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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 476 - (show annotations) (download)
Sun Jan 3 15:37:24 2010 UTC (4 years, 7 months ago) by ph10
File size: 177577 byte(s)
Add casts to avoid compiler warnings.

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66 was out-of-date, and there was no check on the pcre_compile() error code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

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