ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 455 - (show annotations)
Sat Sep 26 19:12:32 2009 UTC (11 years, 5 months ago) by ph10
File size: 174131 byte(s)
Added lower bound length-finding to pcre_study() and use it when matching; make 
the value available via pcre_fullinfo(); also fixed bugs connected with
pcre_study() in pcre_dfa_exec(). 
1 ChangeLog for PCRE
2 ------------------
4 Version 8.00 ??-???-??
5 ----------------------
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8 was out-of-date, and there was no check on the pcre_compile() error code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.
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.
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.
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.
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.
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.
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.
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.
51 9. The restrictions on what a pattern can contain when partial matching is
52 requested for pcre_exec() have been removed. All patterns can now be
53 partially matched by this function. In addition, if there are at least two
54 slots in the offset vector, the offset of the earliest inspected character
55 for the match and the offset of the end of the subject are set in them when
56 PCRE_ERROR_PARTIAL is returned.
58 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
59 synonymous with PCRE_PARTIAL, for backwards compatibility, and
60 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
61 and may be more useful for multi-segment matching, especially with
62 pcre_exec().
64 11. Partial matching with pcre_exec() is now more intuitive. A partial match
65 used to be given if ever the end of the subject was reached; now it is
66 given only if matching could not proceed because another character was
67 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
68 string "Z", which now yields "no match" instead of "partial match". In the
69 case of pcre_dfa_exec(), "no match" is given if every matching path for the
70 final character ended with (*FAIL).
72 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
73 if the pattern had a "must contain" character that was already found in the
74 earlier partial match, unless partial matching was again requested. For
75 example, with the pattern /dog.(body)?/, the "must contain" character is
76 "g". If the first part-match was for the string "dog", restarting with
77 "sbody" failed. This bug has been fixed.
79 13. The string returned by pcre_dfa_exec() after a partial match has been
80 changed so that it starts at the first inspected character rather than the
81 first character of the match. This makes a difference only if the pattern
82 starts with a lookbehind assertion or \b or \B (\K is not supported by
83 pcre_dfa_exec()). It's an incompatible change, but it makes the two
84 matching functions compatible, and I think it's the right thing to do.
86 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
87 so that the demonstration program is easily available in environments where
88 PCRE has not been installed from source.
90 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
91 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
92 library.
94 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
95 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
96 is not the first non-POSIX option to be added. Clearly some people find
97 these options useful.
99 17. If a caller to the POSIX matching function regexec() passes a non-zero
100 value for nmatch with a NULL value for pmatch, the value of
101 nmatch is forced to zero.
103 18. RunGrepTest did not have a test for the availability of the -u option of
104 the diff command, as RunTest does. It now checks in the same way as
105 RunTest, and also checks for the -b option.
107 19. If an odd number of negated classes containing just a single character
108 interposed, within parentheses, between a forward reference to a named
109 subpattern and the definition of the subpattern, compilation crashed with
110 an internal error, complaining that it could not find the referenced
111 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
112 [The bug was that it was starting one character too far in when skipping
113 over the character class, thus treating the ] as data rather than
114 terminating the class. This meant it could skip too much.]
116 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
117 /g option in pcretest when the pattern contains \K, which makes it possible
118 to have an empty string match not at the start, even when the pattern is
119 anchored. Updated pcretest and pcredemo to use this option.
121 21. If the maximum number of capturing subpatterns in a recursion was greater
122 than the maximum at the outer level, the higher number was returned, but
123 with unset values at the outer level. The correct (outer level) value is
124 now given.
126 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
127 PCRE did not set those parentheses (unlike Perl). I have now found a way to
128 make it do so. The string so far is captured, making this feature
129 compatible with Perl.
131 23. The tests have been re-organized, adding tests 11 and 12, to make it
132 possible to check the Perl 5.10 features against Perl 5.10.
134 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
135 pattern matches a fixed length string. PCRE did not allow this; now it
136 does. Neither allows recursion.
138 25. I finally figured out how to implement a request to provide the minimum
139 length of subject string that was needed in order to match a given pattern.
140 (It was back references and recursion that I had previously got hung up
141 on.) This code has now been added to pcre_study(); it finds a lower bound
142 to the length of subject needed. It is not necessarily the greatest lower
143 bound, but using it to avoid searching strings that are too short does give
144 some useful speed-ups. The value is available to calling programs via
145 pcre_fullinfo().
147 26. While implementing 25, I discovered to my embarrassment that pcretest had
148 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
149 study optimizations had never been tested with that matching function.
150 Oops. What is worse, even when it was passed study data, there was a bug in
151 pcre_dfa_exec() that meant it never actually used it. Double oops. There
152 were also very few tests of studied patterns with pcre_dfa_exec().
155 Version 7.9 11-Apr-09
156 ---------------------
158 1. When building with support for bzlib/zlib (pcregrep) and/or readline
159 (pcretest), all targets were linked against these libraries. This included
160 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
161 libraries. This caused unwanted dependencies to be created. This problem
162 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
163 pcretest is linked with readline.
165 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
166 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
167 moved outside it again, because FALSE and TRUE are already defined in AIX,
168 but BOOL is not.
170 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
171 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
173 4. The pcregrep documentation said spaces were inserted as well as colons (or
174 hyphens) following file names and line numbers when outputting matching
175 lines. This is not true; no spaces are inserted. I have also clarified the
176 wording for the --colour (or --color) option.
178 5. In pcregrep, when --colour was used with -o, the list of matching strings
179 was not coloured; this is different to GNU grep, so I have changed it to be
180 the same.
182 6. When --colo(u)r was used in pcregrep, only the first matching substring in
183 each matching line was coloured. Now it goes on to look for further matches
184 of any of the test patterns, which is the same behaviour as GNU grep.
186 7. A pattern that could match an empty string could cause pcregrep to loop; it
187 doesn't make sense to accept an empty string match in pcregrep, so I have
188 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
189 seems to be how GNU grep behaves.
191 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
192 start or after a newline", because the conditional assertion was not being
193 correctly handled. The rule now is that both the assertion and what follows
194 in the first alternative must satisfy the test.
196 9. If auto-callout was enabled in a pattern with a conditional group whose
197 condition was an assertion, PCRE could crash during matching, both with
198 pcre_exec() and pcre_dfa_exec().
200 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
201 used for matching.
203 11. Unicode property support in character classes was not working for
204 characters (bytes) greater than 127 when not in UTF-8 mode.
206 12. Added the -M command line option to pcretest.
208 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
210 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
212 16. Added comments and documentation about mis-use of no_arg in the C++
213 wrapper.
215 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
216 from Martin Jerabek that uses macro names for all relevant character and
217 string constants.
219 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
220 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
221 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
222 these, but not everybody uses configure.
224 19. A conditional group that had only one branch was not being correctly
225 recognized as an item that could match an empty string. This meant that an
226 enclosing group might also not be so recognized, causing infinite looping
227 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
228 with the subject "ab", where knowledge that the repeated group can match
229 nothing is needed in order to break the loop.
231 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
232 exec(), but without supplying a callout function, matching went wrong.
234 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
235 leak if the size of the offset vector was greater than 30. When the vector
236 is smaller, the saved offsets during recursion go onto a local stack
237 vector, but for larger vectors malloc() is used. It was failing to free
238 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
239 error, in fact).
241 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
242 heapframe that is used only when UTF-8 support is enabled. This caused no
243 problem, but was untidy.
245 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
246 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
247 included within another project.
249 24. Steven Van Ingelgem's patches to add more options to the CMake support,
250 slightly modified by me:
252 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
253 not building pcregrep.
255 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
256 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
258 25. Forward references, both numeric and by name, in patterns that made use of
259 duplicate group numbers, could behave incorrectly or give incorrect errors,
260 because when scanning forward to find the reference group, PCRE was not
261 taking into account the duplicate group numbers. A pattern such as
262 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
264 26. Changed a few more instances of "const unsigned char *" to USPTR, making
265 the feature of a custom pointer more persuasive (as requested by a user).
267 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
268 pcretest.c, inside #ifndefs, because it seems they are sometimes already
269 pre-defined.
271 28. Added support for (*UTF8) at the start of a pattern.
273 29. Arrange for flags added by the "release type" setting in CMake to be shown
274 in the configuration summary.
277 Version 7.8 05-Sep-08
278 ---------------------
280 1. Replaced UCP searching code with optimized version as implemented for Ad
281 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
282 stage table and inline lookup instead of a function, giving speed ups of 2
283 to 5 times on some simple patterns that I tested. Permission was given to
284 distribute the MultiStage2.py script that generates the tables (it's not in
285 the tarball, but is in the Subversion repository).
287 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
288 scripts.
290 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
291 a group with a zero qualifier. The result of the study could be incorrect,
292 or the function might crash, depending on the pattern.
294 4. Caseless matching was not working for non-ASCII characters in back
295 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
296 It now works when Unicode Property Support is available.
298 5. In pcretest, an escape such as \x{de} in the data was always generating
299 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
300 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
301 truncation.
303 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
305 7. Added two (int) casts to pcregrep when printing the difference of two
306 pointers, in case they are 64-bit values.
308 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
309 test 2 if it fails.
311 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
312 and a #define of that name to empty if it is not externally set. This is to
313 allow users of MSVC to set it if necessary.
315 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
316 the convenience functions in the pcre_get.c source file.
318 11. An option change at the start of a pattern that had top-level alternatives
319 could cause overwriting and/or a crash. This command provoked a crash in
320 some environments:
322 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
324 This potential security problem was recorded as CVE-2008-2371.
326 12. For a pattern where the match had to start at the beginning or immediately
327 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
328 pcre_dfa_exec() could read past the end of the passed subject if there was
329 no match. To help with detecting such bugs (e.g. with valgrind), I modified
330 pcretest so that it places the subject at the end of its malloc-ed buffer.
332 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
333 exec() might read past the end of the data buffer in UTF-8 mode.
335 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
336 the data contained the byte 0x85 as part of a UTF-8 character within its
337 first line. This applied both to normal and DFA matching.
339 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
340 /^[^d]*?$/8 failed to match "abc".
342 16. Added a missing copyright notice to pcrecpp_internal.h.
344 17. Make it more clear in the documentation that values returned from
345 pcre_exec() in ovector are byte offsets, not character counts.
347 18. Tidied a few places to stop certain compilers from issuing warnings.
349 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
350 supplied by Stefan Weber. I made a further small update for 7.8 because
351 there is a change of source arrangements: the pcre_searchfuncs.c module is
352 replaced by pcre_ucd.c.
355 Version 7.7 07-May-08
356 ---------------------
358 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
359 a string to a long long, pretend we don't even have a long long." This is
360 done by checking for the strtoq, strtoll, and _strtoi64 functions.
362 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
363 pre-7.6 versions, which defined a global no_arg variable instead of putting
364 it in the RE class. (See also #8 below.)
366 3. Remove a line of dead code, identified by coverity and reported by Nuno
367 Lopes.
369 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
371 (1) The include/exclude patterns were being applied to the whole pathnames
372 of files, instead of just to the final components.
374 (2) If there was more than one level of directory, the subdirectories were
375 skipped unless they satisfied the include/exclude conditions. This is
376 inconsistent with GNU grep (and could even be seen as contrary to the
377 pcregrep specification - which I improved to make it absolutely clear).
378 The action now is always to scan all levels of directory, and just
379 apply the include/exclude patterns to regular files.
381 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
382 --exclude_dir in the tests to avoid scanning .svn directories.
384 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
385 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
386 doesn't support NULs in patterns.
388 7. Added some missing "const"s to declarations of static tables in
389 pcre_compile.c and pcre_dfa_exec.c.
391 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
392 caused by fix #2 above. (Subsequently also a second patch to fix the
393 first patch. And a third patch - this was a messy problem.)
395 9. Applied Craig's patch to remove the use of push_back().
397 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
398 matching function regexec().
400 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
401 which, however, unlike Perl's \g{...}, are subroutine calls, not back
402 references. PCRE supports relative numbers with this syntax (I don't think
403 Oniguruma does).
405 12. Previously, a group with a zero repeat such as (...){0} was completely
406 omitted from the compiled regex. However, this means that if the group
407 was called as a subroutine from elsewhere in the pattern, things went wrong
408 (an internal error was given). Such groups are now left in the compiled
409 pattern, with a new opcode that causes them to be skipped at execution
410 time.
412 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
413 to the way PCRE behaves:
415 (a) A lone ] character is dis-allowed (Perl treats it as data).
417 (b) A back reference to an unmatched subpattern matches an empty string
418 (Perl fails the current match path).
420 (c) A data ] in a character class must be notated as \] because if the
421 first data character in a class is ], it defines an empty class. (In
422 Perl it is not possible to have an empty class.) The empty class []
423 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
424 The negative empty class [^] matches any one character, independently
425 of the DOTALL setting.
427 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
428 non-existent subpattern following a character class starting with ']' and
429 containing () gave an internal compiling error instead of "reference to
430 non-existent subpattern". Fortunately, when the pattern did exist, the
431 compiled code was correct. (When scanning forwards to check for the
432 existencd of the subpattern, it was treating the data ']' as terminating
433 the class, so got the count wrong. When actually compiling, the reference
434 was subsequently set up correctly.)
436 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
437 it was being rejected as not supported by pcre_dfa_exec(), even though
438 other assertions are supported. I have made pcre_dfa_exec() support
439 (*FAIL).
441 16. The implementation of 13c above involved the invention of a new opcode,
442 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
443 cannot be changed at match time, I realized I could make a small
444 improvement to matching performance by compiling OP_ALLANY instead of
445 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
446 on the OP_ANY path.
448 17. Compiling pcretest on Windows with readline support failed without the
449 following two fixes: (1) Make the unistd.h include conditional on
450 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
452 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
453 ncurses library to be included for pcretest when ReadLine support is
454 requested, but also to allow for it to be overridden. This patch came from
455 Daniel Bergström.
457 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
458 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
459 any errors with the current Unicode tables. Thanks to Peter Kankowski for
460 spotting this.
463 Version 7.6 28-Jan-08
464 ---------------------
466 1. A character class containing a very large number of characters with
467 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
468 overflow.
470 2. Patch to cut out the "long long" test in pcrecpp_unittest when
471 HAVE_LONG_LONG is not defined.
473 3. Applied Christian Ehrlicher's patch to update the CMake build files to
474 bring them up to date and include new features. This patch includes:
476 - Fixed PH's badly added libz and libbz2 support.
477 - Fixed a problem with static linking.
478 - Added pcredemo. [But later removed - see 7 below.]
479 - Fixed dftables problem and added an option.
480 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
482 - Added readline support for pcretest.
483 - Added an listing of the option settings after cmake has run.
485 4. A user submitted a patch to Makefile that makes it easy to create
486 "pcre.dll" under mingw when using Configure/Make. I added stuff to
487 Makefile.am that cause it to include this special target, without
488 affecting anything else. Note that the same mingw target plus all
489 the other distribution libraries and programs are now supported
490 when configuring with CMake (see 6 below) instead of with
491 Configure/Make.
493 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
494 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
495 exported in the Windows port". It has not yet been confirmed that the patch
496 solves the problem, but it does no harm.
498 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
499 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
500 with CMake, and also correct the comment about stack recursion.
502 7. Remove the automatic building of pcredemo from the ./configure system and
503 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
504 of a program that users should build themselves after PCRE is installed, so
505 building it automatically is not really right. What is more, it gave
506 trouble in some build environments.
508 8. Further tidies to CMakeLists.txt from Sheri and Christian.
511 Version 7.5 10-Jan-08
512 ---------------------
514 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
515 values in parens when parsing an RE using the C++ wrapper."
517 2. Negative specials like \S did not work in character classes in UTF-8 mode.
518 Characters greater than 255 were excluded from the class instead of being
519 included.
521 3. The same bug as (2) above applied to negated POSIX classes such as
522 [:^space:].
524 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
525 defined or documented. It seems to have been a typo for PCRE_STATIC, so
526 I have changed it.
528 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
529 first named subpattern) and a construct such as (?&a) would reference the
530 first named subpattern whose name started with "a" (in other words, the
531 length check was missing). Both these problems are fixed. "Subpattern name
532 expected" is now given for (?&) (a zero-length name), and this patch also
533 makes it give the same error for \k'' (previously it complained that that
534 was a reference to a non-existent subpattern).
536 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
537 this is right because (?- can be followed by option settings as well as by
538 digits. I have, however, made the messages clearer.
540 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
541 than the number used in the conditional) now cause a compile-time error.
542 This is actually not compatible with Perl, which accepts such patterns, but
543 treats the conditional as always being FALSE (as PCRE used to), but it
544 seems to me that giving a diagnostic is better.
546 8. Change "alphameric" to the more common word "alphanumeric" in comments
547 and messages.
549 9. Fix two occurrences of "backslash" in comments that should have been
550 "backspace".
552 10. Remove two redundant lines of code that can never be obeyed (their function
553 was moved elsewhere).
555 11. The program that makes PCRE's Unicode character property table had a bug
556 which caused it to generate incorrect table entries for sequences of
557 characters that have the same character type, but are in different scripts.
558 It amalgamated them into a single range, with the script of the first of
559 them. In other words, some characters were in the wrong script. There were
560 thirteen such cases, affecting characters in the following ranges:
562 U+002b0 - U+002c1
563 U+0060c - U+0060d
564 U+0061e - U+00612
565 U+0064b - U+0065e
566 U+0074d - U+0076d
567 U+01800 - U+01805
568 U+01d00 - U+01d77
569 U+01d9b - U+01dbf
570 U+0200b - U+0200f
571 U+030fc - U+030fe
572 U+03260 - U+0327f
573 U+0fb46 - U+0fbb1
574 U+10450 - U+1049d
576 12. The -o option (show only the matching part of a line) for pcregrep was not
577 compatible with GNU grep in that, if there was more than one match in a
578 line, it showed only the first of them. It now behaves in the same way as
579 GNU grep.
581 13. If the -o and -v options were combined for pcregrep, it printed a blank
582 line for every non-matching line. GNU grep prints nothing, and pcregrep now
583 does the same. The return code can be used to tell if there were any
584 non-matching lines.
586 14. Added --file-offsets and --line-offsets to pcregrep.
588 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
589 infinitely looping recursion. The bug was that positive lookaheads were not
590 being skipped when checking for a possible empty match (negative lookaheads
591 and both kinds of lookbehind were skipped).
593 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
594 inclusion of <windows.h> to before rather than after the definition of
595 INVALID_FILE_ATTRIBUTES (patch from David Byron).
597 17. Specifying a possessive quantifier with a specific limit for a Unicode
598 character property caused pcre_compile() to compile bad code, which led at
599 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
600 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
601 caused the error; without that there was no problem.
603 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
605 19. Added --enable-pcretest-libreadline.
607 20. In pcrecpp.cc, the variable 'count' was incremented twice in
608 RE::GlobalReplace(). As a result, the number of replacements returned was
609 double what it should be. I removed one of the increments, but Craig sent a
610 later patch that removed the other one (the right fix) and added unit tests
611 that check the return values (which was not done before).
613 21. Several CMake things:
615 (1) Arranged that, when cmake is used on Unix, the libraries end up with
616 the names libpcre and libpcreposix, not just pcre and pcreposix.
618 (2) The above change means that pcretest and pcregrep are now correctly
619 linked with the newly-built libraries, not previously installed ones.
623 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
624 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
625 UTF-8 newline character). The key issue is that the pattern starts .*;
626 this means that the match must be either at the beginning, or after a
627 newline. The bug was in the code for advancing after a failed match and
628 checking that the new position followed a newline. It was not taking
629 account of UTF-8 characters correctly.
631 23. PCRE was behaving differently from Perl in the way it recognized POSIX
632 character classes. PCRE was not treating the sequence [:...:] as a
633 character class unless the ... were all letters. Perl, however, seems to
634 allow any characters between [: and :], though of course it rejects as
635 unknown any "names" that contain non-letters, because all the known class
636 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
637 for example, whereas PCRE did not - it did not recognize a POSIX character
638 class. This seemed a bit dangerous, so the code has been changed to be
639 closer to Perl. The behaviour is not identical to Perl, because PCRE will
640 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
641 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
642 Perl does, and where it didn't before.
644 24. Rewrite so as to remove the single use of %n from pcregrep because in some
645 Windows environments %n is disabled by default.
648 Version 7.4 21-Sep-07
649 ---------------------
651 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
652 means that a class such as [\s] counted as "explicit reference to CR or
653 LF". That isn't really right - the whole point of the change was to try to
654 help when there was an actual mention of one of the two characters. So now
655 the change happens only if \r or \n (or a literal CR or LF) character is
656 encountered.
658 2. The 32-bit options word was also used for 6 internal flags, but the numbers
659 of both had grown to the point where there were only 3 bits left.
660 Fortunately, there was spare space in the data structure, and so I have
661 moved the internal flags into a new 16-bit field to free up more option
662 bits.
664 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
665 but did not set the internal JCHANGED flag - either of these is enough to
666 control the way the "get" function works - but the PCRE_INFO_JCHANGED
667 facility is supposed to tell if (?J) was ever used, so now (?J) at the
668 start sets both bits.
670 4. Added options (at build time, compile time, exec time) to change \R from
671 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
673 5. doc/pcresyntax.html was missing from the distribution.
675 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
676 compatibility, even though it is no longer used.
678 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
679 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
680 windows.h file is present (where different names are used). [This was
681 reversed later after testing - see 16 below.]
683 8. Changed all #include <config.h> to #include "config.h". There were also
684 some further <pcre.h> cases that I changed to "pcre.h".
686 9. When pcregrep was used with the --colour option, it missed the line ending
687 sequence off the lines that it output.
689 10. It was pointed out to me that arrays of string pointers cause lots of
690 relocations when a shared library is dynamically loaded. A technique of
691 using a single long string with a table of offsets can drastically reduce
692 these. I have refactored PCRE in four places to do this. The result is
693 dramatic:
695 Originally: 290
696 After changing UCP table: 187
697 After changing error message table: 43
698 After changing table of "verbs" 36
699 After changing table of Posix names 22
701 Thanks to the folks working on Gregex for glib for this insight.
703 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
704 unicode-properties was also set.
706 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
708 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
709 checked only for CRLF.
711 14. Added casts to pcretest.c to avoid compiler warnings.
713 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
715 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
716 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
717 entirely. This removes changes made in 7 above.
719 17. The CMake files have been updated, and there is now more information about
720 building with CMake in the NON-UNIX-USE document.
723 Version 7.3 28-Aug-07
724 ---------------------
726 1. In the rejigging of the build system that eventually resulted in 7.1, the
727 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
728 brackets there is not right, since it causes compilers to look for an
729 installed pcre.h, not the version that is in the source that is being
730 compiled (which of course may be different). I have changed it back to:
732 #include "pcre.h"
734 I have a vague recollection that the change was concerned with compiling in
735 different directories, but in the new build system, that is taken care of
736 by the VPATH setting the Makefile.
738 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
739 when the subject happened to end in the byte 0x85 (e.g. if the last
740 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
741 characters but of course it shouldn't be taken as a newline when it is part
742 of another character. The bug was that, for an unlimited repeat of . in
743 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
744 characters when looking for a newline.
746 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
748 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
749 in debug output.
751 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
752 long printing in the pcrecpp unittest when running under MinGW.
754 6. ESC_K was left out of the EBCDIC table.
756 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
757 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
758 limit also applies to "virtual nesting" when a pattern is recursive, and in
759 this case 1000 isn't so big. I have been able to remove this limit at the
760 expense of backing off one optimization in certain circumstances. Normally,
761 when pcre_exec() would call its internal match() function recursively and
762 immediately return the result unconditionally, it uses a "tail recursion"
763 feature to save stack. However, when a subpattern that can match an empty
764 string has an unlimited repetition quantifier, it no longer makes this
765 optimization. That gives it a stack frame in which to save the data for
766 checking that an empty string has been matched. Previously this was taken
767 from the 1000-entry workspace that had been reserved. So now there is no
768 explicit limit, but more stack is used.
770 8. Applied Daniel's patches to solve problems with the import/export magic
771 syntax that is required for Windows, and which was going wrong for the
772 pcreposix and pcrecpp parts of the library. These were overlooked when this
773 problem was solved for the main library.
775 9. There were some crude static tests to avoid integer overflow when computing
776 the size of patterns that contain repeated groups with explicit upper
777 limits. As the maximum quantifier is 65535, the maximum group length was
778 set at 30,000 so that the product of these two numbers did not overflow a
779 32-bit integer. However, it turns out that people want to use groups that
780 are longer than 30,000 bytes (though not repeat them that many times).
781 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
782 made it possible to implement the integer overflow checks in a much more
783 dynamic way, which I have now done. The artificial limitation on group
784 length has been removed - we now have only the limit on the total length of
785 the compiled pattern, which depends on the LINK_SIZE setting.
787 10. Fixed a bug in the documentation for get/copy named substring when
788 duplicate names are permitted. If none of the named substrings are set, the
789 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
790 empty string.
792 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
793 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
794 because the ] is interpreted as the first data character and the
795 terminating ] is not found. PCRE has been made compatible with Perl in this
796 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
797 cause memory overwriting.
799 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
800 string has been matched (to stop an infinite loop). It was not recognizing
801 a conditional subpattern that could match an empty string if that
802 subpattern was within another subpattern. For example, it looped when
803 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
804 condition was not nested. This bug has been fixed.
806 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
807 past the start of the subject in the presence of bytes with the top bit
808 set, for example "\x8aBCD".
810 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
811 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
813 14. Optimized (?!) to (*FAIL).
815 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
816 This restricts code points to be within the range 0 to 0x10FFFF, excluding
817 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
818 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
819 does: it's just the validity check that is more restrictive.
821 16. Inserted checks for integer overflows during escape sequence (backslash)
822 processing, and also fixed erroneous offset values for syntax errors during
823 backslash processing.
825 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
826 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
828 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
829 caused an overrun.
831 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
832 something other than just ASCII characters) inside a group that had an
833 unlimited repeat caused a loop at compile time (while checking to see
834 whether the group could match an empty string).
836 20. Debugging a pattern containing \p or \P could cause a crash. For example,
837 [\P{Any}] did so. (Error in the code for printing property names.)
839 21. An orphan \E inside a character class could cause a crash.
841 22. A repeated capturing bracket such as (A)? could cause a wild memory
842 reference during compilation.
844 23. There are several functions in pcre_compile() that scan along a compiled
845 expression for various reasons (e.g. to see if it's fixed length for look
846 behind). There were bugs in these functions when a repeated \p or \P was
847 present in the pattern. These operators have additional parameters compared
848 with \d, etc, and these were not being taken into account when moving along
849 the compiled data. Specifically:
851 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
852 length.
854 (b) An item such as \pL+ within a repeated group could cause crashes or
855 loops.
857 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
858 "reference to non-existent subpattern" error.
860 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
862 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
863 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
865 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
867 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
868 character were causing crashes (broken optimization).
870 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
871 \p or \P) caused a compile-time loop.
873 28. More problems have arisen in unanchored patterns when CRLF is a valid line
874 break. For example, the unstudied pattern [\r\n]A does not match the string
875 "\r\nA" because change 7.0/46 below moves the current point on by two
876 characters after failing to match at the start. However, the pattern \nA
877 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
878 the same is true. There doesn't seem any very clean way out of this, but
879 what I have chosen to do makes the common cases work: PCRE now takes note
880 of whether there can be an explicit match for \r or \n anywhere in the
881 pattern, and if so, 7.0/46 no longer applies. As part of this change,
882 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
883 pattern has explicit CR or LF references.
885 29. Added (*CR) etc for changing newline setting at start of pattern.
888 Version 7.2 19-Jun-07
889 ---------------------
891 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
892 which is apparently normally available under Windows.
894 2. Re-jig the pcregrep tests with different newline settings in an attempt
895 to make them independent of the local environment's newline setting.
897 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
899 4. Some of the "internals" tests were previously cut out when the link size
900 was not 2, because the output contained actual offsets. The recent new
901 "Z" feature of pcretest means that these can be cut out, making the tests
902 usable with all link sizes.
904 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
905 stack recursion. This gives a massive performance boost under BSD, but just
906 a small improvement under Linux. However, it saves one field in the frame
907 in all cases.
909 6. Added more features from the forthcoming Perl 5.10:
911 (a) (?-n) (where n is a string of digits) is a relative subroutine or
912 recursion call. It refers to the nth most recently opened parentheses.
914 (b) (?+n) is also a relative subroutine call; it refers to the nth next
915 to be opened parentheses.
917 (c) Conditions that refer to capturing parentheses can be specified
918 relatively, for example, (?(-2)... or (?(+3)...
920 (d) \K resets the start of the current match so that everything before
921 is not part of it.
923 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
925 (f) \g{name} is another synonym - part of Perl 5.10's unification of
926 reference syntax.
928 (g) (?| introduces a group in which the numbering of parentheses in each
929 alternative starts with the same number.
931 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
933 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
936 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
937 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
938 for detecting groups that can match an empty string.
940 9. A pattern with a very large number of alternatives (more than several
941 hundred) was running out of internal workspace during the pre-compile
942 phase, where pcre_compile() figures out how much memory will be needed. A
943 bit of new cunning has reduced the workspace needed for groups with
944 alternatives. The 1000-alternative test pattern now uses 12 bytes of
945 workspace instead of running out of the 4096 that are available.
947 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
949 11. Applied patch from Google to remove an optimization that didn't quite work.
950 The report of the bug said:
952 pcrecpp::RE("a*").FullMatch("aaa") matches, while
953 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
954 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
956 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
957 it matched the wrong number of bytes.
960 Version 7.1 24-Apr-07
961 ---------------------
963 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
964 that is more "standard", making use of automake and other Autotools. There
965 is some re-arrangement of the files and adjustment of comments consequent
966 on this.
968 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
969 for recursive directory scanning broke on some systems because the files
970 are not scanned in any specific order and on different systems the order
971 was different. A call to "sort" has been inserted into RunGrepTest for the
972 approprate test as a short-term fix. In the longer term there may be an
973 alternative.
975 3. I had an email from Eric Raymond about problems translating some of PCRE's
976 man pages to HTML (despite the fact that I distribute HTML pages, some
977 people do their own conversions for various reasons). The problems
978 concerned the use of low-level troff macros .br and .in. I have therefore
979 removed all such uses from the man pages (some were redundant, some could
980 be replaced by .nf/.fi pairs). The 132html script that I use to generate
981 HTML has been updated to handle .nf/.fi and to complain if it encounters
982 .br or .in.
984 4. Updated comments in configure.ac that get placed in config.h.in and also
985 arranged for config.h to be included in the distribution, with the name
986 config.h.generic, for the benefit of those who have to compile without
987 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
989 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
990 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
991 makevp.bat for the current PCRE, using the additional files
992 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
994 6. A Windows user reported a minor discrepancy with test 2, which turned out
995 to be caused by a trailing space on an input line that had got lost in his
996 copy. The trailing space was an accident, so I've just removed it.
998 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
999 that is needed.
1001 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1002 as "const" (a) because they are and (b) because it helps the PHP
1003 maintainers who have recently made a script to detect big data structures
1004 in the php code that should be moved to the .rodata section. I remembered
1005 to update Builducptable as well, so it won't revert if ucptable.h is ever
1006 re-created.
1008 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1009 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1010 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1011 support is not required. This saves 1.5-2K of code, which is important in
1012 some applications.
1014 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1015 so as not to refer to the tables, even though these functions will never be
1016 called when UTF-8 support is disabled. Otherwise there are problems with a
1017 shared library.
1019 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1021 (a) It was defining its arguments as char * instead of void *.
1023 (b) It was assuming that all moves were upwards in memory; this was true
1024 a long time ago when I wrote it, but is no longer the case.
1026 The emulated memove() is provided for those environments that have neither
1027 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1028 is clearly not the case, as these two bugs were recently reported.
1030 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1031 and Detrail to create the HTML documentation, the .txt form of the man
1032 pages, and it removes trailing spaces from listed files. It also creates
1033 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1034 case, it wraps all the #defines with #ifndefs. This script should be run
1035 before "make dist".
1037 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1038 with Unicode property support.
1040 (a) For a maximizing quantifier, if the two different cases of the
1041 character were of different lengths in their UTF-8 codings (there are
1042 some cases like this - I found 11), and the matching function had to
1043 back up over a mixture of the two cases, it incorrectly assumed they
1044 were both the same length.
1046 (b) When PCRE was configured to use the heap rather than the stack for
1047 recursion during matching, it was not correctly preserving the data for
1048 the other case of a UTF-8 character when checking ahead for a match
1049 while processing a minimizing repeat. If the check also involved
1050 matching a wide character, but failed, corruption could cause an
1051 erroneous result when trying to check for a repeat of the original
1052 character.
1054 13. Some tidying changes to the testing mechanism:
1056 (a) The RunTest script now detects the internal link size and whether there
1057 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1058 values substituted by "configure". (The RunGrepTest script already did
1059 this for UTF-8.) The configure.ac script no longer substitutes the
1060 relevant variables.
1062 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1063 with length and offset values. This means that the output is different
1064 for different internal link sizes. Test 2 is skipped for link sizes
1065 other than 2 because of this, bypassing the problem. Unfortunately,
1066 there was also a test in test 3 (the locale tests) that used /B and
1067 failed for link sizes other than 2. Rather than cut the whole test out,
1068 I have added a new /Z option to pcretest that replaces the length and
1069 offset values with spaces. This is now used to make test 3 independent
1070 of link size. (Test 2 will be tidied up later.)
1072 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1073 segmentation fault instead of returning the appropriate error message.
1075 15. In multiline mode when the newline sequence was set to "any", the pattern
1076 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1077 This doesn't seem right; it now treats the CRLF combination as the line
1078 ending, and so does not match in that case. It's only a pattern such as ^$
1079 that would hit this one: something like ^ABC$ would have failed after \r
1080 and then tried again after \r\n.
1082 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1083 in an attempt to make files that differ only in their line terminators
1084 compare equal. This works on Linux.
1086 17. Under certain error circumstances pcregrep might try to free random memory
1087 as it exited. This is now fixed, thanks to valgrind.
1089 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1090 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1091 was because its rules for how to advance for /g after matching an empty
1092 string at the end of a line did not allow for this case. They now check for
1093 it specially.
1095 20. pcretest is supposed to handle patterns and data of any length, by
1096 extending its buffers when necessary. It was getting this wrong when the
1097 buffer for a data line had to be extended.
1099 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1100 CRLF as a newline sequence.
1102 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1103 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1104 I have nevertheless tidied it up.
1106 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1108 24. Added a man page for pcre-config.
1111 Version 7.0 19-Dec-06
1112 ---------------------
1114 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1115 moving to gcc 4.1.1.
1117 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1118 sys/time.h, which is documented as needed for this function. It doesn't
1119 seem to matter on Linux, but it showed up on some releases of OS X.
1121 3. It seems that there are systems where bytes whose values are greater than
1122 127 match isprint() in the "C" locale. The "C" locale should be the
1123 default when a C program starts up. In most systems, only ASCII printing
1124 characters match isprint(). This difference caused the output from pcretest
1125 to vary, making some of the tests fail. I have changed pcretest so that:
1127 (a) When it is outputting text in the compiled version of a pattern, bytes
1128 other than 32-126 are always shown as hex escapes.
1130 (b) When it is outputting text that is a matched part of a subject string,
1131 it does the same, unless a different locale has been set for the match
1132 (using the /L modifier). In this case, it uses isprint() to decide.
1134 4. Fixed a major bug that caused incorrect computation of the amount of memory
1135 required for a compiled pattern when options that changed within the
1136 pattern affected the logic of the preliminary scan that determines the
1137 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1138 that the computed length was too small. The symptoms of this bug were
1139 either the PCRE error "internal error: code overflow" from pcre_compile(),
1140 or a glibc crash with a message such as "pcretest: free(): invalid next
1141 size (fast)". Examples of patterns that provoked this bug (shown in
1142 pcretest format) are:
1144 /(?-x: )/x
1145 /(?x)(?-x: \s*#\s*)/
1146 /((?i)[\x{c0}])/8
1147 /(?i:[\x{c0}])/8
1149 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1150 is now done differently.
1152 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1153 wrapper classes; (b) implement a new function in the C++ scanner that is
1154 more efficient than the old way of doing things because it avoids levels of
1155 recursion in the regex matching; (c) add a paragraph to the documentation
1156 for the FullMatch() function.
1158 6. The escape sequence \n was being treated as whatever was defined as
1159 "newline". Not only was this contrary to the documentation, which states
1160 that \n is character 10 (hex 0A), but it also went horribly wrong when
1161 "newline" was defined as CRLF. This has been fixed.
1163 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1164 was being set to -1 for the "end of line" case (supposedly a value that no
1165 character can have). Though this value is never used (the check for end of
1166 line is "zero bytes in current character"), it caused compiler complaints.
1167 I've changed it to 0xffffffff.
1169 8. In pcre_version.c, the version string was being built by a sequence of
1170 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1171 string (as it is for production releases) called a macro with an empty
1172 argument. The C standard says the result of this is undefined. The gcc
1173 compiler treats it as an empty string (which was what was wanted) but it is
1174 reported that Visual C gives an error. The source has been hacked around to
1175 avoid this problem.
1177 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1178 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1179 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1180 of them did).
1182 10. Originally, pcretest opened its input and output without "b"; then I was
1183 told that "b" was needed in some environments, so it was added for release
1184 5.0 to both the input and output. (It makes no difference on Unix-like
1185 systems.) Later I was told that it is wrong for the input on Windows. I've
1186 now abstracted the modes into two macros, to make it easier to fiddle with
1187 them, and removed "b" from the input mode under Windows.
1189 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1191 12. Added -help and --help to pcretest as an official way of being reminded
1192 of the options.
1194 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1195 and pcrecpp.cc because they annoy compilers at high warning levels.
1197 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1199 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1200 scripts are not C programs :-) and which was not noticed because it works
1201 on Linux.
1203 16. pcretest is supposed to handle any length of pattern and data line (as one
1204 line or as a continued sequence of lines) by extending its input buffer if
1205 necessary. This feature was broken for very long pattern lines, leading to
1206 a string of junk being passed to pcre_compile() if the pattern was longer
1207 than about 50K.
1209 17. I have done a major re-factoring of the way pcre_compile() computes the
1210 amount of memory needed for a compiled pattern. Previously, there was code
1211 that made a preliminary scan of the pattern in order to do this. That was
1212 OK when PCRE was new, but as the facilities have expanded, it has become
1213 harder and harder to keep it in step with the real compile phase, and there
1214 have been a number of bugs (see for example, 4 above). I have now found a
1215 cunning way of running the real compile function in a "fake" mode that
1216 enables it to compute how much memory it would need, while actually only
1217 ever using a few hundred bytes of working memory and without too many
1218 tests of the mode. This should make future maintenance and development
1219 easier. A side effect of this work is that the limit of 200 on the nesting
1220 depth of parentheses has been removed (though this was never a serious
1221 limitation, I suspect). However, there is a downside: pcre_compile() now
1222 runs more slowly than before (30% or more, depending on the pattern). I
1223 hope this isn't a big issue. There is no effect on runtime performance.
1225 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1226 newline (only possible for the last line of a file) and it was a
1227 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1229 19. Added additional timing features to pcretest. (1) The -tm option now times
1230 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1231 separate command line item, by a number that specifies the number of
1232 repeats to use when timing. The default is 50000; this gives better
1233 precision, but takes uncomfortably long for very large patterns.
1235 20. Extended pcre_study() to be more clever in cases where a branch of a
1236 subpattern has no definite first character. For example, (a*|b*)[cd] would
1237 previously give no result from pcre_study(). Now it recognizes that the
1238 first character must be a, b, c, or d.
1240 21. There was an incorrect error "recursive call could loop indefinitely" if
1241 a subpattern (or the entire pattern) that was being tested for matching an
1242 empty string contained only one non-empty item after a nested subpattern.
1243 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1244 incorrectly, because the \d was being skipped in the check.
1246 22. The pcretest program now has a new pattern option /B and a command line
1247 option -b, which is equivalent to adding /B to every pattern. This causes
1248 it to show the compiled bytecode, without the additional information that
1249 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1250 is the same as /B/I).
1252 23. A new optimization is now able automatically to treat some sequences such
1253 as a*b as a*+b. More specifically, if something simple (such as a character
1254 or a simple class like \d) has an unlimited quantifier, and is followed by
1255 something that cannot possibly match the quantified thing, the quantifier
1256 is automatically "possessified".
1258 24. A recursive reference to a subpattern whose number was greater than 39
1259 went wrong under certain circumstances in UTF-8 mode. This bug could also
1260 have affected the operation of pcre_study().
1262 25. Realized that a little bit of performance could be had by replacing
1263 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1265 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1267 27. Possessive quantifiers such as a++ were previously implemented by turning
1268 them into atomic groups such as ($>a+). Now they have their own opcodes,
1269 which improves performance. This includes the automatically created ones
1270 from 23 above.
1272 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1273 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1274 the first matched character to be a colon. This applied both to named and
1275 numbered groups.
1277 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1279 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1280 building PCRE on Windows easier, so I have included it in the distribution.
1282 31. There is now a check in pcretest against a ridiculously large number being
1283 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1284 loop, the loop is abandoned.
1286 32. Forward references to subpatterns in conditions such as (?(2)...) where
1287 subpattern 2 is defined later cause pcre_compile() to search forwards in
1288 the pattern for the relevant set of parentheses. This search went wrong
1289 when there were unescaped parentheses in a character class, parentheses
1290 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1292 33. "Subroutine" calls and backreferences were previously restricted to
1293 referencing subpatterns earlier in the regex. This restriction has now
1294 been removed.
1296 34. Added a number of extra features that are going to be in Perl 5.10. On the
1297 whole, these are just syntactic alternatives for features that PCRE had
1298 previously implemented using the Python syntax or my own invention. The
1299 other formats are all retained for compatibility.
1301 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1302 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1303 also .NET compatible.
1305 (b) A recursion or subroutine call to a named group can now be defined as
1306 (?&name) as well as (?P>name).
1308 (c) A backreference to a named group can now be defined as \k<name> or
1309 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1310 5.10, are also .NET compatible.
1312 (d) A conditional reference to a named group can now use the syntax
1313 (?(<name>) or (?('name') as well as (?(name).
1315 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1316 groups (named and numbered) that are never evaluated inline, but can be
1317 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1318 is always false. There may be only one alternative in such a group.
1320 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1321 as the simple (?(R). The condition is true only if the most recent
1322 recursion is that of the given number or name. It does not search out
1323 through the entire recursion stack.
1325 (g) The escape \gN or \g{N} has been added, where N is a positive or
1326 negative number, specifying an absolute or relative reference.
1328 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1329 some "unreachable code" warnings.
1331 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1332 things, this adds five new scripts.
1334 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1335 There were also incompatibilities regarding the handling of \Q..\E inside
1336 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1337 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1339 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1340 matches an empty string, and forcibly breaks the loop. There were bugs in
1341 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1342 against aaaa the result was just "a" rather than "aaaa", for example. Two
1343 separate and independent bugs (that affected different cases) have been
1344 fixed.
1346 39. Refactored the code to abolish the use of different opcodes for small
1347 capturing bracket numbers. This is a tidy that I avoided doing when I
1348 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1349 The new approach is not only tidier, it makes it possible to reduce the
1350 memory needed to fix the previous bug (38).
1352 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1353 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1354 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1355 mode.
1357 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1358 report.
1360 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1361 copy construction and assignment in the C++ wrapper.
1363 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1364 couple of bugs that could have given wrong results in the "--newline=crlf"
1365 case.
1367 44. Added a number of casts and did some reorganization of signed/unsigned int
1368 variables following suggestions from Dair Grant. Also renamed the variable
1369 "this" as "item" because it is a C++ keyword.
1371 45. Arranged for dftables to add
1373 #include "pcre_internal.h"
1375 to pcre_chartables.c because without it, gcc 4.x may remove the array
1376 definition from the final binary if PCRE is built into a static library and
1377 dead code stripping is activated.
1379 46. For an unanchored pattern, if a match attempt fails at the start of a
1380 newline sequence, and the newline setting is CRLF or ANY, and the next two
1381 characters are CRLF, advance by two characters instead of one.
1384 Version 6.7 04-Jul-06
1385 ---------------------
1387 1. In order to handle tests when input lines are enormously long, pcretest has
1388 been re-factored so that it automatically extends its buffers when
1389 necessary. The code is crude, but this _is_ just a test program. The
1390 default size has been increased from 32K to 50K.
1392 2. The code in pcre_study() was using the value of the re argument before
1393 testing it for NULL. (Of course, in any sensible call of the function, it
1394 won't be NULL.)
1396 3. The memmove() emulation function in pcre_internal.h, which is used on
1397 systems that lack both memmove() and bcopy() - that is, hardly ever -
1398 was missing a "static" storage class specifier.
1400 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1401 containing an extended class (one that cannot be represented by a bitmap
1402 because it contains high-valued characters or Unicode property items, e.g.
1403 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1404 pattern, but PCRE should not loop if you do not (it no longer does).
1405 [Detail: two cases were found: (a) a repeated subpattern containing an
1406 extended class; (b) a recursive reference to a subpattern that followed a
1407 previous extended class. It wasn't skipping over the extended class
1408 correctly when UTF-8 mode was not set.]
1410 5. A negated single-character class was not being recognized as fixed-length
1411 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1412 compile error "lookbehind assertion is not fixed length".
1414 6. The RunPerlTest auxiliary script was showing an unexpected difference
1415 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1416 write a Perl script that can interpret lines of an input file either as
1417 byte characters or as UTF-8, which is what "perltest" was being required to
1418 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1419 can't do is switch easily at run time between having the "use utf8;" pragma
1420 or not. In the end, I fudged it by using the RunPerlTest script to insert
1421 "use utf8;" explicitly for the UTF-8 tests.
1423 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1424 the end of the subject string, contrary to the documentation and to what
1425 Perl does. This was true of both matching functions. Now it matches only at
1426 the start of the subject and immediately after *internal* newlines.
1428 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1429 a pointer to an int instead of a pointer to an unsigned long int. This
1430 caused problems on 64-bit systems.
1432 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1433 instance of the 'standard' template library not being so standard".
1435 10. There was no check on the number of named subpatterns nor the maximum
1436 length of a subpattern name. The product of these values is used to compute
1437 the size of the memory block for a compiled pattern. By supplying a very
1438 long subpattern name and a large number of named subpatterns, the size
1439 computation could be caused to overflow. This is now prevented by limiting
1440 the length of names to 32 characters, and the number of named subpatterns
1441 to 10,000.
1443 11. Subpatterns that are repeated with specific counts have to be replicated in
1444 the compiled pattern. The size of memory for this was computed from the
1445 length of the subpattern and the repeat count. The latter is limited to
1446 65535, but there was no limit on the former, meaning that integer overflow
1447 could in principle occur. The compiled length of a repeated subpattern is
1448 now limited to 30,000 bytes in order to prevent this.
1450 12. Added the optional facility to have named substrings with the same name.
1452 13. Added the ability to use a named substring as a condition, using the
1453 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1454 are numbers (not recommended). Forward references are permitted.
1456 14. Added forward references in named backreferences (if you see what I mean).
1458 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1459 pattern could run off the end of the subject. For example, the pattern
1460 "(?s)(.{1,5})"8 did this with the subject "ab".
1462 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1463 PCRE_CASELESS was set when matching characters that were quantified with ?
1464 or *.
1466 17. A character class other than a single negated character that had a minimum
1467 but no maximum quantifier - for example [ab]{6,} - was not handled
1468 correctly by pce_dfa_exec(). It would match only one character.
1470 18. A valid (though odd) pattern that looked like a POSIX character
1471 class but used an invalid character after [ (for example [[,abc,]]) caused
1472 pcre_compile() to give the error "Failed: internal error: code overflow" or
1473 in some cases to crash with a glibc free() error. This could even happen if
1474 the pattern terminated after [[ but there just happened to be a sequence of
1475 letters, a binary zero, and a closing ] in the memory that followed.
1477 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1478 over the years. Originally (before any Unicode support), just the bottom 8
1479 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1480 output from "man perlunicode" includes this:
1482 The regular expression compiler produces polymorphic opcodes. That
1483 is, the pattern adapts to the data and automatically switches to
1484 the Unicode character scheme when presented with Unicode data--or
1485 instead uses a traditional byte scheme when presented with byte
1486 data.
1488 Sadly, a wide octal escape does not cause a switch, and in a string with
1489 no other multibyte characters, these octal escapes are treated as before.
1490 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1491 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1492 Unicode string.
1494 I have not perpetrated such confusion in PCRE. Up till now, it took just
1495 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1496 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1497 translate to the appropriate multibyte character.
1499 29. Applied some refactoring to reduce the number of warnings from Microsoft
1500 and Borland compilers. This has included removing the fudge introduced
1501 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1502 a warning about an unused variable.
1504 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1505 characters since release 4.0, because Perl (from release 5.004) does not.
1506 [Or at least, is documented not to: some releases seem to be in conflict
1507 with the documentation.] However, when a pattern was studied with
1508 pcre_study() and all its branches started with \s, PCRE still included VT
1509 as a possible starting character. Of course, this did no harm; it just
1510 caused an unnecessary match attempt.
1512 22. Removed a now-redundant internal flag bit that recorded the fact that case
1513 dependency changed within the pattern. This was once needed for "required
1514 byte" processing, but is no longer used. This recovers a now-scarce options
1515 bit. Also moved the least significant internal flag bit to the most-
1516 significant bit of the word, which was not previously used (hangover from
1517 the days when it was an int rather than a uint) to free up another bit for
1518 the future.
1520 23. Added support for CRLF line endings as well as CR and LF. As well as the
1521 default being selectable at build time, it can now be changed at runtime
1522 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1523 specify that it is scanning data with non-default line endings.
1525 24. Changed the definition of CXXLINK to make it agree with the definition of
1526 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1528 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1529 recursions. This makes a big different to stack usage for some patterns.
1531 26. If a subpattern containing a named recursion or subroutine reference such
1532 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1533 the space required for the compiled pattern went wrong and gave too small a
1534 value. Depending on the environment, this could lead to "Failed: internal
1535 error: code overflow at offset 49" or "glibc detected double free or
1536 corruption" errors.
1538 27. Applied patches from Google (a) to support the new newline modes and (b) to
1539 advance over multibyte UTF-8 characters in GlobalReplace.
1541 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1542 difference for some implementation of PCRE in some Windows version.
1544 29. Added some extra testing facilities to pcretest:
1546 \q<number> in a data line sets the "match limit" value
1547 \Q<number> in a data line sets the "match recursion limt" value
1548 -S <number> sets the stack size, where <number> is in megabytes
1550 The -S option isn't available for Windows.
1553 Version 6.6 06-Feb-06
1554 ---------------------
1556 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1557 in pcreposix.h. I have copied the definition from pcre.h.
1559 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1560 because pcre.h is no longer a built file.
1562 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1563 not normally included in the compiled code.
1566 Version 6.5 01-Feb-06
1567 ---------------------
1569 1. When using the partial match feature with pcre_dfa_exec(), it was not
1570 anchoring the second and subsequent partial matches at the new starting
1571 point. This could lead to incorrect results. For example, with the pattern
1572 /1234/, partially matching against "123" and then "a4" gave a match.
1574 2. Changes to pcregrep:
1576 (a) All non-match returns from pcre_exec() were being treated as failures
1577 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1578 error message is output. Some extra information is given for the
1580 probably the only errors that are likely to be caused by users (by
1581 specifying a regex that has nested indefinite repeats, for instance).
1582 If there are more than 20 of these errors, pcregrep is abandoned.
1584 (b) A binary zero was treated as data while matching, but terminated the
1585 output line if it was written out. This has been fixed: binary zeroes
1586 are now no different to any other data bytes.
1588 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1589 used to set a locale for matching. The --locale=xxxx long option has
1590 been added (no short equivalent) to specify a locale explicitly on the
1591 pcregrep command, overriding the environment variables.
1593 (d) When -B was used with -n, some line numbers in the output were one less
1594 than they should have been.
1596 (e) Added the -o (--only-matching) option.
1598 (f) If -A or -C was used with -c (count only), some lines of context were
1599 accidentally printed for the final match.
1601 (g) Added the -H (--with-filename) option.
1603 (h) The combination of options -rh failed to suppress file names for files
1604 that were found from directory arguments.
1606 (i) Added the -D (--devices) and -d (--directories) options.
1608 (j) Added the -F (--fixed-strings) option.
1610 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1612 (l) Added the --colo(u)r option.
1614 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1615 is not present by default.
1617 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1618 items such as (?R) or (?1), when the recursion could match a number of
1619 alternatives. If it matched one of the alternatives, but subsequently,
1620 outside the recursion, there was a failure, the code tried to back up into
1621 the recursion. However, because of the way PCRE is implemented, this is not
1622 possible, and the result was an incorrect result from the match.
1624 In order to prevent this happening, the specification of recursion has
1625 been changed so that all such subpatterns are automatically treated as
1626 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1628 4. I had overlooked the fact that, in some locales, there are characters for
1629 which isalpha() is true but neither isupper() nor islower() are true. In
1630 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1631 and ordfeminine) are like this. This affected the treatment of \w and \W
1632 when they appeared in character classes, but not when they appeared outside
1633 a character class. The bit map for "word" characters is now created
1634 separately from the results of isalnum() instead of just taking it from the
1635 upper, lower, and digit maps. (Plus the underscore character, of course.)
1637 5. The above bug also affected the handling of POSIX character classes such as
1638 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1639 permanent tables. Instead, the bit maps for such a class were previously
1640 created as the appropriate unions of the upper, lower, and digit bitmaps.
1641 Now they are created by subtraction from the [[:word:]] class, which has
1642 its own bitmap.
1644 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1645 It is created by subtracting the vertical space characters (\x09, \x0a,
1646 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1647 subtraction was done in the overall bitmap for a character class, meaning
1648 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1649 be recognized. This bug has been fixed.
1651 7. Patches from the folks at Google:
1653 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1654 real life, but is still worth protecting against".
1656 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1657 regular expressions".
1659 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1660 have it.
1662 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1663 "configure" and the latter not, in order to fix a problem somebody had
1664 with compiling the Arg class on HP-UX.
1666 (e) Improve the error-handling of the C++ wrapper a little bit.
1668 (f) New tests for checking recursion limiting.
1670 8. The pcre_memmove() function, which is used only if the environment does not
1671 have a standard memmove() function (and is therefore rarely compiled),
1672 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1673 returning a result (though PCRE never actually uses the result).
1675 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1676 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1677 returned instead of calling malloc() with an overflowing number that would
1678 most likely cause subsequent chaos.
1680 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1682 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1683 with this option is matched, the nmatch and pmatch options of regexec() are
1684 ignored.
1686 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1687 provided in case anyone wants to the the POSIX interface with UTF-8
1688 strings.
1690 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1691 C++ linking (needed for some HP-UX environments).
1693 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1694 (unused parameter) and in the pcre_printint() function (omitted "default"
1695 switch label when the default is to do nothing).
1697 15. Added some code to make it possible, when PCRE is compiled as a C++
1698 library, to replace subject pointers for pcre_exec() with a smart pointer
1699 class, thus making it possible to process discontinuous strings.
1701 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1702 much the same function. They were added by different people who were trying
1703 to make PCRE easy to compile on non-Unix systems. It has been suggested
1704 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1705 for compiling on Windows systems. I have therefore replaced it with
1706 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1707 defaults to "extern" for C or "extern C" for C++, which works fine on
1708 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1709 SCOPE with something explicit in config.h. In addition:
1711 (a) pcreposix.h still had just "extern" instead of either of these macros;
1712 I have replaced it with PCRE_DATA_SCOPE.
1714 (b) Functions such as _pcre_xclass(), which are internal to the library,
1715 but external in the C sense, all had PCRE_EXPORT in their definitions.
1716 This is apparently wrong for the Windows case, so I have removed it.
1717 (It makes no difference on Unix-like systems.)
1719 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1720 of recursive calls to match(). This is different to MATCH_LIMIT because
1721 that limits the total number of calls to match(), not all of which increase
1722 the depth of recursion. Limiting the recursion depth limits the amount of
1723 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1724 when PCRE is compiled, and changed at run time. A patch from Google adds
1725 this functionality to the C++ interface.
1727 18. Changes to the handling of Unicode character properties:
1729 (a) Updated the table to Unicode 4.1.0.
1731 (b) Recognize characters that are not in the table as "Cn" (undefined).
1733 (c) I revised the way the table is implemented to a much improved format
1734 which includes recognition of ranges. It now supports the ranges that
1735 are defined in UnicodeData.txt, and it also amalgamates other
1736 characters into ranges. This has reduced the number of entries in the
1737 table from around 16,000 to around 3,000, thus reducing its size
1738 considerably. I realized I did not need to use a tree structure after
1739 all - a binary chop search is just as efficient. Having reduced the
1740 number of entries, I extended their size from 6 bytes to 8 bytes to
1741 allow for more data.
1743 (d) Added support for Unicode script names via properties such as \p{Han}.
1745 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1746 matching that character.
1748 20. When matching a repeated Unicode property with a minimum greater than zero,
1749 (for example \pL{2,}), PCRE could look past the end of the subject if it
1750 reached it while seeking the minimum number of characters. This could
1751 happen only if some of the characters were more than one byte long, because
1752 there is a check for at least the minimum number of bytes.
1754 21. Refactored the implementation of \p and \P so as to be more general, to
1755 allow for more different types of property in future. This has changed the
1756 compiled form incompatibly. Anybody with saved compiled patterns that use
1757 \p or \P will have to recompile them.
1759 22. Added "Any" and "L&" to the supported property types.
1761 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1762 but give a compile time error if the value is greater than 0xff.
1764 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1765 accidentally not being installed or uninstalled.
1767 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1768 made were to insert the current release number. This seemed silly, because
1769 it made things harder for people building PCRE on systems that don't run
1770 "configure". I have turned pcre.h into a distributed file, no longer built
1771 by "configure", with the version identification directly included. There is
1772 no longer a pcre.h.in file.
1774 However, this change necessitated a change to the pcre-config script as
1775 well. It is built from pcre-config.in, and one of the substitutions was the
1776 release number. I have updated configure.ac so that ./configure now finds
1777 the release number by grepping pcre.h.
1779 26. Added the ability to run the tests under valgrind.
1782 Version 6.4 05-Sep-05
1783 ---------------------
1785 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1786 "--" to be printed when multiple files were scanned, even when none of the
1787 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1788 consider it to be a bug, and have restored the previous behaviour.
1790 2. A couple of code tidies to get rid of compiler warnings.
1792 3. The pcretest program used to cheat by referring to symbols in the library
1793 whose names begin with _pcre_. These are internal symbols that are not
1794 really supposed to be visible externally, and in some environments it is
1795 possible to suppress them. The cheating is now confined to including
1796 certain files from the library's source, which is a bit cleaner.
1798 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1799 file's purpose clearer.
1801 5. Reorganized pcre_ucp_findchar().
1804 Version 6.3 15-Aug-05
1805 ---------------------
1807 1. The file libpcre.pc.in did not have general read permission in the tarball.
1809 2. There were some problems when building without C++ support:
1811 (a) If C++ support was not built, "make install" and "make test" still
1812 tried to test it.
1814 (b) There were problems when the value of CXX was explicitly set. Some
1815 changes have been made to try to fix these, and ...
1817 (c) --disable-cpp can now be used to explicitly disable C++ support.
1819 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1820 backslash in a target when C++ was disabled. This confuses some
1821 versions of "make", apparently. Using an intermediate variable solves
1822 this. (Same for CPP_LOBJ.)
1824 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1825 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1826 necessary on certain architectures.
1828 4. Added a setting of -export-symbols-regex to the link command to remove
1829 those symbols that are exported in the C sense, but actually are local
1830 within the library, and not documented. Their names all begin with
1831 "_pcre_". This is not a perfect job, because (a) we have to except some
1832 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1833 available (and never for static libraries). I have made a note to try to
1834 find a way round (a) in the future.
1837 Version 6.2 01-Aug-05
1838 ---------------------
1840 1. There was no test for integer overflow of quantifier values. A construction
1841 such as {1111111111111111} would give undefined results. What is worse, if
1842 a minimum quantifier for a parenthesized subpattern overflowed and became
1843 negative, the calculation of the memory size went wrong. This could have
1844 led to memory overwriting.
1846 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1848 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1849 operating environments where this matters.
1851 4. Applied Giuseppe Maxia's patch to add additional features for controlling
1852 PCRE options from within the C++ wrapper.
1854 5. Named capturing subpatterns were not being correctly counted when a pattern
1855 was compiled. This caused two problems: (a) If there were more than 100
1856 such subpatterns, the calculation of the memory needed for the whole
1857 compiled pattern went wrong, leading to an overflow error. (b) Numerical
1858 back references of the form \12, where the number was greater than 9, were
1859 not recognized as back references, even though there were sufficient
1860 previous subpatterns.
1862 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1863 versions of gcc, e.g. 2.95.4.
1866 Version 6.1 21-Jun-05
1867 ---------------------
1869 1. There was one reference to the variable "posix" in pcretest.c that was not
1870 surrounded by "#if !defined NOPOSIX".
1872 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1873 the cross-check on the old pcre_info() function, for the benefit of the
1874 cut-down version of PCRE that is currently imported into Exim.
1876 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1877 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1878 patterns that start (?i) but I don't think that matters. The use of (?i) is
1879 just an example; this all applies to the other options as well.
1881 4. Since libtool seems to echo the compile commands it is issuing, the output
1882 from "make" can be reduced a bit by putting "@" in front of each libtool
1883 compile command.
1885 5. Patch from the folks at Google for configure.in to be a bit more thorough
1886 in checking for a suitable C++ installation before trying to compile the
1887 C++ stuff. This should fix a reported problem when a compiler was present,
1888 but no suitable headers.
1890 6. The man pages all had just "PCRE" as their title. I have changed them to
1891 be the relevant file name. I have also arranged that these names are
1892 retained in the file doc/pcre.txt, which is a concatenation in text format
1893 of all the man pages except the little individual ones for each function.
1895 7. The NON-UNIX-USE file had not been updated for the different set of source
1896 files that come with release 6. I also added a few comments about the C++
1897 wrapper.
1900 Version 6.0 07-Jun-05
1901 ---------------------
1903 1. Some minor internal re-organization to help with my DFA experiments.
1905 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1906 didn't matter for the library itself when fully configured, but did matter
1907 when compiling without UCP support, or within Exim, where the ucp files are
1908 not imported.
1910 3. Refactoring of the library code to split up the various functions into
1911 different source modules. The addition of the new DFA matching code (see
1912 below) to a single monolithic source would have made it really too
1913 unwieldy, quite apart from causing all the code to be include in a
1914 statically linked application, when only some functions are used. This is
1915 relevant even without the DFA addition now that patterns can be compiled in
1916 one application and matched in another.
1918 The downside of splitting up is that there have to be some external
1919 functions and data tables that are used internally in different modules of
1920 the library but which are not part of the API. These have all had their
1921 names changed to start with "_pcre_" so that they are unlikely to clash
1922 with other external names.
1924 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1925 a different (DFA) algorithm. Although it is slower than the original
1926 function, it does have some advantages for certain types of matching
1927 problem.
1929 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1930 including restarting after a partial match.
1932 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1933 defined when compiling for Windows was sent to me. I have put it into the
1934 code, though I have no means of testing or verifying it.
1936 7. Added the pcre_refcount() auxiliary function.
1938 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1939 match before or at the first newline in the subject string. In pcretest,
1940 the /f option on a pattern can be used to set this.
1942 9. A repeated \w when used in UTF-8 mode with characters greater than 256
1943 would behave wrongly. This has been present in PCRE since release 4.0.
1945 10. A number of changes to the pcregrep command:
1947 (a) Refactored how -x works; insert ^(...)$ instead of setting
1948 PCRE_ANCHORED and checking the length, in preparation for adding
1949 something similar for -w.
1951 (b) Added the -w (match as a word) option.
1953 (c) Refactored the way lines are read and buffered so as to have more
1954 than one at a time available.
1956 (d) Implemented a pcregrep test script.
1958 (e) Added the -M (multiline match) option. This allows patterns to match
1959 over several lines of the subject. The buffering ensures that at least
1960 8K, or the rest of the document (whichever is the shorter) is available
1961 for matching (and similarly the previous 8K for lookbehind assertions).
1963 (f) Changed the --help output so that it now says
1965 -w, --word-regex(p)
1967 instead of two lines, one with "regex" and the other with "regexp"
1968 because that confused at least one person since the short forms are the
1969 same. (This required a bit of code, as the output is generated
1970 automatically from a table. It wasn't just a text change.)
1972 (g) -- can be used to terminate pcregrep options if the next thing isn't an
1973 option but starts with a hyphen. Could be a pattern or a path name
1974 starting with a hyphen, for instance.
1976 (h) "-" can be given as a file name to represent stdin.
1978 (i) When file names are being printed, "(standard input)" is used for
1979 the standard input, for compatibility with GNU grep. Previously
1980 "<stdin>" was used.
1982 (j) The option --label=xxx can be used to supply a name to be used for
1983 stdin when file names are being printed. There is no short form.
1985 (k) Re-factored the options decoding logic because we are going to add
1986 two more options that take data. Such options can now be given in four
1987 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
1989 (l) Added the -A, -B, and -C options for requesting that lines of context
1990 around matches be printed.
1992 (m) Added the -L option to print the names of files that do not contain
1993 any matching lines, that is, the complement of -l.
1995 (n) The return code is 2 if any file cannot be opened, but pcregrep does
1996 continue to scan other files.
1998 (o) The -s option was incorrectly implemented. For compatibility with other
1999 greps, it now suppresses the error message for a non-existent or non-
2000 accessible file (but not the return code). There is a new option called
2001 -q that suppresses the output of matching lines, which was what -s was
2002 previously doing.
2004 (p) Added --include and --exclude options to specify files for inclusion
2005 and exclusion when recursing.
2007 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2008 Hopefully, it now does.
2010 12. Missing cast in pcre_study().
2012 13. Added an "uninstall" target to the makefile.
2014 14. Replaced "extern" in the function prototypes in Makefile.in with
2015 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2016 world, but is set differently for Windows.
2018 15. Added a second compiling function called pcre_compile2(). The only
2019 difference is that it has an extra argument, which is a pointer to an
2020 integer error code. When there is a compile-time failure, this is set
2021 non-zero, in addition to the error test pointer being set to point to an
2022 error message. The new argument may be NULL if no error number is required
2023 (but then you may as well call pcre_compile(), which is now just a
2024 wrapper). This facility is provided because some applications need a
2025 numeric error indication, but it has also enabled me to tidy up the way
2026 compile-time errors are handled in the POSIX wrapper.
2028 16. Added VPATH=.libs to the makefile; this should help when building with one
2029 prefix path and installing with another. (Or so I'm told by someone who
2030 knows more about this stuff than I do.)
2032 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2033 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2034 match everything, including newlines. This is not POSIX-compatible, but
2035 somebody wanted the feature. From pcretest it can be activated by using
2036 both the P and the s flags.
2038 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2040 19. libpcre.pc was being incorrectly installed as executable.
2042 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2043 it now also looks for '\r' so that it will work unmodified on Windows.
2045 21. Added Google's contributed C++ wrapper to the distribution.
2047 22. Added some untidy missing memory free() calls in pcretest, to keep
2048 Electric Fence happy when testing.
2052 Version 5.0 13-Sep-04
2053 ---------------------
2055 1. Internal change: literal characters are no longer packed up into items
2056 containing multiple characters in a single byte-string. Each character
2057 is now matched using a separate opcode. However, there may be more than one
2058 byte in the character in UTF-8 mode.
2060 2. The pcre_callout_block structure has two new fields: pattern_position and
2061 next_item_length. These contain the offset in the pattern to the next match
2062 item, and its length, respectively.
2064 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2065 insertion of callouts before each pattern item. Added the /C option to
2066 pcretest to make use of this.
2068 4. On the advice of a Windows user, the lines
2070 #if defined(_WIN32) || defined(WIN32)
2071 _setmode( _fileno( stdout ), 0x8000 );
2072 #endif /* defined(_WIN32) || defined(WIN32) */
2074 have been added to the source of pcretest. This apparently does useful
2075 magic in relation to line terminators.
2077 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2078 for the benefit of those environments where the "b" makes a difference.
2080 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2081 to know about it. I have put a hack into configure.in that adds in code
2082 to set GCC=yes if CC=icc. This seems to end up at a point in the
2083 generated configure script that is early enough to affect the setting of
2084 compiler options, which is what is needed, but I have no means of testing
2085 whether it really works. (The user who reported this had patched the
2086 generated configure script, which of course I cannot do.)
2088 LATER: After change 22 below (new libtool files), the configure script
2089 seems to know about icc (and also ecc). Therefore, I have commented out
2090 this hack in configure.in.
2092 7. Added support for pkg-config (2 patches were sent in).
2094 8. Negated POSIX character classes that used a combination of internal tables
2095 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2096 [[:^ascii]]. Typically, they would match almost any characters. The other
2097 POSIX classes were not broken in this way.
2099 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2100 to find the match, as PCRE was deluded into thinking that the match had to
2101 start at the start point or following a newline. The same bug applied to
2102 patterns with negative forward assertions or any backward assertions
2103 preceding ".*" at the start, unless the pattern required a fixed first
2104 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2106 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2107 starting at the last subject character, bytes beyond the end of the subject
2108 string were read.
2110 11. Renamed the variable "class" as "classbits" to make life easier for C++
2111 users. (Previously there was a macro definition, but it apparently wasn't
2112 enough.)
2114 12. Added the new field "tables" to the extra data so that tables can be passed
2115 in at exec time, or the internal tables can be re-selected. This allows
2116 a compiled regex to be saved and re-used at a later time by a different
2117 program that might have everything at different addresses.
2119 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2120 -R library as well as a -L library.
2122 14. The debugging options of pcretest (-d on the command line or D on a
2123 pattern) showed incorrect output for anything following an extended class
2124 that contained multibyte characters and which was followed by a quantifier.
2126 15. Added optional support for general category Unicode character properties
2127 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2128 support. It adds about 90K to the size of the library. The meanings of the
2129 inbuilt class escapes such as \d and \s have NOT been changed.
2131 16. Updated pcredemo.c to include calls to free() to release the memory for the
2132 compiled pattern.
2134 17. The generated file chartables.c was being created in the source directory
2135 instead of in the building directory. This caused the build to fail if the
2136 source directory was different from the building directory, and was
2137 read-only.
2139 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2140 file. No doubt somebody will tell me if they don't make sense... Also added
2141 Dan Mooney's comments about building on OpenVMS.
2143 19. Added support for partial matching via the PCRE_PARTIAL option for
2144 pcre_exec() and the \P data escape in pcretest.
2146 20. Extended pcretest with 3 new pattern features:
2148 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2149 write the compiled pattern to the file whose name is "rest-of-line".
2150 This is a straight binary dump of the data, with the saved pointer to
2151 the character tables forced to be NULL. The study data, if any, is
2152 written too. After writing, pcretest reads a new pattern.
2154 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2155 compiled pattern from the given file. There must not be any
2156 occurrences of "<" in the file name (pretty unlikely); if there are,
2157 pcretest will instead treat the initial "<" as a pattern delimiter.
2158 After reading in the pattern, pcretest goes on to read data lines as
2159 usual.
2161 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2162 and 16-bit fields in a compiled pattern, to simulate a pattern that
2163 was compiled on a host of opposite endianness.
2165 21. The pcre-exec() function can now cope with patterns that were compiled on
2166 hosts of opposite endianness, with this restriction:
2168 As for any compiled expression that is saved and used later, the tables
2169 pointer field cannot be preserved; the extra_data field in the arguments
2170 to pcre_exec() should be used to pass in a tables address if a value
2171 other than the default internal tables were used at compile time.
2173 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2174 now diagnosed as an error. Previously, most of the time, a negative number
2175 would have been treated as zero, but if in addition "ovector" was passed as
2176 NULL, a crash could occur.
2178 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2179 new versions from the libtool 1.5 distribution (the last one is a copy of
2180 a file called libtool.m4). This seems to have fixed the need to patch
2181 "configure" to support Darwin 1.3 (which I used to do). However, I still
2182 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2183 workstation).
2185 24. Changed the PCRE licence to be the more standard "BSD" licence.
2188 Version 4.5 01-Dec-03
2189 ---------------------
2191 1. There has been some re-arrangement of the code for the match() function so
2192 that it can be compiled in a version that does not call itself recursively.
2193 Instead, it keeps those local variables that need separate instances for
2194 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2195 needs to "recurse". Keeping track of where control must go is done by means
2196 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2197 hide most of the details from the main code, and operates only if
2198 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2199 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2200 operating.
2202 To make it easier for callers to provide specially tailored get/free
2203 functions for this usage, two new functions, pcre_stack_malloc, and
2204 pcre_stack_free, are used. They are always called in strict stacking order,
2205 and the size of block requested is always the same.
2207 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2208 PCRE has been compiled to use the stack or the heap for recursion. The
2209 -C option of pcretest uses this to show which version is compiled.
2211 A new data escape \S, is added to pcretest; it causes the amounts of store
2212 obtained and freed by both kinds of malloc/free at match time to be added
2213 to the output.
2215 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2216 what's available on my current Linux desktop machine.
2218 3. When matching a UTF-8 string, the test for a valid string at the start has
2219 been extended. If start_offset is not zero, PCRE now checks that it points
2220 to a byte that is the start of a UTF-8 character. If not, it returns
2221 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2222 this is necessary because there may be backward assertions in the pattern.
2223 When matching the same subject several times, it may save resources to use
2224 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2226 4. The code for checking the validity of UTF-8 strings has been tightened so
2227 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2228 containing "overlong sequences".
2230 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2231 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2232 should have been "&", but it just so happened that all the cases this let
2233 through by mistake were picked up later in the function.
2235 6. I had used a variable called "isblank" - this is a C99 function, causing
2236 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2238 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2239 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2240 so the version is shown; (c) stop "make test" echoing "./RunTest".
2242 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2244 9. The prototype for memmove() for systems that don't have it was using
2245 size_t, but the inclusion of the header that defines size_t was later. I've
2246 moved the #includes for the C headers earlier to avoid this.
2248 10. Added some adjustments to the code to make it easier to compiler on certain
2249 special systems:
2251 (a) Some "const" qualifiers were missing.
2252 (b) Added the macro EXPORT before all exported functions; by default this
2253 is defined to be empty.
2254 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2255 that it reads its output file name as an argument instead of writing
2256 to the standard output and assuming this can be redirected.
2258 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2259 class containing characters with values greater than 255, PCRE compilation
2260 went into a loop.
2262 12. A recursive reference to a subpattern that was within another subpattern
2263 that had a minimum quantifier of zero caused PCRE to crash. For example,
2264 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2265 recursion. If the recursively-called subpattern itself had a zero repeat,
2266 that was OK.
2268 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2269 buffer into which it was copied (for escape processing) was still set at
2270 1024, so long lines caused crashes.
2272 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2273 "internal error: code overflow...". This applied to any character class
2274 that was followed by a possessive quantifier.
2276 15. Modified the Makefile to add libpcre.la as a prerequisite for
2277 libpcreposix.la because I was told this is needed for a parallel build to
2278 work.
2280 16. If a pattern that contained .* following optional items at the start was
2281 studied, the wrong optimizing data was generated, leading to matching
2282 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2283 matching string must start with a or b or c. The correct conclusion for
2284 this pattern is that a match can start with any character.
2287 Version 4.4 13-Aug-03
2288 ---------------------
2290 1. In UTF-8 mode, a character class containing characters with values between
2291 127 and 255 was not handled correctly if the compiled pattern was studied.
2292 In fixing this, I have also improved the studying algorithm for such
2293 classes (slightly).
2295 2. Three internal functions had redundant arguments passed to them. Removal
2296 might give a very teeny performance improvement.
2298 3. Documentation bug: the value of the capture_top field in a callout is *one
2299 more than* the number of the hightest numbered captured substring.
2301 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2302 in incorrectly linking with a previously installed version. They now link
2303 explicitly with libpcre.la.
2305 5. configure.in no longer needs to recognize Cygwin specially.
2307 6. A problem in pcre.in for Windows platforms is fixed.
2309 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2310 pcretest, it used to include the size of the study block as part of its
2311 output. Unfortunately, the structure contains a field that has a different
2312 size on different hardware architectures. This meant that the tests that
2313 showed this size failed. As the block is currently always of a fixed size,
2314 this information isn't actually particularly useful in pcretest output, so
2315 I have just removed it.
2317 8. Three pre-processor statements accidentally did not start in column 1.
2318 Sadly, there are *still* compilers around that complain, even though
2319 standard C has not required this for well over a decade. Sigh.
2321 9. In pcretest, the code for checking callouts passed small integers in the
2322 callout_data field, which is a void * field. However, some picky compilers
2323 complained about the casts involved for this on 64-bit systems. Now
2324 pcretest passes the address of the small integer instead, which should get
2325 rid of the warnings.
2327 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2328 both compile and run time, and gives an error if an invalid UTF-8 sequence
2329 is found. There is a option for disabling this check in cases where the
2330 string is known to be correct and/or the maximum performance is wanted.
2332 11. In response to a bug report, I changed one line in Makefile.in from
2334 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2335 to
2336 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2338 to look similar to other lines, but I have no way of telling whether this
2339 is the right thing to do, as I do not use Windows. No doubt I'll get told
2340 if it's wrong...
2343 Version 4.3 21-May-03
2344 ---------------------
2346 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2347 Makefile.
2349 2. Some refactoring to improve the quality of the code:
2351 (i) The utf8_table... variables are now declared "const".
2353 (ii) The code for \cx, which used the "case flipping" table to upper case
2354 lower case letters, now just substracts 32. This is ASCII-specific,
2355 but the whole concept of \cx is ASCII-specific, so it seems
2356 reasonable.
2358 (iii) PCRE was using its character types table to recognize decimal and
2359 hexadecimal digits in the pattern. This is silly, because it handles
2360 only 0-9, a-f, and A-F, but the character types table is locale-
2361 specific, which means strange things might happen. A private
2362 table is now used for this - though it costs 256 bytes, a table is
2363 much faster than multiple explicit tests. Of course, the standard
2364 character types table is still used for matching digits in subject
2365 strings against \d.
2367 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2368 ending in _t are). So I've renamed it as ESC_tee.
2370 3. The first argument for regexec() in the POSIX wrapper should have been
2371 defined as "const".
2373 4. Changed pcretest to use malloc() for its buffers so that they can be
2374 Electric Fenced for debugging.
2376 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2377 to read one or more bytes before the start of the subject string. Often this
2378 had no effect on PCRE's behaviour, but in some circumstances it could
2379 provoke a segmentation fault.
2381 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2382 to try to read one or more bytes before the start of the subject string.
2384 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2385 UTF-8 support could misbehave in various ways if the subject string
2386 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2387 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2388 back over UTF-8 characters.)
2391 Version 4.2 14-Apr-03
2392 ---------------------
2394 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2396 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2397 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2398 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2399 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2400 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2402 Note: automatic setting of the BUILD variables is not yet working
2403 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2404 compile-time but not at link-time
2405 [LINK]: use for linking executables only
2406 make different versions for Windows and non-Windows
2407 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2408 libraries
2409 [LINK_FOR_BUILD]: new variable
2410 [OBJEXT]: use throughout
2411 [EXEEXT]: use throughout
2412 <winshared>: new target
2413 <wininstall>: new target
2414 <dftables.o>: use native compiler
2415 <dftables>: use native linker
2416 <install>: handle Windows platform correctly
2417 <clean>: ditto
2418 <check>: ditto
2419 copy DLL to top builddir before testing
2421 As part of these changes, -no-undefined was removed again. This was reported
2422 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2423 in any case.
2425 3. Some tidies to get rid of compiler warnings:
2427 . In the match_data structure, match_limit was an unsigned long int, whereas
2428 match_call_count was an int. I've made them both unsigned long ints.
2430 . In pcretest the fact that a const uschar * doesn't automatically cast to
2431 a void * provoked a warning.
2433 . Turning on some more compiler warnings threw up some "shadow" variables
2434 and a few more missing casts.
2436 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2437 option, a class that contained a single character with a value between 128
2438 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2440 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2441 option, a class that contained several characters, but with at least one
2442 whose value was between 128 and 255 caused PCRE to crash.
2445 Version 4.1 12-Mar-03
2446 ---------------------
2448 1. Compiling with gcc -pedantic found a couple of places where casts were
2449 needed, and a string in dftables.c that was longer than standard compilers are
2450 required to support.
2452 2. Compiling with Sun's compiler found a few more places where the code could
2453 be tidied up in order to avoid warnings.
2455 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2456 first of these names is deprecated in the latest Autoconf in favour of the name
2457 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2458 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2459 analogy I have changed it to CFLAGS_FOR_BUILD.
2461 4. Added -no-undefined to the linking command in the Makefile, because this is
2462 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2463 linking step for the pcreposix library.
2465 5. PCRE was failing to diagnose the case of two named groups with the same
2466 name.
2468 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2469 literal character that is needed in the subject for a match, and scans along to
2470 ensure that it is present before embarking on the full matching process. This
2471 saves time in cases of nested unlimited repeats that are never going to match.
2472 Problem: the scan can take a lot of time if the subject is very long (e.g.
2473 megabytes), thus penalizing straightforward matches. It is now done only if the
2474 amount of subject to be scanned is less than 1000 bytes.
2476 7. A lesser problem with the same optimization is that it was recording the
2477 first character of an anchored pattern as "needed", thus provoking a search
2478 right along the subject, even when the first match of the pattern was going to
2479 fail. The "needed" character is now not set for anchored patterns, unless it
2480 follows something in the pattern that is of non-fixed length. Thus, it still
2481 fulfils its original purpose of finding quick non-matches in cases of nested
2482 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2485 Version 4.0 17-Feb-03
2486 ---------------------
2488 1. If a comment in an extended regex that started immediately after a meta-item
2489 extended to the end of string, PCRE compiled incorrect data. This could lead to
2490 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2492 2. Moved to autoconf 2.53 and libtool 1.4.2.
2494 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2495 the special perltest8 script is no longer needed - all the tests can be run
2496 from a single perltest script.
2498 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2499 by \s. It has now been removed in PCRE. This means it isn't recognized as
2500 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2501 class [:space:] *does* include VT, thereby creating a mess.
2503 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2504 space and tab.
2506 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2507 its new features into the main test script, reducing the number of scripts.
2509 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2510 were backward compatible, and made the (?i) apply to the whole pattern, as if
2511 /i were given. Now it behaves more logically, and applies the option setting
2512 only to what follows. PCRE has been changed to follow suit. However, if it
2513 finds options settings right at the start of the pattern, it extracts them into
2514 the global options, as before. Thus, they show up in the info data.
2516 8. Added support for the \Q...\E escape sequence. Characters in between are
2517 treated as literals. This is slightly different from Perl in that $ and @ are
2518 also handled as literals inside the quotes. In Perl, they will cause variable
2519 interpolation. Note the following examples:
2521 Pattern PCRE matches Perl matches
2523 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2524 \Qabc\$xyz\E abc\$xyz abc\$xyz
2525 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2527 For compatibility with Perl, \Q...\E sequences are recognized inside character
2528 classes as well as outside them.
2530 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2531 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2532 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2533 signed/unsigned warnings.
2535 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2536 option for pcretest, so I've replaced it by a simple function that does just
2537 that job.
2539 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2540 "pcregrep -".
2542 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2543 Java package. This provides some syntactic sugar for simple cases of what my
2544 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2545 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2546 item, you can use this simplified notation. Note that only makes sense with
2547 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2548 greediness, whatever the setting of the PCRE_UNGREEDY option.
2550 13. A change of greediness default within a pattern was not taking effect at
2551 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2552 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2553 was abstracted outside.
2555 14. PCRE now supports the \G assertion. It is true when the current matching
2556 position is at the start point of the match. This differs from \A when the
2557 starting offset is non-zero. Used with the /g option of pcretest (or similar
2558 code), it works in the same way as it does for Perl's /g option. If all
2559 alternatives of a regex begin with \G, the expression is anchored to the start
2560 match position, and the "anchored" flag is set in the compiled expression.
2562 15. Some bugs concerning the handling of certain option changes within patterns
2563 have been fixed. These applied to options other than (?ims). For example,
2564 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2565 been the other way round. Some of this was related to change 7 above.
2567 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2568 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2569 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2570 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2572 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2573 mode. Unlike ".", it always matches newline, whatever the setting of
2574 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2575 assertions. Perl allows it, but it doesn't (in general) work because it can't
2576 calculate the length of the lookbehind. At least, that's the case for Perl
2577 5.8.0 - I've been told they are going to document that it doesn't work in
2578 future.
2580 18. Added an error diagnosis for escapes that PCRE does not support: these are
2581 \L, \l, \N, \P, \p, \U, \u, and \X.
2583 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2584 reading past the end of the pattern in cases such as /[abcd/.
2586 20. PCRE was getting more memory than necessary for patterns with classes that
2587 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2589 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2590 compiling PCRE for use with Virtual Pascal.
2592 22. Small fix to the Makefile to make it work properly if the build is done
2593 outside the source tree.
2595 23. Added a new extension: a condition to go with recursion. If a conditional
2596 subpattern starts with (?(R) the "true" branch is used if recursion has
2597 happened, whereas the "false" branch is used only at the top level.
2599 24. When there was a very long string of literal characters (over 255 bytes
2600 without UTF support, over 250 bytes with UTF support), the computation of how
2601 much memory was required could be incorrect, leading to segfaults or other
2602 strange effects.
2604 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2605 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2606 there was a subsequent back reference to those brackets. This meant that, for
2607 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2608 possible to check for precisely this case. All we can do is abandon the
2609 optimization if .* occurs inside capturing brackets when there are any back
2610 references whatsoever. (See below for a better fix that came later.)
2612 26. The handling of the optimization for finding the first character of a
2613 non-anchored pattern, and for finding a character that is required later in the
2614 match were failing in some cases. This didn't break the matching; it just
2615 failed to optimize when it could. The way this is done has been re-implemented.
2617 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2619 28. Added a new feature that provides some of the functionality that Perl
2620 provides with (?{...}). The facility is termed a "callout". The way it is done
2621 in PCRE is for the caller to provide an optional function, by setting
2622 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2623 global variable. By default it is unset, which disables all calling out. To get
2624 the function called, the regex must include (?C) at appropriate points. This
2625 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2626 This provides a means of identifying different callout points. When PCRE
2627 reaches such a point in the regex, if pcre_callout has been set, the external
2628 function is called. It is provided with data in a structure called
2629 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2630 matching continues; if it returns a non-zero value, the match at the current
2631 point fails. However, backtracking will occur if possible. [This was changed
2632 later and other features added - see item 49 below.]
2634 29. pcretest is upgraded to test the callout functionality. It provides a
2635 callout function that displays information. By default, it shows the start of
2636 the match and the current position in the text. There are some new data escapes
2637 to vary what happens:
2639 \C+ in addition, show current contents of captured substrings
2640 \C- do not supply a callout function
2641 \C!n return 1 when callout number n is reached
2642 \C!n!m return 1 when callout number n is reached for the mth time
2644 30. If pcregrep was called with the -l option and just a single file name, it
2645 output "<stdin>" if a match was found, instead of the file name.
2647 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2648 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2649 pcre_exec(). This saves a malloc/free per call. The default value of
2650 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2651 when configuring.
2653 32. The default maximum size of a compiled pattern is 64K. There have been a
2654 few cases of people hitting this limit. The code now uses macros to handle the
2655 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2656 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2657 configuring. Tests 2 and 5 work only with 2-byte links because they output
2658 debugging information about compiled patterns.
2660 33. Internal code re-arrangements:
2662 (a) Moved the debugging function for printing out a compiled regex into
2663 its own source file (printint.c) and used #include to pull it into
2664 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2665 separate copies.
2667 (b) Defined the list of op-code names for debugging as a macro in
2668 internal.h so that it is next to the definition of the opcodes.
2670 (c) Defined a table of op-code lengths for simpler skipping along compiled
2671 code. This is again a macro in internal.h so that it is next to the
2672 definition of the opcodes.
2674 34. Added support for recursive calls to individual subpatterns, along the
2675 lines of Robin Houston's patch (but implemented somewhat differently).
2677 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2678 allow it to read and process whole directories in Win32. This code was
2679 contributed by Lionel Fourquaux; it has not been tested by me.
2681 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2682 used to name a group. Names consist of alphanumerics and underscores, and must
2683 be unique. Back references use the syntax (?P=name) and recursive calls use
2684 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2685 numbers. The function pcre_fullinfo() can be used after compilation to extract
2686 a name/number map. There are three relevant calls:
2688 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2689 PCRE_INFO_NAMECOUNT yields the number of entries
2690 PCRE_INFO_NAMETABLE yields a pointer to the map.
2692 The map is a vector of fixed-size entries. The size of each entry depends on
2693 the length of the longest name used. The first two bytes of each entry are the
2694 group number, most significant byte first. There follows the corresponding
2695 name, zero terminated. The names are in alphabetical order.
2697 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2698 case instead of 255. Making it the same both with and without UTF-8 support
2699 means that the same test output works with both.
2701 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2702 calling malloc() with a zero argument.
2704 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2705 optimization. I've improved things by keeping a bitmap of backreferences with
2706 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2707 fact referenced, the optimization can be applied. It is unlikely that a
2708 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2709 the match to follow \n) will appear inside brackets with a number greater than
2710 31, but if it does, any back reference > 31 suppresses the optimization.
2712 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2713 of disabling numbered capturing parentheses. Any opening parenthesis that is
2714 not followed by ? behaves as if it were followed by ?: but named parentheses
2715 can still be used for capturing (and they will acquire numbers in the usual
2716 way).
2718 41. Redesigned the return codes from the match() function into yes/no/error so
2719 that errors can be passed back from deep inside the nested calls. A malloc
2720 failure while inside a recursive subpattern call now causes the
2721 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2723 42. It is now possible to set a limit on the number of times the match()
2724 function is called in a call to pcre_exec(). This facility makes it possible to
2725 limit the amount of recursion and backtracking, though not in a directly
2726 obvious way, because the match() function is used in a number of different
2727 circumstances. The count starts from zero for each position in the subject
2728 string (for non-anchored patterns). The default limit is, for compatibility, a
2729 large number, namely 10 000 000. You can change this in two ways:
2731 (a) When configuring PCRE before making, you can use --with-match-limit=n
2732 to set a default value for the compiled library.
2734 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2735 a different value is set. See 45 below.
2737 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2739 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2740 of things that can be changed at compile time. The first argument specifies
2741 what is wanted and the second points to where the information is to be placed.
2742 The current list of available information is:
2746 The output is an integer that is set to one if UTF-8 support is available;
2747 otherwise it is set to zero.
2751 The output is an integer that it set to the value of the code that is used for
2752 newline. It is either LF (10) or CR (13).
2756 The output is an integer that contains the number of bytes used for internal
2757 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2761 The output is an integer that contains the threshold above which the POSIX
2762 interface uses malloc() for output vectors. See item 31 above.
2766 The output is an unsigned integer that contains the default limit of the number
2767 of match() calls in a pcre_exec() execution. See 42 above.
2769 44. pcretest has been upgraded by the addition of the -C option. This causes it
2770 to extract all the available output from the new pcre_config() function, and to
2771 output it. The program then exits immediately.
2773 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2774 order to support additional features. One way would have been to define
2775 pcre_exec2() (for example) with extra arguments, but this would not have been
2776 extensible, and would also have required all calls to the original function to
2777 be mapped to the new one. Instead, I have chosen to extend the mechanism that
2778 is used for passing in "extra" data from pcre_study().
2780 The pcre_extra structure is now exposed and defined in pcre.h. It currently
2781 contains the following fields:
2783 flags a bitmap indicating which of the following fields are set
2784 study_data opaque data from pcre_study()
2785 match_limit a way of specifying a limit on match() calls for a specific
2786 call to pcre_exec()
2787 callout_data data for callouts (see 49 below)
2789 The flag bits are also defined in pcre.h, and are
2795 The pcre_study() function now returns one of these new pcre_extra blocks, with
2796 the actual study data pointed to by the study_data field, and the
2797 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2798 before. That is, this change is entirely upwards-compatible and requires no
2799 change to existing code.
2801 If you want to pass in additional data to pcre_exec(), you can either place it
2802 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2803 block.
2805 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2806 data string contains the escape sequence \M, pcretest calls pcre_exec() several
2807 times with different match limits, until it finds the minimum value needed for
2808 pcre_exec() to complete. The value is then output. This can be instructive; for
2809 most simple matches the number is quite small, but for pathological cases it
2810 gets very large very quickly.
2812 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2813 returns the size of the data block pointed to by the study_data field in a
2814 pcre_extra block, that is, the value that was passed as the argument to
2815 pcre_malloc() when PCRE was getting memory in which to place the information
2816 created by pcre_study(). The fourth argument should point to a size_t variable.
2817 pcretest has been extended so that this information is shown after a successful
2818 pcre_study() call when information about the compiled regex is being displayed.
2820 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2821 because what follows is always an absolute path. (Later: it turns out that this
2822 is more than cosmetic for MinGW, because it doesn't like empty path
2823 components.)
2825 49. Some changes have been made to the callout feature (see 28 above):
2827 (i) A callout function now has three choices for what it returns:
2829 0 => success, carry on matching
2830 > 0 => failure at this point, but backtrack if possible
2831 < 0 => serious error, return this value from pcre_exec()
2833 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2834 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2835 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2836 use by callout functions. It will never be used by PCRE itself.
2838 (ii) The pcre_extra structure (see 45 above) has a void * field called
2839 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2840 pcre_callout_block structure has a field of the same name. The contents of
2841 the field passed in the pcre_extra structure are passed to the callout
2842 function in the corresponding field in the callout block. This makes it
2843 easier to use the same callout-containing regex from multiple threads. For
2844 testing, the pcretest program has a new data escape
2846 \C*n pass the number n (may be negative) as callout_data
2848 If the callout function in pcretest receives a non-zero value as
2849 callout_data, it returns that value.
2851 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2852 there were some redundant $(CFLAGS) in commands that are now specified as
2853 $(LINK), which already includes $(CFLAGS).
2855 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2856 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2857 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2858 one-byte characters throughout. Note that case-insensitive matching applies
2859 only to characters whose values are less than 256. PCRE doesn't support the
2860 notion of cases for higher-valued characters.
2862 (i) A character class whose characters are all within 0-255 is handled as
2863 a bit map, and the map is inverted for negative classes. Previously, a
2864 character > 255 always failed to match such a class; however it should
2865 match if the class was a negative one (e.g. [^ab]). This has been fixed.
2867 (ii) A negated character class with a single character < 255 is coded as
2868 "not this character" (OP_NOT). This wasn't working properly when the test
2869 character was multibyte, either singly or repeated.
2871 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2872 mode, for example: \x{100}{2,3}.
2874 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2875 singly or repeated) now correctly test multibyte characters. However,
2876 PCRE doesn't recognize any characters with values greater than 255 as
2877 digits, spaces, or word characters. Such characters always match \D, \S,
2878 and \W, and never match \d, \s, or \w.
2880 (v) Classes may now contain characters and character ranges with values
2881 greater than 255. For example: [ab\x{100}-\x{400}].
2883 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
2884 PCRE in UTF-8 mode.
2886 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2887 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2888 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2889 value.)
2891 53. The single man page has become too large. I have therefore split it up into
2892 a number of separate man pages. These also give rise to individual HTML pages;
2893 these are now put in a separate directory, and there is an index.html page that
2894 lists them all. Some hyperlinking between the pages has been installed.
2896 54. Added convenience functions for handling named capturing parentheses.
2898 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2899 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2900 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2901 are faulted.
2903 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2904 calling configure. These values are used when compiling the dftables.c program
2905 which is run to generate the source of the default character tables. They
2906 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2907 you will need to set these values.
2909 57. Updated the building process for Windows DLL, as provided by Fred Cox.
2912 Version 3.9 02-Jan-02
2913 ---------------------
2915 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
2917 2. If --disable-static was given, the building process failed when trying to
2918 build pcretest and pcregrep. (For some reason it was using libtool to compile
2919 them, which is not right, as they aren't part of the library.)
2922 Version 3.8 18-Dec-01
2923 ---------------------
2925 1. The experimental UTF-8 code was completely screwed up. It was packing the
2926 bytes in the wrong order. How dumb can you get?
2929 Version 3.7 29-Oct-01
2930 ---------------------
2932 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
2933 This caused pcretest, when used on the test data, to segfault. Unfortunately,
2934 this didn't happen under Solaris 8, where I normally test things.
2936 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
2937 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
2938 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
2939 this fix an hour or so after the initial 3.7 release.)
2942 Version 3.6 23-Oct-01
2943 ---------------------
2945 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
2946 offsets passed as NULL with zero offset count.
2948 2. The config.guess and config.sub files had not been updated when I moved to
2949 the latest autoconf.
2952 Version 3.5 15-Aug-01
2953 ---------------------
2955 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
2956 had been forgotten.
2958 2. By using declared but undefined structures, we can avoid using "void"
2959 definitions in pcre.h while keeping the internal definitions of the structures
2960 private.
2962 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
2963 user point of view, this means that both static and shared libraries are built
2964 by default, but this can be individually controlled. More of the work of
2965 handling this static/shared cases is now inside libtool instead of PCRE's make
2966 file.
2968 4. The pcretest utility is now installed along with pcregrep because it is
2969 useful for users (to test regexs) and by doing this, it automatically gets
2970 relinked by libtool. The documentation has been turned into a man page, so
2971 there are now .1, .txt, and .html versions in /doc.
2973 5. Upgrades to pcregrep:
2974 (i) Added long-form option names like gnu grep.
2975 (ii) Added --help to list all options with an explanatory phrase.
2976 (iii) Added -r, --recursive to recurse into sub-directories.
2977 (iv) Added -f, --file to read patterns from a file.
2979 6. pcre_exec() was referring to its "code" argument before testing that
2980 argument for NULL (and giving an error if it was NULL).
2982 7. Upgraded Makefile.in to allow for compiling in a different directory from
2983 the source directory.
2985 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
2986 options bits, the pointer it was passed was to an int instead of to an unsigned
2987 long int. This mattered only on 64-bit systems.
2989 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
2990 generated) instead of pcre.in, which it its source. Also made the same change
2991 in several of the .c files.
2993 10. A new release of gcc defines printf() as a macro, which broke pcretest
2994 because it had an ifdef in the middle of a string argument for printf(). Fixed
2995 by using separate calls to printf().
2997 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
2998 script, to force use of CR or LF instead of \n in the source. On non-Unix
2999 systems, the value can be set in config.h.
3001 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3002 absolute limit. Changed the text of the error message to make this clear, and
3003 likewise updated the man page.
3005 13. The limit of 99 on the number of capturing subpatterns has been removed.
3006 The new limit is 65535, which I hope will not be a "real" limit.
3009 Version 3.4 22-Aug-00
3010 ---------------------
3012 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3014 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3017 Version 3.3 01-Aug-00
3018 ---------------------
3020 1. If an octal character was given, but the value was greater than \377, it
3021 was not getting masked to the least significant bits, as documented. This could
3022 lead to crashes in some systems.
3024 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3025 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3027 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3028 These just pass their arguments on to (pcre_free)(), but they are provided
3029 because some uses of PCRE bind it to non-C systems that can call its functions,
3030 but cannot call free() or pcre_free() directly.
3032 4. Add "make test" as a synonym for "make check". Corrected some comments in
3033 the Makefile.
3035 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3036 Makefile.
3038 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3039 command called pgrep for grepping around the active processes.
3041 7. Added the beginnings of support for UTF-8 character strings.
3043 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3044 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3045 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3046 out for the ar command.)
3049 Version 3.2 12-May-00
3050 ---------------------
3052 This is purely a bug fixing release.
3054 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3055 of ZA. This was just one example of several cases that could provoke this bug,
3056 which was introduced by change 9 of version 2.00. The code for breaking
3057 infinite loops after an iteration that matches an empty string was't working
3058 correctly.
3060 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3061 when matched against abbab (for example). After matching an empty string, it
3062 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3063 caused it to match further down the string than it should.
3065 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3066 was there because it doesn't seem to be needed, and it causes trouble on some
3067 systems, as it is not a Standard C header. It has been removed.
3069 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3070 were reported on the Macintosh. The changes were from
3072 while ((c = *(++ptr)) != 0 && c != '\n');
3073 to
3074 while ((c = *(++ptr)) != 0 && c != '\n') ;
3076 Totally extraordinary, but if that's what it takes...
3078 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3079 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3080 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3081 assumes the way PCRE uses memmove() (always moving upwards).
3083 6. PCRE is being used in one environment where strchr() is not available. There
3084 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3085 faster code anyway.
3088 Version 3.1 09-Feb-00
3089 ---------------------
3091 The only change in this release is the fixing of some bugs in Makefile.in for
3092 the "install" target:
3094 (1) It was failing to install pcreposix.h.
3096 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3099 Version 3.0 01-Feb-00
3100 ---------------------
3102 1. Add support for the /+ modifier to perltest (to output $` like it does in
3103 pcretest).
3105 2. Add support for the /g modifier to perltest.
3107 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3108 matches null strings.
3110 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3111 pattern. Perl treats empty patterns specially - it reuses the most recent
3112 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3113 effect.
3115 5. The POSIX interface was broken in that it was just handing over the POSIX
3116 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3117 required a bigger vector, with some working space on the end. This means that
3118 the POSIX wrapper now has to get and free some memory, and copy the results.
3120 6. Added some simple autoconf support, placing the test data and the
3121 documentation in separate directories, re-organizing some of the
3122 information files, and making it build pcre-config (a GNU standard). Also added
3123 libtool support for building PCRE as a shared library, which is now the
3124 default.
3126 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3127 09 are not valid octal constants. Single digits will be used for minor values
3128 less than 10.
3130 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3131 existing programs that set these in the POSIX interface can use PCRE without
3132 modification.
3134 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3135 return all that pcre_info() returns, plus additional data. The pcre_info()
3136 function is retained for compatibility, but is considered to be obsolete.
3138 10. Added experimental recursion feature (?R) to handle one common case that
3139 Perl 5.6 will be able to do with (?p{...}).
3141 11. Added support for POSIX character classes like [:alpha:], which Perl is
3142 adopting.
3145 Version 2.08 31-Aug-99
3146 ----------------------
3148 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3149 trying to match at the startoffset position, but instead was moving forward to
3150 the next newline as if a previous match had failed.
3152 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3153 and could get into a loop if a null string was matched other than at the start
3154 of the subject.
3156 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3157 be distinguished at compile time, and for completeness also added PCRE_DATE.
3159 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3160 in GnuWin32 environments.
3163 Version 2.07 29-Jul-99
3164 ----------------------
3166 1. The documentation is now supplied in plain text form and HTML as well as in
3167 the form of man page sources.
3169 2. C++ compilers don't like assigning (void *) values to other pointer types.
3170 In particular this affects malloc(). Although there is no problem in Standard
3171 C, I've put in casts to keep C++ compilers happy.
3173 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3174 should be (const char *).
3176 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3177 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3178 However, I haven't made this a standard facility. The documentation doesn't
3179 mention it, and the Makefile doesn't support it.
3181 5. The Makefile now contains an "install" target, with editable destinations at
3182 the top of the file. The pcretest program is not installed.
3184 6. pgrep -V now gives the PCRE version number and date.
3186 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3187 causing the entire string to be ignored, instead of just the last character.
3189 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3190 non-matching string, it can take a very, very long time, even for strings of
3191 quite modest length, because of the nested recursion. PCRE now does better in
3192 some of these cases. It does this by remembering the last required literal
3193 character in the pattern, and pre-searching the subject to ensure it is present
3194 before running the real match. In other words, it applies a heuristic to detect
3195 some types of certain failure quickly, and in the above example, if presented
3196 with a string that has no trailing " it gives "no match" very quickly.
3198 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3199 other alternatives are tried instead.
3202 Version 2.06 09-Jun-99
3203 ----------------------
3205 1. Change pcretest's output for amount of store used to show just the code
3206 space, because the remainder (the data block) varies in size between 32-bit and
3207 64-bit systems.
3209 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3210 start matching at. This allows lookbehinds to work when searching for multiple
3211 occurrences in a string.
3213 3. Added additional options to pcretest for testing multiple occurrences:
3215 /+ outputs the rest of the string that follows a match
3216 /g loops for multiple occurrences, using the new startoffset argument
3217 /G loops for multiple occurrences by passing an incremented pointer
3219 4. PCRE wasn't doing the "first character" optimization for patterns starting
3220 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3221 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3222 the letter 'x'. On long subject strings, this gives a significant speed-up.
3225 Version 2.05 21-Apr-99
3226 ----------------------
3228 1. Changed the type of magic_number from int to long int so that it works
3229 properly on 16-bit systems.
3231 2. Fixed a bug which caused patterns starting with .* not to work correctly
3232 when the subject string contained newline characters. PCRE was assuming
3233 anchoring for such patterns in all cases, which is not correct because .* will
3234 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3235 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3236 must be retried after every newline in the subject.
3239 Version 2.04 18-Feb-99
3240 ----------------------
3242 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3243 computation of the store needed to hold the pattern was incorrect (too large).
3244 If such patterns were nested a few deep, this could multiply and become a real
3245 problem.
3247 2. Added /M option to pcretest to show the memory requirement of a specific
3248 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3250 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3251 compiled in such a way that the backtracking after subsequent failure was
3252 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3253 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3256 Version 2.03 02-Feb-99
3257 ----------------------
3259 1. Fixed typo and small mistake in man page.
3261 2. Added 4th condition (GPL supersedes if conflict) and created separate
3262 LICENCE file containing the conditions.
3264 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3265 Perl, that is the internal \ allows the delimiter to be included in the
3266 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3267 the final delimiter, add \ to the end of the pattern (to test the error).
3269 4. Added the convenience functions for extracting substrings after a successful
3270 match. Updated pcretest to make it able to test these functions.
3273 Version 2.02 14-Jan-99
3274 ----------------------
3276 1. Initialized the working variables associated with each extraction so that
3277 their saving and restoring doesn't refer to uninitialized store.
3279 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3280 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3281 fix the problem.
3283 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3284 calls, and wasn't printing the correct value for compiling calls. Increased the
3285 default value of LOOPREPEAT, and the number of significant figures in the
3286 times.
3288 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3290 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3291 a building problem on Windows NT with a FAT file system.
3294 Version 2.01 21-Oct-98
3295 ----------------------
3297 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3298 to character tables built by pcre_maketables() in the current locale. If NULL
3299 is passed, the default tables are used.
3302 Version 2.00 24-Sep-98
3303 ----------------------
3305 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3306 it any more.
3308 2. Allow quantification of (?>) groups, and make it work correctly.
3310 3. The first character computation wasn't working for (?>) groups.
3312 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3313 end of the subject) and add 5.005's \z, which really does match only at the
3314 very end of the subject.
3316 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3318 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3319 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3320 localized options. All options to pcre_study() were also removed.
3322 7. Add other new features from 5.005:
3324 $(?<= positive lookbehind
3325 $(?<! negative lookbehind
3326 (?imsx-imsx) added the unsetting capability
3327 such a setting is global if at outer level; local otherwise
3328 (?imsx-imsx:) non-capturing groups with option setting
3329 (?(cond)re|re) conditional pattern matching
3331 A backreference to itself in a repeated group matches the previous
3332 captured string.
3334 8. General tidying up of studying (both automatic and via "study")
3335 consequential on the addition of new assertions.
3337 9. As in 5.005, unlimited repeated groups that could match an empty substring
3338 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3339 runtime if any iteration does actually match an empty substring.
3341 10. Include the RunTest script in the distribution.
3343 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3344 discrepancies, some of which were old and were also with respect to 5.004. They
3345 have now been fixed.
3348 Version 1.09 28-Apr-98
3349 ----------------------
3351 1. A negated single character class followed by a quantifier with a minimum
3352 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3353 program crashes, or just wrong answers. This did not apply to negated classes
3354 containing more than one character, or to minima other than one.
3357 Version 1.08 27-Mar-98
3358 ----------------------
3360 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3362 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3363 latter must appear before anything that relies on it in the pattern.
3366 Version 1.07 16-Feb-98
3367 ----------------------
3369 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3370 repeat of a potentially empty string).
3373 Version 1.06 23-Jan-98
3374 ----------------------
3376 1. Added Markus Oberhumer's little patches for C++.
3378 2. Literal strings longer than 255 characters were broken.
3381 Version 1.05 23-Dec-97
3382 ----------------------
3384 1. Negated character classes containing more than one character were failing if
3385 PCRE_CASELESS was set at run time.
3388 Version 1.04 19-Dec-97
3389 ----------------------
3391 1. Corrected the man page, where some "const" qualifiers had been omitted.
3393 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3394 input syntax.
3396 3. Fixed memory leak which occurred when a regex with back references was
3397 matched with an offsets vector that wasn't big enough. The temporary memory
3398 that is used in this case wasn't being freed if the match failed.
3400 4. Tidied pcretest to ensure it frees memory that it gets.
3402 5. Temporary memory was being obtained in the case where the passed offsets
3403 vector was exactly big enough.
3405 6. Corrected definition of offsetof() from change 5 below.
3407 7. I had screwed up change 6 below and broken the rules for the use of
3408 setjmp(). Now fixed.
3411 Version 1.03 18-Dec-97
3412 ----------------------
3414 1. A erroneous regex with a missing opening parenthesis was correctly
3415 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3416 on some systems.
3418 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3419 it was reported that one broken compiler failed on the former because "code" is
3420 also an independent variable.
3422 3. The erroneous regex a[]b caused an array overrun reference.
3424 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3425 fail on data ending with that character. (It was going on too far, and checking
3426 the next character, typically a binary zero.) This was specific to the
3427 optimized code for single-character negative classes.
3429 5. Added a contributed patch from the TIN world which does the following:
3431 + Add an undef for memmove, in case the the system defines a macro for it.
3433 + Add a definition of offsetof(), in case there isn't one. (I don't know
3434 the reason behind this - offsetof() is part of the ANSI standard - but
3435 it does no harm).
3437 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3438 most of the places where whitespace preceded '#'. I have given up and
3439 allowed the remaining 2 cases to be at the margin.
3441 + Rename some variables in pcre to eliminate shadowing. This seems very
3442 pedantic, but does no harm, of course.
3444 6. Moved the call to setjmp() into its own function, to get rid of warnings
3445 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3447 7. Constructs such as \d{8,} were compiling into the equivalent of
3448 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3449 outcome, but in this particular case used more store than had been allocated,
3450 which caused the bug to be discovered because it threw up an internal error.
3452 8. The debugging code in both pcre and pcretest for outputting the compiled
3453 form of a regex was going wrong in the case of back references followed by
3454 curly-bracketed repeats.
3457 Version 1.02 12-Dec-97
3458 ----------------------
3460 1. Typos in pcre.3 and comments in the source fixed.
3462 2. Applied a contributed patch to get rid of places where it used to remove
3463 'const' from variables, and fixed some signed/unsigned and uninitialized
3464 variable warnings.
3466 3. Added the "runtest" target to Makefile.
3468 4. Set default compiler flag to -O2 rather than just -O.
3471 Version 1.01 19-Nov-97
3472 ----------------------
3474 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3475 like /([ab]*)*/, that is, for classes with more than one character in them.
3477 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3478 as /((?>a*))*/ (a PCRE_EXTRA facility).
3481 Version 1.00 18-Nov-97
3482 ----------------------
3484 1. Added compile-time macros to support systems such as SunOS4 which don't have
3485 memmove() or strerror() but have other things that can be used instead.
3487 2. Arranged that "make clean" removes the executables.
3490 Version 0.99 27-Oct-97
3491 ----------------------
3493 1. Fixed bug in code for optimizing classes with only one character. It was
3494 initializing a 32-byte map regardless, which could cause it to run off the end
3495 of the memory it had got.
3497 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3500 Version 0.98 22-Oct-97
3501 ----------------------
3503 1. Fixed bug in code for handling temporary memory usage when there are more
3504 back references than supplied space in the ovector. This could cause segfaults.
3507 Version 0.97 21-Oct-97
3508 ----------------------
3510 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3512 2. Optimized negated single characters not to use a bit map.
3514 3. Brought error texts together as macro definitions; clarified some of them;
3515 fixed one that was wrong - it said "range out of order" when it meant "invalid
3516 escape sequence".
3518 4. Changed some char * arguments to const char *.
3520 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3522 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3523 pcretest.
3526 Version 0.96 16-Oct-97
3527 ----------------------
3529 1. Added a simple "pgrep" utility to the distribution.
3531 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3532 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3533 where "ddd" means "one or more decimal digits".
3535 3. Fixed serious bug. If a pattern had a back reference, but the call to
3536 pcre_exec() didn't supply a large enough ovector to record the related
3537 identifying subpattern, the match always failed. PCRE now remembers the number
3538 of the largest back reference, and gets some temporary memory in which to save
3539 the offsets during matching if necessary, in order to ensure that
3540 backreferences always work.
3542 4. Increased the compatibility with Perl in a number of ways:
3544 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3545 to request this handling. The option can be set at compile or exec time.
3547 (b) $ matches before a terminating newline by default; an option
3548 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3549 mode). The option can be set at compile or exec time.
3551 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3552 the same as Perl's. If the decimal number it represents is less than 10
3553 or there aren't that many previous left capturing parentheses, an octal
3554 escape is read. Inside a character class, it's always an octal escape,
3555 even if it is a single digit.
3557 (d) An escaped but undefined alphabetic character is taken as a literal,
3558 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3559 escapes.
3561 (e) {0} is now permitted. (The previous item is removed from the compiled
3562 pattern).
3564 5. Changed all the names of code files so that the basic parts are no longer
3565 than 10 characters, and abolished the teeny "globals.c" file.
3567 6. Changed the handling of character classes; they are now done with a 32-byte
3568 bit map always.
3570 7. Added the -d and /D options to pcretest to make it possible to look at the
3571 internals of compilation without having to recompile pcre.
3574 Version 0.95 23-Sep-97
3575 ----------------------
3577 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3578 \x20 at the start of a run of normal characters. These were being treated as
3579 real characters, instead of the source characters being re-checked.
3582 Version 0.94 18-Sep-97
3583 ----------------------
3585 1. The functions are now thread-safe, with the caveat that the global variables
3586 containing pointers to malloc() and free() or alternative functions are the
3587 same for all threads.
3589 2. Get pcre_study() to generate a bitmap of initial characters for non-
3590 anchored patterns when this is possible, and use it if passed to pcre_exec().
3593 Version 0.93 15-Sep-97
3594 ----------------------
3596 1. /(b)|(:+)/ was computing an incorrect first character.
3598 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3599 but not actually doing anything yet.
3601 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3602 as Perl does (e.g. [-az] or [az-]).
3604 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3605 all possible positions.
3607 5. Split up into different modules to avoid including unneeded functions in a
3608 compiled binary. However, compile and exec are still in one module. The "study"
3609 function is split off.
3611 6. The character tables are now in a separate module whose source is generated
3612 by an auxiliary program - but can then be edited by hand if required. There are
3613 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3614 toupper() in the code.
3616 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3617 make them global. Abolish the function for setting them, as the caller can now
3618 set them directly.
3621 Version 0.92 11-Sep-97
3622 ----------------------
3624 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3625 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3627 2. Caseless matching was not working in character classes if the characters in
3628 the pattern were in upper case.
3630 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3632 4. Make PCRE_ANCHORED public and accept as a compile option.
3634 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3635 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3636 pass them.
3638 6. Give an error if bad option bits passed at compile or run time.
3640 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3641 pcretest to cause it to pass that flag.
3643 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3644 options, and the first character, if set.
3646 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3649 Version 0.91 10-Sep-97
3650 ----------------------
3652 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3653 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3655 2. PCRE was looping on encountering an indefinitely repeated back reference to
3656 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3657 Perl does - treats the match as successful.
3659 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5