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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 437 - (show annotations)
Sun Sep 6 16:34:42 2009 UTC (5 years, 9 months ago) by ph10
File size: 171359 byte(s)
Error occurred while calculating annotation data.
Add test for diff -u to RunGrepTest.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.00 ??-???-??
5 ----------------------
6
7 1. The table for translating pcre_compile() error codes into POSIX error codes
8 was out-of-date, and there was no check on the pcre_compile() error code
9 being within the table. This could lead to an OK return being given in
10 error.
11
12 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
13 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
14 in a Windows environment.
15
16 3. The pcregrep --count option prints the count for each file even when it is
17 zero, as does GNU grep. However, pcregrep was also printing all files when
18 --files-with-matches was added. Now, when both options are given, it prints
19 counts only for those files that have at least one match. (GNU grep just
20 prints the file name in this circumstance, but including the count seems
21 more useful - otherwise, why use --count?) Also ensured that the
22 combination -clh just lists non-zero counts, with no names.
23
24 4. The long form of the pcregrep -F option was incorrectly implemented as
25 --fixed_strings instead of --fixed-strings. This is an incompatible change,
26 but it seems right to fix it, and I didn't think it was worth preserving
27 the old behaviour.
28
29 5. The command line items --regex=pattern and --regexp=pattern were not
30 recognized by pcregrep, which required --regex pattern or --regexp pattern
31 (with a space rather than an '='). The man page documented the '=' forms,
32 which are compatible with GNU grep; these now work.
33
34 6. No libpcreposix.pc file was created for pkg-config; there was just
35 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
36
37 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
38 when UCP support is not needed, by modifying the Python script that
39 generates it from Unicode data files. This should not matter if the module
40 is correctly used as a library, but I received one complaint about 50K of
41 unwanted data. My guess is that the person linked everything into his
42 program rather than using a library. Anyway, it does no harm.
43
44 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
45 was a minimum greater than 1 for a wide character in a possessive
46 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
47 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
48 character. Chaos in the form of incorrect output or a compiling loop could
49 result.
50
51 9. The restrictions on what a pattern can contain when partial matching is
52 requested for pcre_exec() have been removed. All patterns can now be
53 partially matched by this function. In addition, if there are at least two
54 slots in the offset vector, the offset of the earliest inspected character
55 for the match and the offset of the end of the subject are set in them when
56 PCRE_ERROR_PARTIAL is returned.
57
58 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
59 synonymous with PCRE_PARTIAL, for backwards compatibility, and
60 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
61 and may be more useful for multi-segment matching, especially with
62 pcre_exec().
63
64 11. Partial matching with pcre_exec() is now more intuitive. A partial match
65 used to be given if ever the end of the subject was reached; now it is
66 given only if matching could not proceed because another character was
67 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
68 string "Z", which now yields "no match" instead of "partial match". In the
69 case of pcre_dfa_exec(), "no match" is given if every matching path for the
70 final character ended with (*FAIL).
71
72 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
73 if the pattern had a "must contain" character that was already found in the
74 earlier partial match, unless partial matching was again requested. For
75 example, with the pattern /dog.(body)?/, the "must contain" character is
76 "g". If the first part-match was for the string "dog", restarting with
77 "sbody" failed. This bug has been fixed.
78
79 13. The string returned by pcre_dfa_exec() after a partial match has been
80 changed so that it starts at the first inspected character rather than the
81 first character of the match. This makes a difference only if the pattern
82 starts with a lookbehind assertion or \b or \B (\K is not supported by
83 pcre_dfa_exec()). It's an incompatible change, but it makes the two
84 matching functions compatible, and I think it's the right thing to do.
85
86 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
87 so that the demonstration program is easily available in environments where
88 PCRE has not been installed from source.
89
90 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
91 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
92 library.
93
94 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
95 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
96 is not the first non-POSIX option to be added. Clearly some people find
97 these options useful.
98
99 17. If a caller to the POSIX matching function regexec() passes a non-zero
100 value for \fInmatch\fP with a NULL value for \fIpmatch\fP, the value of
101 \fInmatch\fP is forced to zero.
102
103 18. RunGrepTest did not have a test for the availability of the -u option of
104 the diff command, as RunTest does. It now checks in the same way as
105 RunTest, and also checks for the -b option.
106
107
108 Version 7.9 11-Apr-09
109 ---------------------
110
111 1. When building with support for bzlib/zlib (pcregrep) and/or readline
112 (pcretest), all targets were linked against these libraries. This included
113 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
114 libraries. This caused unwanted dependencies to be created. This problem
115 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
116 pcretest is linked with readline.
117
118 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
119 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
120 moved outside it again, because FALSE and TRUE are already defined in AIX,
121 but BOOL is not.
122
123 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
124 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
125
126 4. The pcregrep documentation said spaces were inserted as well as colons (or
127 hyphens) following file names and line numbers when outputting matching
128 lines. This is not true; no spaces are inserted. I have also clarified the
129 wording for the --colour (or --color) option.
130
131 5. In pcregrep, when --colour was used with -o, the list of matching strings
132 was not coloured; this is different to GNU grep, so I have changed it to be
133 the same.
134
135 6. When --colo(u)r was used in pcregrep, only the first matching substring in
136 each matching line was coloured. Now it goes on to look for further matches
137 of any of the test patterns, which is the same behaviour as GNU grep.
138
139 7. A pattern that could match an empty string could cause pcregrep to loop; it
140 doesn't make sense to accept an empty string match in pcregrep, so I have
141 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
142 seems to be how GNU grep behaves.
143
144 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
145 start or after a newline", because the conditional assertion was not being
146 correctly handled. The rule now is that both the assertion and what follows
147 in the first alternative must satisfy the test.
148
149 9. If auto-callout was enabled in a pattern with a conditional group whose
150 condition was an assertion, PCRE could crash during matching, both with
151 pcre_exec() and pcre_dfa_exec().
152
153 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
154 used for matching.
155
156 11. Unicode property support in character classes was not working for
157 characters (bytes) greater than 127 when not in UTF-8 mode.
158
159 12. Added the -M command line option to pcretest.
160
161 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
162
163 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
164
165 16. Added comments and documentation about mis-use of no_arg in the C++
166 wrapper.
167
168 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
169 from Martin Jerabek that uses macro names for all relevant character and
170 string constants.
171
172 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
173 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
174 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
175 these, but not everybody uses configure.
176
177 19. A conditional group that had only one branch was not being correctly
178 recognized as an item that could match an empty string. This meant that an
179 enclosing group might also not be so recognized, causing infinite looping
180 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
181 with the subject "ab", where knowledge that the repeated group can match
182 nothing is needed in order to break the loop.
183
184 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
185 exec(), but without supplying a callout function, matching went wrong.
186
187 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
188 leak if the size of the offset vector was greater than 30. When the vector
189 is smaller, the saved offsets during recursion go onto a local stack
190 vector, but for larger vectors malloc() is used. It was failing to free
191 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
192 error, in fact).
193
194 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
195 heapframe that is used only when UTF-8 support is enabled. This caused no
196 problem, but was untidy.
197
198 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
199 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
200 included within another project.
201
202 24. Steven Van Ingelgem's patches to add more options to the CMake support,
203 slightly modified by me:
204
205 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
206 not building pcregrep.
207
208 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
209 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
210
211 25. Forward references, both numeric and by name, in patterns that made use of
212 duplicate group numbers, could behave incorrectly or give incorrect errors,
213 because when scanning forward to find the reference group, PCRE was not
214 taking into account the duplicate group numbers. A pattern such as
215 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
216
217 26. Changed a few more instances of "const unsigned char *" to USPTR, making
218 the feature of a custom pointer more persuasive (as requested by a user).
219
220 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
221 pcretest.c, inside #ifndefs, because it seems they are sometimes already
222 pre-defined.
223
224 28. Added support for (*UTF8) at the start of a pattern.
225
226 29. Arrange for flags added by the "release type" setting in CMake to be shown
227 in the configuration summary.
228
229
230 Version 7.8 05-Sep-08
231 ---------------------
232
233 1. Replaced UCP searching code with optimized version as implemented for Ad
234 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
235 stage table and inline lookup instead of a function, giving speed ups of 2
236 to 5 times on some simple patterns that I tested. Permission was given to
237 distribute the MultiStage2.py script that generates the tables (it's not in
238 the tarball, but is in the Subversion repository).
239
240 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
241 scripts.
242
243 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
244 a group with a zero qualifier. The result of the study could be incorrect,
245 or the function might crash, depending on the pattern.
246
247 4. Caseless matching was not working for non-ASCII characters in back
248 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
249 It now works when Unicode Property Support is available.
250
251 5. In pcretest, an escape such as \x{de} in the data was always generating
252 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
253 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
254 truncation.
255
256 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
257
258 7. Added two (int) casts to pcregrep when printing the difference of two
259 pointers, in case they are 64-bit values.
260
261 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
262 test 2 if it fails.
263
264 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
265 and a #define of that name to empty if it is not externally set. This is to
266 allow users of MSVC to set it if necessary.
267
268 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
269 the convenience functions in the pcre_get.c source file.
270
271 11. An option change at the start of a pattern that had top-level alternatives
272 could cause overwriting and/or a crash. This command provoked a crash in
273 some environments:
274
275 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
276
277 This potential security problem was recorded as CVE-2008-2371.
278
279 12. For a pattern where the match had to start at the beginning or immediately
280 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
281 pcre_dfa_exec() could read past the end of the passed subject if there was
282 no match. To help with detecting such bugs (e.g. with valgrind), I modified
283 pcretest so that it places the subject at the end of its malloc-ed buffer.
284
285 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
286 exec() might read past the end of the data buffer in UTF-8 mode.
287
288 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
289 the data contained the byte 0x85 as part of a UTF-8 character within its
290 first line. This applied both to normal and DFA matching.
291
292 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
293 /^[^d]*?$/8 failed to match "abc".
294
295 16. Added a missing copyright notice to pcrecpp_internal.h.
296
297 17. Make it more clear in the documentation that values returned from
298 pcre_exec() in ovector are byte offsets, not character counts.
299
300 18. Tidied a few places to stop certain compilers from issuing warnings.
301
302 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
303 supplied by Stefan Weber. I made a further small update for 7.8 because
304 there is a change of source arrangements: the pcre_searchfuncs.c module is
305 replaced by pcre_ucd.c.
306
307
308 Version 7.7 07-May-08
309 ---------------------
310
311 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
312 a string to a long long, pretend we don't even have a long long." This is
313 done by checking for the strtoq, strtoll, and _strtoi64 functions.
314
315 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
316 pre-7.6 versions, which defined a global no_arg variable instead of putting
317 it in the RE class. (See also #8 below.)
318
319 3. Remove a line of dead code, identified by coverity and reported by Nuno
320 Lopes.
321
322 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
323
324 (1) The include/exclude patterns were being applied to the whole pathnames
325 of files, instead of just to the final components.
326
327 (2) If there was more than one level of directory, the subdirectories were
328 skipped unless they satisfied the include/exclude conditions. This is
329 inconsistent with GNU grep (and could even be seen as contrary to the
330 pcregrep specification - which I improved to make it absolutely clear).
331 The action now is always to scan all levels of directory, and just
332 apply the include/exclude patterns to regular files.
333
334 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
335 --exclude_dir in the tests to avoid scanning .svn directories.
336
337 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
338 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
339 doesn't support NULs in patterns.
340
341 7. Added some missing "const"s to declarations of static tables in
342 pcre_compile.c and pcre_dfa_exec.c.
343
344 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
345 caused by fix #2 above. (Subsequently also a second patch to fix the
346 first patch. And a third patch - this was a messy problem.)
347
348 9. Applied Craig's patch to remove the use of push_back().
349
350 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
351 matching function regexec().
352
353 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
354 which, however, unlike Perl's \g{...}, are subroutine calls, not back
355 references. PCRE supports relative numbers with this syntax (I don't think
356 Oniguruma does).
357
358 12. Previously, a group with a zero repeat such as (...){0} was completely
359 omitted from the compiled regex. However, this means that if the group
360 was called as a subroutine from elsewhere in the pattern, things went wrong
361 (an internal error was given). Such groups are now left in the compiled
362 pattern, with a new opcode that causes them to be skipped at execution
363 time.
364
365 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
366 to the way PCRE behaves:
367
368 (a) A lone ] character is dis-allowed (Perl treats it as data).
369
370 (b) A back reference to an unmatched subpattern matches an empty string
371 (Perl fails the current match path).
372
373 (c) A data ] in a character class must be notated as \] because if the
374 first data character in a class is ], it defines an empty class. (In
375 Perl it is not possible to have an empty class.) The empty class []
376 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
377 The negative empty class [^] matches any one character, independently
378 of the DOTALL setting.
379
380 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
381 non-existent subpattern following a character class starting with ']' and
382 containing () gave an internal compiling error instead of "reference to
383 non-existent subpattern". Fortunately, when the pattern did exist, the
384 compiled code was correct. (When scanning forwards to check for the
385 existencd of the subpattern, it was treating the data ']' as terminating
386 the class, so got the count wrong. When actually compiling, the reference
387 was subsequently set up correctly.)
388
389 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
390 it was being rejected as not supported by pcre_dfa_exec(), even though
391 other assertions are supported. I have made pcre_dfa_exec() support
392 (*FAIL).
393
394 16. The implementation of 13c above involved the invention of a new opcode,
395 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
396 cannot be changed at match time, I realized I could make a small
397 improvement to matching performance by compiling OP_ALLANY instead of
398 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
399 on the OP_ANY path.
400
401 17. Compiling pcretest on Windows with readline support failed without the
402 following two fixes: (1) Make the unistd.h include conditional on
403 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
404
405 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
406 ncurses library to be included for pcretest when ReadLine support is
407 requested, but also to allow for it to be overridden. This patch came from
408 Daniel Bergström.
409
410 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
411 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
412 any errors with the current Unicode tables. Thanks to Peter Kankowski for
413 spotting this.
414
415
416 Version 7.6 28-Jan-08
417 ---------------------
418
419 1. A character class containing a very large number of characters with
420 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
421 overflow.
422
423 2. Patch to cut out the "long long" test in pcrecpp_unittest when
424 HAVE_LONG_LONG is not defined.
425
426 3. Applied Christian Ehrlicher's patch to update the CMake build files to
427 bring them up to date and include new features. This patch includes:
428
429 - Fixed PH's badly added libz and libbz2 support.
430 - Fixed a problem with static linking.
431 - Added pcredemo. [But later removed - see 7 below.]
432 - Fixed dftables problem and added an option.
433 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
434 HAVE_LONG_LONG.
435 - Added readline support for pcretest.
436 - Added an listing of the option settings after cmake has run.
437
438 4. A user submitted a patch to Makefile that makes it easy to create
439 "pcre.dll" under mingw when using Configure/Make. I added stuff to
440 Makefile.am that cause it to include this special target, without
441 affecting anything else. Note that the same mingw target plus all
442 the other distribution libraries and programs are now supported
443 when configuring with CMake (see 6 below) instead of with
444 Configure/Make.
445
446 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
447 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
448 exported in the Windows port". It has not yet been confirmed that the patch
449 solves the problem, but it does no harm.
450
451 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
452 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
453 with CMake, and also correct the comment about stack recursion.
454
455 7. Remove the automatic building of pcredemo from the ./configure system and
456 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
457 of a program that users should build themselves after PCRE is installed, so
458 building it automatically is not really right. What is more, it gave
459 trouble in some build environments.
460
461 8. Further tidies to CMakeLists.txt from Sheri and Christian.
462
463
464 Version 7.5 10-Jan-08
465 ---------------------
466
467 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
468 values in parens when parsing an RE using the C++ wrapper."
469
470 2. Negative specials like \S did not work in character classes in UTF-8 mode.
471 Characters greater than 255 were excluded from the class instead of being
472 included.
473
474 3. The same bug as (2) above applied to negated POSIX classes such as
475 [:^space:].
476
477 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
478 defined or documented. It seems to have been a typo for PCRE_STATIC, so
479 I have changed it.
480
481 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
482 first named subpattern) and a construct such as (?&a) would reference the
483 first named subpattern whose name started with "a" (in other words, the
484 length check was missing). Both these problems are fixed. "Subpattern name
485 expected" is now given for (?&) (a zero-length name), and this patch also
486 makes it give the same error for \k'' (previously it complained that that
487 was a reference to a non-existent subpattern).
488
489 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
490 this is right because (?- can be followed by option settings as well as by
491 digits. I have, however, made the messages clearer.
492
493 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
494 than the number used in the conditional) now cause a compile-time error.
495 This is actually not compatible with Perl, which accepts such patterns, but
496 treats the conditional as always being FALSE (as PCRE used to), but it
497 seems to me that giving a diagnostic is better.
498
499 8. Change "alphameric" to the more common word "alphanumeric" in comments
500 and messages.
501
502 9. Fix two occurrences of "backslash" in comments that should have been
503 "backspace".
504
505 10. Remove two redundant lines of code that can never be obeyed (their function
506 was moved elsewhere).
507
508 11. The program that makes PCRE's Unicode character property table had a bug
509 which caused it to generate incorrect table entries for sequences of
510 characters that have the same character type, but are in different scripts.
511 It amalgamated them into a single range, with the script of the first of
512 them. In other words, some characters were in the wrong script. There were
513 thirteen such cases, affecting characters in the following ranges:
514
515 U+002b0 - U+002c1
516 U+0060c - U+0060d
517 U+0061e - U+00612
518 U+0064b - U+0065e
519 U+0074d - U+0076d
520 U+01800 - U+01805
521 U+01d00 - U+01d77
522 U+01d9b - U+01dbf
523 U+0200b - U+0200f
524 U+030fc - U+030fe
525 U+03260 - U+0327f
526 U+0fb46 - U+0fbb1
527 U+10450 - U+1049d
528
529 12. The -o option (show only the matching part of a line) for pcregrep was not
530 compatible with GNU grep in that, if there was more than one match in a
531 line, it showed only the first of them. It now behaves in the same way as
532 GNU grep.
533
534 13. If the -o and -v options were combined for pcregrep, it printed a blank
535 line for every non-matching line. GNU grep prints nothing, and pcregrep now
536 does the same. The return code can be used to tell if there were any
537 non-matching lines.
538
539 14. Added --file-offsets and --line-offsets to pcregrep.
540
541 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
542 infinitely looping recursion. The bug was that positive lookaheads were not
543 being skipped when checking for a possible empty match (negative lookaheads
544 and both kinds of lookbehind were skipped).
545
546 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
547 inclusion of <windows.h> to before rather than after the definition of
548 INVALID_FILE_ATTRIBUTES (patch from David Byron).
549
550 17. Specifying a possessive quantifier with a specific limit for a Unicode
551 character property caused pcre_compile() to compile bad code, which led at
552 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
553 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
554 caused the error; without that there was no problem.
555
556 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
557
558 19. Added --enable-pcretest-libreadline.
559
560 20. In pcrecpp.cc, the variable 'count' was incremented twice in
561 RE::GlobalReplace(). As a result, the number of replacements returned was
562 double what it should be. I removed one of the increments, but Craig sent a
563 later patch that removed the other one (the right fix) and added unit tests
564 that check the return values (which was not done before).
565
566 21. Several CMake things:
567
568 (1) Arranged that, when cmake is used on Unix, the libraries end up with
569 the names libpcre and libpcreposix, not just pcre and pcreposix.
570
571 (2) The above change means that pcretest and pcregrep are now correctly
572 linked with the newly-built libraries, not previously installed ones.
573
574 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
575
576 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
577 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
578 UTF-8 newline character). The key issue is that the pattern starts .*;
579 this means that the match must be either at the beginning, or after a
580 newline. The bug was in the code for advancing after a failed match and
581 checking that the new position followed a newline. It was not taking
582 account of UTF-8 characters correctly.
583
584 23. PCRE was behaving differently from Perl in the way it recognized POSIX
585 character classes. PCRE was not treating the sequence [:...:] as a
586 character class unless the ... were all letters. Perl, however, seems to
587 allow any characters between [: and :], though of course it rejects as
588 unknown any "names" that contain non-letters, because all the known class
589 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
590 for example, whereas PCRE did not - it did not recognize a POSIX character
591 class. This seemed a bit dangerous, so the code has been changed to be
592 closer to Perl. The behaviour is not identical to Perl, because PCRE will
593 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
594 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
595 Perl does, and where it didn't before.
596
597 24. Rewrite so as to remove the single use of %n from pcregrep because in some
598 Windows environments %n is disabled by default.
599
600
601 Version 7.4 21-Sep-07
602 ---------------------
603
604 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
605 means that a class such as [\s] counted as "explicit reference to CR or
606 LF". That isn't really right - the whole point of the change was to try to
607 help when there was an actual mention of one of the two characters. So now
608 the change happens only if \r or \n (or a literal CR or LF) character is
609 encountered.
610
611 2. The 32-bit options word was also used for 6 internal flags, but the numbers
612 of both had grown to the point where there were only 3 bits left.
613 Fortunately, there was spare space in the data structure, and so I have
614 moved the internal flags into a new 16-bit field to free up more option
615 bits.
616
617 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
618 but did not set the internal JCHANGED flag - either of these is enough to
619 control the way the "get" function works - but the PCRE_INFO_JCHANGED
620 facility is supposed to tell if (?J) was ever used, so now (?J) at the
621 start sets both bits.
622
623 4. Added options (at build time, compile time, exec time) to change \R from
624 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
625
626 5. doc/pcresyntax.html was missing from the distribution.
627
628 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
629 compatibility, even though it is no longer used.
630
631 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
632 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
633 windows.h file is present (where different names are used). [This was
634 reversed later after testing - see 16 below.]
635
636 8. Changed all #include <config.h> to #include "config.h". There were also
637 some further <pcre.h> cases that I changed to "pcre.h".
638
639 9. When pcregrep was used with the --colour option, it missed the line ending
640 sequence off the lines that it output.
641
642 10. It was pointed out to me that arrays of string pointers cause lots of
643 relocations when a shared library is dynamically loaded. A technique of
644 using a single long string with a table of offsets can drastically reduce
645 these. I have refactored PCRE in four places to do this. The result is
646 dramatic:
647
648 Originally: 290
649 After changing UCP table: 187
650 After changing error message table: 43
651 After changing table of "verbs" 36
652 After changing table of Posix names 22
653
654 Thanks to the folks working on Gregex for glib for this insight.
655
656 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
657 unicode-properties was also set.
658
659 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
660
661 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
662 checked only for CRLF.
663
664 14. Added casts to pcretest.c to avoid compiler warnings.
665
666 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
667
668 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
669 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
670 entirely. This removes changes made in 7 above.
671
672 17. The CMake files have been updated, and there is now more information about
673 building with CMake in the NON-UNIX-USE document.
674
675
676 Version 7.3 28-Aug-07
677 ---------------------
678
679 1. In the rejigging of the build system that eventually resulted in 7.1, the
680 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
681 brackets there is not right, since it causes compilers to look for an
682 installed pcre.h, not the version that is in the source that is being
683 compiled (which of course may be different). I have changed it back to:
684
685 #include "pcre.h"
686
687 I have a vague recollection that the change was concerned with compiling in
688 different directories, but in the new build system, that is taken care of
689 by the VPATH setting the Makefile.
690
691 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
692 when the subject happened to end in the byte 0x85 (e.g. if the last
693 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
694 characters but of course it shouldn't be taken as a newline when it is part
695 of another character. The bug was that, for an unlimited repeat of . in
696 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
697 characters when looking for a newline.
698
699 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
700
701 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
702 in debug output.
703
704 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
705 long printing in the pcrecpp unittest when running under MinGW.
706
707 6. ESC_K was left out of the EBCDIC table.
708
709 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
710 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
711 limit also applies to "virtual nesting" when a pattern is recursive, and in
712 this case 1000 isn't so big. I have been able to remove this limit at the
713 expense of backing off one optimization in certain circumstances. Normally,
714 when pcre_exec() would call its internal match() function recursively and
715 immediately return the result unconditionally, it uses a "tail recursion"
716 feature to save stack. However, when a subpattern that can match an empty
717 string has an unlimited repetition quantifier, it no longer makes this
718 optimization. That gives it a stack frame in which to save the data for
719 checking that an empty string has been matched. Previously this was taken
720 from the 1000-entry workspace that had been reserved. So now there is no
721 explicit limit, but more stack is used.
722
723 8. Applied Daniel's patches to solve problems with the import/export magic
724 syntax that is required for Windows, and which was going wrong for the
725 pcreposix and pcrecpp parts of the library. These were overlooked when this
726 problem was solved for the main library.
727
728 9. There were some crude static tests to avoid integer overflow when computing
729 the size of patterns that contain repeated groups with explicit upper
730 limits. As the maximum quantifier is 65535, the maximum group length was
731 set at 30,000 so that the product of these two numbers did not overflow a
732 32-bit integer. However, it turns out that people want to use groups that
733 are longer than 30,000 bytes (though not repeat them that many times).
734 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
735 made it possible to implement the integer overflow checks in a much more
736 dynamic way, which I have now done. The artificial limitation on group
737 length has been removed - we now have only the limit on the total length of
738 the compiled pattern, which depends on the LINK_SIZE setting.
739
740 10. Fixed a bug in the documentation for get/copy named substring when
741 duplicate names are permitted. If none of the named substrings are set, the
742 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
743 empty string.
744
745 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
746 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
747 because the ] is interpreted as the first data character and the
748 terminating ] is not found. PCRE has been made compatible with Perl in this
749 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
750 cause memory overwriting.
751
752 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
753 string has been matched (to stop an infinite loop). It was not recognizing
754 a conditional subpattern that could match an empty string if that
755 subpattern was within another subpattern. For example, it looped when
756 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
757 condition was not nested. This bug has been fixed.
758
759 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
760 past the start of the subject in the presence of bytes with the top bit
761 set, for example "\x8aBCD".
762
763 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
764 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
765
766 14. Optimized (?!) to (*FAIL).
767
768 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
769 This restricts code points to be within the range 0 to 0x10FFFF, excluding
770 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
771 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
772 does: it's just the validity check that is more restrictive.
773
774 16. Inserted checks for integer overflows during escape sequence (backslash)
775 processing, and also fixed erroneous offset values for syntax errors during
776 backslash processing.
777
778 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
779 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
780
781 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
782 caused an overrun.
783
784 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
785 something other than just ASCII characters) inside a group that had an
786 unlimited repeat caused a loop at compile time (while checking to see
787 whether the group could match an empty string).
788
789 20. Debugging a pattern containing \p or \P could cause a crash. For example,
790 [\P{Any}] did so. (Error in the code for printing property names.)
791
792 21. An orphan \E inside a character class could cause a crash.
793
794 22. A repeated capturing bracket such as (A)? could cause a wild memory
795 reference during compilation.
796
797 23. There are several functions in pcre_compile() that scan along a compiled
798 expression for various reasons (e.g. to see if it's fixed length for look
799 behind). There were bugs in these functions when a repeated \p or \P was
800 present in the pattern. These operators have additional parameters compared
801 with \d, etc, and these were not being taken into account when moving along
802 the compiled data. Specifically:
803
804 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
805 length.
806
807 (b) An item such as \pL+ within a repeated group could cause crashes or
808 loops.
809
810 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
811 "reference to non-existent subpattern" error.
812
813 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
814
815 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
816 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
817
818 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
819
820 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
821 character were causing crashes (broken optimization).
822
823 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
824 \p or \P) caused a compile-time loop.
825
826 28. More problems have arisen in unanchored patterns when CRLF is a valid line
827 break. For example, the unstudied pattern [\r\n]A does not match the string
828 "\r\nA" because change 7.0/46 below moves the current point on by two
829 characters after failing to match at the start. However, the pattern \nA
830 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
831 the same is true. There doesn't seem any very clean way out of this, but
832 what I have chosen to do makes the common cases work: PCRE now takes note
833 of whether there can be an explicit match for \r or \n anywhere in the
834 pattern, and if so, 7.0/46 no longer applies. As part of this change,
835 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
836 pattern has explicit CR or LF references.
837
838 29. Added (*CR) etc for changing newline setting at start of pattern.
839
840
841 Version 7.2 19-Jun-07
842 ---------------------
843
844 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
845 which is apparently normally available under Windows.
846
847 2. Re-jig the pcregrep tests with different newline settings in an attempt
848 to make them independent of the local environment's newline setting.
849
850 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
851
852 4. Some of the "internals" tests were previously cut out when the link size
853 was not 2, because the output contained actual offsets. The recent new
854 "Z" feature of pcretest means that these can be cut out, making the tests
855 usable with all link sizes.
856
857 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
858 stack recursion. This gives a massive performance boost under BSD, but just
859 a small improvement under Linux. However, it saves one field in the frame
860 in all cases.
861
862 6. Added more features from the forthcoming Perl 5.10:
863
864 (a) (?-n) (where n is a string of digits) is a relative subroutine or
865 recursion call. It refers to the nth most recently opened parentheses.
866
867 (b) (?+n) is also a relative subroutine call; it refers to the nth next
868 to be opened parentheses.
869
870 (c) Conditions that refer to capturing parentheses can be specified
871 relatively, for example, (?(-2)... or (?(+3)...
872
873 (d) \K resets the start of the current match so that everything before
874 is not part of it.
875
876 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
877
878 (f) \g{name} is another synonym - part of Perl 5.10's unification of
879 reference syntax.
880
881 (g) (?| introduces a group in which the numbering of parentheses in each
882 alternative starts with the same number.
883
884 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
885
886 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
887 PCRE_INFO_JCHANGED.
888
889 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
890 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
891 for detecting groups that can match an empty string.
892
893 9. A pattern with a very large number of alternatives (more than several
894 hundred) was running out of internal workspace during the pre-compile
895 phase, where pcre_compile() figures out how much memory will be needed. A
896 bit of new cunning has reduced the workspace needed for groups with
897 alternatives. The 1000-alternative test pattern now uses 12 bytes of
898 workspace instead of running out of the 4096 that are available.
899
900 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
901
902 11. Applied patch from Google to remove an optimization that didn't quite work.
903 The report of the bug said:
904
905 pcrecpp::RE("a*").FullMatch("aaa") matches, while
906 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
907 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
908
909 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
910 it matched the wrong number of bytes.
911
912
913 Version 7.1 24-Apr-07
914 ---------------------
915
916 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
917 that is more "standard", making use of automake and other Autotools. There
918 is some re-arrangement of the files and adjustment of comments consequent
919 on this.
920
921 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
922 for recursive directory scanning broke on some systems because the files
923 are not scanned in any specific order and on different systems the order
924 was different. A call to "sort" has been inserted into RunGrepTest for the
925 approprate test as a short-term fix. In the longer term there may be an
926 alternative.
927
928 3. I had an email from Eric Raymond about problems translating some of PCRE's
929 man pages to HTML (despite the fact that I distribute HTML pages, some
930 people do their own conversions for various reasons). The problems
931 concerned the use of low-level troff macros .br and .in. I have therefore
932 removed all such uses from the man pages (some were redundant, some could
933 be replaced by .nf/.fi pairs). The 132html script that I use to generate
934 HTML has been updated to handle .nf/.fi and to complain if it encounters
935 .br or .in.
936
937 4. Updated comments in configure.ac that get placed in config.h.in and also
938 arranged for config.h to be included in the distribution, with the name
939 config.h.generic, for the benefit of those who have to compile without
940 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
941
942 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
943 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
944 makevp.bat for the current PCRE, using the additional files
945 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
946
947 6. A Windows user reported a minor discrepancy with test 2, which turned out
948 to be caused by a trailing space on an input line that had got lost in his
949 copy. The trailing space was an accident, so I've just removed it.
950
951 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
952 that is needed.
953
954 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
955 as "const" (a) because they are and (b) because it helps the PHP
956 maintainers who have recently made a script to detect big data structures
957 in the php code that should be moved to the .rodata section. I remembered
958 to update Builducptable as well, so it won't revert if ucptable.h is ever
959 re-created.
960
961 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
962 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
963 order to be able to cut out the UTF-8 tables in the latter when UTF-8
964 support is not required. This saves 1.5-2K of code, which is important in
965 some applications.
966
967 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
968 so as not to refer to the tables, even though these functions will never be
969 called when UTF-8 support is disabled. Otherwise there are problems with a
970 shared library.
971
972 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
973
974 (a) It was defining its arguments as char * instead of void *.
975
976 (b) It was assuming that all moves were upwards in memory; this was true
977 a long time ago when I wrote it, but is no longer the case.
978
979 The emulated memove() is provided for those environments that have neither
980 memmove() nor bcopy(). I didn't think anyone used it these days, but that
981 is clearly not the case, as these two bugs were recently reported.
982
983 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
984 and Detrail to create the HTML documentation, the .txt form of the man
985 pages, and it removes trailing spaces from listed files. It also creates
986 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
987 case, it wraps all the #defines with #ifndefs. This script should be run
988 before "make dist".
989
990 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
991 with Unicode property support.
992
993 (a) For a maximizing quantifier, if the two different cases of the
994 character were of different lengths in their UTF-8 codings (there are
995 some cases like this - I found 11), and the matching function had to
996 back up over a mixture of the two cases, it incorrectly assumed they
997 were both the same length.
998
999 (b) When PCRE was configured to use the heap rather than the stack for
1000 recursion during matching, it was not correctly preserving the data for
1001 the other case of a UTF-8 character when checking ahead for a match
1002 while processing a minimizing repeat. If the check also involved
1003 matching a wide character, but failed, corruption could cause an
1004 erroneous result when trying to check for a repeat of the original
1005 character.
1006
1007 13. Some tidying changes to the testing mechanism:
1008
1009 (a) The RunTest script now detects the internal link size and whether there
1010 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1011 values substituted by "configure". (The RunGrepTest script already did
1012 this for UTF-8.) The configure.ac script no longer substitutes the
1013 relevant variables.
1014
1015 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1016 with length and offset values. This means that the output is different
1017 for different internal link sizes. Test 2 is skipped for link sizes
1018 other than 2 because of this, bypassing the problem. Unfortunately,
1019 there was also a test in test 3 (the locale tests) that used /B and
1020 failed for link sizes other than 2. Rather than cut the whole test out,
1021 I have added a new /Z option to pcretest that replaces the length and
1022 offset values with spaces. This is now used to make test 3 independent
1023 of link size. (Test 2 will be tidied up later.)
1024
1025 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1026 segmentation fault instead of returning the appropriate error message.
1027
1028 15. In multiline mode when the newline sequence was set to "any", the pattern
1029 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1030 This doesn't seem right; it now treats the CRLF combination as the line
1031 ending, and so does not match in that case. It's only a pattern such as ^$
1032 that would hit this one: something like ^ABC$ would have failed after \r
1033 and then tried again after \r\n.
1034
1035 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1036 in an attempt to make files that differ only in their line terminators
1037 compare equal. This works on Linux.
1038
1039 17. Under certain error circumstances pcregrep might try to free random memory
1040 as it exited. This is now fixed, thanks to valgrind.
1041
1042 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1043 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1044 was because its rules for how to advance for /g after matching an empty
1045 string at the end of a line did not allow for this case. They now check for
1046 it specially.
1047
1048 20. pcretest is supposed to handle patterns and data of any length, by
1049 extending its buffers when necessary. It was getting this wrong when the
1050 buffer for a data line had to be extended.
1051
1052 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1053 CRLF as a newline sequence.
1054
1055 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1056 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1057 I have nevertheless tidied it up.
1058
1059 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1060
1061 24. Added a man page for pcre-config.
1062
1063
1064 Version 7.0 19-Dec-06
1065 ---------------------
1066
1067 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1068 moving to gcc 4.1.1.
1069
1070 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1071 sys/time.h, which is documented as needed for this function. It doesn't
1072 seem to matter on Linux, but it showed up on some releases of OS X.
1073
1074 3. It seems that there are systems where bytes whose values are greater than
1075 127 match isprint() in the "C" locale. The "C" locale should be the
1076 default when a C program starts up. In most systems, only ASCII printing
1077 characters match isprint(). This difference caused the output from pcretest
1078 to vary, making some of the tests fail. I have changed pcretest so that:
1079
1080 (a) When it is outputting text in the compiled version of a pattern, bytes
1081 other than 32-126 are always shown as hex escapes.
1082
1083 (b) When it is outputting text that is a matched part of a subject string,
1084 it does the same, unless a different locale has been set for the match
1085 (using the /L modifier). In this case, it uses isprint() to decide.
1086
1087 4. Fixed a major bug that caused incorrect computation of the amount of memory
1088 required for a compiled pattern when options that changed within the
1089 pattern affected the logic of the preliminary scan that determines the
1090 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1091 that the computed length was too small. The symptoms of this bug were
1092 either the PCRE error "internal error: code overflow" from pcre_compile(),
1093 or a glibc crash with a message such as "pcretest: free(): invalid next
1094 size (fast)". Examples of patterns that provoked this bug (shown in
1095 pcretest format) are:
1096
1097 /(?-x: )/x
1098 /(?x)(?-x: \s*#\s*)/
1099 /((?i)[\x{c0}])/8
1100 /(?i:[\x{c0}])/8
1101
1102 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1103 is now done differently.
1104
1105 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1106 wrapper classes; (b) implement a new function in the C++ scanner that is
1107 more efficient than the old way of doing things because it avoids levels of
1108 recursion in the regex matching; (c) add a paragraph to the documentation
1109 for the FullMatch() function.
1110
1111 6. The escape sequence \n was being treated as whatever was defined as
1112 "newline". Not only was this contrary to the documentation, which states
1113 that \n is character 10 (hex 0A), but it also went horribly wrong when
1114 "newline" was defined as CRLF. This has been fixed.
1115
1116 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1117 was being set to -1 for the "end of line" case (supposedly a value that no
1118 character can have). Though this value is never used (the check for end of
1119 line is "zero bytes in current character"), it caused compiler complaints.
1120 I've changed it to 0xffffffff.
1121
1122 8. In pcre_version.c, the version string was being built by a sequence of
1123 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1124 string (as it is for production releases) called a macro with an empty
1125 argument. The C standard says the result of this is undefined. The gcc
1126 compiler treats it as an empty string (which was what was wanted) but it is
1127 reported that Visual C gives an error. The source has been hacked around to
1128 avoid this problem.
1129
1130 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1131 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1132 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1133 of them did).
1134
1135 10. Originally, pcretest opened its input and output without "b"; then I was
1136 told that "b" was needed in some environments, so it was added for release
1137 5.0 to both the input and output. (It makes no difference on Unix-like
1138 systems.) Later I was told that it is wrong for the input on Windows. I've
1139 now abstracted the modes into two macros, to make it easier to fiddle with
1140 them, and removed "b" from the input mode under Windows.
1141
1142 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1143
1144 12. Added -help and --help to pcretest as an official way of being reminded
1145 of the options.
1146
1147 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1148 and pcrecpp.cc because they annoy compilers at high warning levels.
1149
1150 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1151
1152 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1153 scripts are not C programs :-) and which was not noticed because it works
1154 on Linux.
1155
1156 16. pcretest is supposed to handle any length of pattern and data line (as one
1157 line or as a continued sequence of lines) by extending its input buffer if
1158 necessary. This feature was broken for very long pattern lines, leading to
1159 a string of junk being passed to pcre_compile() if the pattern was longer
1160 than about 50K.
1161
1162 17. I have done a major re-factoring of the way pcre_compile() computes the
1163 amount of memory needed for a compiled pattern. Previously, there was code
1164 that made a preliminary scan of the pattern in order to do this. That was
1165 OK when PCRE was new, but as the facilities have expanded, it has become
1166 harder and harder to keep it in step with the real compile phase, and there
1167 have been a number of bugs (see for example, 4 above). I have now found a
1168 cunning way of running the real compile function in a "fake" mode that
1169 enables it to compute how much memory it would need, while actually only
1170 ever using a few hundred bytes of working memory and without too many
1171 tests of the mode. This should make future maintenance and development
1172 easier. A side effect of this work is that the limit of 200 on the nesting
1173 depth of parentheses has been removed (though this was never a serious
1174 limitation, I suspect). However, there is a downside: pcre_compile() now
1175 runs more slowly than before (30% or more, depending on the pattern). I
1176 hope this isn't a big issue. There is no effect on runtime performance.
1177
1178 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1179 newline (only possible for the last line of a file) and it was a
1180 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1181
1182 19. Added additional timing features to pcretest. (1) The -tm option now times
1183 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1184 separate command line item, by a number that specifies the number of
1185 repeats to use when timing. The default is 50000; this gives better
1186 precision, but takes uncomfortably long for very large patterns.
1187
1188 20. Extended pcre_study() to be more clever in cases where a branch of a
1189 subpattern has no definite first character. For example, (a*|b*)[cd] would
1190 previously give no result from pcre_study(). Now it recognizes that the
1191 first character must be a, b, c, or d.
1192
1193 21. There was an incorrect error "recursive call could loop indefinitely" if
1194 a subpattern (or the entire pattern) that was being tested for matching an
1195 empty string contained only one non-empty item after a nested subpattern.
1196 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1197 incorrectly, because the \d was being skipped in the check.
1198
1199 22. The pcretest program now has a new pattern option /B and a command line
1200 option -b, which is equivalent to adding /B to every pattern. This causes
1201 it to show the compiled bytecode, without the additional information that
1202 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1203 is the same as /B/I).
1204
1205 23. A new optimization is now able automatically to treat some sequences such
1206 as a*b as a*+b. More specifically, if something simple (such as a character
1207 or a simple class like \d) has an unlimited quantifier, and is followed by
1208 something that cannot possibly match the quantified thing, the quantifier
1209 is automatically "possessified".
1210
1211 24. A recursive reference to a subpattern whose number was greater than 39
1212 went wrong under certain circumstances in UTF-8 mode. This bug could also
1213 have affected the operation of pcre_study().
1214
1215 25. Realized that a little bit of performance could be had by replacing
1216 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1217
1218 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1219
1220 27. Possessive quantifiers such as a++ were previously implemented by turning
1221 them into atomic groups such as ($>a+). Now they have their own opcodes,
1222 which improves performance. This includes the automatically created ones
1223 from 23 above.
1224
1225 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1226 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1227 the first matched character to be a colon. This applied both to named and
1228 numbered groups.
1229
1230 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1231
1232 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1233 building PCRE on Windows easier, so I have included it in the distribution.
1234
1235 31. There is now a check in pcretest against a ridiculously large number being
1236 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1237 loop, the loop is abandoned.
1238
1239 32. Forward references to subpatterns in conditions such as (?(2)...) where
1240 subpattern 2 is defined later cause pcre_compile() to search forwards in
1241 the pattern for the relevant set of parentheses. This search went wrong
1242 when there were unescaped parentheses in a character class, parentheses
1243 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1244
1245 33. "Subroutine" calls and backreferences were previously restricted to
1246 referencing subpatterns earlier in the regex. This restriction has now
1247 been removed.
1248
1249 34. Added a number of extra features that are going to be in Perl 5.10. On the
1250 whole, these are just syntactic alternatives for features that PCRE had
1251 previously implemented using the Python syntax or my own invention. The
1252 other formats are all retained for compatibility.
1253
1254 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1255 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1256 also .NET compatible.
1257
1258 (b) A recursion or subroutine call to a named group can now be defined as
1259 (?&name) as well as (?P>name).
1260
1261 (c) A backreference to a named group can now be defined as \k<name> or
1262 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1263 5.10, are also .NET compatible.
1264
1265 (d) A conditional reference to a named group can now use the syntax
1266 (?(<name>) or (?('name') as well as (?(name).
1267
1268 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1269 groups (named and numbered) that are never evaluated inline, but can be
1270 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1271 is always false. There may be only one alternative in such a group.
1272
1273 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1274 as the simple (?(R). The condition is true only if the most recent
1275 recursion is that of the given number or name. It does not search out
1276 through the entire recursion stack.
1277
1278 (g) The escape \gN or \g{N} has been added, where N is a positive or
1279 negative number, specifying an absolute or relative reference.
1280
1281 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1282 some "unreachable code" warnings.
1283
1284 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1285 things, this adds five new scripts.
1286
1287 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1288 There were also incompatibilities regarding the handling of \Q..\E inside
1289 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1290 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1291
1292 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1293 matches an empty string, and forcibly breaks the loop. There were bugs in
1294 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1295 against aaaa the result was just "a" rather than "aaaa", for example. Two
1296 separate and independent bugs (that affected different cases) have been
1297 fixed.
1298
1299 39. Refactored the code to abolish the use of different opcodes for small
1300 capturing bracket numbers. This is a tidy that I avoided doing when I
1301 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1302 The new approach is not only tidier, it makes it possible to reduce the
1303 memory needed to fix the previous bug (38).
1304
1305 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1306 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1307 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1308 mode.
1309
1310 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1311 report.
1312
1313 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1314 copy construction and assignment in the C++ wrapper.
1315
1316 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1317 couple of bugs that could have given wrong results in the "--newline=crlf"
1318 case.
1319
1320 44. Added a number of casts and did some reorganization of signed/unsigned int
1321 variables following suggestions from Dair Grant. Also renamed the variable
1322 "this" as "item" because it is a C++ keyword.
1323
1324 45. Arranged for dftables to add
1325
1326 #include "pcre_internal.h"
1327
1328 to pcre_chartables.c because without it, gcc 4.x may remove the array
1329 definition from the final binary if PCRE is built into a static library and
1330 dead code stripping is activated.
1331
1332 46. For an unanchored pattern, if a match attempt fails at the start of a
1333 newline sequence, and the newline setting is CRLF or ANY, and the next two
1334 characters are CRLF, advance by two characters instead of one.
1335
1336
1337 Version 6.7 04-Jul-06
1338 ---------------------
1339
1340 1. In order to handle tests when input lines are enormously long, pcretest has
1341 been re-factored so that it automatically extends its buffers when
1342 necessary. The code is crude, but this _is_ just a test program. The
1343 default size has been increased from 32K to 50K.
1344
1345 2. The code in pcre_study() was using the value of the re argument before
1346 testing it for NULL. (Of course, in any sensible call of the function, it
1347 won't be NULL.)
1348
1349 3. The memmove() emulation function in pcre_internal.h, which is used on
1350 systems that lack both memmove() and bcopy() - that is, hardly ever -
1351 was missing a "static" storage class specifier.
1352
1353 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1354 containing an extended class (one that cannot be represented by a bitmap
1355 because it contains high-valued characters or Unicode property items, e.g.
1356 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1357 pattern, but PCRE should not loop if you do not (it no longer does).
1358 [Detail: two cases were found: (a) a repeated subpattern containing an
1359 extended class; (b) a recursive reference to a subpattern that followed a
1360 previous extended class. It wasn't skipping over the extended class
1361 correctly when UTF-8 mode was not set.]
1362
1363 5. A negated single-character class was not being recognized as fixed-length
1364 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1365 compile error "lookbehind assertion is not fixed length".
1366
1367 6. The RunPerlTest auxiliary script was showing an unexpected difference
1368 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1369 write a Perl script that can interpret lines of an input file either as
1370 byte characters or as UTF-8, which is what "perltest" was being required to
1371 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1372 can't do is switch easily at run time between having the "use utf8;" pragma
1373 or not. In the end, I fudged it by using the RunPerlTest script to insert
1374 "use utf8;" explicitly for the UTF-8 tests.
1375
1376 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1377 the end of the subject string, contrary to the documentation and to what
1378 Perl does. This was true of both matching functions. Now it matches only at
1379 the start of the subject and immediately after *internal* newlines.
1380
1381 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1382 a pointer to an int instead of a pointer to an unsigned long int. This
1383 caused problems on 64-bit systems.
1384
1385 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1386 instance of the 'standard' template library not being so standard".
1387
1388 10. There was no check on the number of named subpatterns nor the maximum
1389 length of a subpattern name. The product of these values is used to compute
1390 the size of the memory block for a compiled pattern. By supplying a very
1391 long subpattern name and a large number of named subpatterns, the size
1392 computation could be caused to overflow. This is now prevented by limiting
1393 the length of names to 32 characters, and the number of named subpatterns
1394 to 10,000.
1395
1396 11. Subpatterns that are repeated with specific counts have to be replicated in
1397 the compiled pattern. The size of memory for this was computed from the
1398 length of the subpattern and the repeat count. The latter is limited to
1399 65535, but there was no limit on the former, meaning that integer overflow
1400 could in principle occur. The compiled length of a repeated subpattern is
1401 now limited to 30,000 bytes in order to prevent this.
1402
1403 12. Added the optional facility to have named substrings with the same name.
1404
1405 13. Added the ability to use a named substring as a condition, using the
1406 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1407 are numbers (not recommended). Forward references are permitted.
1408
1409 14. Added forward references in named backreferences (if you see what I mean).
1410
1411 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1412 pattern could run off the end of the subject. For example, the pattern
1413 "(?s)(.{1,5})"8 did this with the subject "ab".
1414
1415 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1416 PCRE_CASELESS was set when matching characters that were quantified with ?
1417 or *.
1418
1419 17. A character class other than a single negated character that had a minimum
1420 but no maximum quantifier - for example [ab]{6,} - was not handled
1421 correctly by pce_dfa_exec(). It would match only one character.
1422
1423 18. A valid (though odd) pattern that looked like a POSIX character
1424 class but used an invalid character after [ (for example [[,abc,]]) caused
1425 pcre_compile() to give the error "Failed: internal error: code overflow" or
1426 in some cases to crash with a glibc free() error. This could even happen if
1427 the pattern terminated after [[ but there just happened to be a sequence of
1428 letters, a binary zero, and a closing ] in the memory that followed.
1429
1430 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1431 over the years. Originally (before any Unicode support), just the bottom 8
1432 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1433 output from "man perlunicode" includes this:
1434
1435 The regular expression compiler produces polymorphic opcodes. That
1436 is, the pattern adapts to the data and automatically switches to
1437 the Unicode character scheme when presented with Unicode data--or
1438 instead uses a traditional byte scheme when presented with byte
1439 data.
1440
1441 Sadly, a wide octal escape does not cause a switch, and in a string with
1442 no other multibyte characters, these octal escapes are treated as before.
1443 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1444 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1445 Unicode string.
1446
1447 I have not perpetrated such confusion in PCRE. Up till now, it took just
1448 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1449 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1450 translate to the appropriate multibyte character.
1451
1452 29. Applied some refactoring to reduce the number of warnings from Microsoft
1453 and Borland compilers. This has included removing the fudge introduced
1454 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1455 a warning about an unused variable.
1456
1457 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1458 characters since release 4.0, because Perl (from release 5.004) does not.
1459 [Or at least, is documented not to: some releases seem to be in conflict
1460 with the documentation.] However, when a pattern was studied with
1461 pcre_study() and all its branches started with \s, PCRE still included VT
1462 as a possible starting character. Of course, this did no harm; it just
1463 caused an unnecessary match attempt.
1464
1465 22. Removed a now-redundant internal flag bit that recorded the fact that case
1466 dependency changed within the pattern. This was once needed for "required
1467 byte" processing, but is no longer used. This recovers a now-scarce options
1468 bit. Also moved the least significant internal flag bit to the most-
1469 significant bit of the word, which was not previously used (hangover from
1470 the days when it was an int rather than a uint) to free up another bit for
1471 the future.
1472
1473 23. Added support for CRLF line endings as well as CR and LF. As well as the
1474 default being selectable at build time, it can now be changed at runtime
1475 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1476 specify that it is scanning data with non-default line endings.
1477
1478 24. Changed the definition of CXXLINK to make it agree with the definition of
1479 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1480
1481 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1482 recursions. This makes a big different to stack usage for some patterns.
1483
1484 26. If a subpattern containing a named recursion or subroutine reference such
1485 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1486 the space required for the compiled pattern went wrong and gave too small a
1487 value. Depending on the environment, this could lead to "Failed: internal
1488 error: code overflow at offset 49" or "glibc detected double free or
1489 corruption" errors.
1490
1491 27. Applied patches from Google (a) to support the new newline modes and (b) to
1492 advance over multibyte UTF-8 characters in GlobalReplace.
1493
1494 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1495 difference for some implementation of PCRE in some Windows version.
1496
1497 29. Added some extra testing facilities to pcretest:
1498
1499 \q<number> in a data line sets the "match limit" value
1500 \Q<number> in a data line sets the "match recursion limt" value
1501 -S <number> sets the stack size, where <number> is in megabytes
1502
1503 The -S option isn't available for Windows.
1504
1505
1506 Version 6.6 06-Feb-06
1507 ---------------------
1508
1509 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1510 in pcreposix.h. I have copied the definition from pcre.h.
1511
1512 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1513 because pcre.h is no longer a built file.
1514
1515 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1516 not normally included in the compiled code.
1517
1518
1519 Version 6.5 01-Feb-06
1520 ---------------------
1521
1522 1. When using the partial match feature with pcre_dfa_exec(), it was not
1523 anchoring the second and subsequent partial matches at the new starting
1524 point. This could lead to incorrect results. For example, with the pattern
1525 /1234/, partially matching against "123" and then "a4" gave a match.
1526
1527 2. Changes to pcregrep:
1528
1529 (a) All non-match returns from pcre_exec() were being treated as failures
1530 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1531 error message is output. Some extra information is given for the
1532 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1533 probably the only errors that are likely to be caused by users (by
1534 specifying a regex that has nested indefinite repeats, for instance).
1535 If there are more than 20 of these errors, pcregrep is abandoned.
1536
1537 (b) A binary zero was treated as data while matching, but terminated the
1538 output line if it was written out. This has been fixed: binary zeroes
1539 are now no different to any other data bytes.
1540
1541 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1542 used to set a locale for matching. The --locale=xxxx long option has
1543 been added (no short equivalent) to specify a locale explicitly on the
1544 pcregrep command, overriding the environment variables.
1545
1546 (d) When -B was used with -n, some line numbers in the output were one less
1547 than they should have been.
1548
1549 (e) Added the -o (--only-matching) option.
1550
1551 (f) If -A or -C was used with -c (count only), some lines of context were
1552 accidentally printed for the final match.
1553
1554 (g) Added the -H (--with-filename) option.
1555
1556 (h) The combination of options -rh failed to suppress file names for files
1557 that were found from directory arguments.
1558
1559 (i) Added the -D (--devices) and -d (--directories) options.
1560
1561 (j) Added the -F (--fixed-strings) option.
1562
1563 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1564
1565 (l) Added the --colo(u)r option.
1566
1567 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1568 is not present by default.
1569
1570 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1571 items such as (?R) or (?1), when the recursion could match a number of
1572 alternatives. If it matched one of the alternatives, but subsequently,
1573 outside the recursion, there was a failure, the code tried to back up into
1574 the recursion. However, because of the way PCRE is implemented, this is not
1575 possible, and the result was an incorrect result from the match.
1576
1577 In order to prevent this happening, the specification of recursion has
1578 been changed so that all such subpatterns are automatically treated as
1579 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1580
1581 4. I had overlooked the fact that, in some locales, there are characters for
1582 which isalpha() is true but neither isupper() nor islower() are true. In
1583 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1584 and ordfeminine) are like this. This affected the treatment of \w and \W
1585 when they appeared in character classes, but not when they appeared outside
1586 a character class. The bit map for "word" characters is now created
1587 separately from the results of isalnum() instead of just taking it from the
1588 upper, lower, and digit maps. (Plus the underscore character, of course.)
1589
1590 5. The above bug also affected the handling of POSIX character classes such as
1591 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1592 permanent tables. Instead, the bit maps for such a class were previously
1593 created as the appropriate unions of the upper, lower, and digit bitmaps.
1594 Now they are created by subtraction from the [[:word:]] class, which has
1595 its own bitmap.
1596
1597 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1598 It is created by subtracting the vertical space characters (\x09, \x0a,
1599 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1600 subtraction was done in the overall bitmap for a character class, meaning
1601 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1602 be recognized. This bug has been fixed.
1603
1604 7. Patches from the folks at Google:
1605
1606 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1607 real life, but is still worth protecting against".
1608
1609 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1610 regular expressions".
1611
1612 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1613 have it.
1614
1615 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1616 "configure" and the latter not, in order to fix a problem somebody had
1617 with compiling the Arg class on HP-UX.
1618
1619 (e) Improve the error-handling of the C++ wrapper a little bit.
1620
1621 (f) New tests for checking recursion limiting.
1622
1623 8. The pcre_memmove() function, which is used only if the environment does not
1624 have a standard memmove() function (and is therefore rarely compiled),
1625 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1626 returning a result (though PCRE never actually uses the result).
1627
1628 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1629 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1630 returned instead of calling malloc() with an overflowing number that would
1631 most likely cause subsequent chaos.
1632
1633 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1634
1635 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1636 with this option is matched, the nmatch and pmatch options of regexec() are
1637 ignored.
1638
1639 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1640 provided in case anyone wants to the the POSIX interface with UTF-8
1641 strings.
1642
1643 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1644 C++ linking (needed for some HP-UX environments).
1645
1646 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1647 (unused parameter) and in the pcre_printint() function (omitted "default"
1648 switch label when the default is to do nothing).
1649
1650 15. Added some code to make it possible, when PCRE is compiled as a C++
1651 library, to replace subject pointers for pcre_exec() with a smart pointer
1652 class, thus making it possible to process discontinuous strings.
1653
1654 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1655 much the same function. They were added by different people who were trying
1656 to make PCRE easy to compile on non-Unix systems. It has been suggested
1657 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1658 for compiling on Windows systems. I have therefore replaced it with
1659 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1660 defaults to "extern" for C or "extern C" for C++, which works fine on
1661 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1662 SCOPE with something explicit in config.h. In addition:
1663
1664 (a) pcreposix.h still had just "extern" instead of either of these macros;
1665 I have replaced it with PCRE_DATA_SCOPE.
1666
1667 (b) Functions such as _pcre_xclass(), which are internal to the library,
1668 but external in the C sense, all had PCRE_EXPORT in their definitions.
1669 This is apparently wrong for the Windows case, so I have removed it.
1670 (It makes no difference on Unix-like systems.)
1671
1672 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1673 of recursive calls to match(). This is different to MATCH_LIMIT because
1674 that limits the total number of calls to match(), not all of which increase
1675 the depth of recursion. Limiting the recursion depth limits the amount of
1676 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1677 when PCRE is compiled, and changed at run time. A patch from Google adds
1678 this functionality to the C++ interface.
1679
1680 18. Changes to the handling of Unicode character properties:
1681
1682 (a) Updated the table to Unicode 4.1.0.
1683
1684 (b) Recognize characters that are not in the table as "Cn" (undefined).
1685
1686 (c) I revised the way the table is implemented to a much improved format
1687 which includes recognition of ranges. It now supports the ranges that
1688 are defined in UnicodeData.txt, and it also amalgamates other
1689 characters into ranges. This has reduced the number of entries in the
1690 table from around 16,000 to around 3,000, thus reducing its size
1691 considerably. I realized I did not need to use a tree structure after
1692 all - a binary chop search is just as efficient. Having reduced the
1693 number of entries, I extended their size from 6 bytes to 8 bytes to
1694 allow for more data.
1695
1696 (d) Added support for Unicode script names via properties such as \p{Han}.
1697
1698 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1699 matching that character.
1700
1701 20. When matching a repeated Unicode property with a minimum greater than zero,
1702 (for example \pL{2,}), PCRE could look past the end of the subject if it
1703 reached it while seeking the minimum number of characters. This could
1704 happen only if some of the characters were more than one byte long, because
1705 there is a check for at least the minimum number of bytes.
1706
1707 21. Refactored the implementation of \p and \P so as to be more general, to
1708 allow for more different types of property in future. This has changed the
1709 compiled form incompatibly. Anybody with saved compiled patterns that use
1710 \p or \P will have to recompile them.
1711
1712 22. Added "Any" and "L&" to the supported property types.
1713
1714 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1715 but give a compile time error if the value is greater than 0xff.
1716
1717 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1718 accidentally not being installed or uninstalled.
1719
1720 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1721 made were to insert the current release number. This seemed silly, because
1722 it made things harder for people building PCRE on systems that don't run
1723 "configure". I have turned pcre.h into a distributed file, no longer built
1724 by "configure", with the version identification directly included. There is
1725 no longer a pcre.h.in file.
1726
1727 However, this change necessitated a change to the pcre-config script as
1728 well. It is built from pcre-config.in, and one of the substitutions was the
1729 release number. I have updated configure.ac so that ./configure now finds
1730 the release number by grepping pcre.h.
1731
1732 26. Added the ability to run the tests under valgrind.
1733
1734
1735 Version 6.4 05-Sep-05
1736 ---------------------
1737
1738 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1739 "--" to be printed when multiple files were scanned, even when none of the
1740 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1741 consider it to be a bug, and have restored the previous behaviour.
1742
1743 2. A couple of code tidies to get rid of compiler warnings.
1744
1745 3. The pcretest program used to cheat by referring to symbols in the library
1746 whose names begin with _pcre_. These are internal symbols that are not
1747 really supposed to be visible externally, and in some environments it is
1748 possible to suppress them. The cheating is now confined to including
1749 certain files from the library's source, which is a bit cleaner.
1750
1751 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1752 file's purpose clearer.
1753
1754 5. Reorganized pcre_ucp_findchar().
1755
1756
1757 Version 6.3 15-Aug-05
1758 ---------------------
1759
1760 1. The file libpcre.pc.in did not have general read permission in the tarball.
1761
1762 2. There were some problems when building without C++ support:
1763
1764 (a) If C++ support was not built, "make install" and "make test" still
1765 tried to test it.
1766
1767 (b) There were problems when the value of CXX was explicitly set. Some
1768 changes have been made to try to fix these, and ...
1769
1770 (c) --disable-cpp can now be used to explicitly disable C++ support.
1771
1772 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1773 backslash in a target when C++ was disabled. This confuses some
1774 versions of "make", apparently. Using an intermediate variable solves
1775 this. (Same for CPP_LOBJ.)
1776
1777 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1778 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1779 necessary on certain architectures.
1780
1781 4. Added a setting of -export-symbols-regex to the link command to remove
1782 those symbols that are exported in the C sense, but actually are local
1783 within the library, and not documented. Their names all begin with
1784 "_pcre_". This is not a perfect job, because (a) we have to except some
1785 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1786 available (and never for static libraries). I have made a note to try to
1787 find a way round (a) in the future.
1788
1789
1790 Version 6.2 01-Aug-05
1791 ---------------------
1792
1793 1. There was no test for integer overflow of quantifier values. A construction
1794 such as {1111111111111111} would give undefined results. What is worse, if
1795 a minimum quantifier for a parenthesized subpattern overflowed and became
1796 negative, the calculation of the memory size went wrong. This could have
1797 led to memory overwriting.
1798
1799 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1800
1801 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1802 operating environments where this matters.
1803
1804 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1805 PCRE options from within the C++ wrapper.
1806
1807 5. Named capturing subpatterns were not being correctly counted when a pattern
1808 was compiled. This caused two problems: (a) If there were more than 100
1809 such subpatterns, the calculation of the memory needed for the whole
1810 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1811 back references of the form \12, where the number was greater than 9, were
1812 not recognized as back references, even though there were sufficient
1813 previous subpatterns.
1814
1815 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1816 versions of gcc, e.g. 2.95.4.
1817
1818
1819 Version 6.1 21-Jun-05
1820 ---------------------
1821
1822 1. There was one reference to the variable "posix" in pcretest.c that was not
1823 surrounded by "#if !defined NOPOSIX".
1824
1825 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1826 the cross-check on the old pcre_info() function, for the benefit of the
1827 cut-down version of PCRE that is currently imported into Exim.
1828
1829 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1830 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1831 patterns that start (?i) but I don't think that matters. The use of (?i) is
1832 just an example; this all applies to the other options as well.
1833
1834 4. Since libtool seems to echo the compile commands it is issuing, the output
1835 from "make" can be reduced a bit by putting "@" in front of each libtool
1836 compile command.
1837
1838 5. Patch from the folks at Google for configure.in to be a bit more thorough
1839 in checking for a suitable C++ installation before trying to compile the
1840 C++ stuff. This should fix a reported problem when a compiler was present,
1841 but no suitable headers.
1842
1843 6. The man pages all had just "PCRE" as their title. I have changed them to
1844 be the relevant file name. I have also arranged that these names are
1845 retained in the file doc/pcre.txt, which is a concatenation in text format
1846 of all the man pages except the little individual ones for each function.
1847
1848 7. The NON-UNIX-USE file had not been updated for the different set of source
1849 files that come with release 6. I also added a few comments about the C++
1850 wrapper.
1851
1852
1853 Version 6.0 07-Jun-05
1854 ---------------------
1855
1856 1. Some minor internal re-organization to help with my DFA experiments.
1857
1858 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1859 didn't matter for the library itself when fully configured, but did matter
1860 when compiling without UCP support, or within Exim, where the ucp files are
1861 not imported.
1862
1863 3. Refactoring of the library code to split up the various functions into
1864 different source modules. The addition of the new DFA matching code (see
1865 below) to a single monolithic source would have made it really too
1866 unwieldy, quite apart from causing all the code to be include in a
1867 statically linked application, when only some functions are used. This is
1868 relevant even without the DFA addition now that patterns can be compiled in
1869 one application and matched in another.
1870
1871 The downside of splitting up is that there have to be some external
1872 functions and data tables that are used internally in different modules of
1873 the library but which are not part of the API. These have all had their
1874 names changed to start with "_pcre_" so that they are unlikely to clash
1875 with other external names.
1876
1877 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1878 a different (DFA) algorithm. Although it is slower than the original
1879 function, it does have some advantages for certain types of matching
1880 problem.
1881
1882 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1883 including restarting after a partial match.
1884
1885 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1886 defined when compiling for Windows was sent to me. I have put it into the
1887 code, though I have no means of testing or verifying it.
1888
1889 7. Added the pcre_refcount() auxiliary function.
1890
1891 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1892 match before or at the first newline in the subject string. In pcretest,
1893 the /f option on a pattern can be used to set this.
1894
1895 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1896 would behave wrongly. This has been present in PCRE since release 4.0.
1897
1898 10. A number of changes to the pcregrep command:
1899
1900 (a) Refactored how -x works; insert ^(...)$ instead of setting
1901 PCRE_ANCHORED and checking the length, in preparation for adding
1902 something similar for -w.
1903
1904 (b) Added the -w (match as a word) option.
1905
1906 (c) Refactored the way lines are read and buffered so as to have more
1907 than one at a time available.
1908
1909 (d) Implemented a pcregrep test script.
1910
1911 (e) Added the -M (multiline match) option. This allows patterns to match
1912 over several lines of the subject. The buffering ensures that at least
1913 8K, or the rest of the document (whichever is the shorter) is available
1914 for matching (and similarly the previous 8K for lookbehind assertions).
1915
1916 (f) Changed the --help output so that it now says
1917
1918 -w, --word-regex(p)
1919
1920 instead of two lines, one with "regex" and the other with "regexp"
1921 because that confused at least one person since the short forms are the
1922 same. (This required a bit of code, as the output is generated
1923 automatically from a table. It wasn't just a text change.)
1924
1925 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1926 option but starts with a hyphen. Could be a pattern or a path name
1927 starting with a hyphen, for instance.
1928
1929 (h) "-" can be given as a file name to represent stdin.
1930
1931 (i) When file names are being printed, "(standard input)" is used for
1932 the standard input, for compatibility with GNU grep. Previously
1933 "<stdin>" was used.
1934
1935 (j) The option --label=xxx can be used to supply a name to be used for
1936 stdin when file names are being printed. There is no short form.
1937
1938 (k) Re-factored the options decoding logic because we are going to add
1939 two more options that take data. Such options can now be given in four
1940 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1941
1942 (l) Added the -A, -B, and -C options for requesting that lines of context
1943 around matches be printed.
1944
1945 (m) Added the -L option to print the names of files that do not contain
1946 any matching lines, that is, the complement of -l.
1947
1948 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1949 continue to scan other files.
1950
1951 (o) The -s option was incorrectly implemented. For compatibility with other
1952 greps, it now suppresses the error message for a non-existent or non-
1953 accessible file (but not the return code). There is a new option called
1954 -q that suppresses the output of matching lines, which was what -s was
1955 previously doing.
1956
1957 (p) Added --include and --exclude options to specify files for inclusion
1958 and exclusion when recursing.
1959
1960 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
1961 Hopefully, it now does.
1962
1963 12. Missing cast in pcre_study().
1964
1965 13. Added an "uninstall" target to the makefile.
1966
1967 14. Replaced "extern" in the function prototypes in Makefile.in with
1968 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
1969 world, but is set differently for Windows.
1970
1971 15. Added a second compiling function called pcre_compile2(). The only
1972 difference is that it has an extra argument, which is a pointer to an
1973 integer error code. When there is a compile-time failure, this is set
1974 non-zero, in addition to the error test pointer being set to point to an
1975 error message. The new argument may be NULL if no error number is required
1976 (but then you may as well call pcre_compile(), which is now just a
1977 wrapper). This facility is provided because some applications need a
1978 numeric error indication, but it has also enabled me to tidy up the way
1979 compile-time errors are handled in the POSIX wrapper.
1980
1981 16. Added VPATH=.libs to the makefile; this should help when building with one
1982 prefix path and installing with another. (Or so I'm told by someone who
1983 knows more about this stuff than I do.)
1984
1985 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
1986 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
1987 match everything, including newlines. This is not POSIX-compatible, but
1988 somebody wanted the feature. From pcretest it can be activated by using
1989 both the P and the s flags.
1990
1991 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
1992
1993 19. libpcre.pc was being incorrectly installed as executable.
1994
1995 20. A couple of places in pcretest check for end-of-line by looking for '\n';
1996 it now also looks for '\r' so that it will work unmodified on Windows.
1997
1998 21. Added Google's contributed C++ wrapper to the distribution.
1999
2000 22. Added some untidy missing memory free() calls in pcretest, to keep
2001 Electric Fence happy when testing.
2002
2003
2004
2005 Version 5.0 13-Sep-04
2006 ---------------------
2007
2008 1. Internal change: literal characters are no longer packed up into items
2009 containing multiple characters in a single byte-string. Each character
2010 is now matched using a separate opcode. However, there may be more than one
2011 byte in the character in UTF-8 mode.
2012
2013 2. The pcre_callout_block structure has two new fields: pattern_position and
2014 next_item_length. These contain the offset in the pattern to the next match
2015 item, and its length, respectively.
2016
2017 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2018 insertion of callouts before each pattern item. Added the /C option to
2019 pcretest to make use of this.
2020
2021 4. On the advice of a Windows user, the lines
2022
2023 #if defined(_WIN32) || defined(WIN32)
2024 _setmode( _fileno( stdout ), 0x8000 );
2025 #endif /* defined(_WIN32) || defined(WIN32) */
2026
2027 have been added to the source of pcretest. This apparently does useful
2028 magic in relation to line terminators.
2029
2030 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2031 for the benefit of those environments where the "b" makes a difference.
2032
2033 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2034 to know about it. I have put a hack into configure.in that adds in code
2035 to set GCC=yes if CC=icc. This seems to end up at a point in the
2036 generated configure script that is early enough to affect the setting of
2037 compiler options, which is what is needed, but I have no means of testing
2038 whether it really works. (The user who reported this had patched the
2039 generated configure script, which of course I cannot do.)
2040
2041 LATER: After change 22 below (new libtool files), the configure script
2042 seems to know about icc (and also ecc). Therefore, I have commented out
2043 this hack in configure.in.
2044
2045 7. Added support for pkg-config (2 patches were sent in).
2046
2047 8. Negated POSIX character classes that used a combination of internal tables
2048 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2049 [[:^ascii]]. Typically, they would match almost any characters. The other
2050 POSIX classes were not broken in this way.
2051
2052 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2053 to find the match, as PCRE was deluded into thinking that the match had to
2054 start at the start point or following a newline. The same bug applied to
2055 patterns with negative forward assertions or any backward assertions
2056 preceding ".*" at the start, unless the pattern required a fixed first
2057 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2058
2059 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2060 starting at the last subject character, bytes beyond the end of the subject
2061 string were read.
2062
2063 11. Renamed the variable "class" as "classbits" to make life easier for C++
2064 users. (Previously there was a macro definition, but it apparently wasn't
2065 enough.)
2066
2067 12. Added the new field "tables" to the extra data so that tables can be passed
2068 in at exec time, or the internal tables can be re-selected. This allows
2069 a compiled regex to be saved and re-used at a later time by a different
2070 program that might have everything at different addresses.
2071
2072 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2073 -R library as well as a -L library.
2074
2075 14. The debugging options of pcretest (-d on the command line or D on a
2076 pattern) showed incorrect output for anything following an extended class
2077 that contained multibyte characters and which was followed by a quantifier.
2078
2079 15. Added optional support for general category Unicode character properties
2080 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2081 support. It adds about 90K to the size of the library. The meanings of the
2082 inbuilt class escapes such as \d and \s have NOT been changed.
2083
2084 16. Updated pcredemo.c to include calls to free() to release the memory for the
2085 compiled pattern.
2086
2087 17. The generated file chartables.c was being created in the source directory
2088 instead of in the building directory. This caused the build to fail if the
2089 source directory was different from the building directory, and was
2090 read-only.
2091
2092 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2093 file. No doubt somebody will tell me if they don't make sense... Also added
2094 Dan Mooney's comments about building on OpenVMS.
2095
2096 19. Added support for partial matching via the PCRE_PARTIAL option for
2097 pcre_exec() and the \P data escape in pcretest.
2098
2099 20. Extended pcretest with 3 new pattern features:
2100
2101 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2102 write the compiled pattern to the file whose name is "rest-of-line".
2103 This is a straight binary dump of the data, with the saved pointer to
2104 the character tables forced to be NULL. The study data, if any, is
2105 written too. After writing, pcretest reads a new pattern.
2106
2107 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2108 compiled pattern from the given file. There must not be any
2109 occurrences of "<" in the file name (pretty unlikely); if there are,
2110 pcretest will instead treat the initial "<" as a pattern delimiter.
2111 After reading in the pattern, pcretest goes on to read data lines as
2112 usual.
2113
2114 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2115 and 16-bit fields in a compiled pattern, to simulate a pattern that
2116 was compiled on a host of opposite endianness.
2117
2118 21. The pcre-exec() function can now cope with patterns that were compiled on
2119 hosts of opposite endianness, with this restriction:
2120
2121 As for any compiled expression that is saved and used later, the tables
2122 pointer field cannot be preserved; the extra_data field in the arguments
2123 to pcre_exec() should be used to pass in a tables address if a value
2124 other than the default internal tables were used at compile time.
2125
2126 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2127 now diagnosed as an error. Previously, most of the time, a negative number
2128 would have been treated as zero, but if in addition "ovector" was passed as
2129 NULL, a crash could occur.
2130
2131 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2132 new versions from the libtool 1.5 distribution (the last one is a copy of
2133 a file called libtool.m4). This seems to have fixed the need to patch
2134 "configure" to support Darwin 1.3 (which I used to do). However, I still
2135 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2136 workstation).
2137
2138 24. Changed the PCRE licence to be the more standard "BSD" licence.
2139
2140
2141 Version 4.5 01-Dec-03
2142 ---------------------
2143
2144 1. There has been some re-arrangement of the code for the match() function so
2145 that it can be compiled in a version that does not call itself recursively.
2146 Instead, it keeps those local variables that need separate instances for
2147 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2148 needs to "recurse". Keeping track of where control must go is done by means
2149 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2150 hide most of the details from the main code, and operates only if
2151 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2152 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2153 operating.
2154
2155 To make it easier for callers to provide specially tailored get/free
2156 functions for this usage, two new functions, pcre_stack_malloc, and
2157 pcre_stack_free, are used. They are always called in strict stacking order,
2158 and the size of block requested is always the same.
2159
2160 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2161 PCRE has been compiled to use the stack or the heap for recursion. The
2162 -C option of pcretest uses this to show which version is compiled.
2163
2164 A new data escape \S, is added to pcretest; it causes the amounts of store
2165 obtained and freed by both kinds of malloc/free at match time to be added
2166 to the output.
2167
2168 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2169 what's available on my current Linux desktop machine.
2170
2171 3. When matching a UTF-8 string, the test for a valid string at the start has
2172 been extended. If start_offset is not zero, PCRE now checks that it points
2173 to a byte that is the start of a UTF-8 character. If not, it returns
2174 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2175 this is necessary because there may be backward assertions in the pattern.
2176 When matching the same subject several times, it may save resources to use
2177 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2178
2179 4. The code for checking the validity of UTF-8 strings has been tightened so
2180 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2181 containing "overlong sequences".
2182
2183 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2184 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2185 should have been "&", but it just so happened that all the cases this let
2186 through by mistake were picked up later in the function.
2187
2188 6. I had used a variable called "isblank" - this is a C99 function, causing
2189 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2190
2191 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2192 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2193 so the version is shown; (c) stop "make test" echoing "./RunTest".
2194
2195 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2196
2197 9. The prototype for memmove() for systems that don't have it was using
2198 size_t, but the inclusion of the header that defines size_t was later. I've
2199 moved the #includes for the C headers earlier to avoid this.
2200
2201 10. Added some adjustments to the code to make it easier to compiler on certain
2202 special systems:
2203
2204 (a) Some "const" qualifiers were missing.
2205 (b) Added the macro EXPORT before all exported functions; by default this
2206 is defined to be empty.
2207 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2208 that it reads its output file name as an argument instead of writing
2209 to the standard output and assuming this can be redirected.
2210
2211 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2212 class containing characters with values greater than 255, PCRE compilation
2213 went into a loop.
2214
2215 12. A recursive reference to a subpattern that was within another subpattern
2216 that had a minimum quantifier of zero caused PCRE to crash. For example,
2217 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2218 recursion. If the recursively-called subpattern itself had a zero repeat,
2219 that was OK.
2220
2221 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2222 buffer into which it was copied (for escape processing) was still set at
2223 1024, so long lines caused crashes.
2224
2225 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2226 "internal error: code overflow...". This applied to any character class
2227 that was followed by a possessive quantifier.
2228
2229 15. Modified the Makefile to add libpcre.la as a prerequisite for
2230 libpcreposix.la because I was told this is needed for a parallel build to
2231 work.
2232
2233 16. If a pattern that contained .* following optional items at the start was
2234 studied, the wrong optimizing data was generated, leading to matching
2235 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2236 matching string must start with a or b or c. The correct conclusion for
2237 this pattern is that a match can start with any character.
2238
2239
2240 Version 4.4 13-Aug-03
2241 ---------------------
2242
2243 1. In UTF-8 mode, a character class containing characters with values between
2244 127 and 255 was not handled correctly if the compiled pattern was studied.
2245 In fixing this, I have also improved the studying algorithm for such
2246 classes (slightly).
2247
2248 2. Three internal functions had redundant arguments passed to them. Removal
2249 might give a very teeny performance improvement.
2250
2251 3. Documentation bug: the value of the capture_top field in a callout is *one
2252 more than* the number of the hightest numbered captured substring.
2253
2254 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2255 in incorrectly linking with a previously installed version. They now link
2256 explicitly with libpcre.la.
2257
2258 5. configure.in no longer needs to recognize Cygwin specially.
2259
2260 6. A problem in pcre.in for Windows platforms is fixed.
2261
2262 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2263 pcretest, it used to include the size of the study block as part of its
2264 output. Unfortunately, the structure contains a field that has a different
2265 size on different hardware architectures. This meant that the tests that
2266 showed this size failed. As the block is currently always of a fixed size,
2267 this information isn't actually particularly useful in pcretest output, so
2268 I have just removed it.
2269
2270 8. Three pre-processor statements accidentally did not start in column 1.
2271 Sadly, there are *still* compilers around that complain, even though
2272 standard C has not required this for well over a decade. Sigh.
2273
2274 9. In pcretest, the code for checking callouts passed small integers in the
2275 callout_data field, which is a void * field. However, some picky compilers
2276 complained about the casts involved for this on 64-bit systems. Now
2277 pcretest passes the address of the small integer instead, which should get
2278 rid of the warnings.
2279
2280 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2281 both compile and run time, and gives an error if an invalid UTF-8 sequence
2282 is found. There is a option for disabling this check in cases where the
2283 string is known to be correct and/or the maximum performance is wanted.
2284
2285 11. In response to a bug report, I changed one line in Makefile.in from
2286
2287 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2288 to
2289 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2290
2291 to look similar to other lines, but I have no way of telling whether this
2292 is the right thing to do, as I do not use Windows. No doubt I'll get told
2293 if it's wrong...
2294
2295
2296 Version 4.3 21-May-03
2297 ---------------------
2298
2299 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2300 Makefile.
2301
2302 2. Some refactoring to improve the quality of the code:
2303
2304 (i) The utf8_table... variables are now declared "const".
2305
2306 (ii) The code for \cx, which used the "case flipping" table to upper case
2307 lower case letters, now just substracts 32. This is ASCII-specific,
2308 but the whole concept of \cx is ASCII-specific, so it seems
2309 reasonable.
2310
2311 (iii) PCRE was using its character types table to recognize decimal and
2312 hexadecimal digits in the pattern. This is silly, because it handles
2313 only 0-9, a-f, and A-F, but the character types table is locale-
2314 specific, which means strange things might happen. A private
2315 table is now used for this - though it costs 256 bytes, a table is
2316 much faster than multiple explicit tests. Of course, the standard
2317 character types table is still used for matching digits in subject
2318 strings against \d.
2319
2320 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2321 ending in _t are). So I've renamed it as ESC_tee.
2322
2323 3. The first argument for regexec() in the POSIX wrapper should have been
2324 defined as "const".
2325
2326 4. Changed pcretest to use malloc() for its buffers so that they can be
2327 Electric Fenced for debugging.
2328
2329 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2330 to read one or more bytes before the start of the subject string. Often this
2331 had no effect on PCRE's behaviour, but in some circumstances it could
2332 provoke a segmentation fault.
2333
2334 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2335 to try to read one or more bytes before the start of the subject string.
2336
2337 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2338 UTF-8 support could misbehave in various ways if the subject string
2339 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2340 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2341 back over UTF-8 characters.)
2342
2343
2344 Version 4.2 14-Apr-03
2345 ---------------------
2346
2347 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2348
2349 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2350 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2351 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2352 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2353 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2354 and BUILD_EXEEXT
2355 Note: automatic setting of the BUILD variables is not yet working
2356 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2357 compile-time but not at link-time
2358 [LINK]: use for linking executables only
2359 make different versions for Windows and non-Windows
2360 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2361 libraries
2362 [LINK_FOR_BUILD]: new variable
2363 [OBJEXT]: use throughout
2364 [EXEEXT]: use throughout
2365 <winshared>: new target
2366 <wininstall>: new target
2367 <dftables.o>: use native compiler
2368 <dftables>: use native linker
2369 <install>: handle Windows platform correctly
2370 <clean>: ditto
2371 <check>: ditto
2372 copy DLL to top builddir before testing
2373
2374 As part of these changes, -no-undefined was removed again. This was reported
2375 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2376 in any case.
2377
2378 3. Some tidies to get rid of compiler warnings:
2379
2380 . In the match_data structure, match_limit was an unsigned long int, whereas
2381 match_call_count was an int. I've made them both unsigned long ints.
2382
2383 . In pcretest the fact that a const uschar * doesn't automatically cast to
2384 a void * provoked a warning.
2385
2386 . Turning on some more compiler warnings threw up some "shadow" variables
2387 and a few more missing casts.
2388
2389 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2390 option, a class that contained a single character with a value between 128
2391 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2392
2393 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2394 option, a class that contained several characters, but with at least one
2395 whose value was between 128 and 255 caused PCRE to crash.
2396
2397
2398 Version 4.1 12-Mar-03
2399 ---------------------
2400
2401 1. Compiling with gcc -pedantic found a couple of places where casts were
2402 needed, and a string in dftables.c that was longer than standard compilers are
2403 required to support.
2404
2405 2. Compiling with Sun's compiler found a few more places where the code could
2406 be tidied up in order to avoid warnings.
2407
2408 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2409 first of these names is deprecated in the latest Autoconf in favour of the name
2410 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2411 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2412 analogy I have changed it to CFLAGS_FOR_BUILD.
2413
2414 4. Added -no-undefined to the linking command in the Makefile, because this is
2415 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2416 linking step for the pcreposix library.
2417
2418 5. PCRE was failing to diagnose the case of two named groups with the same
2419 name.
2420
2421 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2422 literal character that is needed in the subject for a match, and scans along to
2423 ensure that it is present before embarking on the full matching process. This
2424 saves time in cases of nested unlimited repeats that are never going to match.
2425 Problem: the scan can take a lot of time if the subject is very long (e.g.
2426 megabytes), thus penalizing straightforward matches. It is now done only if the
2427 amount of subject to be scanned is less than 1000 bytes.
2428
2429 7. A lesser problem with the same optimization is that it was recording the
2430 first character of an anchored pattern as "needed", thus provoking a search
2431 right along the subject, even when the first match of the pattern was going to
2432 fail. The "needed" character is now not set for anchored patterns, unless it
2433 follows something in the pattern that is of non-fixed length. Thus, it still
2434 fulfils its original purpose of finding quick non-matches in cases of nested
2435 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2436
2437
2438 Version 4.0 17-Feb-03
2439 ---------------------
2440
2441 1. If a comment in an extended regex that started immediately after a meta-item
2442 extended to the end of string, PCRE compiled incorrect data. This could lead to
2443 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2444
2445 2. Moved to autoconf 2.53 and libtool 1.4.2.
2446
2447 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2448 the special perltest8 script is no longer needed - all the tests can be run
2449 from a single perltest script.
2450
2451 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2452 by \s. It has now been removed in PCRE. This means it isn't recognized as
2453 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2454 class [:space:] *does* include VT, thereby creating a mess.
2455
2456 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2457 space and tab.
2458
2459 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2460 its new features into the main test script, reducing the number of scripts.
2461
2462 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2463 were backward compatible, and made the (?i) apply to the whole pattern, as if
2464 /i were given. Now it behaves more logically, and applies the option setting
2465 only to what follows. PCRE has been changed to follow suit. However, if it
2466 finds options settings right at the start of the pattern, it extracts them into
2467 the global options, as before. Thus, they show up in the info data.
2468
2469 8. Added support for the \Q...\E escape sequence. Characters in between are
2470 treated as literals. This is slightly different from Perl in that $ and @ are
2471 also handled as literals inside the quotes. In Perl, they will cause variable
2472 interpolation. Note the following examples:
2473
2474 Pattern PCRE matches Perl matches
2475
2476 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2477 \Qabc\$xyz\E abc\$xyz abc\$xyz
2478 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2479
2480 For compatibility with Perl, \Q...\E sequences are recognized inside character
2481 classes as well as outside them.
2482
2483 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2484 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2485 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2486 signed/unsigned warnings.
2487
2488 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2489 option for pcretest, so I've replaced it by a simple function that does just
2490 that job.
2491
2492 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2493 "pcregrep -".
2494
2495 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2496 Java package. This provides some syntactic sugar for simple cases of what my
2497 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2498 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2499 item, you can use this simplified notation. Note that only makes sense with
2500 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2501 greediness, whatever the setting of the PCRE_UNGREEDY option.
2502
2503 13. A change of greediness default within a pattern was not taking effect at
2504 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2505 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2506 was abstracted outside.
2507
2508 14. PCRE now supports the \G assertion. It is true when the current matching
2509 position is at the start point of the match. This differs from \A when the
2510 starting offset is non-zero. Used with the /g option of pcretest (or similar
2511 code), it works in the same way as it does for Perl's /g option. If all
2512 alternatives of a regex begin with \G, the expression is anchored to the start
2513 match position, and the "anchored" flag is set in the compiled expression.
2514
2515 15. Some bugs concerning the handling of certain option changes within patterns
2516 have been fixed. These applied to options other than (?ims). For example,
2517 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2518 been the other way round. Some of this was related to change 7 above.
2519
2520 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2521 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2522 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2523 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2524
2525 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2526 mode. Unlike ".", it always matches newline, whatever the setting of
2527 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2528 assertions. Perl allows it, but it doesn't (in general) work because it can't
2529 calculate the length of the lookbehind. At least, that's the case for Perl
2530 5.8.0 - I've been told they are going to document that it doesn't work in
2531 future.
2532
2533 18. Added an error diagnosis for escapes that PCRE does not support: these are
2534 \L, \l, \N, \P, \p, \U, \u, and \X.
2535
2536 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2537 reading past the end of the pattern in cases such as /[abcd/.
2538
2539 20. PCRE was getting more memory than necessary for patterns with classes that
2540 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2541
2542 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2543 compiling PCRE for use with Virtual Pascal.
2544
2545 22. Small fix to the Makefile to make it work properly if the build is done
2546 outside the source tree.
2547
2548 23. Added a new extension: a condition to go with recursion. If a conditional
2549 subpattern starts with (?(R) the "true" branch is used if recursion has
2550 happened, whereas the "false" branch is used only at the top level.
2551
2552 24. When there was a very long string of literal characters (over 255 bytes
2553 without UTF support, over 250 bytes with UTF support), the computation of how
2554 much memory was required could be incorrect, leading to segfaults or other
2555 strange effects.
2556
2557 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2558 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2559 there was a subsequent back reference to those brackets. This meant that, for
2560 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2561 possible to check for precisely this case. All we can do is abandon the
2562 optimization if .* occurs inside capturing brackets when there are any back
2563 references whatsoever. (See below for a better fix that came later.)
2564
2565 26. The handling of the optimization for finding the first character of a
2566 non-anchored pattern, and for finding a character that is required later in the
2567 match were failing in some cases. This didn't break the matching; it just
2568 failed to optimize when it could. The way this is done has been re-implemented.
2569
2570 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2571
2572 28. Added a new feature that provides some of the functionality that Perl
2573 provides with (?{...}). The facility is termed a "callout". The way it is done
2574 in PCRE is for the caller to provide an optional function, by setting
2575 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2576 global variable. By default it is unset, which disables all calling out. To get
2577 the function called, the regex must include (?C) at appropriate points. This
2578 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2579 This provides a means of identifying different callout points. When PCRE
2580 reaches such a point in the regex, if pcre_callout has been set, the external
2581 function is called. It is provided with data in a structure called
2582 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2583 matching continues; if it returns a non-zero value, the match at the current
2584 point fails. However, backtracking will occur if possible. [This was changed
2585 later and other features added - see item 49 below.]
2586
2587 29. pcretest is upgraded to test the callout functionality. It provides a
2588 callout function that displays information. By default, it shows the start of
2589 the match and the current position in the text. There are some new data escapes
2590 to vary what happens:
2591
2592 \C+ in addition, show current contents of captured substrings
2593 \C- do not supply a callout function
2594 \C!n return 1 when callout number n is reached
2595 \C!n!m return 1 when callout number n is reached for the mth time
2596
2597 30. If pcregrep was called with the -l option and just a single file name, it
2598 output "<stdin>" if a match was found, instead of the file name.
2599
2600 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2601 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2602 pcre_exec(). This saves a malloc/free per call. The default value of
2603 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2604 when configuring.
2605
2606 32. The default maximum size of a compiled pattern is 64K. There have been a
2607 few cases of people hitting this limit. The code now uses macros to handle the
2608 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2609 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2610 configuring. Tests 2 and 5 work only with 2-byte links because they output
2611 debugging information about compiled patterns.
2612
2613 33. Internal code re-arrangements:
2614
2615 (a) Moved the debugging function for printing out a compiled regex into
2616 its own source file (printint.c) and used #include to pull it into
2617 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2618 separate copies.
2619
2620 (b) Defined the list of op-code names for debugging as a macro in
2621 internal.h so that it is next to the definition of the opcodes.
2622
2623 (c) Defined a table of op-code lengths for simpler skipping along compiled
2624 code. This is again a macro in internal.h so that it is next to the
2625 definition of the opcodes.
2626
2627 34. Added support for recursive calls to individual subpatterns, along the
2628 lines of Robin Houston's patch (but implemented somewhat differently).
2629
2630 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2631 allow it to read and process whole directories in Win32. This code was
2632 contributed by Lionel Fourquaux; it has not been tested by me.
2633
2634 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2635 used to name a group. Names consist of alphanumerics and underscores, and must
2636 be unique. Back references use the syntax (?P=name) and recursive calls use
2637 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2638 numbers. The function pcre_fullinfo() can be used after compilation to extract
2639 a name/number map. There are three relevant calls:
2640
2641 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2642 PCRE_INFO_NAMECOUNT yields the number of entries
2643 PCRE_INFO_NAMETABLE yields a pointer to the map.
2644
2645 The map is a vector of fixed-size entries. The size of each entry depends on
2646 the length of the longest name used. The first two bytes of each entry are the
2647 group number, most significant byte first. There follows the corresponding
2648 name, zero terminated. The names are in alphabetical order.
2649
2650 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2651 case instead of 255. Making it the same both with and without UTF-8 support
2652 means that the same test output works with both.
2653
2654 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2655 calling malloc() with a zero argument.
2656
2657 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2658 optimization. I've improved things by keeping a bitmap of backreferences with
2659 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2660 fact referenced, the optimization can be applied. It is unlikely that a
2661 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2662 the match to follow \n) will appear inside brackets with a number greater than
2663 31, but if it does, any back reference > 31 suppresses the optimization.
2664
2665 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2666 of disabling numbered capturing parentheses. Any opening parenthesis that is
2667 not followed by ? behaves as if it were followed by ?: but named parentheses
2668 can still be used for capturing (and they will acquire numbers in the usual
2669 way).
2670
2671 41. Redesigned the return codes from the match() function into yes/no/error so
2672 that errors can be passed back from deep inside the nested calls. A malloc
2673 failure while inside a recursive subpattern call now causes the
2674 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2675
2676 42. It is now possible to set a limit on the number of times the match()
2677 function is called in a call to pcre_exec(). This facility makes it possible to
2678 limit the amount of recursion and backtracking, though not in a directly
2679 obvious way, because the match() function is used in a number of different
2680 circumstances. The count starts from zero for each position in the subject
2681 string (for non-anchored patterns). The default limit is, for compatibility, a
2682 large number, namely 10 000 000. You can change this in two ways:
2683
2684 (a) When configuring PCRE before making, you can use --with-match-limit=n
2685 to set a default value for the compiled library.
2686
2687 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2688 a different value is set. See 45 below.
2689
2690 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2691
2692 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2693 of things that can be changed at compile time. The first argument specifies
2694 what is wanted and the second points to where the information is to be placed.
2695 The current list of available information is:
2696
2697 PCRE_CONFIG_UTF8
2698
2699 The output is an integer that is set to one if UTF-8 support is available;
2700 otherwise it is set to zero.
2701
2702 PCRE_CONFIG_NEWLINE
2703
2704 The output is an integer that it set to the value of the code that is used for
2705 newline. It is either LF (10) or CR (13).
2706
2707 PCRE_CONFIG_LINK_SIZE
2708
2709 The output is an integer that contains the number of bytes used for internal
2710 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2711
2712 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2713
2714 The output is an integer that contains the threshold above which the POSIX
2715 interface uses malloc() for output vectors. See item 31 above.
2716
2717 PCRE_CONFIG_MATCH_LIMIT
2718
2719 The output is an unsigned integer that contains the default limit of the number
2720 of match() calls in a pcre_exec() execution. See 42 above.
2721
2722 44. pcretest has been upgraded by the addition of the -C option. This causes it
2723 to extract all the available output from the new pcre_config() function, and to
2724 output it. The program then exits immediately.
2725
2726 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2727 order to support additional features. One way would have been to define
2728 pcre_exec2() (for example) with extra arguments, but this would not have been
2729 extensible, and would also have required all calls to the original function to
2730 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2731 is used for passing in "extra" data from pcre_study().
2732
2733 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2734 contains the following fields:
2735
2736 flags a bitmap indicating which of the following fields are set
2737 study_data opaque data from pcre_study()
2738 match_limit a way of specifying a limit on match() calls for a specific
2739 call to pcre_exec()
2740 callout_data data for callouts (see 49 below)
2741
2742 The flag bits are also defined in pcre.h, and are
2743
2744 PCRE_EXTRA_STUDY_DATA
2745 PCRE_EXTRA_MATCH_LIMIT
2746 PCRE_EXTRA_CALLOUT_DATA
2747
2748 The pcre_study() function now returns one of these new pcre_extra blocks, with
2749 the actual study data pointed to by the study_data field, and the
2750 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2751 before. That is, this change is entirely upwards-compatible and requires no
2752 change to existing code.
2753
2754 If you want to pass in additional data to pcre_exec(), you can either place it
2755 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2756 block.
2757
2758 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2759 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2760 times with different match limits, until it finds the minimum value needed for
2761 pcre_exec() to complete. The value is then output. This can be instructive; for
2762 most simple matches the number is quite small, but for pathological cases it
2763 gets very large very quickly.
2764
2765 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2766 returns the size of the data block pointed to by the study_data field in a
2767 pcre_extra block, that is, the value that was passed as the argument to
2768 pcre_malloc() when PCRE was getting memory in which to place the information
2769 created by pcre_study(). The fourth argument should point to a size_t variable.
2770 pcretest has been extended so that this information is shown after a successful
2771 pcre_study() call when information about the compiled regex is being displayed.
2772
2773 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2774 because what follows is always an absolute path. (Later: it turns out that this
2775 is more than cosmetic for MinGW, because it doesn't like empty path
2776 components.)
2777
2778 49. Some changes have been made to the callout feature (see 28 above):
2779
2780 (i) A callout function now has three choices for what it returns:
2781
2782 0 => success, carry on matching
2783 > 0 => failure at this point, but backtrack if possible
2784 < 0 => serious error, return this value from pcre_exec()
2785
2786 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2787 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2788 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2789 use by callout functions. It will never be used by PCRE itself.
2790
2791 (ii) The pcre_extra structure (see 45 above) has a void * field called
2792 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2793 pcre_callout_block structure has a field of the same name. The contents of
2794 the field passed in the pcre_extra structure are passed to the callout
2795 function in the corresponding field in the callout block. This makes it
2796 easier to use the same callout-containing regex from multiple threads. For
2797 testing, the pcretest program has a new data escape
2798
2799 \C*n pass the number n (may be negative) as callout_data
2800
2801 If the callout function in pcretest receives a non-zero value as
2802 callout_data, it returns that value.
2803
2804 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2805 there were some redundant $(CFLAGS) in commands that are now specified as
2806 $(LINK), which already includes $(CFLAGS).
2807
2808 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2809 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2810 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2811 one-byte characters throughout. Note that case-insensitive matching applies
2812 only to characters whose values are less than 256. PCRE doesn't support the
2813 notion of cases for higher-valued characters.
2814
2815 (i) A character class whose characters are all within 0-255 is handled as
2816 a bit map, and the map is inverted for negative classes. Previously, a
2817 character > 255 always failed to match such a class; however it should
2818 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2819
2820 (ii) A negated character class with a single character < 255 is coded as
2821 "not this character" (OP_NOT). This wasn't working properly when the test
2822 character was multibyte, either singly or repeated.
2823
2824 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2825 mode, for example: \x{100}{2,3}.
2826
2827 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2828 singly or repeated) now correctly test multibyte characters. However,
2829 PCRE doesn't recognize any characters with values greater than 255 as
2830 digits, spaces, or word characters. Such characters always match \D, \S,
2831 and \W, and never match \d, \s, or \w.
2832
2833 (v) Classes may now contain characters and character ranges with values
2834 greater than 255. For example: [ab\x{100}-\x{400}].
2835
2836 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2837 PCRE in UTF-8 mode.
2838
2839 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2840 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2841 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2842 value.)
2843
2844 53. The single man page has become too large. I have therefore split it up into
2845 a number of separate man pages. These also give rise to individual HTML pages;
2846 these are now put in a separate directory, and there is an index.html page that
2847 lists them all. Some hyperlinking between the pages has been installed.
2848
2849 54. Added convenience functions for handling named capturing parentheses.
2850
2851 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2852 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2853 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2854 are faulted.
2855
2856 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2857 calling configure. These values are used when compiling the dftables.c program
2858 which is run to generate the source of the default character tables. They
2859 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2860 you will need to set these values.
2861
2862 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2863
2864
2865 Version 3.9 02-Jan-02
2866 ---------------------
2867
2868 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2869
2870 2. If --disable-static was given, the building process failed when trying to
2871 build pcretest and pcregrep. (For some reason it was using libtool to compile
2872 them, which is not right, as they aren't part of the library.)
2873
2874
2875 Version 3.8 18-Dec-01
2876 ---------------------
2877
2878 1. The experimental UTF-8 code was completely screwed up. It was packing the
2879 bytes in the wrong order. How dumb can you get?
2880
2881
2882 Version 3.7 29-Oct-01
2883 ---------------------
2884
2885 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2886 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2887 this didn't happen under Solaris 8, where I normally test things.
2888
2889 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2890 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2891 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2892 this fix an hour or so after the initial 3.7 release.)
2893
2894
2895 Version 3.6 23-Oct-01
2896 ---------------------
2897
2898 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2899 offsets passed as NULL with zero offset count.
2900
2901 2. The config.guess and config.sub files had not been updated when I moved to
2902 the latest autoconf.
2903
2904
2905 Version 3.5 15-Aug-01
2906 ---------------------
2907
2908 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2909 had been forgotten.
2910
2911 2. By using declared but undefined structures, we can avoid using "void"
2912 definitions in pcre.h while keeping the internal definitions of the structures
2913 private.
2914
2915 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2916 user point of view, this means that both static and shared libraries are built
2917 by default, but this can be individually controlled. More of the work of
2918 handling this static/shared cases is now inside libtool instead of PCRE's make
2919 file.
2920
2921 4. The pcretest utility is now installed along with pcregrep because it is
2922 useful for users (to test regexs) and by doing this, it automatically gets
2923 relinked by libtool. The documentation has been turned into a man page, so
2924 there are now .1, .txt, and .html versions in /doc.
2925
2926 5. Upgrades to pcregrep:
2927 (i) Added long-form option names like gnu grep.
2928 (ii) Added --help to list all options with an explanatory phrase.
2929 (iii) Added -r, --recursive to recurse into sub-directories.
2930 (iv) Added -f, --file to read patterns from a file.
2931
2932 6. pcre_exec() was referring to its "code" argument before testing that
2933 argument for NULL (and giving an error if it was NULL).
2934
2935 7. Upgraded Makefile.in to allow for compiling in a different directory from
2936 the source directory.
2937
2938 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2939 options bits, the pointer it was passed was to an int instead of to an unsigned
2940 long int. This mattered only on 64-bit systems.
2941
2942 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2943 generated) instead of pcre.in, which it its source. Also made the same change
2944 in several of the .c files.
2945
2946 10. A new release of gcc defines printf() as a macro, which broke pcretest
2947 because it had an ifdef in the middle of a string argument for printf(). Fixed
2948 by using separate calls to printf().
2949
2950 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2951 script, to force use of CR or LF instead of \n in the source. On non-Unix
2952 systems, the value can be set in config.h.
2953
2954 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
2955 absolute limit. Changed the text of the error message to make this clear, and
2956 likewise updated the man page.
2957
2958 13. The limit of 99 on the number of capturing subpatterns has been removed.
2959 The new limit is 65535, which I hope will not be a "real" limit.
2960
2961
2962 Version 3.4 22-Aug-00
2963 ---------------------
2964
2965 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
2966
2967 2. Diagnose condition (?(0) as an error instead of crashing on matching.
2968
2969
2970 Version 3.3 01-Aug-00
2971 ---------------------
2972
2973 1. If an octal character was given, but the value was greater than \377, it
2974 was not getting masked to the least significant bits, as documented. This could
2975 lead to crashes in some systems.
2976
2977 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
2978 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
2979
2980 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
2981 These just pass their arguments on to (pcre_free)(), but they are provided
2982 because some uses of PCRE bind it to non-C systems that can call its functions,
2983 but cannot call free() or pcre_free() directly.
2984
2985 4. Add "make test" as a synonym for "make check". Corrected some comments in
2986 the Makefile.
2987
2988 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
2989 Makefile.
2990
2991 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
2992 command called pgrep for grepping around the active processes.
2993
2994 7. Added the beginnings of support for UTF-8 character strings.
2995
2996 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
2997 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
2998 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
2999 out for the ar command.)
3000
3001
3002 Version 3.2 12-May-00
3003 ---------------------
3004
3005 This is purely a bug fixing release.
3006
3007 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3008 of ZA. This was just one example of several cases that could provoke this bug,
3009 which was introduced by change 9 of version 2.00. The code for breaking
3010 infinite loops after an iteration that matches an empty string was't working
3011 correctly.
3012
3013 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3014 when matched against abbab (for example). After matching an empty string, it
3015 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3016 caused it to match further down the string than it should.
3017
3018 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3019 was there because it doesn't seem to be needed, and it causes trouble on some
3020 systems, as it is not a Standard C header. It has been removed.
3021
3022 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3023 were reported on the Macintosh. The changes were from
3024
3025 while ((c = *(++ptr)) != 0 && c != '\n');
3026 to
3027 while ((c = *(++ptr)) != 0 && c != '\n') ;
3028
3029 Totally extraordinary, but if that's what it takes...
3030
3031 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3032 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3033 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3034 assumes the way PCRE uses memmove() (always moving upwards).
3035
3036 6. PCRE is being used in one environment where strchr() is not available. There
3037 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3038 faster code anyway.
3039
3040
3041 Version 3.1 09-Feb-00
3042 ---------------------
3043
3044 The only change in this release is the fixing of some bugs in Makefile.in for
3045 the "install" target:
3046
3047 (1) It was failing to install pcreposix.h.
3048
3049 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3050
3051
3052 Version 3.0 01-Feb-00
3053 ---------------------
3054
3055 1. Add support for the /+ modifier to perltest (to output $` like it does in
3056 pcretest).
3057
3058 2. Add support for the /g modifier to perltest.
3059
3060 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3061 matches null strings.
3062
3063 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3064 pattern. Perl treats empty patterns specially - it reuses the most recent
3065 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3066 effect.
3067
3068 5. The POSIX interface was broken in that it was just handing over the POSIX
3069 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3070 required a bigger vector, with some working space on the end. This means that
3071 the POSIX wrapper now has to get and free some memory, and copy the results.
3072
3073 6. Added some simple autoconf support, placing the test data and the
3074 documentation in separate directories, re-organizing some of the
3075 information files, and making it build pcre-config (a GNU standard). Also added
3076 libtool support for building PCRE as a shared library, which is now the
3077 default.
3078
3079 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3080 09 are not valid octal constants. Single digits will be used for minor values
3081 less than 10.
3082
3083 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3084 existing programs that set these in the POSIX interface can use PCRE without
3085 modification.
3086
3087 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3088 return all that pcre_info() returns, plus additional data. The pcre_info()
3089 function is retained for compatibility, but is considered to be obsolete.
3090
3091 10. Added experimental recursion feature (?R) to handle one common case that
3092 Perl 5.6 will be able to do with (?p{...}).
3093
3094 11. Added support for POSIX character classes like [:alpha:], which Perl is
3095 adopting.
3096
3097
3098 Version 2.08 31-Aug-99
3099 ----------------------
3100
3101 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3102 trying to match at the startoffset position, but instead was moving forward to
3103 the next newline as if a previous match had failed.
3104
3105 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3106 and could get into a loop if a null string was matched other than at the start
3107 of the subject.
3108
3109 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3110 be distinguished at compile time, and for completeness also added PCRE_DATE.
3111
3112 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3113 in GnuWin32 environments.
3114
3115
3116 Version 2.07 29-Jul-99
3117 ----------------------
3118
3119 1. The documentation is now supplied in plain text form and HTML as well as in
3120 the form of man page sources.
3121
3122 2. C++ compilers don't like assigning (void *) values to other pointer types.
3123 In particular this affects malloc(). Although there is no problem in Standard
3124 C, I've put in casts to keep C++ compilers happy.
3125
3126 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3127 should be (const char *).
3128
3129 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3130 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3131 However, I haven't made this a standard facility. The documentation doesn't
3132 mention it, and the Makefile doesn't support it.
3133
3134 5. The Makefile now contains an "install" target, with editable destinations at
3135 the top of the file. The pcretest program is not installed.
3136
3137 6. pgrep -V now gives the PCRE version number and date.
3138
3139 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3140 causing the entire string to be ignored, instead of just the last character.
3141
3142 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3143 non-matching string, it can take a very, very long time, even for strings of
3144 quite modest length, because of the nested recursion. PCRE now does better in
3145 some of these cases. It does this by remembering the last required literal
3146 character in the pattern, and pre-searching the subject to ensure it is present
3147 before running the real match. In other words, it applies a heuristic to detect
3148 some types of certain failure quickly, and in the above example, if presented
3149 with a string that has no trailing " it gives "no match" very quickly.
3150
3151 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3152 other alternatives are tried instead.
3153
3154
3155 Version 2.06 09-Jun-99
3156 ----------------------
3157
3158 1. Change pcretest's output for amount of store used to show just the code
3159 space, because the remainder (the data block) varies in size between 32-bit and
3160 64-bit systems.
3161
3162 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3163 start matching at. This allows lookbehinds to work when searching for multiple
3164 occurrences in a string.
3165
3166 3. Added additional options to pcretest for testing multiple occurrences:
3167
3168 /+ outputs the rest of the string that follows a match
3169 /g loops for multiple occurrences, using the new startoffset argument
3170 /G loops for multiple occurrences by passing an incremented pointer
3171
3172 4. PCRE wasn't doing the "first character" optimization for patterns starting
3173 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3174 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3175 the letter 'x'. On long subject strings, this gives a significant speed-up.
3176
3177
3178 Version 2.05 21-Apr-99
3179 ----------------------
3180
3181 1. Changed the type of magic_number from int to long int so that it works
3182 properly on 16-bit systems.
3183
3184 2. Fixed a bug which caused patterns starting with .* not to work correctly
3185 when the subject string contained newline characters. PCRE was assuming
3186 anchoring for such patterns in all cases, which is not correct because .* will
3187 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3188 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3189 must be retried after every newline in the subject.
3190
3191
3192 Version 2.04 18-Feb-99
3193 ----------------------
3194
3195 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3196 computation of the store needed to hold the pattern was incorrect (too large).
3197 If such patterns were nested a few deep, this could multiply and become a real
3198 problem.
3199
3200 2. Added /M option to pcretest to show the memory requirement of a specific
3201 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3202
3203 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3204 compiled in such a way that the backtracking after subsequent failure was
3205 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3206 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3207
3208
3209 Version 2.03 02-Feb-99
3210 ----------------------
3211
3212 1. Fixed typo and small mistake in man page.
3213
3214 2. Added 4th condition (GPL supersedes if conflict) and created separate
3215 LICENCE file containing the conditions.
3216
3217 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3218 Perl, that is the internal \ allows the delimiter to be included in the
3219 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3220 the final delimiter, add \ to the end of the pattern (to test the error).
3221
3222 4. Added the convenience functions for extracting substrings after a successful
3223 match. Updated pcretest to make it able to test these functions.
3224
3225
3226 Version 2.02 14-Jan-99
3227 ----------------------
3228
3229 1. Initialized the working variables associated with each extraction so that
3230 their saving and restoring doesn't refer to uninitialized store.
3231
3232 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3233 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3234 fix the problem.
3235
3236 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3237 calls, and wasn't printing the correct value for compiling calls. Increased the
3238 default value of LOOPREPEAT, and the number of significant figures in the
3239 times.
3240
3241 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3242
3243 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3244 a building problem on Windows NT with a FAT file system.
3245
3246
3247 Version 2.01 21-Oct-98
3248 ----------------------
3249
3250 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3251 to character tables built by pcre_maketables() in the current locale. If NULL
3252 is passed, the default tables are used.
3253
3254
3255 Version 2.00 24-Sep-98
3256 ----------------------
3257
3258 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3259 it any more.
3260
3261 2. Allow quantification of (?>) groups, and make it work correctly.
3262
3263 3. The first character computation wasn't working for (?>) groups.
3264
3265 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3266 end of the subject) and add 5.005's \z, which really does match only at the
3267 very end of the subject.
3268
3269 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3270
3271 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3272 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3273 localized options. All options to pcre_study() were also removed.
3274
3275 7. Add other new features from 5.005:
3276
3277 $(?<= positive lookbehind
3278 $(?<! negative lookbehind
3279 (?imsx-imsx) added the unsetting capability
3280 such a setting is global if at outer level; local otherwise
3281 (?imsx-imsx:) non-capturing groups with option setting
3282 (?(cond)re|re) conditional pattern matching
3283
3284 A backreference to itself in a repeated group matches the previous
3285 captured string.
3286
3287 8. General tidying up of studying (both automatic and via "study")
3288 consequential on the addition of new assertions.
3289
3290 9. As in 5.005, unlimited repeated groups that could match an empty substring
3291 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3292 runtime if any iteration does actually match an empty substring.
3293
3294 10. Include the RunTest script in the distribution.
3295
3296 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3297 discrepancies, some of which were old and were also with respect to 5.004. They
3298 have now been fixed.
3299
3300
3301 Version 1.09 28-Apr-98
3302 ----------------------
3303
3304 1. A negated single character class followed by a quantifier with a minimum
3305 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3306 program crashes, or just wrong answers. This did not apply to negated classes
3307 containing more than one character, or to minima other than one.
3308
3309
3310 Version 1.08 27-Mar-98
3311 ----------------------
3312
3313 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3314
3315 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3316 latter must appear before anything that relies on it in the pattern.
3317
3318
3319 Version 1.07 16-Feb-98
3320 ----------------------
3321
3322 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3323 repeat of a potentially empty string).
3324
3325
3326 Version 1.06 23-Jan-98
3327 ----------------------
3328
3329 1. Added Markus Oberhumer's little patches for C++.
3330
3331 2. Literal strings longer than 255 characters were broken.
3332
3333
3334 Version 1.05 23-Dec-97
3335 ----------------------
3336
3337 1. Negated character classes containing more than one character were failing if
3338 PCRE_CASELESS was set at run time.
3339
3340
3341 Version 1.04 19-Dec-97
3342 ----------------------
3343
3344 1. Corrected the man page, where some "const" qualifiers had been omitted.
3345
3346 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3347 input syntax.
3348
3349 3. Fixed memory leak which occurred when a regex with back references was
3350 matched with an offsets vector that wasn't big enough. The temporary memory
3351 that is used in this case wasn't being freed if the match failed.
3352
3353 4. Tidied pcretest to ensure it frees memory that it gets.
3354
3355 5. Temporary memory was being obtained in the case where the passed offsets
3356 vector was exactly big enough.
3357
3358 6. Corrected definition of offsetof() from change 5 below.
3359
3360 7. I had screwed up change 6 below and broken the rules for the use of
3361 setjmp(). Now fixed.
3362
3363
3364 Version 1.03 18-Dec-97
3365 ----------------------
3366
3367 1. A erroneous regex with a missing opening parenthesis was correctly
3368 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3369 on some systems.
3370
3371 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3372 it was reported that one broken compiler failed on the former because "code" is
3373 also an independent variable.
3374
3375 3. The erroneous regex a[]b caused an array overrun reference.
3376
3377 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3378 fail on data ending with that character. (It was going on too far, and checking
3379 the next character, typically a binary zero.) This was specific to the
3380 optimized code for single-character negative classes.
3381
3382 5. Added a contributed patch from the TIN world which does the following:
3383
3384 + Add an undef for memmove, in case the the system defines a macro for it.
3385
3386 + Add a definition of offsetof(), in case there isn't one. (I don't know
3387 the reason behind this - offsetof() is part of the ANSI standard - but
3388 it does no harm).
3389
3390 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3391 most of the places where whitespace preceded '#'. I have given up and
3392 allowed the remaining 2 cases to be at the margin.
3393
3394 + Rename some variables in pcre to eliminate shadowing. This seems very
3395 pedantic, but does no harm, of course.
3396
3397 6. Moved the call to setjmp() into its own function, to get rid of warnings
3398 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3399
3400 7. Constructs such as \d{8,} were compiling into the equivalent of
3401 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3402 outcome, but in this particular case used more store than had been allocated,
3403 which caused the bug to be discovered because it threw up an internal error.
3404
3405 8. The debugging code in both pcre and pcretest for outputting the compiled
3406 form of a regex was going wrong in the case of back references followed by
3407 curly-bracketed repeats.
3408
3409
3410 Version 1.02 12-Dec-97
3411 ----------------------
3412
3413 1. Typos in pcre.3 and comments in the source fixed.
3414
3415 2. Applied a contributed patch to get rid of places where it used to remove
3416 'const' from variables, and fixed some signed/unsigned and uninitialized
3417 variable warnings.
3418
3419 3. Added the "runtest" target to Makefile.
3420
3421 4. Set default compiler flag to -O2 rather than just -O.
3422
3423
3424 Version 1.01 19-Nov-97
3425 ----------------------
3426
3427 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3428 like /([ab]*)*/, that is, for classes with more than one character in them.
3429
3430 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3431 as /((?>a*))*/ (a PCRE_EXTRA facility).
3432
3433
3434 Version 1.00 18-Nov-97
3435 ----------------------
3436
3437 1. Added compile-time macros to support systems such as SunOS4 which don't have
3438 memmove() or strerror() but have other things that can be used instead.
3439
3440 2. Arranged that "make clean" removes the executables.
3441
3442
3443 Version 0.99 27-Oct-97
3444 ----------------------
3445
3446 1. Fixed bug in code for optimizing classes with only one character. It was
3447 initializing a 32-byte map regardless, which could cause it to run off the end
3448 of the memory it had got.
3449
3450 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3451
3452
3453 Version 0.98 22-Oct-97
3454 ----------------------
3455
3456 1. Fixed bug in code for handling temporary memory usage when there are more
3457 back references than supplied space in the ovector. This could cause segfaults.
3458
3459
3460 Version 0.97 21-Oct-97
3461 ----------------------
3462
3463 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3464
3465 2. Optimized negated single characters not to use a bit map.
3466
3467 3. Brought error texts together as macro definitions; clarified some of them;
3468 fixed one that was wrong - it said "range out of order" when it meant "invalid
3469 escape sequence".
3470
3471 4. Changed some char * arguments to const char *.
3472
3473 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3474
3475 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3476 pcretest.
3477
3478
3479 Version 0.96 16-Oct-97
3480 ----------------------
3481
3482 1. Added a simple "pgrep" utility to the distribution.
3483
3484 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3485 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3486 where "ddd" means "one or more decimal digits".
3487
3488 3. Fixed serious bug. If a pattern had a back reference, but the call to
3489 pcre_exec() didn't supply a large enough ovector to record the related
3490 identifying subpattern, the match always failed. PCRE now remembers the number
3491 of the largest back reference, and gets some temporary memory in which to save
3492 the offsets during matching if necessary, in order to ensure that
3493 backreferences always work.
3494
3495 4. Increased the compatibility with Perl in a number of ways:
3496
3497 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3498 to request this handling. The option can be set at compile or exec time.
3499
3500 (b) $ matches before a terminating newline by default; an option
3501 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3502 mode). The option can be set at compile or exec time.
3503
3504 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3505 the same as Perl's. If the decimal number it represents is less than 10
3506 or there aren't that many previous left capturing parentheses, an octal
3507 escape is read. Inside a character class, it's always an octal escape,
3508 even if it is a single digit.
3509
3510 (d) An escaped but undefined alphabetic character is taken as a literal,
3511 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3512 escapes.
3513
3514 (e) {0} is now permitted. (The previous item is removed from the compiled
3515 pattern).
3516
3517 5. Changed all the names of code files so that the basic parts are no longer
3518 than 10 characters, and abolished the teeny "globals.c" file.
3519
3520 6. Changed the handling of character classes; they are now done with a 32-byte
3521 bit map always.
3522
3523 7. Added the -d and /D options to pcretest to make it possible to look at the
3524 internals of compilation without having to recompile pcre.
3525
3526
3527 Version 0.95 23-Sep-97
3528 ----------------------
3529
3530 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3531 \x20 at the start of a run of normal characters. These were being treated as
3532 real characters, instead of the source characters being re-checked.
3533
3534
3535 Version 0.94 18-Sep-97
3536 ----------------------
3537
3538 1. The functions are now thread-safe, with the caveat that the global variables
3539 containing pointers to malloc() and free() or alternative functions are the
3540 same for all threads.
3541
3542 2. Get pcre_study() to generate a bitmap of initial characters for non-
3543 anchored patterns when this is possible, and use it if passed to pcre_exec().
3544
3545
3546 Version 0.93 15-Sep-97
3547 ----------------------
3548
3549 1. /(b)|(:+)/ was computing an incorrect first character.
3550
3551 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3552 but not actually doing anything yet.
3553
3554 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3555 as Perl does (e.g. [-az] or [az-]).
3556
3557 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3558 all possible positions.
3559
3560 5. Split up into different modules to avoid including unneeded functions in a
3561 compiled binary. However, compile and exec are still in one module. The "study"
3562 function is split off.
3563
3564 6. The character tables are now in a separate module whose source is generated
3565 by an auxiliary program - but can then be edited by hand if required. There are
3566 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3567 toupper() in the code.
3568
3569 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3570 make them global. Abolish the function for setting them, as the caller can now
3571 set them directly.
3572
3573
3574 Version 0.92 11-Sep-97
3575 ----------------------
3576
3577 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3578 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3579
3580 2. Caseless matching was not working in character classes if the characters in
3581 the pattern were in upper case.
3582
3583 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3584
3585 4. Make PCRE_ANCHORED public and accept as a compile option.
3586
3587 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3588 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3589 pass them.
3590
3591 6. Give an error if bad option bits passed at compile or run time.
3592
3593 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3594 pcretest to cause it to pass that flag.
3595
3596 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3597 options, and the first character, if set.
3598
3599 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3600
3601
3602 Version 0.91 10-Sep-97
3603 ----------------------
3604
3605 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3606 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3607
3608 2. PCRE was looping on encountering an indefinitely repeated back reference to
3609 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3610 Perl does - treats the match as successful.
3611
3612 ****

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

  ViewVC Help
Powered by ViewVC 1.1.5