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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 432 - (show annotations) (download)
Wed Sep 2 16:02:56 2009 UTC (5 years, 1 month ago) by ph10
File size: 170420 byte(s)
Add REG_UNGREEDY (non-standard) to the POSIX interface.

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8 was out-of-date, and there was no check on the pcre_compile() error code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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12