167 Totally extraordinary, but if that's what it takes...
3168
3169 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3170 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3171 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3172 assumes the way PCRE uses memmove() (always moving upwards).
3173
3174 6. PCRE is being used in one environment where strchr() is not available. There
3175 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3176 faster code anyway.
3177
3178
3179 Version 3.1 09-Feb-00
3180 ---------------------
3181
3182 The only change in this release is the fixing of some bugs in Makefile.in for
3183 the "install" target:
3184
3185 (1) It was failing to install pcreposix.h.
3186
3187 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3188
3189
3190 Version 3.0 01-Feb-00
3191 ---------------------
3192
3193 1. Add support for the /+ modifier to perltest (to output $` like it does in
3194 pcretest).
3195
3196 2. Add support for the /g modifier to perltest.
3197
3198 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3199 matches null strings.
3200
3201 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3202 pattern. Perl treats empty patterns specially - it reuses the most recent
3203 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3204 effect.
3205
3206 5. The POSIX interface was broken in that it was just handing over the POSIX
3207 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3208 required a bigger vector, with some working space on the end. This means that
3209 the POSIX wrapper now has to get and free some memory, and copy the results.
3210
3211 6. Added some simple autoconf support, placing the test data and the
3212 documentation in separate directories, re-organizing some of the
3213 information files, and making it build pcre-config (a GNU standard). Also added
3214 libtool support for building PCRE as a shared library, which is now the
3215 default.
3216
3217 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3218 09 are not valid octal constants. Single digits will be used for minor values
3219 less than 10.
3220
3221 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3222 existing programs that set these in the POSIX interface can use PCRE without
3223 modification.
3224
3225 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3226 return all that pcre_info() returns, plus additional data. The pcre_info()
3227 function is retained for compatibility, but is considered to be obsolete.
3228
3229 10. Added experimental recursion feature (?R) to handle one common case that
3230 Perl 5.6 will be able to do with (?p{...}).
3231
3232 11. Added support for POSIX character classes like [:alpha:], which Perl is
3233 adopting.
3234
3235
3236 Version 2.08 31-Aug-99
3237 ----------------------
3238
3239 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3240 trying to match at the startoffset position, but instead was moving forward to
3241 the next newline as if a previous match had failed.
3242
3243 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3244 and could get into a loop if a null string was matched other than at the start
3245 of the subject.
3246
3247 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3248 be distinguished at compile time, and for completeness also added PCRE_DATE.
3249
3250 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3251 in GnuWin32 environments.
3252
3253
3254 Version 2.07 29-Jul-99
3255 ----------------------
3256
3257 1. The documentation is now supplied in plain text form and HTML as well as in
3258 the form of man page sources.
3259
3260 2. C++ compilers don't like assigning (void *) values to other pointer types.
3261 In particular this affects malloc(). Although there is no problem in Standard
3262 C, I've put in casts to keep C++ compilers happy.
3263
3264 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3265 should be (const char *).
3266
3267 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3268 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3269 However, I haven't made this a standard facility. The documentation doesn't
3270 mention it, and the Makefile doesn't support it.
3271
3272 5. The Makefile now contains an "install" target, with editable destinations at
3273 the top of the file. The pcretest program is not installed.
3274
3275 6. pgrep -V now gives the PCRE version number and date.
3276
3277 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3278 causing the entire string to be ignored, instead of just the last character.
3279
3280 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3281 non-matching string, it can take a very, very long time, even for strings of
3282 quite modest length, because of the nested recursion. PCRE now does better in
3283 some of these cases. It does this by remembering the last required literal
3284 character in the pattern, and pre-searching the subject to ensure it is present
3285 before running the real match. In other words, it applies a heuristic to detect
3286 some types of certain failure quickly, and in the above example, if presented
3287 with a string that has no trailing " it gives "no match" very quickly.
3288
3289 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3290 other alternatives are tried instead.
3291
3292
3293 Version 2.06 09-Jun-99
3294 ----------------------
3295
3296 1. Change pcretest's output for amount of store used to show just the code
3297 space, because the remainder (the data block) varies in size between 32-bit and
3298 64-bit systems.
3299
3300 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3301 start matching at. This allows lookbehinds to work when searching for multiple
3302 occurrences in a string.
3303
3304 3. Added additional options to pcretest for testing multiple occurrences:
3305
3306 /+ outputs the rest of the string that follows a match
3307 /g loops for multiple occurrences, using the new startoffset argument
3308 /G loops for multiple occurrences by passing an incremented pointer
3309
3310 4. PCRE wasn't doing the "first character" optimization for patterns starting
3311 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3312 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3313 the letter 'x'. On long subject strings, this gives a significant speed-up.
3314
3315
3316 Version 2.05 21-Apr-99
3317 ----------------------
3318
3319 1. Changed the type of magic_number from int to long int so that it works
3320 properly on 16-bit systems.
3321
3322 2. Fixed a bug which caused patterns starting with .* not to work correctly
3323 when the subject string contained newline characters. PCRE was assuming
3324 anchoring for such patterns in all cases, which is not correct because .* will
3325 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3326 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3327 must be retried after every newline in the subject.
3328
3329
3330 Version 2.04 18-Feb-99
3331 ----------------------
3332
3333 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3334 computation of the store needed to hold the pattern was incorrect (too large).
3335 If such patterns were nested a few deep, this could multiply and become a real
3336 problem.
3337
3338 2. Added /M option to pcretest to show the memory requirement of a specific
3339 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3340
3341 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3342 compiled in such a way that the backtracking after subsequent failure was
3343 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3344 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3345
3346
3347 Version 2.03 02-Feb-99
3348 ----------------------
3349
3350 1. Fixed typo and small mistake in man page.
3351
3352 2. Added 4th condition (GPL supersedes if conflict) and created separate
3353 LICENCE file containing the conditions.
3354
3355 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3356 Perl, that is the internal \ allows the delimiter to be included in the
3357 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3358 the final delimiter, add \ to the end of the pattern (to test the error).
3359
3360 4. Added the convenience functions for extracting substrings after a successful
3361 match. Updated pcretest to make it able to test these functions.
3362
3363
3364 Version 2.02 14-Jan-99
3365 ----------------------
3366
3367 1. Initialized the working variables associated with each extraction so that
3368 their saving and restoring doesn't refer to uninitialized store.
3369
3370 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3371 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3372 fix the problem.
3373
3374 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3375 calls, and wasn't printing the correct value for compiling calls. Increased the
3376 default value of LOOPREPEAT, and the number of significant figures in the
3377 times.
3378
3379 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3380
3381 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3382 a building problem on Windows NT with a FAT file system.
3383
3384
3385 Version 2.01 21-Oct-98
3386 ----------------------
3387
3388 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3389 to character tables built by pcre_maketables() in the current locale. If NULL
3390 is passed, the default tables are used.
3391
3392
3393 Version 2.00 24-Sep-98
3394 ----------------------
3395
3396 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3397 it any more.
3398
3399 2. Allow quantification of (?>) groups, and make it work correctly.
3400
3401 3. The first character computation wasn't working for (?>) groups.
3402
3403 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3404 end of the subject) and add 5.005's \z, which really does match only at the
3405 very end of the subject.
3406
3407 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3408
3409 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3410 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3411 localized options. All options to pcre_study() were also removed.
3412
3413 7. Add other new features from 5.005:
3414
3415 $(?<= positive lookbehind
3416 $(?<! negative lookbehind
3417 (?imsx-imsx) added the unsetting capability
3418 such a setting is global if at outer level; local otherwise
3419 (?imsx-imsx:) non-capturing groups with option setting
3420 (?(cond)re|re) conditional pattern matching
3421
3422 A backreference to itself in a repeated group matches the previous
3423 captured string.
3424
3425 8. General tidying up of studying (both automatic and via "study")
3426 consequential on the addition of new assertions.
3427
3428 9. As in 5.005, unlimited repeated groups that could match an empty substring
3429 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3430 runtime if any iteration does actually match an empty substring.
3431
3432 10. Include the RunTest script in the distribution.
3433
3434 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3435 discrepancies, some of which were old and were also with respect to 5.004. They
3436 have now been fixed.
3437
3438
3439 Version 1.09 28-Apr-98
3440 ----------------------
3441
3442 1. A negated single character class followed by a quantifier with a minimum
3443 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3444 program crashes, or just wrong answers. This did not apply to negated classes
3445 containing more than one character, or to minima other than one.
3446
3447
3448 Version 1.08 27-Mar-98
3449 ----------------------
3450
3451 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3452
3453 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3454 latter must appear before anything that relies on it in the pattern.
3455
3456
3457 Version 1.07 16-Feb-98
3458 ----------------------
3459
3460 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3461 repeat of a potentially empty string).
3462
3463
3464 Version 1.06 23-Jan-98
3465 ----------------------
3466
3467 1. Added Markus Oberhumer's little patches for C++.
3468
3469 2. Literal strings longer than 255 characters were broken.
3470
3471
3472 Version 1.05 23-Dec-97
3473 ----------------------
3474
3475 1. Negated character classes containing more than one character were failing if
3476 PCRE_CASELESS was set at run time.
3477
3478
3479 Version 1.04 19-Dec-97
3480 ----------------------
3481
3482 1. Corrected the man page, where some "const" qualifiers had been omitted.
3483
3484 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3485 input syntax.
3486
3487 3. Fixed memory leak which occurred when a regex with back references was
3488 matched with an offsets vector that wasn't big enough. The temporary memory
3489 that is used in this case wasn't being freed if the match failed.
3490
3491 4. Tidied pcretest to ensure it frees memory that it gets.
3492
3493 5. Temporary memory was being obtained in the case where the passed offsets
3494 vector was exactly big enough.
3495
3496 6. Corrected definition of offsetof() from change 5 below.
3497
3498 7. I had screwed up change 6 below and broken the rules for the use of
3499 setjmp(). Now fixed.
3500
3501
3502 Version 1.03 18-Dec-97
3503 ----------------------
3504
3505 1. A erroneous regex with a missing opening parenthesis was correctly
3506 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3507 on some systems.
3508
3509 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3510 it was reported that one broken compiler failed on the former because "code" is
3511 also an independent variable.
3512
3513 3. The erroneous regex a[]b caused an array overrun reference.
3514
3515 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3516 fail on data ending with that character. (It was going on too far, and checking
3517 the next character, typically a binary zero.) This was specific to the
3518 optimized code for single-character negative classes.
3519
3520 5. Added a contributed patch from the TIN world which does the following:
3521
3522 + Add an undef for memmove, in case the the system defines a macro for it.
3523
3524 + Add a definition of offsetof(), in case there isn't one. (I don't know
3525 the reason behind this - offsetof() is part of the ANSI standard - but
3526 it does no harm).
3527
3528 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3529 most of the places where whitespace preceded '#'. I have given up and
3530 allowed the remaining 2 cases to be at the margin.
3531
3532 + Rename some variables in pcre to eliminate shadowing. This seems very
3533 pedantic, but does no harm, of course.
3534
3535 6. Moved the call to setjmp() into its own function, to get rid of warnings
3536 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3537
3538 7. Constructs such as \d{8,} were compiling into the equivalent of
3539 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3540 outcome, but in this particular case used more store than had been allocated,
3541 which caused the bug to be discovered because it threw up an internal error.
3542
3543 8. The debugging code in both pcre and pcretest for outputting the compiled
3544 form of a regex was going wrong in the case of back references followed by
3545 curly-bracketed repeats.
3546
3547
3548 Version 1.02 12-Dec-97
3549 ----------------------
3550
3551 1. Typos in pcre.3 and comments in the source fixed.
3552
3553 2. Applied a contributed patch to get rid of places where it used to remove
3554 'const' from variables, and fixed some signed/unsigned and uninitialized
3555 variable warnings.
3556
3557 3. Added the "runtest" target to Makefile.
3558
3559 4. Set default compiler flag to -O2 rather than just -O.
3560
3561
3562 Version 1.01 19-Nov-97
3563 ----------------------
3564
3565 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3566 like /([ab]*)*/, that is, for classes with more than one character in them.
3567
3568 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3569 as /((?>a*))*/ (a PCRE_EXTRA facility).
3570
3571
3572 Version 1.00 18-Nov-97
3573 ----------------------
3574
3575 1. Added compile-time macros to support systems such as SunOS4 which don't have
3576 memmove() or strerror() but have other things that can be used instead.
3577
3578 2. Arranged that "make clean" removes the executables.
3579
3580
3581 Version 0.99 27-Oct-97
3582 ----------------------
3583
3584 1. Fixed bug in code for optimizing classes with only one character. It was
3585 initializing a 32-byte map regardless, which could cause it to run off the end
3586 of the memory it had got.
3587
3588 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3589
3590
3591 Version 0.98 22-Oct-97
3592 ----------------------
3593
3594 1. Fixed bug in code for handling temporary memory usage when there are more
3595 back references than supplied space in the ovector. This could cause segfaults.
3596
3597
3598 Version 0.97 21-Oct-97
3599 ----------------------
3600
3601 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3602
3603 2. Optimized negated single characters not to use a bit map.
3604
3605 3. Brought error texts together as macro definitions; clarified some of them;
3606 fixed one that was wrong - it said "range out of order" when it meant "invalid
3607 escape sequence".
3608
3609 4. Changed some char * arguments to const char *.
3610
3611 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3612
3613 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3614 pcretest.
3615
3616
3617 Version 0.96 16-Oct-97
3618 ----------------------
3619
3620 1. Added a simple "pgrep" utility to the distribution.
3621
3622 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3623 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3624 where "ddd" means "one or more decimal digits".
3625
3626 3. Fixed serious bug. If a pattern had a back reference, but the call to
3627 pcre_exec() didn't supply a large enough ovector to record the related
3628 identifying subpattern, the match always failed. PCRE now remembers the number
3629 of the largest back reference, and gets some temporary memory in which to save
3630 the offsets during matching if necessary, in order to ensure that
3631 backreferences always work.
3632
3633 4. Increased the compatibility with Perl in a number of ways:
3634
3635 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3636 to request this handling. The option can be set at compile or exec time.
3637
3638 (b) $ matches before a terminating newline by default; an option
3639 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3640 mode). The option can be set at compile or exec time.
3641
3642 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3643 the same as Perl's. If the decimal number it represents is less than 10
3644 or there aren't that many previous left capturing parentheses, an octal
3645 escape is read. Inside a character class, it's always an octal escape,
3646 even if it is a single digit.
3647
3648 (d) An escaped but undefined alphabetic character is taken as a literal,
3649 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3650 escapes.
3651
3652 (e) {0} is now permitted. (The previous item is removed from the compiled
3653 pattern).
3654
3655 5. Changed all the names of code files so that the basic parts are no longer
3656 than 10 characters, and abolished the teeny "globals.c" file.
3657
3658 6. Changed the handling of character classes; they are now done with a 32-byte
3659 bit map always.
3660
3661 7. Added the -d and /D options to pcretest to make it possible to look at the
3662 internals of compilation without having to recompile pcre.
3663
3664
3665 Version 0.95 23-Sep-97
3666 ----------------------
3667
3668 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3669 \x20 at the start of a run of normal characters. These were being treated as
3670 real characters, instead of the source characters being re-checked.
3671
3672
3673 Version 0.94 18-Sep-97
3674 ----------------------
3675
3676 1. The functions are now thread-safe, with the caveat that the global variables
3677 containing pointers to malloc() and free() or alternative functions are the
3678 same for all threads.
3679
3680 2. Get pcre_study() to generate a bitmap of initial characters for non-
3681 anchored patterns when this is possible, and use it if passed to pcre_exec().
3682
3683
3684 Version 0.93 15-Sep-97
3685 ----------------------
3686
3687 1. /(b)|(:+)/ was computing an incorrect first character.
3688
3689 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3690 but not actually doing anything yet.
3691
3692 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3693 as Perl does (e.g. [-az] or [az-]).
3694
3695 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3696 all possible positions.
3697
3698 5. Split up into different modules to avoid including unneeded functions in a
3699 compiled binary. However, compile and exec are still in one module. The "study"
3700 function is split off.
3701
3702 6. The character tables are now in a separate module whose source is generated
3703 by an auxiliary program - but can then be edited by hand if required. There are
3704 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3705 toupper() in the code.
3706
3707 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3708 make them global. Abolish the function for setting them, as the caller can now
3709 set them directly.
3710
3711
3712 Version 0.92 11-Sep-97
3713 ----------------------
3714
3715 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3716 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3717
3718 2. Caseless matching was not working in character classes if the characters in
3719 the pattern were in upper case.
3720
3721 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3722
3723 4. Make PCRE_ANCHORED public and accept as a compile option.
3724
3725 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3726 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3727 pass them.
3728
3729 6. Give an error if bad option bits passed at compile or run time.
3730
3731 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3732 pcretest to cause it to pass that flag.
3733
3734 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3735 options, and the first character, if set.
3736
3737 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3738
3739
3740 Version 0.91 10-Sep-97
3741 ----------------------
3742
3743 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3744 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3745
3746 2. PCRE was looping on encountering an indefinitely repeated back reference to
3747 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3748 Perl does - treats the match as successful.
3749
3750 ****

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