ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 518 - (show annotations)
Tue May 18 15:47:01 2010 UTC (5 years, 6 months ago) by ph10
File size: 186444 byte(s)
Added PCRE_UCP and related stuff to make \w etc use Unicode properties.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.10 03 May-2010
5 ------------------------
7 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
10 2. (*ACCEPT) was not working when inside an atomic group.
12 3. Inside a character class, \B is treated as a literal by default, but
13 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
14 causes the error). The code is unchanged, but I tidied the documentation.
16 4. Inside a character class, PCRE always treated \R and \X as literals,
17 whereas Perl faults them if its -w option is set. I have changed PCRE so
18 that it faults them when PCRE_EXTRA is set.
20 5. Added support for \N, which always matches any character other than
21 newline. (It is the same as "." when PCRE_DOTALL is not set.)
23 6. When compiling pcregrep with newer versions of gcc which may have
24 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
25 declared with attribute warn_unused_result" were given. Just casting the
26 result to (void) does not stop the warnings; a more elaborate fudge is
27 needed. I've used a macro to implement this.
29 7. Minor change to pcretest.c to avoid a compiler warning.
31 8. Added four artifical Unicode properties to help with an option to make
32 \s etc use properties (see next item). The new properties are: Xan
33 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
35 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
36 use Unicode properties. (*UCP) at the start of a pattern can be used to set
37 this option. Modified pcretest to add /W to test this facility. Added
38 REG_UCP to make it available via the POSIX interface.
41 Version 8.02 19-Mar-2010
42 ------------------------
44 1. The Unicode data tables have been updated to Unicode 5.2.0.
46 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
47 configured.
49 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
50 original author of that file, following a query about its status.
52 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
53 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
55 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
56 quantifier applied to a forward-referencing subroutine call, could compile
57 incorrect code or give the error "internal error: previously-checked
58 referenced subpattern not found".
60 6. Both MS Visual Studio and Symbian OS have problems with initializing
61 variables to point to external functions. For these systems, therefore,
62 pcre_malloc etc. are now initialized to local functions that call the
63 relevant global functions.
65 7. There were two entries missing in the vectors called coptable and poptable
66 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
67 I've fixed the data, and added a kludgy way of testing at compile time that
68 the lengths are correct (equal to the number of opcodes).
70 8. Following on from 7, I added a similar kludge to check the length of the
71 eint vector in pcreposix.c.
73 9. Error texts for pcre_compile() are held as one long string to avoid too
74 much relocation at load time. To find a text, the string is searched,
75 counting zeros. There was no check for running off the end of the string,
76 which could happen if a new error number was added without updating the
77 string.
79 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
81 11. \K was not working if it appeared in an atomic group or in a group that
82 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
83 \K is "not well defined" if used in an assertion. PCRE now accepts it if
84 the assertion is positive, but not if it is negative.
86 12. Change 11 fortuitously reduced the size of the stack frame used in the
87 "match()" function of pcre_exec.c by one pointer. Forthcoming
88 implementation of support for (*MARK) will need an extra pointer on the
89 stack; I have reserved it now, so that the stack frame size does not
90 decrease.
92 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
93 item in branch that calls a recursion is a subroutine call - as in the
94 second branch in the above example - was incorrectly given the compile-
95 time error "recursive call could loop indefinitely" because pcre_compile()
96 was not correctly checking the subroutine for matching a non-empty string.
98 14. The checks for overrunning compiling workspace could trigger after an
99 overrun had occurred. This is a "should never occur" error, but it can be
100 triggered by pathological patterns such as hundreds of nested parentheses.
101 The checks now trigger 100 bytes before the end of the workspace.
103 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
106 Version 8.01 19-Jan-2010
107 ------------------------
109 1. If a pattern contained a conditional subpattern with only one branch (in
110 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
111 computed the wrong minimum data length (which is of course zero for such
112 subpatterns). This could cause incorrect "no match" results.
114 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
115 the pattern is reset in the first branch, pcre_compile() failed with
116 "internal error: code overflow at offset...". This happened only when
117 the reset was to the original external option setting. (An optimization
118 abstracts leading options settings into an external setting, which was the
119 cause of this.)
121 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
122 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
123 assertion pattern did not match (meaning that the assertion was true), it
124 was incorrectly treated as false if the SKIP had been reached during the
125 matching. This also applied to assertions used as conditions.
127 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
128 assertion subpattern, including such a pattern used as a condition,
129 unpredictable results occurred, instead of the error return
132 5. The C++ GlobalReplace function was not working like Perl for the special
133 situation when an empty string is matched. It now does the fancy magic
134 stuff that is necessary.
136 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
137 removed. (These were left over from very, very early versions of PCRE.)
139 7. Some cosmetic changes to the code to make life easier when compiling it
140 as part of something else:
142 (a) Change DEBUG to PCRE_DEBUG.
144 (b) In pcre_compile(), rename the member of the "branch_chain" structure
145 called "current" as "current_branch", to prevent a collision with the
146 Linux macro when compiled as a kernel module.
148 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
149 prevent a collision with the Linux macro when compiled as a kernel
150 module.
152 8. In pcre_compile() there are some checks for integer overflows that used to
153 cast potentially large values to (double). This has been changed to that
154 when building, a check for int64_t is made, and if it is found, it is used
155 instead, thus avoiding the use of floating point arithmetic. (There is no
156 other use of FP in PCRE.) If int64_t is not found, the fallback is to
157 double.
159 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
160 2005 (difference between two addresses compared to an unsigned value).
162 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
163 custom one, because of the following reported problem in Windows:
165 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
166 under Win32.
167 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
168 therefore missing the function definition.
169 - The compiler thus generates a "C" signature for the test function.
170 - The linker fails to find the "C" function.
171 - PCRE fails to configure if asked to do so against libbz2.
173 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
174 messages were output:
176 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
177 rerunning libtoolize, to keep the correct libtool macros in-tree.
178 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
180 I have done both of these things.
182 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
183 most of the time, it *can* run out if it is given a pattern that contains a
184 runaway infinite recursion. I updated the discussion in the pcrestack man
185 page.
187 13. Now that we have gone to the x.xx style of version numbers, the minor
188 version may start with zero. Using 08 or 09 is a bad idea because users
189 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
190 interpreted as invalid octal numbers. I've updated the previous comment in
191 configure.ac, and also added a check that gives an error if 08 or 09 are
192 used.
194 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
195 causing partial matching to fail when the end of the subject matched \W
196 in a UTF-8 pattern where \W was quantified with a minimum of 3.
198 15. There were some discrepancies between the declarations in pcre_internal.h
199 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
200 their definitions. The declarations used "const uschar *" and the
201 definitions used USPTR. Even though USPTR is normally defined as "const
202 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
203 reported that: "This difference in casting confuses some C++ compilers, for
204 example, SunCC recognizes above declarations as different functions and
205 generates broken code for hbpcre." I have changed the declarations to use
206 USPTR.
208 16. GNU libtool is named differently on some systems. The autogen.sh script now
209 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
210 (FreeBSD).
212 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
213 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
214 comment: "Figure out how to create a longlong from a string: strtoll and
215 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
216 instance, but it only takes 2 args instead of 3!"
218 18. A subtle bug concerned with back references has been fixed by a change of
219 specification, with a corresponding code fix. A pattern such as
220 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
221 refers, was giving matches when it shouldn't. For example, xa=xaaa would
222 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
223 same bug. Such groups have to be quantified to be useful, or contained
224 inside another quantified group. (If there's no repetition, the reference
225 can never match.) The problem arises because, having left the group and
226 moved on to the rest of the pattern, a later failure that backtracks into
227 the group uses the captured value from the final iteration of the group
228 rather than the correct earlier one. I have fixed this in PCRE by forcing
229 any group that contains a reference to itself to be an atomic group; that
230 is, there cannot be any backtracking into it once it has completed. This is
231 similar to recursive and subroutine calls.
234 Version 8.00 19-Oct-09
235 ----------------------
237 1. The table for translating pcre_compile() error codes into POSIX error codes
238 was out-of-date, and there was no check on the pcre_compile() error code
239 being within the table. This could lead to an OK return being given in
240 error.
242 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
243 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
244 in a Windows environment.
246 3. The pcregrep --count option prints the count for each file even when it is
247 zero, as does GNU grep. However, pcregrep was also printing all files when
248 --files-with-matches was added. Now, when both options are given, it prints
249 counts only for those files that have at least one match. (GNU grep just
250 prints the file name in this circumstance, but including the count seems
251 more useful - otherwise, why use --count?) Also ensured that the
252 combination -clh just lists non-zero counts, with no names.
254 4. The long form of the pcregrep -F option was incorrectly implemented as
255 --fixed_strings instead of --fixed-strings. This is an incompatible change,
256 but it seems right to fix it, and I didn't think it was worth preserving
257 the old behaviour.
259 5. The command line items --regex=pattern and --regexp=pattern were not
260 recognized by pcregrep, which required --regex pattern or --regexp pattern
261 (with a space rather than an '='). The man page documented the '=' forms,
262 which are compatible with GNU grep; these now work.
264 6. No libpcreposix.pc file was created for pkg-config; there was just
265 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
267 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
268 when UCP support is not needed, by modifying the Python script that
269 generates it from Unicode data files. This should not matter if the module
270 is correctly used as a library, but I received one complaint about 50K of
271 unwanted data. My guess is that the person linked everything into his
272 program rather than using a library. Anyway, it does no harm.
274 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
275 was a minimum greater than 1 for a wide character in a possessive
276 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
277 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
278 character. Chaos in the form of incorrect output or a compiling loop could
279 result.
281 9. The restrictions on what a pattern can contain when partial matching is
282 requested for pcre_exec() have been removed. All patterns can now be
283 partially matched by this function. In addition, if there are at least two
284 slots in the offset vector, the offset of the earliest inspected character
285 for the match and the offset of the end of the subject are set in them when
286 PCRE_ERROR_PARTIAL is returned.
288 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
289 synonymous with PCRE_PARTIAL, for backwards compatibility, and
290 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
291 and may be more useful for multi-segment matching.
293 11. Partial matching with pcre_exec() is now more intuitive. A partial match
294 used to be given if ever the end of the subject was reached; now it is
295 given only if matching could not proceed because another character was
296 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
297 string "Z", which now yields "no match" instead of "partial match". In the
298 case of pcre_dfa_exec(), "no match" is given if every matching path for the
299 final character ended with (*FAIL).
301 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
302 if the pattern had a "must contain" character that was already found in the
303 earlier partial match, unless partial matching was again requested. For
304 example, with the pattern /dog.(body)?/, the "must contain" character is
305 "g". If the first part-match was for the string "dog", restarting with
306 "sbody" failed. This bug has been fixed.
308 13. The string returned by pcre_dfa_exec() after a partial match has been
309 changed so that it starts at the first inspected character rather than the
310 first character of the match. This makes a difference only if the pattern
311 starts with a lookbehind assertion or \b or \B (\K is not supported by
312 pcre_dfa_exec()). It's an incompatible change, but it makes the two
313 matching functions compatible, and I think it's the right thing to do.
315 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
316 so that the demonstration program is easily available in environments where
317 PCRE has not been installed from source.
319 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
320 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
321 library.
323 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
324 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
325 is not the first non-POSIX option to be added. Clearly some people find
326 these options useful.
328 17. If a caller to the POSIX matching function regexec() passes a non-zero
329 value for nmatch with a NULL value for pmatch, the value of
330 nmatch is forced to zero.
332 18. RunGrepTest did not have a test for the availability of the -u option of
333 the diff command, as RunTest does. It now checks in the same way as
334 RunTest, and also checks for the -b option.
336 19. If an odd number of negated classes containing just a single character
337 interposed, within parentheses, between a forward reference to a named
338 subpattern and the definition of the subpattern, compilation crashed with
339 an internal error, complaining that it could not find the referenced
340 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
341 [The bug was that it was starting one character too far in when skipping
342 over the character class, thus treating the ] as data rather than
343 terminating the class. This meant it could skip too much.]
345 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
346 /g option in pcretest when the pattern contains \K, which makes it possible
347 to have an empty string match not at the start, even when the pattern is
348 anchored. Updated pcretest and pcredemo to use this option.
350 21. If the maximum number of capturing subpatterns in a recursion was greater
351 than the maximum at the outer level, the higher number was returned, but
352 with unset values at the outer level. The correct (outer level) value is
353 now given.
355 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
356 PCRE did not set those parentheses (unlike Perl). I have now found a way to
357 make it do so. The string so far is captured, making this feature
358 compatible with Perl.
360 23. The tests have been re-organized, adding tests 11 and 12, to make it
361 possible to check the Perl 5.10 features against Perl 5.10.
363 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
364 pattern matches a fixed length string. PCRE did not allow this; now it
365 does. Neither allows recursion.
367 25. I finally figured out how to implement a request to provide the minimum
368 length of subject string that was needed in order to match a given pattern.
369 (It was back references and recursion that I had previously got hung up
370 on.) This code has now been added to pcre_study(); it finds a lower bound
371 to the length of subject needed. It is not necessarily the greatest lower
372 bound, but using it to avoid searching strings that are too short does give
373 some useful speed-ups. The value is available to calling programs via
374 pcre_fullinfo().
376 26. While implementing 25, I discovered to my embarrassment that pcretest had
377 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
378 study optimizations had never been tested with that matching function.
379 Oops. What is worse, even when it was passed study data, there was a bug in
380 pcre_dfa_exec() that meant it never actually used it. Double oops. There
381 were also very few tests of studied patterns with pcre_dfa_exec().
383 27. If (?| is used to create subpatterns with duplicate numbers, they are now
384 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
385 on the other side of the coin, they are no longer allowed to have different
386 names, because these cannot be distinguished in PCRE, and this has caused
387 confusion. (This is a difference from Perl.)
389 28. When duplicate subpattern names are present (necessarily with different
390 numbers, as required by 27 above), and a test is made by name in a
391 conditional pattern, either for a subpattern having been matched, or for
392 recursion in such a pattern, all the associated numbered subpatterns are
393 tested, and the overall condition is true if the condition is true for any
394 one of them. This is the way Perl works, and is also more like the way
395 testing by number works.
398 Version 7.9 11-Apr-09
399 ---------------------
401 1. When building with support for bzlib/zlib (pcregrep) and/or readline
402 (pcretest), all targets were linked against these libraries. This included
403 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
404 libraries. This caused unwanted dependencies to be created. This problem
405 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
406 pcretest is linked with readline.
408 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
409 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
410 moved outside it again, because FALSE and TRUE are already defined in AIX,
411 but BOOL is not.
413 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
414 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
416 4. The pcregrep documentation said spaces were inserted as well as colons (or
417 hyphens) following file names and line numbers when outputting matching
418 lines. This is not true; no spaces are inserted. I have also clarified the
419 wording for the --colour (or --color) option.
421 5. In pcregrep, when --colour was used with -o, the list of matching strings
422 was not coloured; this is different to GNU grep, so I have changed it to be
423 the same.
425 6. When --colo(u)r was used in pcregrep, only the first matching substring in
426 each matching line was coloured. Now it goes on to look for further matches
427 of any of the test patterns, which is the same behaviour as GNU grep.
429 7. A pattern that could match an empty string could cause pcregrep to loop; it
430 doesn't make sense to accept an empty string match in pcregrep, so I have
431 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
432 seems to be how GNU grep behaves.
434 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
435 start or after a newline", because the conditional assertion was not being
436 correctly handled. The rule now is that both the assertion and what follows
437 in the first alternative must satisfy the test.
439 9. If auto-callout was enabled in a pattern with a conditional group whose
440 condition was an assertion, PCRE could crash during matching, both with
441 pcre_exec() and pcre_dfa_exec().
443 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
444 used for matching.
446 11. Unicode property support in character classes was not working for
447 characters (bytes) greater than 127 when not in UTF-8 mode.
449 12. Added the -M command line option to pcretest.
451 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
453 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
455 16. Added comments and documentation about mis-use of no_arg in the C++
456 wrapper.
458 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
459 from Martin Jerabek that uses macro names for all relevant character and
460 string constants.
462 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
463 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
464 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
465 these, but not everybody uses configure.
467 19. A conditional group that had only one branch was not being correctly
468 recognized as an item that could match an empty string. This meant that an
469 enclosing group might also not be so recognized, causing infinite looping
470 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
471 with the subject "ab", where knowledge that the repeated group can match
472 nothing is needed in order to break the loop.
474 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
475 exec(), but without supplying a callout function, matching went wrong.
477 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
478 leak if the size of the offset vector was greater than 30. When the vector
479 is smaller, the saved offsets during recursion go onto a local stack
480 vector, but for larger vectors malloc() is used. It was failing to free
481 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
482 error, in fact).
484 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
485 heapframe that is used only when UTF-8 support is enabled. This caused no
486 problem, but was untidy.
488 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
489 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
490 included within another project.
492 24. Steven Van Ingelgem's patches to add more options to the CMake support,
493 slightly modified by me:
495 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
496 not building pcregrep.
498 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
499 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
501 25. Forward references, both numeric and by name, in patterns that made use of
502 duplicate group numbers, could behave incorrectly or give incorrect errors,
503 because when scanning forward to find the reference group, PCRE was not
504 taking into account the duplicate group numbers. A pattern such as
505 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
507 26. Changed a few more instances of "const unsigned char *" to USPTR, making
508 the feature of a custom pointer more persuasive (as requested by a user).
510 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
511 pcretest.c, inside #ifndefs, because it seems they are sometimes already
512 pre-defined.
514 28. Added support for (*UTF8) at the start of a pattern.
516 29. Arrange for flags added by the "release type" setting in CMake to be shown
517 in the configuration summary.
520 Version 7.8 05-Sep-08
521 ---------------------
523 1. Replaced UCP searching code with optimized version as implemented for Ad
524 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
525 stage table and inline lookup instead of a function, giving speed ups of 2
526 to 5 times on some simple patterns that I tested. Permission was given to
527 distribute the MultiStage2.py script that generates the tables (it's not in
528 the tarball, but is in the Subversion repository).
530 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
531 scripts.
533 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
534 a group with a zero qualifier. The result of the study could be incorrect,
535 or the function might crash, depending on the pattern.
537 4. Caseless matching was not working for non-ASCII characters in back
538 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
539 It now works when Unicode Property Support is available.
541 5. In pcretest, an escape such as \x{de} in the data was always generating
542 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
543 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
544 truncation.
546 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
548 7. Added two (int) casts to pcregrep when printing the difference of two
549 pointers, in case they are 64-bit values.
551 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
552 test 2 if it fails.
554 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
555 and a #define of that name to empty if it is not externally set. This is to
556 allow users of MSVC to set it if necessary.
558 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
559 the convenience functions in the pcre_get.c source file.
561 11. An option change at the start of a pattern that had top-level alternatives
562 could cause overwriting and/or a crash. This command provoked a crash in
563 some environments:
565 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
567 This potential security problem was recorded as CVE-2008-2371.
569 12. For a pattern where the match had to start at the beginning or immediately
570 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
571 pcre_dfa_exec() could read past the end of the passed subject if there was
572 no match. To help with detecting such bugs (e.g. with valgrind), I modified
573 pcretest so that it places the subject at the end of its malloc-ed buffer.
575 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
576 exec() might read past the end of the data buffer in UTF-8 mode.
578 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
579 the data contained the byte 0x85 as part of a UTF-8 character within its
580 first line. This applied both to normal and DFA matching.
582 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
583 /^[^d]*?$/8 failed to match "abc".
585 16. Added a missing copyright notice to pcrecpp_internal.h.
587 17. Make it more clear in the documentation that values returned from
588 pcre_exec() in ovector are byte offsets, not character counts.
590 18. Tidied a few places to stop certain compilers from issuing warnings.
592 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
593 supplied by Stefan Weber. I made a further small update for 7.8 because
594 there is a change of source arrangements: the pcre_searchfuncs.c module is
595 replaced by pcre_ucd.c.
598 Version 7.7 07-May-08
599 ---------------------
601 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
602 a string to a long long, pretend we don't even have a long long." This is
603 done by checking for the strtoq, strtoll, and _strtoi64 functions.
605 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
606 pre-7.6 versions, which defined a global no_arg variable instead of putting
607 it in the RE class. (See also #8 below.)
609 3. Remove a line of dead code, identified by coverity and reported by Nuno
610 Lopes.
612 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
614 (1) The include/exclude patterns were being applied to the whole pathnames
615 of files, instead of just to the final components.
617 (2) If there was more than one level of directory, the subdirectories were
618 skipped unless they satisfied the include/exclude conditions. This is
619 inconsistent with GNU grep (and could even be seen as contrary to the
620 pcregrep specification - which I improved to make it absolutely clear).
621 The action now is always to scan all levels of directory, and just
622 apply the include/exclude patterns to regular files.
624 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
625 --exclude_dir in the tests to avoid scanning .svn directories.
627 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
628 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
629 doesn't support NULs in patterns.
631 7. Added some missing "const"s to declarations of static tables in
632 pcre_compile.c and pcre_dfa_exec.c.
634 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
635 caused by fix #2 above. (Subsequently also a second patch to fix the
636 first patch. And a third patch - this was a messy problem.)
638 9. Applied Craig's patch to remove the use of push_back().
640 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
641 matching function regexec().
643 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
644 which, however, unlike Perl's \g{...}, are subroutine calls, not back
645 references. PCRE supports relative numbers with this syntax (I don't think
646 Oniguruma does).
648 12. Previously, a group with a zero repeat such as (...){0} was completely
649 omitted from the compiled regex. However, this means that if the group
650 was called as a subroutine from elsewhere in the pattern, things went wrong
651 (an internal error was given). Such groups are now left in the compiled
652 pattern, with a new opcode that causes them to be skipped at execution
653 time.
655 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
656 to the way PCRE behaves:
658 (a) A lone ] character is dis-allowed (Perl treats it as data).
660 (b) A back reference to an unmatched subpattern matches an empty string
661 (Perl fails the current match path).
663 (c) A data ] in a character class must be notated as \] because if the
664 first data character in a class is ], it defines an empty class. (In
665 Perl it is not possible to have an empty class.) The empty class []
666 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
667 The negative empty class [^] matches any one character, independently
668 of the DOTALL setting.
670 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
671 non-existent subpattern following a character class starting with ']' and
672 containing () gave an internal compiling error instead of "reference to
673 non-existent subpattern". Fortunately, when the pattern did exist, the
674 compiled code was correct. (When scanning forwards to check for the
675 existencd of the subpattern, it was treating the data ']' as terminating
676 the class, so got the count wrong. When actually compiling, the reference
677 was subsequently set up correctly.)
679 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
680 it was being rejected as not supported by pcre_dfa_exec(), even though
681 other assertions are supported. I have made pcre_dfa_exec() support
682 (*FAIL).
684 16. The implementation of 13c above involved the invention of a new opcode,
685 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
686 cannot be changed at match time, I realized I could make a small
687 improvement to matching performance by compiling OP_ALLANY instead of
688 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
689 on the OP_ANY path.
691 17. Compiling pcretest on Windows with readline support failed without the
692 following two fixes: (1) Make the unistd.h include conditional on
693 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
695 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
696 ncurses library to be included for pcretest when ReadLine support is
697 requested, but also to allow for it to be overridden. This patch came from
698 Daniel Bergström.
700 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
701 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
702 any errors with the current Unicode tables. Thanks to Peter Kankowski for
703 spotting this.
706 Version 7.6 28-Jan-08
707 ---------------------
709 1. A character class containing a very large number of characters with
710 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
711 overflow.
713 2. Patch to cut out the "long long" test in pcrecpp_unittest when
714 HAVE_LONG_LONG is not defined.
716 3. Applied Christian Ehrlicher's patch to update the CMake build files to
717 bring them up to date and include new features. This patch includes:
719 - Fixed PH's badly added libz and libbz2 support.
720 - Fixed a problem with static linking.
721 - Added pcredemo. [But later removed - see 7 below.]
722 - Fixed dftables problem and added an option.
723 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
725 - Added readline support for pcretest.
726 - Added an listing of the option settings after cmake has run.
728 4. A user submitted a patch to Makefile that makes it easy to create
729 "pcre.dll" under mingw when using Configure/Make. I added stuff to
730 Makefile.am that cause it to include this special target, without
731 affecting anything else. Note that the same mingw target plus all
732 the other distribution libraries and programs are now supported
733 when configuring with CMake (see 6 below) instead of with
734 Configure/Make.
736 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
737 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
738 exported in the Windows port". It has not yet been confirmed that the patch
739 solves the problem, but it does no harm.
741 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
742 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
743 with CMake, and also correct the comment about stack recursion.
745 7. Remove the automatic building of pcredemo from the ./configure system and
746 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
747 of a program that users should build themselves after PCRE is installed, so
748 building it automatically is not really right. What is more, it gave
749 trouble in some build environments.
751 8. Further tidies to CMakeLists.txt from Sheri and Christian.
754 Version 7.5 10-Jan-08
755 ---------------------
757 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
758 values in parens when parsing an RE using the C++ wrapper."
760 2. Negative specials like \S did not work in character classes in UTF-8 mode.
761 Characters greater than 255 were excluded from the class instead of being
762 included.
764 3. The same bug as (2) above applied to negated POSIX classes such as
765 [:^space:].
767 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
768 defined or documented. It seems to have been a typo for PCRE_STATIC, so
769 I have changed it.
771 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
772 first named subpattern) and a construct such as (?&a) would reference the
773 first named subpattern whose name started with "a" (in other words, the
774 length check was missing). Both these problems are fixed. "Subpattern name
775 expected" is now given for (?&) (a zero-length name), and this patch also
776 makes it give the same error for \k'' (previously it complained that that
777 was a reference to a non-existent subpattern).
779 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
780 this is right because (?- can be followed by option settings as well as by
781 digits. I have, however, made the messages clearer.
783 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
784 than the number used in the conditional) now cause a compile-time error.
785 This is actually not compatible with Perl, which accepts such patterns, but
786 treats the conditional as always being FALSE (as PCRE used to), but it
787 seems to me that giving a diagnostic is better.
789 8. Change "alphameric" to the more common word "alphanumeric" in comments
790 and messages.
792 9. Fix two occurrences of "backslash" in comments that should have been
793 "backspace".
795 10. Remove two redundant lines of code that can never be obeyed (their function
796 was moved elsewhere).
798 11. The program that makes PCRE's Unicode character property table had a bug
799 which caused it to generate incorrect table entries for sequences of
800 characters that have the same character type, but are in different scripts.
801 It amalgamated them into a single range, with the script of the first of
802 them. In other words, some characters were in the wrong script. There were
803 thirteen such cases, affecting characters in the following ranges:
805 U+002b0 - U+002c1
806 U+0060c - U+0060d
807 U+0061e - U+00612
808 U+0064b - U+0065e
809 U+0074d - U+0076d
810 U+01800 - U+01805
811 U+01d00 - U+01d77
812 U+01d9b - U+01dbf
813 U+0200b - U+0200f
814 U+030fc - U+030fe
815 U+03260 - U+0327f
816 U+0fb46 - U+0fbb1
817 U+10450 - U+1049d
819 12. The -o option (show only the matching part of a line) for pcregrep was not
820 compatible with GNU grep in that, if there was more than one match in a
821 line, it showed only the first of them. It now behaves in the same way as
822 GNU grep.
824 13. If the -o and -v options were combined for pcregrep, it printed a blank
825 line for every non-matching line. GNU grep prints nothing, and pcregrep now
826 does the same. The return code can be used to tell if there were any
827 non-matching lines.
829 14. Added --file-offsets and --line-offsets to pcregrep.
831 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
832 infinitely looping recursion. The bug was that positive lookaheads were not
833 being skipped when checking for a possible empty match (negative lookaheads
834 and both kinds of lookbehind were skipped).
836 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
837 inclusion of <windows.h> to before rather than after the definition of
838 INVALID_FILE_ATTRIBUTES (patch from David Byron).
840 17. Specifying a possessive quantifier with a specific limit for a Unicode
841 character property caused pcre_compile() to compile bad code, which led at
842 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
843 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
844 caused the error; without that there was no problem.
846 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
848 19. Added --enable-pcretest-libreadline.
850 20. In pcrecpp.cc, the variable 'count' was incremented twice in
851 RE::GlobalReplace(). As a result, the number of replacements returned was
852 double what it should be. I removed one of the increments, but Craig sent a
853 later patch that removed the other one (the right fix) and added unit tests
854 that check the return values (which was not done before).
856 21. Several CMake things:
858 (1) Arranged that, when cmake is used on Unix, the libraries end up with
859 the names libpcre and libpcreposix, not just pcre and pcreposix.
861 (2) The above change means that pcretest and pcregrep are now correctly
862 linked with the newly-built libraries, not previously installed ones.
866 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
867 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
868 UTF-8 newline character). The key issue is that the pattern starts .*;
869 this means that the match must be either at the beginning, or after a
870 newline. The bug was in the code for advancing after a failed match and
871 checking that the new position followed a newline. It was not taking
872 account of UTF-8 characters correctly.
874 23. PCRE was behaving differently from Perl in the way it recognized POSIX
875 character classes. PCRE was not treating the sequence [:...:] as a
876 character class unless the ... were all letters. Perl, however, seems to
877 allow any characters between [: and :], though of course it rejects as
878 unknown any "names" that contain non-letters, because all the known class
879 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
880 for example, whereas PCRE did not - it did not recognize a POSIX character
881 class. This seemed a bit dangerous, so the code has been changed to be
882 closer to Perl. The behaviour is not identical to Perl, because PCRE will
883 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
884 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
885 Perl does, and where it didn't before.
887 24. Rewrite so as to remove the single use of %n from pcregrep because in some
888 Windows environments %n is disabled by default.
891 Version 7.4 21-Sep-07
892 ---------------------
894 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
895 means that a class such as [\s] counted as "explicit reference to CR or
896 LF". That isn't really right - the whole point of the change was to try to
897 help when there was an actual mention of one of the two characters. So now
898 the change happens only if \r or \n (or a literal CR or LF) character is
899 encountered.
901 2. The 32-bit options word was also used for 6 internal flags, but the numbers
902 of both had grown to the point where there were only 3 bits left.
903 Fortunately, there was spare space in the data structure, and so I have
904 moved the internal flags into a new 16-bit field to free up more option
905 bits.
907 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
908 but did not set the internal JCHANGED flag - either of these is enough to
909 control the way the "get" function works - but the PCRE_INFO_JCHANGED
910 facility is supposed to tell if (?J) was ever used, so now (?J) at the
911 start sets both bits.
913 4. Added options (at build time, compile time, exec time) to change \R from
914 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
916 5. doc/pcresyntax.html was missing from the distribution.
918 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
919 compatibility, even though it is no longer used.
921 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
922 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
923 windows.h file is present (where different names are used). [This was
924 reversed later after testing - see 16 below.]
926 8. Changed all #include <config.h> to #include "config.h". There were also
927 some further <pcre.h> cases that I changed to "pcre.h".
929 9. When pcregrep was used with the --colour option, it missed the line ending
930 sequence off the lines that it output.
932 10. It was pointed out to me that arrays of string pointers cause lots of
933 relocations when a shared library is dynamically loaded. A technique of
934 using a single long string with a table of offsets can drastically reduce
935 these. I have refactored PCRE in four places to do this. The result is
936 dramatic:
938 Originally: 290
939 After changing UCP table: 187
940 After changing error message table: 43
941 After changing table of "verbs" 36
942 After changing table of Posix names 22
944 Thanks to the folks working on Gregex for glib for this insight.
946 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
947 unicode-properties was also set.
949 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
951 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
952 checked only for CRLF.
954 14. Added casts to pcretest.c to avoid compiler warnings.
956 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
958 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
959 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
960 entirely. This removes changes made in 7 above.
962 17. The CMake files have been updated, and there is now more information about
963 building with CMake in the NON-UNIX-USE document.
966 Version 7.3 28-Aug-07
967 ---------------------
969 1. In the rejigging of the build system that eventually resulted in 7.1, the
970 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
971 brackets there is not right, since it causes compilers to look for an
972 installed pcre.h, not the version that is in the source that is being
973 compiled (which of course may be different). I have changed it back to:
975 #include "pcre.h"
977 I have a vague recollection that the change was concerned with compiling in
978 different directories, but in the new build system, that is taken care of
979 by the VPATH setting the Makefile.
981 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
982 when the subject happened to end in the byte 0x85 (e.g. if the last
983 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
984 characters but of course it shouldn't be taken as a newline when it is part
985 of another character. The bug was that, for an unlimited repeat of . in
986 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
987 characters when looking for a newline.
989 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
991 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
992 in debug output.
994 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
995 long printing in the pcrecpp unittest when running under MinGW.
997 6. ESC_K was left out of the EBCDIC table.
999 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1000 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1001 limit also applies to "virtual nesting" when a pattern is recursive, and in
1002 this case 1000 isn't so big. I have been able to remove this limit at the
1003 expense of backing off one optimization in certain circumstances. Normally,
1004 when pcre_exec() would call its internal match() function recursively and
1005 immediately return the result unconditionally, it uses a "tail recursion"
1006 feature to save stack. However, when a subpattern that can match an empty
1007 string has an unlimited repetition quantifier, it no longer makes this
1008 optimization. That gives it a stack frame in which to save the data for
1009 checking that an empty string has been matched. Previously this was taken
1010 from the 1000-entry workspace that had been reserved. So now there is no
1011 explicit limit, but more stack is used.
1013 8. Applied Daniel's patches to solve problems with the import/export magic
1014 syntax that is required for Windows, and which was going wrong for the
1015 pcreposix and pcrecpp parts of the library. These were overlooked when this
1016 problem was solved for the main library.
1018 9. There were some crude static tests to avoid integer overflow when computing
1019 the size of patterns that contain repeated groups with explicit upper
1020 limits. As the maximum quantifier is 65535, the maximum group length was
1021 set at 30,000 so that the product of these two numbers did not overflow a
1022 32-bit integer. However, it turns out that people want to use groups that
1023 are longer than 30,000 bytes (though not repeat them that many times).
1024 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1025 made it possible to implement the integer overflow checks in a much more
1026 dynamic way, which I have now done. The artificial limitation on group
1027 length has been removed - we now have only the limit on the total length of
1028 the compiled pattern, which depends on the LINK_SIZE setting.
1030 10. Fixed a bug in the documentation for get/copy named substring when
1031 duplicate names are permitted. If none of the named substrings are set, the
1032 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1033 empty string.
1035 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1036 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1037 because the ] is interpreted as the first data character and the
1038 terminating ] is not found. PCRE has been made compatible with Perl in this
1039 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1040 cause memory overwriting.
1042 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1043 string has been matched (to stop an infinite loop). It was not recognizing
1044 a conditional subpattern that could match an empty string if that
1045 subpattern was within another subpattern. For example, it looped when
1046 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1047 condition was not nested. This bug has been fixed.
1049 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1050 past the start of the subject in the presence of bytes with the top bit
1051 set, for example "\x8aBCD".
1053 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1054 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1056 14. Optimized (?!) to (*FAIL).
1058 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1059 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1060 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1061 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1062 does: it's just the validity check that is more restrictive.
1064 16. Inserted checks for integer overflows during escape sequence (backslash)
1065 processing, and also fixed erroneous offset values for syntax errors during
1066 backslash processing.
1068 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1069 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1071 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1072 caused an overrun.
1074 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1075 something other than just ASCII characters) inside a group that had an
1076 unlimited repeat caused a loop at compile time (while checking to see
1077 whether the group could match an empty string).
1079 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1080 [\P{Any}] did so. (Error in the code for printing property names.)
1082 21. An orphan \E inside a character class could cause a crash.
1084 22. A repeated capturing bracket such as (A)? could cause a wild memory
1085 reference during compilation.
1087 23. There are several functions in pcre_compile() that scan along a compiled
1088 expression for various reasons (e.g. to see if it's fixed length for look
1089 behind). There were bugs in these functions when a repeated \p or \P was
1090 present in the pattern. These operators have additional parameters compared
1091 with \d, etc, and these were not being taken into account when moving along
1092 the compiled data. Specifically:
1094 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1095 length.
1097 (b) An item such as \pL+ within a repeated group could cause crashes or
1098 loops.
1100 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1101 "reference to non-existent subpattern" error.
1103 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1105 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1106 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1108 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1110 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1111 character were causing crashes (broken optimization).
1113 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1114 \p or \P) caused a compile-time loop.
1116 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1117 break. For example, the unstudied pattern [\r\n]A does not match the string
1118 "\r\nA" because change 7.0/46 below moves the current point on by two
1119 characters after failing to match at the start. However, the pattern \nA
1120 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1121 the same is true. There doesn't seem any very clean way out of this, but
1122 what I have chosen to do makes the common cases work: PCRE now takes note
1123 of whether there can be an explicit match for \r or \n anywhere in the
1124 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1125 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1126 pattern has explicit CR or LF references.
1128 29. Added (*CR) etc for changing newline setting at start of pattern.
1131 Version 7.2 19-Jun-07
1132 ---------------------
1134 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1135 which is apparently normally available under Windows.
1137 2. Re-jig the pcregrep tests with different newline settings in an attempt
1138 to make them independent of the local environment's newline setting.
1140 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1142 4. Some of the "internals" tests were previously cut out when the link size
1143 was not 2, because the output contained actual offsets. The recent new
1144 "Z" feature of pcretest means that these can be cut out, making the tests
1145 usable with all link sizes.
1147 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1148 stack recursion. This gives a massive performance boost under BSD, but just
1149 a small improvement under Linux. However, it saves one field in the frame
1150 in all cases.
1152 6. Added more features from the forthcoming Perl 5.10:
1154 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1155 recursion call. It refers to the nth most recently opened parentheses.
1157 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1158 to be opened parentheses.
1160 (c) Conditions that refer to capturing parentheses can be specified
1161 relatively, for example, (?(-2)... or (?(+3)...
1163 (d) \K resets the start of the current match so that everything before
1164 is not part of it.
1166 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1168 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1169 reference syntax.
1171 (g) (?| introduces a group in which the numbering of parentheses in each
1172 alternative starts with the same number.
1174 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1176 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1179 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1180 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1181 for detecting groups that can match an empty string.
1183 9. A pattern with a very large number of alternatives (more than several
1184 hundred) was running out of internal workspace during the pre-compile
1185 phase, where pcre_compile() figures out how much memory will be needed. A
1186 bit of new cunning has reduced the workspace needed for groups with
1187 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1188 workspace instead of running out of the 4096 that are available.
1190 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1192 11. Applied patch from Google to remove an optimization that didn't quite work.
1193 The report of the bug said:
1195 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1196 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1197 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1199 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1200 it matched the wrong number of bytes.
1203 Version 7.1 24-Apr-07
1204 ---------------------
1206 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1207 that is more "standard", making use of automake and other Autotools. There
1208 is some re-arrangement of the files and adjustment of comments consequent
1209 on this.
1211 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1212 for recursive directory scanning broke on some systems because the files
1213 are not scanned in any specific order and on different systems the order
1214 was different. A call to "sort" has been inserted into RunGrepTest for the
1215 approprate test as a short-term fix. In the longer term there may be an
1216 alternative.
1218 3. I had an email from Eric Raymond about problems translating some of PCRE's
1219 man pages to HTML (despite the fact that I distribute HTML pages, some
1220 people do their own conversions for various reasons). The problems
1221 concerned the use of low-level troff macros .br and .in. I have therefore
1222 removed all such uses from the man pages (some were redundant, some could
1223 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1224 HTML has been updated to handle .nf/.fi and to complain if it encounters
1225 .br or .in.
1227 4. Updated comments in configure.ac that get placed in config.h.in and also
1228 arranged for config.h to be included in the distribution, with the name
1229 config.h.generic, for the benefit of those who have to compile without
1230 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1232 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1233 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1234 makevp.bat for the current PCRE, using the additional files
1235 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1237 6. A Windows user reported a minor discrepancy with test 2, which turned out
1238 to be caused by a trailing space on an input line that had got lost in his
1239 copy. The trailing space was an accident, so I've just removed it.
1241 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1242 that is needed.
1244 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1245 as "const" (a) because they are and (b) because it helps the PHP
1246 maintainers who have recently made a script to detect big data structures
1247 in the php code that should be moved to the .rodata section. I remembered
1248 to update Builducptable as well, so it won't revert if ucptable.h is ever
1249 re-created.
1251 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1252 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1253 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1254 support is not required. This saves 1.5-2K of code, which is important in
1255 some applications.
1257 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1258 so as not to refer to the tables, even though these functions will never be
1259 called when UTF-8 support is disabled. Otherwise there are problems with a
1260 shared library.
1262 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1264 (a) It was defining its arguments as char * instead of void *.
1266 (b) It was assuming that all moves were upwards in memory; this was true
1267 a long time ago when I wrote it, but is no longer the case.
1269 The emulated memove() is provided for those environments that have neither
1270 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1271 is clearly not the case, as these two bugs were recently reported.
1273 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1274 and Detrail to create the HTML documentation, the .txt form of the man
1275 pages, and it removes trailing spaces from listed files. It also creates
1276 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1277 case, it wraps all the #defines with #ifndefs. This script should be run
1278 before "make dist".
1280 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1281 with Unicode property support.
1283 (a) For a maximizing quantifier, if the two different cases of the
1284 character were of different lengths in their UTF-8 codings (there are
1285 some cases like this - I found 11), and the matching function had to
1286 back up over a mixture of the two cases, it incorrectly assumed they
1287 were both the same length.
1289 (b) When PCRE was configured to use the heap rather than the stack for
1290 recursion during matching, it was not correctly preserving the data for
1291 the other case of a UTF-8 character when checking ahead for a match
1292 while processing a minimizing repeat. If the check also involved
1293 matching a wide character, but failed, corruption could cause an
1294 erroneous result when trying to check for a repeat of the original
1295 character.
1297 13. Some tidying changes to the testing mechanism:
1299 (a) The RunTest script now detects the internal link size and whether there
1300 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1301 values substituted by "configure". (The RunGrepTest script already did
1302 this for UTF-8.) The configure.ac script no longer substitutes the
1303 relevant variables.
1305 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1306 with length and offset values. This means that the output is different
1307 for different internal link sizes. Test 2 is skipped for link sizes
1308 other than 2 because of this, bypassing the problem. Unfortunately,
1309 there was also a test in test 3 (the locale tests) that used /B and
1310 failed for link sizes other than 2. Rather than cut the whole test out,
1311 I have added a new /Z option to pcretest that replaces the length and
1312 offset values with spaces. This is now used to make test 3 independent
1313 of link size. (Test 2 will be tidied up later.)
1315 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1316 segmentation fault instead of returning the appropriate error message.
1318 15. In multiline mode when the newline sequence was set to "any", the pattern
1319 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1320 This doesn't seem right; it now treats the CRLF combination as the line
1321 ending, and so does not match in that case. It's only a pattern such as ^$
1322 that would hit this one: something like ^ABC$ would have failed after \r
1323 and then tried again after \r\n.
1325 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1326 in an attempt to make files that differ only in their line terminators
1327 compare equal. This works on Linux.
1329 17. Under certain error circumstances pcregrep might try to free random memory
1330 as it exited. This is now fixed, thanks to valgrind.
1332 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1333 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1334 was because its rules for how to advance for /g after matching an empty
1335 string at the end of a line did not allow for this case. They now check for
1336 it specially.
1338 20. pcretest is supposed to handle patterns and data of any length, by
1339 extending its buffers when necessary. It was getting this wrong when the
1340 buffer for a data line had to be extended.
1342 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1343 CRLF as a newline sequence.
1345 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1346 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1347 I have nevertheless tidied it up.
1349 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1351 24. Added a man page for pcre-config.
1354 Version 7.0 19-Dec-06
1355 ---------------------
1357 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1358 moving to gcc 4.1.1.
1360 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1361 sys/time.h, which is documented as needed for this function. It doesn't
1362 seem to matter on Linux, but it showed up on some releases of OS X.
1364 3. It seems that there are systems where bytes whose values are greater than
1365 127 match isprint() in the "C" locale. The "C" locale should be the
1366 default when a C program starts up. In most systems, only ASCII printing
1367 characters match isprint(). This difference caused the output from pcretest
1368 to vary, making some of the tests fail. I have changed pcretest so that:
1370 (a) When it is outputting text in the compiled version of a pattern, bytes
1371 other than 32-126 are always shown as hex escapes.
1373 (b) When it is outputting text that is a matched part of a subject string,
1374 it does the same, unless a different locale has been set for the match
1375 (using the /L modifier). In this case, it uses isprint() to decide.
1377 4. Fixed a major bug that caused incorrect computation of the amount of memory
1378 required for a compiled pattern when options that changed within the
1379 pattern affected the logic of the preliminary scan that determines the
1380 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1381 that the computed length was too small. The symptoms of this bug were
1382 either the PCRE error "internal error: code overflow" from pcre_compile(),
1383 or a glibc crash with a message such as "pcretest: free(): invalid next
1384 size (fast)". Examples of patterns that provoked this bug (shown in
1385 pcretest format) are:
1387 /(?-x: )/x
1388 /(?x)(?-x: \s*#\s*)/
1389 /((?i)[\x{c0}])/8
1390 /(?i:[\x{c0}])/8
1392 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1393 is now done differently.
1395 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1396 wrapper classes; (b) implement a new function in the C++ scanner that is
1397 more efficient than the old way of doing things because it avoids levels of
1398 recursion in the regex matching; (c) add a paragraph to the documentation
1399 for the FullMatch() function.
1401 6. The escape sequence \n was being treated as whatever was defined as
1402 "newline". Not only was this contrary to the documentation, which states
1403 that \n is character 10 (hex 0A), but it also went horribly wrong when
1404 "newline" was defined as CRLF. This has been fixed.
1406 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1407 was being set to -1 for the "end of line" case (supposedly a value that no
1408 character can have). Though this value is never used (the check for end of
1409 line is "zero bytes in current character"), it caused compiler complaints.
1410 I've changed it to 0xffffffff.
1412 8. In pcre_version.c, the version string was being built by a sequence of
1413 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1414 string (as it is for production releases) called a macro with an empty
1415 argument. The C standard says the result of this is undefined. The gcc
1416 compiler treats it as an empty string (which was what was wanted) but it is
1417 reported that Visual C gives an error. The source has been hacked around to
1418 avoid this problem.
1420 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1421 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1422 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1423 of them did).
1425 10. Originally, pcretest opened its input and output without "b"; then I was
1426 told that "b" was needed in some environments, so it was added for release
1427 5.0 to both the input and output. (It makes no difference on Unix-like
1428 systems.) Later I was told that it is wrong for the input on Windows. I've
1429 now abstracted the modes into two macros, to make it easier to fiddle with
1430 them, and removed "b" from the input mode under Windows.
1432 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1434 12. Added -help and --help to pcretest as an official way of being reminded
1435 of the options.
1437 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1438 and pcrecpp.cc because they annoy compilers at high warning levels.
1440 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1442 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1443 scripts are not C programs :-) and which was not noticed because it works
1444 on Linux.
1446 16. pcretest is supposed to handle any length of pattern and data line (as one
1447 line or as a continued sequence of lines) by extending its input buffer if
1448 necessary. This feature was broken for very long pattern lines, leading to
1449 a string of junk being passed to pcre_compile() if the pattern was longer
1450 than about 50K.
1452 17. I have done a major re-factoring of the way pcre_compile() computes the
1453 amount of memory needed for a compiled pattern. Previously, there was code
1454 that made a preliminary scan of the pattern in order to do this. That was
1455 OK when PCRE was new, but as the facilities have expanded, it has become
1456 harder and harder to keep it in step with the real compile phase, and there
1457 have been a number of bugs (see for example, 4 above). I have now found a
1458 cunning way of running the real compile function in a "fake" mode that
1459 enables it to compute how much memory it would need, while actually only
1460 ever using a few hundred bytes of working memory and without too many
1461 tests of the mode. This should make future maintenance and development
1462 easier. A side effect of this work is that the limit of 200 on the nesting
1463 depth of parentheses has been removed (though this was never a serious
1464 limitation, I suspect). However, there is a downside: pcre_compile() now
1465 runs more slowly than before (30% or more, depending on the pattern). I
1466 hope this isn't a big issue. There is no effect on runtime performance.
1468 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1469 newline (only possible for the last line of a file) and it was a
1470 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1472 19. Added additional timing features to pcretest. (1) The -tm option now times
1473 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1474 separate command line item, by a number that specifies the number of
1475 repeats to use when timing. The default is 50000; this gives better
1476 precision, but takes uncomfortably long for very large patterns.
1478 20. Extended pcre_study() to be more clever in cases where a branch of a
1479 subpattern has no definite first character. For example, (a*|b*)[cd] would
1480 previously give no result from pcre_study(). Now it recognizes that the
1481 first character must be a, b, c, or d.
1483 21. There was an incorrect error "recursive call could loop indefinitely" if
1484 a subpattern (or the entire pattern) that was being tested for matching an
1485 empty string contained only one non-empty item after a nested subpattern.
1486 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1487 incorrectly, because the \d was being skipped in the check.
1489 22. The pcretest program now has a new pattern option /B and a command line
1490 option -b, which is equivalent to adding /B to every pattern. This causes
1491 it to show the compiled bytecode, without the additional information that
1492 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1493 is the same as /B/I).
1495 23. A new optimization is now able automatically to treat some sequences such
1496 as a*b as a*+b. More specifically, if something simple (such as a character
1497 or a simple class like \d) has an unlimited quantifier, and is followed by
1498 something that cannot possibly match the quantified thing, the quantifier
1499 is automatically "possessified".
1501 24. A recursive reference to a subpattern whose number was greater than 39
1502 went wrong under certain circumstances in UTF-8 mode. This bug could also
1503 have affected the operation of pcre_study().
1505 25. Realized that a little bit of performance could be had by replacing
1506 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1508 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1510 27. Possessive quantifiers such as a++ were previously implemented by turning
1511 them into atomic groups such as ($>a+). Now they have their own opcodes,
1512 which improves performance. This includes the automatically created ones
1513 from 23 above.
1515 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1516 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1517 the first matched character to be a colon. This applied both to named and
1518 numbered groups.
1520 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1522 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1523 building PCRE on Windows easier, so I have included it in the distribution.
1525 31. There is now a check in pcretest against a ridiculously large number being
1526 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1527 loop, the loop is abandoned.
1529 32. Forward references to subpatterns in conditions such as (?(2)...) where
1530 subpattern 2 is defined later cause pcre_compile() to search forwards in
1531 the pattern for the relevant set of parentheses. This search went wrong
1532 when there were unescaped parentheses in a character class, parentheses
1533 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1535 33. "Subroutine" calls and backreferences were previously restricted to
1536 referencing subpatterns earlier in the regex. This restriction has now
1537 been removed.
1539 34. Added a number of extra features that are going to be in Perl 5.10. On the
1540 whole, these are just syntactic alternatives for features that PCRE had
1541 previously implemented using the Python syntax or my own invention. The
1542 other formats are all retained for compatibility.
1544 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1545 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1546 also .NET compatible.
1548 (b) A recursion or subroutine call to a named group can now be defined as
1549 (?&name) as well as (?P>name).
1551 (c) A backreference to a named group can now be defined as \k<name> or
1552 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1553 5.10, are also .NET compatible.
1555 (d) A conditional reference to a named group can now use the syntax
1556 (?(<name>) or (?('name') as well as (?(name).
1558 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1559 groups (named and numbered) that are never evaluated inline, but can be
1560 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1561 is always false. There may be only one alternative in such a group.
1563 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1564 as the simple (?(R). The condition is true only if the most recent
1565 recursion is that of the given number or name. It does not search out
1566 through the entire recursion stack.
1568 (g) The escape \gN or \g{N} has been added, where N is a positive or
1569 negative number, specifying an absolute or relative reference.
1571 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1572 some "unreachable code" warnings.
1574 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1575 things, this adds five new scripts.
1577 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1578 There were also incompatibilities regarding the handling of \Q..\E inside
1579 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1580 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1582 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1583 matches an empty string, and forcibly breaks the loop. There were bugs in
1584 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1585 against aaaa the result was just "a" rather than "aaaa", for example. Two
1586 separate and independent bugs (that affected different cases) have been
1587 fixed.
1589 39. Refactored the code to abolish the use of different opcodes for small
1590 capturing bracket numbers. This is a tidy that I avoided doing when I
1591 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1592 The new approach is not only tidier, it makes it possible to reduce the
1593 memory needed to fix the previous bug (38).
1595 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1596 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1597 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1598 mode.
1600 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1601 report.
1603 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1604 copy construction and assignment in the C++ wrapper.
1606 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1607 couple of bugs that could have given wrong results in the "--newline=crlf"
1608 case.
1610 44. Added a number of casts and did some reorganization of signed/unsigned int
1611 variables following suggestions from Dair Grant. Also renamed the variable
1612 "this" as "item" because it is a C++ keyword.
1614 45. Arranged for dftables to add
1616 #include "pcre_internal.h"
1618 to pcre_chartables.c because without it, gcc 4.x may remove the array
1619 definition from the final binary if PCRE is built into a static library and
1620 dead code stripping is activated.
1622 46. For an unanchored pattern, if a match attempt fails at the start of a
1623 newline sequence, and the newline setting is CRLF or ANY, and the next two
1624 characters are CRLF, advance by two characters instead of one.
1627 Version 6.7 04-Jul-06
1628 ---------------------
1630 1. In order to handle tests when input lines are enormously long, pcretest has
1631 been re-factored so that it automatically extends its buffers when
1632 necessary. The code is crude, but this _is_ just a test program. The
1633 default size has been increased from 32K to 50K.
1635 2. The code in pcre_study() was using the value of the re argument before
1636 testing it for NULL. (Of course, in any sensible call of the function, it
1637 won't be NULL.)
1639 3. The memmove() emulation function in pcre_internal.h, which is used on
1640 systems that lack both memmove() and bcopy() - that is, hardly ever -
1641 was missing a "static" storage class specifier.
1643 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1644 containing an extended class (one that cannot be represented by a bitmap
1645 because it contains high-valued characters or Unicode property items, e.g.
1646 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1647 pattern, but PCRE should not loop if you do not (it no longer does).
1648 [Detail: two cases were found: (a) a repeated subpattern containing an
1649 extended class; (b) a recursive reference to a subpattern that followed a
1650 previous extended class. It wasn't skipping over the extended class
1651 correctly when UTF-8 mode was not set.]
1653 5. A negated single-character class was not being recognized as fixed-length
1654 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1655 compile error "lookbehind assertion is not fixed length".
1657 6. The RunPerlTest auxiliary script was showing an unexpected difference
1658 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1659 write a Perl script that can interpret lines of an input file either as
1660 byte characters or as UTF-8, which is what "perltest" was being required to
1661 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1662 can't do is switch easily at run time between having the "use utf8;" pragma
1663 or not. In the end, I fudged it by using the RunPerlTest script to insert
1664 "use utf8;" explicitly for the UTF-8 tests.
1666 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1667 the end of the subject string, contrary to the documentation and to what
1668 Perl does. This was true of both matching functions. Now it matches only at
1669 the start of the subject and immediately after *internal* newlines.
1671 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1672 a pointer to an int instead of a pointer to an unsigned long int. This
1673 caused problems on 64-bit systems.
1675 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1676 instance of the 'standard' template library not being so standard".
1678 10. There was no check on the number of named subpatterns nor the maximum
1679 length of a subpattern name. The product of these values is used to compute
1680 the size of the memory block for a compiled pattern. By supplying a very
1681 long subpattern name and a large number of named subpatterns, the size
1682 computation could be caused to overflow. This is now prevented by limiting
1683 the length of names to 32 characters, and the number of named subpatterns
1684 to 10,000.
1686 11. Subpatterns that are repeated with specific counts have to be replicated in
1687 the compiled pattern. The size of memory for this was computed from the
1688 length of the subpattern and the repeat count. The latter is limited to
1689 65535, but there was no limit on the former, meaning that integer overflow
1690 could in principle occur. The compiled length of a repeated subpattern is
1691 now limited to 30,000 bytes in order to prevent this.
1693 12. Added the optional facility to have named substrings with the same name.
1695 13. Added the ability to use a named substring as a condition, using the
1696 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1697 are numbers (not recommended). Forward references are permitted.
1699 14. Added forward references in named backreferences (if you see what I mean).
1701 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1702 pattern could run off the end of the subject. For example, the pattern
1703 "(?s)(.{1,5})"8 did this with the subject "ab".
1705 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1706 PCRE_CASELESS was set when matching characters that were quantified with ?
1707 or *.
1709 17. A character class other than a single negated character that had a minimum
1710 but no maximum quantifier - for example [ab]{6,} - was not handled
1711 correctly by pce_dfa_exec(). It would match only one character.
1713 18. A valid (though odd) pattern that looked like a POSIX character
1714 class but used an invalid character after [ (for example [[,abc,]]) caused
1715 pcre_compile() to give the error "Failed: internal error: code overflow" or
1716 in some cases to crash with a glibc free() error. This could even happen if
1717 the pattern terminated after [[ but there just happened to be a sequence of
1718 letters, a binary zero, and a closing ] in the memory that followed.
1720 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1721 over the years. Originally (before any Unicode support), just the bottom 8
1722 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1723 output from "man perlunicode" includes this:
1725 The regular expression compiler produces polymorphic opcodes. That
1726 is, the pattern adapts to the data and automatically switches to
1727 the Unicode character scheme when presented with Unicode data--or
1728 instead uses a traditional byte scheme when presented with byte
1729 data.
1731 Sadly, a wide octal escape does not cause a switch, and in a string with
1732 no other multibyte characters, these octal escapes are treated as before.
1733 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1734 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1735 Unicode string.
1737 I have not perpetrated such confusion in PCRE. Up till now, it took just
1738 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1739 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1740 translate to the appropriate multibyte character.
1742 29. Applied some refactoring to reduce the number of warnings from Microsoft
1743 and Borland compilers. This has included removing the fudge introduced
1744 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1745 a warning about an unused variable.
1747 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1748 characters since release 4.0, because Perl (from release 5.004) does not.
1749 [Or at least, is documented not to: some releases seem to be in conflict
1750 with the documentation.] However, when a pattern was studied with
1751 pcre_study() and all its branches started with \s, PCRE still included VT
1752 as a possible starting character. Of course, this did no harm; it just
1753 caused an unnecessary match attempt.
1755 22. Removed a now-redundant internal flag bit that recorded the fact that case
1756 dependency changed within the pattern. This was once needed for "required
1757 byte" processing, but is no longer used. This recovers a now-scarce options
1758 bit. Also moved the least significant internal flag bit to the most-
1759 significant bit of the word, which was not previously used (hangover from
1760 the days when it was an int rather than a uint) to free up another bit for
1761 the future.
1763 23. Added support for CRLF line endings as well as CR and LF. As well as the
1764 default being selectable at build time, it can now be changed at runtime
1765 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1766 specify that it is scanning data with non-default line endings.
1768 24. Changed the definition of CXXLINK to make it agree with the definition of
1769 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1771 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1772 recursions. This makes a big different to stack usage for some patterns.
1774 26. If a subpattern containing a named recursion or subroutine reference such
1775 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1776 the space required for the compiled pattern went wrong and gave too small a
1777 value. Depending on the environment, this could lead to "Failed: internal
1778 error: code overflow at offset 49" or "glibc detected double free or
1779 corruption" errors.
1781 27. Applied patches from Google (a) to support the new newline modes and (b) to
1782 advance over multibyte UTF-8 characters in GlobalReplace.
1784 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1785 difference for some implementation of PCRE in some Windows version.
1787 29. Added some extra testing facilities to pcretest:
1789 \q<number> in a data line sets the "match limit" value
1790 \Q<number> in a data line sets the "match recursion limt" value
1791 -S <number> sets the stack size, where <number> is in megabytes
1793 The -S option isn't available for Windows.
1796 Version 6.6 06-Feb-06
1797 ---------------------
1799 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1800 in pcreposix.h. I have copied the definition from pcre.h.
1802 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1803 because pcre.h is no longer a built file.
1805 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1806 not normally included in the compiled code.
1809 Version 6.5 01-Feb-06
1810 ---------------------
1812 1. When using the partial match feature with pcre_dfa_exec(), it was not
1813 anchoring the second and subsequent partial matches at the new starting
1814 point. This could lead to incorrect results. For example, with the pattern
1815 /1234/, partially matching against "123" and then "a4" gave a match.
1817 2. Changes to pcregrep:
1819 (a) All non-match returns from pcre_exec() were being treated as failures
1820 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1821 error message is output. Some extra information is given for the
1823 probably the only errors that are likely to be caused by users (by
1824 specifying a regex that has nested indefinite repeats, for instance).
1825 If there are more than 20 of these errors, pcregrep is abandoned.
1827 (b) A binary zero was treated as data while matching, but terminated the
1828 output line if it was written out. This has been fixed: binary zeroes
1829 are now no different to any other data bytes.
1831 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1832 used to set a locale for matching. The --locale=xxxx long option has
1833 been added (no short equivalent) to specify a locale explicitly on the
1834 pcregrep command, overriding the environment variables.
1836 (d) When -B was used with -n, some line numbers in the output were one less
1837 than they should have been.
1839 (e) Added the -o (--only-matching) option.
1841 (f) If -A or -C was used with -c (count only), some lines of context were
1842 accidentally printed for the final match.
1844 (g) Added the -H (--with-filename) option.
1846 (h) The combination of options -rh failed to suppress file names for files
1847 that were found from directory arguments.
1849 (i) Added the -D (--devices) and -d (--directories) options.
1851 (j) Added the -F (--fixed-strings) option.
1853 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1855 (l) Added the --colo(u)r option.
1857 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1858 is not present by default.
1860 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1861 items such as (?R) or (?1), when the recursion could match a number of
1862 alternatives. If it matched one of the alternatives, but subsequently,
1863 outside the recursion, there was a failure, the code tried to back up into
1864 the recursion. However, because of the way PCRE is implemented, this is not
1865 possible, and the result was an incorrect result from the match.
1867 In order to prevent this happening, the specification of recursion has
1868 been changed so that all such subpatterns are automatically treated as
1869 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1871 4. I had overlooked the fact that, in some locales, there are characters for
1872 which isalpha() is true but neither isupper() nor islower() are true. In
1873 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1874 and ordfeminine) are like this. This affected the treatment of \w and \W
1875 when they appeared in character classes, but not when they appeared outside
1876 a character class. The bit map for "word" characters is now created
1877 separately from the results of isalnum() instead of just taking it from the
1878 upper, lower, and digit maps. (Plus the underscore character, of course.)
1880 5. The above bug also affected the handling of POSIX character classes such as
1881 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1882 permanent tables. Instead, the bit maps for such a class were previously
1883 created as the appropriate unions of the upper, lower, and digit bitmaps.
1884 Now they are created by subtraction from the [[:word:]] class, which has
1885 its own bitmap.
1887 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1888 It is created by subtracting the vertical space characters (\x09, \x0a,
1889 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1890 subtraction was done in the overall bitmap for a character class, meaning
1891 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1892 be recognized. This bug has been fixed.
1894 7. Patches from the folks at Google:
1896 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1897 real life, but is still worth protecting against".
1899 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1900 regular expressions".
1902 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1903 have it.
1905 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1906 "configure" and the latter not, in order to fix a problem somebody had
1907 with compiling the Arg class on HP-UX.
1909 (e) Improve the error-handling of the C++ wrapper a little bit.
1911 (f) New tests for checking recursion limiting.
1913 8. The pcre_memmove() function, which is used only if the environment does not
1914 have a standard memmove() function (and is therefore rarely compiled),
1915 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1916 returning a result (though PCRE never actually uses the result).
1918 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1919 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1920 returned instead of calling malloc() with an overflowing number that would
1921 most likely cause subsequent chaos.
1923 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1925 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1926 with this option is matched, the nmatch and pmatch options of regexec() are
1927 ignored.
1929 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1930 provided in case anyone wants to the the POSIX interface with UTF-8
1931 strings.
1933 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1934 C++ linking (needed for some HP-UX environments).
1936 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1937 (unused parameter) and in the pcre_printint() function (omitted "default"
1938 switch label when the default is to do nothing).
1940 15. Added some code to make it possible, when PCRE is compiled as a C++
1941 library, to replace subject pointers for pcre_exec() with a smart pointer
1942 class, thus making it possible to process discontinuous strings.
1944 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1945 much the same function. They were added by different people who were trying
1946 to make PCRE easy to compile on non-Unix systems. It has been suggested
1947 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1948 for compiling on Windows systems. I have therefore replaced it with
1949 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1950 defaults to "extern" for C or "extern C" for C++, which works fine on
1951 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1952 SCOPE with something explicit in config.h. In addition:
1954 (a) pcreposix.h still had just "extern" instead of either of these macros;
1955 I have replaced it with PCRE_DATA_SCOPE.
1957 (b) Functions such as _pcre_xclass(), which are internal to the library,
1958 but external in the C sense, all had PCRE_EXPORT in their definitions.
1959 This is apparently wrong for the Windows case, so I have removed it.
1960 (It makes no difference on Unix-like systems.)
1962 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1963 of recursive calls to match(). This is different to MATCH_LIMIT because
1964 that limits the total number of calls to match(), not all of which increase
1965 the depth of recursion. Limiting the recursion depth limits the amount of
1966 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1967 when PCRE is compiled, and changed at run time. A patch from Google adds
1968 this functionality to the C++ interface.
1970 18. Changes to the handling of Unicode character properties:
1972 (a) Updated the table to Unicode 4.1.0.
1974 (b) Recognize characters that are not in the table as "Cn" (undefined).
1976 (c) I revised the way the table is implemented to a much improved format
1977 which includes recognition of ranges. It now supports the ranges that
1978 are defined in UnicodeData.txt, and it also amalgamates other
1979 characters into ranges. This has reduced the number of entries in the
1980 table from around 16,000 to around 3,000, thus reducing its size
1981 considerably. I realized I did not need to use a tree structure after
1982 all - a binary chop search is just as efficient. Having reduced the
1983 number of entries, I extended their size from 6 bytes to 8 bytes to
1984 allow for more data.
1986 (d) Added support for Unicode script names via properties such as \p{Han}.
1988 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1989 matching that character.
1991 20. When matching a repeated Unicode property with a minimum greater than zero,
1992 (for example \pL{2,}), PCRE could look past the end of the subject if it
1993 reached it while seeking the minimum number of characters. This could
1994 happen only if some of the characters were more than one byte long, because
1995 there is a check for at least the minimum number of bytes.
1997 21. Refactored the implementation of \p and \P so as to be more general, to
1998 allow for more different types of property in future. This has changed the
1999 compiled form incompatibly. Anybody with saved compiled patterns that use
2000 \p or \P will have to recompile them.
2002 22. Added "Any" and "L&" to the supported property types.
2004 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2005 but give a compile time error if the value is greater than 0xff.
2007 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2008 accidentally not being installed or uninstalled.
2010 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2011 made were to insert the current release number. This seemed silly, because
2012 it made things harder for people building PCRE on systems that don't run
2013 "configure". I have turned pcre.h into a distributed file, no longer built
2014 by "configure", with the version identification directly included. There is
2015 no longer a pcre.h.in file.
2017 However, this change necessitated a change to the pcre-config script as
2018 well. It is built from pcre-config.in, and one of the substitutions was the
2019 release number. I have updated configure.ac so that ./configure now finds
2020 the release number by grepping pcre.h.
2022 26. Added the ability to run the tests under valgrind.
2025 Version 6.4 05-Sep-05
2026 ---------------------
2028 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2029 "--" to be printed when multiple files were scanned, even when none of the
2030 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2031 consider it to be a bug, and have restored the previous behaviour.
2033 2. A couple of code tidies to get rid of compiler warnings.
2035 3. The pcretest program used to cheat by referring to symbols in the library
2036 whose names begin with _pcre_. These are internal symbols that are not
2037 really supposed to be visible externally, and in some environments it is
2038 possible to suppress them. The cheating is now confined to including
2039 certain files from the library's source, which is a bit cleaner.
2041 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2042 file's purpose clearer.
2044 5. Reorganized pcre_ucp_findchar().
2047 Version 6.3 15-Aug-05
2048 ---------------------
2050 1. The file libpcre.pc.in did not have general read permission in the tarball.
2052 2. There were some problems when building without C++ support:
2054 (a) If C++ support was not built, "make install" and "make test" still
2055 tried to test it.
2057 (b) There were problems when the value of CXX was explicitly set. Some
2058 changes have been made to try to fix these, and ...
2060 (c) --disable-cpp can now be used to explicitly disable C++ support.
2062 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2063 backslash in a target when C++ was disabled. This confuses some
2064 versions of "make", apparently. Using an intermediate variable solves
2065 this. (Same for CPP_LOBJ.)
2067 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2068 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2069 necessary on certain architectures.
2071 4. Added a setting of -export-symbols-regex to the link command to remove
2072 those symbols that are exported in the C sense, but actually are local
2073 within the library, and not documented. Their names all begin with
2074 "_pcre_". This is not a perfect job, because (a) we have to except some
2075 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2076 available (and never for static libraries). I have made a note to try to
2077 find a way round (a) in the future.
2080 Version 6.2 01-Aug-05
2081 ---------------------
2083 1. There was no test for integer overflow of quantifier values. A construction
2084 such as {1111111111111111} would give undefined results. What is worse, if
2085 a minimum quantifier for a parenthesized subpattern overflowed and became
2086 negative, the calculation of the memory size went wrong. This could have
2087 led to memory overwriting.
2089 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2091 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2092 operating environments where this matters.
2094 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2095 PCRE options from within the C++ wrapper.
2097 5. Named capturing subpatterns were not being correctly counted when a pattern
2098 was compiled. This caused two problems: (a) If there were more than 100
2099 such subpatterns, the calculation of the memory needed for the whole
2100 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2101 back references of the form \12, where the number was greater than 9, were
2102 not recognized as back references, even though there were sufficient
2103 previous subpatterns.
2105 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2106 versions of gcc, e.g. 2.95.4.
2109 Version 6.1 21-Jun-05
2110 ---------------------
2112 1. There was one reference to the variable "posix" in pcretest.c that was not
2113 surrounded by "#if !defined NOPOSIX".
2115 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2116 the cross-check on the old pcre_info() function, for the benefit of the
2117 cut-down version of PCRE that is currently imported into Exim.
2119 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2120 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2121 patterns that start (?i) but I don't think that matters. The use of (?i) is
2122 just an example; this all applies to the other options as well.
2124 4. Since libtool seems to echo the compile commands it is issuing, the output
2125 from "make" can be reduced a bit by putting "@" in front of each libtool
2126 compile command.
2128 5. Patch from the folks at Google for configure.in to be a bit more thorough
2129 in checking for a suitable C++ installation before trying to compile the
2130 C++ stuff. This should fix a reported problem when a compiler was present,
2131 but no suitable headers.
2133 6. The man pages all had just "PCRE" as their title. I have changed them to
2134 be the relevant file name. I have also arranged that these names are
2135 retained in the file doc/pcre.txt, which is a concatenation in text format
2136 of all the man pages except the little individual ones for each function.
2138 7. The NON-UNIX-USE file had not been updated for the different set of source
2139 files that come with release 6. I also added a few comments about the C++
2140 wrapper.
2143 Version 6.0 07-Jun-05
2144 ---------------------
2146 1. Some minor internal re-organization to help with my DFA experiments.
2148 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2149 didn't matter for the library itself when fully configured, but did matter
2150 when compiling without UCP support, or within Exim, where the ucp files are
2151 not imported.
2153 3. Refactoring of the library code to split up the various functions into
2154 different source modules. The addition of the new DFA matching code (see
2155 below) to a single monolithic source would have made it really too
2156 unwieldy, quite apart from causing all the code to be include in a
2157 statically linked application, when only some functions are used. This is
2158 relevant even without the DFA addition now that patterns can be compiled in
2159 one application and matched in another.
2161 The downside of splitting up is that there have to be some external
2162 functions and data tables that are used internally in different modules of
2163 the library but which are not part of the API. These have all had their
2164 names changed to start with "_pcre_" so that they are unlikely to clash
2165 with other external names.
2167 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2168 a different (DFA) algorithm. Although it is slower than the original
2169 function, it does have some advantages for certain types of matching
2170 problem.
2172 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2173 including restarting after a partial match.
2175 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2176 defined when compiling for Windows was sent to me. I have put it into the
2177 code, though I have no means of testing or verifying it.
2179 7. Added the pcre_refcount() auxiliary function.
2181 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2182 match before or at the first newline in the subject string. In pcretest,
2183 the /f option on a pattern can be used to set this.
2185 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2186 would behave wrongly. This has been present in PCRE since release 4.0.
2188 10. A number of changes to the pcregrep command:
2190 (a) Refactored how -x works; insert ^(...)$ instead of setting
2191 PCRE_ANCHORED and checking the length, in preparation for adding
2192 something similar for -w.
2194 (b) Added the -w (match as a word) option.
2196 (c) Refactored the way lines are read and buffered so as to have more
2197 than one at a time available.
2199 (d) Implemented a pcregrep test script.
2201 (e) Added the -M (multiline match) option. This allows patterns to match
2202 over several lines of the subject. The buffering ensures that at least
2203 8K, or the rest of the document (whichever is the shorter) is available
2204 for matching (and similarly the previous 8K for lookbehind assertions).
2206 (f) Changed the --help output so that it now says
2208 -w, --word-regex(p)
2210 instead of two lines, one with "regex" and the other with "regexp"
2211 because that confused at least one person since the short forms are the
2212 same. (This required a bit of code, as the output is generated
2213 automatically from a table. It wasn't just a text change.)
2215 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2216 option but starts with a hyphen. Could be a pattern or a path name
2217 starting with a hyphen, for instance.
2219 (h) "-" can be given as a file name to represent stdin.
2221 (i) When file names are being printed, "(standard input)" is used for
2222 the standard input, for compatibility with GNU grep. Previously
2223 "<stdin>" was used.
2225 (j) The option --label=xxx can be used to supply a name to be used for
2226 stdin when file names are being printed. There is no short form.
2228 (k) Re-factored the options decoding logic because we are going to add
2229 two more options that take data. Such options can now be given in four
2230 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2232 (l) Added the -A, -B, and -C options for requesting that lines of context
2233 around matches be printed.
2235 (m) Added the -L option to print the names of files that do not contain
2236 any matching lines, that is, the complement of -l.
2238 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2239 continue to scan other files.
2241 (o) The -s option was incorrectly implemented. For compatibility with other
2242 greps, it now suppresses the error message for a non-existent or non-
2243 accessible file (but not the return code). There is a new option called
2244 -q that suppresses the output of matching lines, which was what -s was
2245 previously doing.
2247 (p) Added --include and --exclude options to specify files for inclusion
2248 and exclusion when recursing.
2250 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2251 Hopefully, it now does.
2253 12. Missing cast in pcre_study().
2255 13. Added an "uninstall" target to the makefile.
2257 14. Replaced "extern" in the function prototypes in Makefile.in with
2258 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2259 world, but is set differently for Windows.
2261 15. Added a second compiling function called pcre_compile2(). The only
2262 difference is that it has an extra argument, which is a pointer to an
2263 integer error code. When there is a compile-time failure, this is set
2264 non-zero, in addition to the error test pointer being set to point to an
2265 error message. The new argument may be NULL if no error number is required
2266 (but then you may as well call pcre_compile(), which is now just a
2267 wrapper). This facility is provided because some applications need a
2268 numeric error indication, but it has also enabled me to tidy up the way
2269 compile-time errors are handled in the POSIX wrapper.
2271 16. Added VPATH=.libs to the makefile; this should help when building with one
2272 prefix path and installing with another. (Or so I'm told by someone who
2273 knows more about this stuff than I do.)
2275 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2276 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2277 match everything, including newlines. This is not POSIX-compatible, but
2278 somebody wanted the feature. From pcretest it can be activated by using
2279 both the P and the s flags.
2281 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2283 19. libpcre.pc was being incorrectly installed as executable.
2285 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2286 it now also looks for '\r' so that it will work unmodified on Windows.
2288 21. Added Google's contributed C++ wrapper to the distribution.
2290 22. Added some untidy missing memory free() calls in pcretest, to keep
2291 Electric Fence happy when testing.
2295 Version 5.0 13-Sep-04
2296 ---------------------
2298 1. Internal change: literal characters are no longer packed up into items
2299 containing multiple characters in a single byte-string. Each character
2300 is now matched using a separate opcode. However, there may be more than one
2301 byte in the character in UTF-8 mode.
2303 2. The pcre_callout_block structure has two new fields: pattern_position and
2304 next_item_length. These contain the offset in the pattern to the next match
2305 item, and its length, respectively.
2307 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2308 insertion of callouts before each pattern item. Added the /C option to
2309 pcretest to make use of this.
2311 4. On the advice of a Windows user, the lines
2313 #if defined(_WIN32) || defined(WIN32)
2314 _setmode( _fileno( stdout ), 0x8000 );
2315 #endif /* defined(_WIN32) || defined(WIN32) */
2317 have been added to the source of pcretest. This apparently does useful
2318 magic in relation to line terminators.
2320 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2321 for the benefit of those environments where the "b" makes a difference.
2323 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2324 to know about it. I have put a hack into configure.in that adds in code
2325 to set GCC=yes if CC=icc. This seems to end up at a point in the
2326 generated configure script that is early enough to affect the setting of
2327 compiler options, which is what is needed, but I have no means of testing
2328 whether it really works. (The user who reported this had patched the
2329 generated configure script, which of course I cannot do.)
2331 LATER: After change 22 below (new libtool files), the configure script
2332 seems to know about icc (and also ecc). Therefore, I have commented out
2333 this hack in configure.in.
2335 7. Added support for pkg-config (2 patches were sent in).
2337 8. Negated POSIX character classes that used a combination of internal tables
2338 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2339 [[:^ascii]]. Typically, they would match almost any characters. The other
2340 POSIX classes were not broken in this way.
2342 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2343 to find the match, as PCRE was deluded into thinking that the match had to
2344 start at the start point or following a newline. The same bug applied to
2345 patterns with negative forward assertions or any backward assertions
2346 preceding ".*" at the start, unless the pattern required a fixed first
2347 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2349 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2350 starting at the last subject character, bytes beyond the end of the subject
2351 string were read.
2353 11. Renamed the variable "class" as "classbits" to make life easier for C++
2354 users. (Previously there was a macro definition, but it apparently wasn't
2355 enough.)
2357 12. Added the new field "tables" to the extra data so that tables can be passed
2358 in at exec time, or the internal tables can be re-selected. This allows
2359 a compiled regex to be saved and re-used at a later time by a different
2360 program that might have everything at different addresses.
2362 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2363 -R library as well as a -L library.
2365 14. The debugging options of pcretest (-d on the command line or D on a
2366 pattern) showed incorrect output for anything following an extended class
2367 that contained multibyte characters and which was followed by a quantifier.
2369 15. Added optional support for general category Unicode character properties
2370 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2371 support. It adds about 90K to the size of the library. The meanings of the
2372 inbuilt class escapes such as \d and \s have NOT been changed.
2374 16. Updated pcredemo.c to include calls to free() to release the memory for the
2375 compiled pattern.
2377 17. The generated file chartables.c was being created in the source directory
2378 instead of in the building directory. This caused the build to fail if the
2379 source directory was different from the building directory, and was
2380 read-only.
2382 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2383 file. No doubt somebody will tell me if they don't make sense... Also added
2384 Dan Mooney's comments about building on OpenVMS.
2386 19. Added support for partial matching via the PCRE_PARTIAL option for
2387 pcre_exec() and the \P data escape in pcretest.
2389 20. Extended pcretest with 3 new pattern features:
2391 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2392 write the compiled pattern to the file whose name is "rest-of-line".
2393 This is a straight binary dump of the data, with the saved pointer to
2394 the character tables forced to be NULL. The study data, if any, is
2395 written too. After writing, pcretest reads a new pattern.
2397 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2398 compiled pattern from the given file. There must not be any
2399 occurrences of "<" in the file name (pretty unlikely); if there are,
2400 pcretest will instead treat the initial "<" as a pattern delimiter.
2401 After reading in the pattern, pcretest goes on to read data lines as
2402 usual.
2404 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2405 and 16-bit fields in a compiled pattern, to simulate a pattern that
2406 was compiled on a host of opposite endianness.
2408 21. The pcre-exec() function can now cope with patterns that were compiled on
2409 hosts of opposite endianness, with this restriction:
2411 As for any compiled expression that is saved and used later, the tables
2412 pointer field cannot be preserved; the extra_data field in the arguments
2413 to pcre_exec() should be used to pass in a tables address if a value
2414 other than the default internal tables were used at compile time.
2416 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2417 now diagnosed as an error. Previously, most of the time, a negative number
2418 would have been treated as zero, but if in addition "ovector" was passed as
2419 NULL, a crash could occur.
2421 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2422 new versions from the libtool 1.5 distribution (the last one is a copy of
2423 a file called libtool.m4). This seems to have fixed the need to patch
2424 "configure" to support Darwin 1.3 (which I used to do). However, I still
2425 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2426 workstation).
2428 24. Changed the PCRE licence to be the more standard "BSD" licence.
2431 Version 4.5 01-Dec-03
2432 ---------------------
2434 1. There has been some re-arrangement of the code for the match() function so
2435 that it can be compiled in a version that does not call itself recursively.
2436 Instead, it keeps those local variables that need separate instances for
2437 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2438 needs to "recurse". Keeping track of where control must go is done by means
2439 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2440 hide most of the details from the main code, and operates only if
2441 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2442 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2443 operating.
2445 To make it easier for callers to provide specially tailored get/free
2446 functions for this usage, two new functions, pcre_stack_malloc, and
2447 pcre_stack_free, are used. They are always called in strict stacking order,
2448 and the size of block requested is always the same.
2450 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2451 PCRE has been compiled to use the stack or the heap for recursion. The
2452 -C option of pcretest uses this to show which version is compiled.
2454 A new data escape \S, is added to pcretest; it causes the amounts of store
2455 obtained and freed by both kinds of malloc/free at match time to be added
2456 to the output.
2458 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2459 what's available on my current Linux desktop machine.
2461 3. When matching a UTF-8 string, the test for a valid string at the start has
2462 been extended. If start_offset is not zero, PCRE now checks that it points
2463 to a byte that is the start of a UTF-8 character. If not, it returns
2464 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2465 this is necessary because there may be backward assertions in the pattern.
2466 When matching the same subject several times, it may save resources to use
2467 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2469 4. The code for checking the validity of UTF-8 strings has been tightened so
2470 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2471 containing "overlong sequences".
2473 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2474 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2475 should have been "&", but it just so happened that all the cases this let
2476 through by mistake were picked up later in the function.
2478 6. I had used a variable called "isblank" - this is a C99 function, causing
2479 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2481 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2482 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2483 so the version is shown; (c) stop "make test" echoing "./RunTest".
2485 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2487 9. The prototype for memmove() for systems that don't have it was using
2488 size_t, but the inclusion of the header that defines size_t was later. I've
2489 moved the #includes for the C headers earlier to avoid this.
2491 10. Added some adjustments to the code to make it easier to compiler on certain
2492 special systems:
2494 (a) Some "const" qualifiers were missing.
2495 (b) Added the macro EXPORT before all exported functions; by default this
2496 is defined to be empty.
2497 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2498 that it reads its output file name as an argument instead of writing
2499 to the standard output and assuming this can be redirected.
2501 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2502 class containing characters with values greater than 255, PCRE compilation
2503 went into a loop.
2505 12. A recursive reference to a subpattern that was within another subpattern
2506 that had a minimum quantifier of zero caused PCRE to crash. For example,
2507 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2508 recursion. If the recursively-called subpattern itself had a zero repeat,
2509 that was OK.
2511 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2512 buffer into which it was copied (for escape processing) was still set at
2513 1024, so long lines caused crashes.
2515 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2516 "internal error: code overflow...". This applied to any character class
2517 that was followed by a possessive quantifier.
2519 15. Modified the Makefile to add libpcre.la as a prerequisite for
2520 libpcreposix.la because I was told this is needed for a parallel build to
2521 work.
2523 16. If a pattern that contained .* following optional items at the start was
2524 studied, the wrong optimizing data was generated, leading to matching
2525 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2526 matching string must start with a or b or c. The correct conclusion for
2527 this pattern is that a match can start with any character.
2530 Version 4.4 13-Aug-03
2531 ---------------------
2533 1. In UTF-8 mode, a character class containing characters with values between
2534 127 and 255 was not handled correctly if the compiled pattern was studied.
2535 In fixing this, I have also improved the studying algorithm for such
2536 classes (slightly).
2538 2. Three internal functions had redundant arguments passed to them. Removal
2539 might give a very teeny performance improvement.
2541 3. Documentation bug: the value of the capture_top field in a callout is *one
2542 more than* the number of the hightest numbered captured substring.
2544 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2545 in incorrectly linking with a previously installed version. They now link
2546 explicitly with libpcre.la.
2548 5. configure.in no longer needs to recognize Cygwin specially.
2550 6. A problem in pcre.in for Windows platforms is fixed.
2552 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2553 pcretest, it used to include the size of the study block as part of its
2554 output. Unfortunately, the structure contains a field that has a different
2555 size on different hardware architectures. This meant that the tests that
2556 showed this size failed. As the block is currently always of a fixed size,
2557 this information isn't actually particularly useful in pcretest output, so
2558 I have just removed it.
2560 8. Three pre-processor statements accidentally did not start in column 1.
2561 Sadly, there are *still* compilers around that complain, even though
2562 standard C has not required this for well over a decade. Sigh.
2564 9. In pcretest, the code for checking callouts passed small integers in the
2565 callout_data field, which is a void * field. However, some picky compilers
2566 complained about the casts involved for this on 64-bit systems. Now
2567 pcretest passes the address of the small integer instead, which should get
2568 rid of the warnings.
2570 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2571 both compile and run time, and gives an error if an invalid UTF-8 sequence
2572 is found. There is a option for disabling this check in cases where the
2573 string is known to be correct and/or the maximum performance is wanted.
2575 11. In response to a bug report, I changed one line in Makefile.in from
2577 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2578 to
2579 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2581 to look similar to other lines, but I have no way of telling whether this
2582 is the right thing to do, as I do not use Windows. No doubt I'll get told
2583 if it's wrong...
2586 Version 4.3 21-May-03
2587 ---------------------
2589 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2590 Makefile.
2592 2. Some refactoring to improve the quality of the code:
2594 (i) The utf8_table... variables are now declared "const".
2596 (ii) The code for \cx, which used the "case flipping" table to upper case
2597 lower case letters, now just substracts 32. This is ASCII-specific,
2598 but the whole concept of \cx is ASCII-specific, so it seems
2599 reasonable.
2601 (iii) PCRE was using its character types table to recognize decimal and
2602 hexadecimal digits in the pattern. This is silly, because it handles
2603 only 0-9, a-f, and A-F, but the character types table is locale-
2604 specific, which means strange things might happen. A private
2605 table is now used for this - though it costs 256 bytes, a table is
2606 much faster than multiple explicit tests. Of course, the standard
2607 character types table is still used for matching digits in subject
2608 strings against \d.
2610 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2611 ending in _t are). So I've renamed it as ESC_tee.
2613 3. The first argument for regexec() in the POSIX wrapper should have been
2614 defined as "const".
2616 4. Changed pcretest to use malloc() for its buffers so that they can be
2617 Electric Fenced for debugging.
2619 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2620 to read one or more bytes before the start of the subject string. Often this
2621 had no effect on PCRE's behaviour, but in some circumstances it could
2622 provoke a segmentation fault.
2624 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2625 to try to read one or more bytes before the start of the subject string.
2627 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2628 UTF-8 support could misbehave in various ways if the subject string
2629 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2630 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2631 back over UTF-8 characters.)
2634 Version 4.2 14-Apr-03
2635 ---------------------
2637 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2639 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2640 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2641 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2642 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2643 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2645 Note: automatic setting of the BUILD variables is not yet working
2646 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2647 compile-time but not at link-time
2648 [LINK]: use for linking executables only
2649 make different versions for Windows and non-Windows
2650 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2651 libraries
2652 [LINK_FOR_BUILD]: new variable
2653 [OBJEXT]: use throughout
2654 [EXEEXT]: use throughout
2655 <winshared>: new target
2656 <wininstall>: new target
2657 <dftables.o>: use native compiler
2658 <dftables>: use native linker
2659 <install>: handle Windows platform correctly
2660 <clean>: ditto
2661 <check>: ditto
2662 copy DLL to top builddir before testing
2664 As part of these changes, -no-undefined was removed again. This was reported
2665 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2666 in any case.
2668 3. Some tidies to get rid of compiler warnings:
2670 . In the match_data structure, match_limit was an unsigned long int, whereas
2671 match_call_count was an int. I've made them both unsigned long ints.
2673 . In pcretest the fact that a const uschar * doesn't automatically cast to
2674 a void * provoked a warning.
2676 . Turning on some more compiler warnings threw up some "shadow" variables
2677 and a few more missing casts.
2679 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2680 option, a class that contained a single character with a value between 128
2681 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2683 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2684 option, a class that contained several characters, but with at least one
2685 whose value was between 128 and 255 caused PCRE to crash.
2688 Version 4.1 12-Mar-03
2689 ---------------------
2691 1. Compiling with gcc -pedantic found a couple of places where casts were
2692 needed, and a string in dftables.c that was longer than standard compilers are
2693 required to support.
2695 2. Compiling with Sun's compiler found a few more places where the code could
2696 be tidied up in order to avoid warnings.
2698 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2699 first of these names is deprecated in the latest Autoconf in favour of the name
2700 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2701 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2702 analogy I have changed it to CFLAGS_FOR_BUILD.
2704 4. Added -no-undefined to the linking command in the Makefile, because this is
2705 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2706 linking step for the pcreposix library.
2708 5. PCRE was failing to diagnose the case of two named groups with the same
2709 name.
2711 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2712 literal character that is needed in the subject for a match, and scans along to
2713 ensure that it is present before embarking on the full matching process. This
2714 saves time in cases of nested unlimited repeats that are never going to match.
2715 Problem: the scan can take a lot of time if the subject is very long (e.g.
2716 megabytes), thus penalizing straightforward matches. It is now done only if the
2717 amount of subject to be scanned is less than 1000 bytes.
2719 7. A lesser problem with the same optimization is that it was recording the
2720 first character of an anchored pattern as "needed", thus provoking a search
2721 right along the subject, even when the first match of the pattern was going to
2722 fail. The "needed" character is now not set for anchored patterns, unless it
2723 follows something in the pattern that is of non-fixed length. Thus, it still
2724 fulfils its original purpose of finding quick non-matches in cases of nested
2725 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2728 Version 4.0 17-Feb-03
2729 ---------------------
2731 1. If a comment in an extended regex that started immediately after a meta-item
2732 extended to the end of string, PCRE compiled incorrect data. This could lead to
2733 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2735 2. Moved to autoconf 2.53 and libtool 1.4.2.
2737 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2738 the special perltest8 script is no longer needed - all the tests can be run
2739 from a single perltest script.
2741 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2742 by \s. It has now been removed in PCRE. This means it isn't recognized as
2743 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2744 class [:space:] *does* include VT, thereby creating a mess.
2746 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2747 space and tab.
2749 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2750 its new features into the main test script, reducing the number of scripts.
2752 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2753 were backward compatible, and made the (?i) apply to the whole pattern, as if
2754 /i were given. Now it behaves more logically, and applies the option setting
2755 only to what follows. PCRE has been changed to follow suit. However, if it
2756 finds options settings right at the start of the pattern, it extracts them into
2757 the global options, as before. Thus, they show up in the info data.
2759 8. Added support for the \Q...\E escape sequence. Characters in between are
2760 treated as literals. This is slightly different from Perl in that $ and @ are
2761 also handled as literals inside the quotes. In Perl, they will cause variable
2762 interpolation. Note the following examples:
2764 Pattern PCRE matches Perl matches
2766 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2767 \Qabc\$xyz\E abc\$xyz abc\$xyz
2768 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2770 For compatibility with Perl, \Q...\E sequences are recognized inside character
2771 classes as well as outside them.
2773 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2774 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2775 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2776 signed/unsigned warnings.
2778 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2779 option for pcretest, so I've replaced it by a simple function that does just
2780 that job.
2782 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2783 "pcregrep -".
2785 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2786 Java package. This provides some syntactic sugar for simple cases of what my
2787 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2788 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2789 item, you can use this simplified notation. Note that only makes sense with
2790 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2791 greediness, whatever the setting of the PCRE_UNGREEDY option.
2793 13. A change of greediness default within a pattern was not taking effect at
2794 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2795 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2796 was abstracted outside.
2798 14. PCRE now supports the \G assertion. It is true when the current matching
2799 position is at the start point of the match. This differs from \A when the
2800 starting offset is non-zero. Used with the /g option of pcretest (or similar
2801 code), it works in the same way as it does for Perl's /g option. If all
2802 alternatives of a regex begin with \G, the expression is anchored to the start
2803 match position, and the "anchored" flag is set in the compiled expression.
2805 15. Some bugs concerning the handling of certain option changes within patterns
2806 have been fixed. These applied to options other than (?ims). For example,
2807 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2808 been the other way round. Some of this was related to change 7 above.
2810 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2811 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2812 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2813 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2815 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2816 mode. Unlike ".", it always matches newline, whatever the setting of
2817 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2818 assertions. Perl allows it, but it doesn't (in general) work because it can't
2819 calculate the length of the lookbehind. At least, that's the case for Perl
2820 5.8.0 - I've been told they are going to document that it doesn't work in
2821 future.
2823 18. Added an error diagnosis for escapes that PCRE does not support: these are
2824 \L, \l, \N, \P, \p, \U, \u, and \X.
2826 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2827 reading past the end of the pattern in cases such as /[abcd/.
2829 20. PCRE was getting more memory than necessary for patterns with classes that
2830 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2832 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2833 compiling PCRE for use with Virtual Pascal.
2835 22. Small fix to the Makefile to make it work properly if the build is done
2836 outside the source tree.
2838 23. Added a new extension: a condition to go with recursion. If a conditional
2839 subpattern starts with (?(R) the "true" branch is used if recursion has
2840 happened, whereas the "false" branch is used only at the top level.
2842 24. When there was a very long string of literal characters (over 255 bytes
2843 without UTF support, over 250 bytes with UTF support), the computation of how
2844 much memory was required could be incorrect, leading to segfaults or other
2845 strange effects.
2847 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2848 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2849 there was a subsequent back reference to those brackets. This meant that, for
2850 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2851 possible to check for precisely this case. All we can do is abandon the
2852 optimization if .* occurs inside capturing brackets when there are any back
2853 references whatsoever. (See below for a better fix that came later.)
2855 26. The handling of the optimization for finding the first character of a
2856 non-anchored pattern, and for finding a character that is required later in the
2857 match were failing in some cases. This didn't break the matching; it just
2858 failed to optimize when it could. The way this is done has been re-implemented.
2860 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2862 28. Added a new feature that provides some of the functionality that Perl
2863 provides with (?{...}). The facility is termed a "callout". The way it is done
2864 in PCRE is for the caller to provide an optional function, by setting
2865 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2866 global variable. By default it is unset, which disables all calling out. To get
2867 the function called, the regex must include (?C) at appropriate points. This
2868 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2869 This provides a means of identifying different callout points. When PCRE
2870 reaches such a point in the regex, if pcre_callout has been set, the external
2871 function is called. It is provided with data in a structure called
2872 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2873 matching continues; if it returns a non-zero value, the match at the current
2874 point fails. However, backtracking will occur if possible. [This was changed
2875 later and other features added - see item 49 below.]
2877 29. pcretest is upgraded to test the callout functionality. It provides a
2878 callout function that displays information. By default, it shows the start of
2879 the match and the current position in the text. There are some new data escapes
2880 to vary what happens:
2882 \C+ in addition, show current contents of captured substrings
2883 \C- do not supply a callout function
2884 \C!n return 1 when callout number n is reached
2885 \C!n!m return 1 when callout number n is reached for the mth time
2887 30. If pcregrep was called with the -l option and just a single file name, it
2888 output "<stdin>" if a match was found, instead of the file name.
2890 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2891 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2892 pcre_exec(). This saves a malloc/free per call. The default value of
2893 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2894 when configuring.
2896 32. The default maximum size of a compiled pattern is 64K. There have been a
2897 few cases of people hitting this limit. The code now uses macros to handle the
2898 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2899 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2900 configuring. Tests 2 and 5 work only with 2-byte links because they output
2901 debugging information about compiled patterns.
2903 33. Internal code re-arrangements:
2905 (a) Moved the debugging function for printing out a compiled regex into
2906 its own source file (printint.c) and used #include to pull it into
2907 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2908 separate copies.
2910 (b) Defined the list of op-code names for debugging as a macro in
2911 internal.h so that it is next to the definition of the opcodes.
2913 (c) Defined a table of op-code lengths for simpler skipping along compiled
2914 code. This is again a macro in internal.h so that it is next to the
2915 definition of the opcodes.
2917 34. Added support for recursive calls to individual subpatterns, along the
2918 lines of Robin Houston's patch (but implemented somewhat differently).
2920 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2921 allow it to read and process whole directories in Win32. This code was
2922 contributed by Lionel Fourquaux; it has not been tested by me.
2924 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2925 used to name a group. Names consist of alphanumerics and underscores, and must
2926 be unique. Back references use the syntax (?P=name) and recursive calls use
2927 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2928 numbers. The function pcre_fullinfo() can be used after compilation to extract
2929 a name/number map. There are three relevant calls:
2931 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2932 PCRE_INFO_NAMECOUNT yields the number of entries
2933 PCRE_INFO_NAMETABLE yields a pointer to the map.
2935 The map is a vector of fixed-size entries. The size of each entry depends on
2936 the length of the longest name used. The first two bytes of each entry are the
2937 group number, most significant byte first. There follows the corresponding
2938 name, zero terminated. The names are in alphabetical order.
2940 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2941 case instead of 255. Making it the same both with and without UTF-8 support
2942 means that the same test output works with both.
2944 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2945 calling malloc() with a zero argument.
2947 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2948 optimization. I've improved things by keeping a bitmap of backreferences with
2949 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2950 fact referenced, the optimization can be applied. It is unlikely that a
2951 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2952 the match to follow \n) will appear inside brackets with a number greater than
2953 31, but if it does, any back reference > 31 suppresses the optimization.
2955 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2956 of disabling numbered capturing parentheses. Any opening parenthesis that is
2957 not followed by ? behaves as if it were followed by ?: but named parentheses
2958 can still be used for capturing (and they will acquire numbers in the usual
2959 way).
2961 41. Redesigned the return codes from the match() function into yes/no/error so
2962 that errors can be passed back from deep inside the nested calls. A malloc
2963 failure while inside a recursive subpattern call now causes the
2964 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2966 42. It is now possible to set a limit on the number of times the match()
2967 function is called in a call to pcre_exec(). This facility makes it possible to
2968 limit the amount of recursion and backtracking, though not in a directly
2969 obvious way, because the match() function is used in a number of different
2970 circumstances. The count starts from zero for each position in the subject
2971 string (for non-anchored patterns). The default limit is, for compatibility, a
2972 large number, namely 10 000 000. You can change this in two ways:
2974 (a) When configuring PCRE before making, you can use --with-match-limit=n
2975 to set a default value for the compiled library.
2977 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2978 a different value is set. See 45 below.
2980 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2982 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2983 of things that can be changed at compile time. The first argument specifies
2984 what is wanted and the second points to where the information is to be placed.
2985 The current list of available information is:
2989 The output is an integer that is set to one if UTF-8 support is available;
2990 otherwise it is set to zero.
2994 The output is an integer that it set to the value of the code that is used for
2995 newline. It is either LF (10) or CR (13).
2999 The output is an integer that contains the number of bytes used for internal
3000 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3004 The output is an integer that contains the threshold above which the POSIX
3005 interface uses malloc() for output vectors. See item 31 above.
3009 The output is an unsigned integer that contains the default limit of the number
3010 of match() calls in a pcre_exec() execution. See 42 above.
3012 44. pcretest has been upgraded by the addition of the -C option. This causes it
3013 to extract all the available output from the new pcre_config() function, and to
3014 output it. The program then exits immediately.
3016 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3017 order to support additional features. One way would have been to define
3018 pcre_exec2() (for example) with extra arguments, but this would not have been
3019 extensible, and would also have required all calls to the original function to
3020 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3021 is used for passing in "extra" data from pcre_study().
3023 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3024 contains the following fields:
3026 flags a bitmap indicating which of the following fields are set
3027 study_data opaque data from pcre_study()
3028 match_limit a way of specifying a limit on match() calls for a specific
3029 call to pcre_exec()
3030 callout_data data for callouts (see 49 below)
3032 The flag bits are also defined in pcre.h, and are
3038 The pcre_study() function now returns one of these new pcre_extra blocks, with
3039 the actual study data pointed to by the study_data field, and the
3040 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3041 before. That is, this change is entirely upwards-compatible and requires no
3042 change to existing code.
3044 If you want to pass in additional data to pcre_exec(), you can either place it
3045 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3046 block.
3048 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3049 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3050 times with different match limits, until it finds the minimum value needed for
3051 pcre_exec() to complete. The value is then output. This can be instructive; for
3052 most simple matches the number is quite small, but for pathological cases it
3053 gets very large very quickly.
3055 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3056 returns the size of the data block pointed to by the study_data field in a
3057 pcre_extra block, that is, the value that was passed as the argument to
3058 pcre_malloc() when PCRE was getting memory in which to place the information
3059 created by pcre_study(). The fourth argument should point to a size_t variable.
3060 pcretest has been extended so that this information is shown after a successful
3061 pcre_study() call when information about the compiled regex is being displayed.
3063 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3064 because what follows is always an absolute path. (Later: it turns out that this
3065 is more than cosmetic for MinGW, because it doesn't like empty path
3066 components.)
3068 49. Some changes have been made to the callout feature (see 28 above):
3070 (i) A callout function now has three choices for what it returns:
3072 0 => success, carry on matching
3073 > 0 => failure at this point, but backtrack if possible
3074 < 0 => serious error, return this value from pcre_exec()
3076 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3077 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3078 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3079 use by callout functions. It will never be used by PCRE itself.
3081 (ii) The pcre_extra structure (see 45 above) has a void * field called
3082 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3083 pcre_callout_block structure has a field of the same name. The contents of
3084 the field passed in the pcre_extra structure are passed to the callout
3085 function in the corresponding field in the callout block. This makes it
3086 easier to use the same callout-containing regex from multiple threads. For
3087 testing, the pcretest program has a new data escape
3089 \C*n pass the number n (may be negative) as callout_data
3091 If the callout function in pcretest receives a non-zero value as
3092 callout_data, it returns that value.
3094 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3095 there were some redundant $(CFLAGS) in commands that are now specified as
3096 $(LINK), which already includes $(CFLAGS).
3098 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3099 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3100 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3101 one-byte characters throughout. Note that case-insensitive matching applies
3102 only to characters whose values are less than 256. PCRE doesn't support the
3103 notion of cases for higher-valued characters.
3105 (i) A character class whose characters are all within 0-255 is handled as
3106 a bit map, and the map is inverted for negative classes. Previously, a
3107 character > 255 always failed to match such a class; however it should
3108 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3110 (ii) A negated character class with a single character < 255 is coded as
3111 "not this character" (OP_NOT). This wasn't working properly when the test
3112 character was multibyte, either singly or repeated.
3114 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3115 mode, for example: \x{100}{2,3}.
3117 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3118 singly or repeated) now correctly test multibyte characters. However,
3119 PCRE doesn't recognize any characters with values greater than 255 as
3120 digits, spaces, or word characters. Such characters always match \D, \S,
3121 and \W, and never match \d, \s, or \w.
3123 (v) Classes may now contain characters and character ranges with values
3124 greater than 255. For example: [ab\x{100}-\x{400}].
3126 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3127 PCRE in UTF-8 mode.
3129 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3130 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3131 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3132 value.)
3134 53. The single man page has become too large. I have therefore split it up into
3135 a number of separate man pages. These also give rise to individual HTML pages;
3136 these are now put in a separate directory, and there is an index.html page that
3137 lists them all. Some hyperlinking between the pages has been installed.
3139 54. Added convenience functions for handling named capturing parentheses.
3141 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3142 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3143 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3144 are faulted.
3146 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3147 calling configure. These values are used when compiling the dftables.c program
3148 which is run to generate the source of the default character tables. They
3149 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3150 you will need to set these values.
3152 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3155 Version 3.9 02-Jan-02
3156 ---------------------
3158 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3160 2. If --disable-static was given, the building process failed when trying to
3161 build pcretest and pcregrep. (For some reason it was using libtool to compile
3162 them, which is not right, as they aren't part of the library.)
3165 Version 3.8 18-Dec-01
3166 ---------------------
3168 1. The experimental UTF-8 code was completely screwed up. It was packing the
3169 bytes in the wrong order. How dumb can you get?
3172 Version 3.7 29-Oct-01
3173 ---------------------
3175 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3176 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3177 this didn't happen under Solaris 8, where I normally test things.
3179 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3180 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3181 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3182 this fix an hour or so after the initial 3.7 release.)
3185 Version 3.6 23-Oct-01
3186 ---------------------
3188 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3189 offsets passed as NULL with zero offset count.
3191 2. The config.guess and config.sub files had not been updated when I moved to
3192 the latest autoconf.
3195 Version 3.5 15-Aug-01
3196 ---------------------
3198 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3199 had been forgotten.
3201 2. By using declared but undefined structures, we can avoid using "void"
3202 definitions in pcre.h while keeping the internal definitions of the structures
3203 private.
3205 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3206 user point of view, this means that both static and shared libraries are built
3207 by default, but this can be individually controlled. More of the work of
3208 handling this static/shared cases is now inside libtool instead of PCRE's make
3209 file.
3211 4. The pcretest utility is now installed along with pcregrep because it is
3212 useful for users (to test regexs) and by doing this, it automatically gets
3213 relinked by libtool. The documentation has been turned into a man page, so
3214 there are now .1, .txt, and .html versions in /doc.
3216 5. Upgrades to pcregrep:
3217 (i) Added long-form option names like gnu grep.
3218 (ii) Added --help to list all options with an explanatory phrase.
3219 (iii) Added -r, --recursive to recurse into sub-directories.
3220 (iv) Added -f, --file to read patterns from a file.
3222 6. pcre_exec() was referring to its "code" argument before testing that
3223 argument for NULL (and giving an error if it was NULL).
3225 7. Upgraded Makefile.in to allow for compiling in a different directory from
3226 the source directory.
3228 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3229 options bits, the pointer it was passed was to an int instead of to an unsigned
3230 long int. This mattered only on 64-bit systems.
3232 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3233 generated) instead of pcre.in, which it its source. Also made the same change
3234 in several of the .c files.
3236 10. A new release of gcc defines printf() as a macro, which broke pcretest
3237 because it had an ifdef in the middle of a string argument for printf(). Fixed
3238 by using separate calls to printf().
3240 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3241 script, to force use of CR or LF instead of \n in the source. On non-Unix
3242 systems, the value can be set in config.h.
3244 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3245 absolute limit. Changed the text of the error message to make this clear, and
3246 likewise updated the man page.
3248 13. The limit of 99 on the number of capturing subpatterns has been removed.
3249 The new limit is 65535, which I hope will not be a "real" limit.
3252 Version 3.4 22-Aug-00
3253 ---------------------
3255 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3257 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3260 Version 3.3 01-Aug-00
3261 ---------------------
3263 1. If an octal character was given, but the value was greater than \377, it
3264 was not getting masked to the least significant bits, as documented. This could
3265 lead to crashes in some systems.
3267 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3268 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3270 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3271 These just pass their arguments on to (pcre_free)(), but they are provided
3272 because some uses of PCRE bind it to non-C systems that can call its functions,
3273 but cannot call free() or pcre_free() directly.
3275 4. Add "make test" as a synonym for "make check". Corrected some comments in
3276 the Makefile.
3278 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3279 Makefile.
3281 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3282 command called pgrep for grepping around the active processes.
3284 7. Added the beginnings of support for UTF-8 character strings.
3286 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3287 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3288 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3289 out for the ar command.)
3292 Version 3.2 12-May-00
3293 ---------------------
3295 This is purely a bug fixing release.
3297 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3298 of ZA. This was just one example of several cases that could provoke this bug,
3299 which was introduced by change 9 of version 2.00. The code for breaking
3300 infinite loops after an iteration that matches an empty string was't working
3301 correctly.
3303 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3304 when matched against abbab (for example). After matching an empty string, it
3305 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3306 caused it to match further down the string than it should.
3308 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3309 was there because it doesn't seem to be needed, and it causes trouble on some
3310 systems, as it is not a Standard C header. It has been removed.
3312 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3313 were reported on the Macintosh. The changes were from
3315 while ((c = *(++ptr)) != 0 && c != '\n');
3316 to
3317 while ((c = *(++ptr)) != 0 && c != '\n') ;
3319 Totally extraordinary, but if that's what it takes...
3321 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3322 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3323 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3324 assumes the way PCRE uses memmove() (always moving upwards).
3326 6. PCRE is being used in one environment where strchr() is not available. There
3327 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3328 faster code anyway.
3331 Version 3.1 09-Feb-00
3332 ---------------------
3334 The only change in this release is the fixing of some bugs in Makefile.in for
3335 the "install" target:
3337 (1) It was failing to install pcreposix.h.
3339 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3342 Version 3.0 01-Feb-00
3343 ---------------------
3345 1. Add support for the /+ modifier to perltest (to output $` like it does in
3346 pcretest).
3348 2. Add support for the /g modifier to perltest.
3350 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3351 matches null strings.
3353 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3354 pattern. Perl treats empty patterns specially - it reuses the most recent
3355 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3356 effect.
3358 5. The POSIX interface was broken in that it was just handing over the POSIX
3359 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3360 required a bigger vector, with some working space on the end. This means that
3361 the POSIX wrapper now has to get and free some memory, and copy the results.
3363 6. Added some simple autoconf support, placing the test data and the
3364 documentation in separate directories, re-organizing some of the
3365 information files, and making it build pcre-config (a GNU standard). Also added
3366 libtool support for building PCRE as a shared library, which is now the
3367 default.
3369 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3370 09 are not valid octal constants. Single digits will be used for minor values
3371 less than 10.
3373 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3374 existing programs that set these in the POSIX interface can use PCRE without
3375 modification.
3377 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3378 return all that pcre_info() returns, plus additional data. The pcre_info()
3379 function is retained for compatibility, but is considered to be obsolete.
3381 10. Added experimental recursion feature (?R) to handle one common case that
3382 Perl 5.6 will be able to do with (?p{...}).
3384 11. Added support for POSIX character classes like [:alpha:], which Perl is
3385 adopting.
3388 Version 2.08 31-Aug-99
3389 ----------------------
3391 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3392 trying to match at the startoffset position, but instead was moving forward to
3393 the next newline as if a previous match had failed.
3395 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3396 and could get into a loop if a null string was matched other than at the start
3397 of the subject.
3399 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3400 be distinguished at compile time, and for completeness also added PCRE_DATE.
3402 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3403 in GnuWin32 environments.
3406 Version 2.07 29-Jul-99
3407 ----------------------
3409 1. The documentation is now supplied in plain text form and HTML as well as in
3410 the form of man page sources.
3412 2. C++ compilers don't like assigning (void *) values to other pointer types.
3413 In particular this affects malloc(). Although there is no problem in Standard
3414 C, I've put in casts to keep C++ compilers happy.
3416 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3417 should be (const char *).
3419 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3420 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3421 However, I haven't made this a standard facility. The documentation doesn't
3422 mention it, and the Makefile doesn't support it.
3424 5. The Makefile now contains an "install" target, with editable destinations at
3425 the top of the file. The pcretest program is not installed.
3427 6. pgrep -V now gives the PCRE version number and date.
3429 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3430 causing the entire string to be ignored, instead of just the last character.
3432 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3433 non-matching string, it can take a very, very long time, even for strings of
3434 quite modest length, because of the nested recursion. PCRE now does better in
3435 some of these cases. It does this by remembering the last required literal
3436 character in the pattern, and pre-searching the subject to ensure it is present
3437 before running the real match. In other words, it applies a heuristic to detect
3438 some types of certain failure quickly, and in the above example, if presented
3439 with a string that has no trailing " it gives "no match" very quickly.
3441 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3442 other alternatives are tried instead.
3445 Version 2.06 09-Jun-99
3446 ----------------------
3448 1. Change pcretest's output for amount of store used to show just the code
3449 space, because the remainder (the data block) varies in size between 32-bit and
3450 64-bit systems.
3452 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3453 start matching at. This allows lookbehinds to work when searching for multiple
3454 occurrences in a string.
3456 3. Added additional options to pcretest for testing multiple occurrences:
3458 /+ outputs the rest of the string that follows a match
3459 /g loops for multiple occurrences, using the new startoffset argument
3460 /G loops for multiple occurrences by passing an incremented pointer
3462 4. PCRE wasn't doing the "first character" optimization for patterns starting
3463 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3464 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3465 the letter 'x'. On long subject strings, this gives a significant speed-up.
3468 Version 2.05 21-Apr-99
3469 ----------------------
3471 1. Changed the type of magic_number from int to long int so that it works
3472 properly on 16-bit systems.
3474 2. Fixed a bug which caused patterns starting with .* not to work correctly
3475 when the subject string contained newline characters. PCRE was assuming
3476 anchoring for such patterns in all cases, which is not correct because .* will
3477 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3478 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3479 must be retried after every newline in the subject.
3482 Version 2.04 18-Feb-99
3483 ----------------------
3485 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3486 computation of the store needed to hold the pattern was incorrect (too large).
3487 If such patterns were nested a few deep, this could multiply and become a real
3488 problem.
3490 2. Added /M option to pcretest to show the memory requirement of a specific
3491 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3493 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3494 compiled in such a way that the backtracking after subsequent failure was
3495 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3496 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3499 Version 2.03 02-Feb-99
3500 ----------------------
3502 1. Fixed typo and small mistake in man page.
3504 2. Added 4th condition (GPL supersedes if conflict) and created separate
3505 LICENCE file containing the conditions.
3507 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3508 Perl, that is the internal \ allows the delimiter to be included in the
3509 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3510 the final delimiter, add \ to the end of the pattern (to test the error).
3512 4. Added the convenience functions for extracting substrings after a successful
3513 match. Updated pcretest to make it able to test these functions.
3516 Version 2.02 14-Jan-99
3517 ----------------------
3519 1. Initialized the working variables associated with each extraction so that
3520 their saving and restoring doesn't refer to uninitialized store.
3522 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3523 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3524 fix the problem.
3526 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3527 calls, and wasn't printing the correct value for compiling calls. Increased the
3528 default value of LOOPREPEAT, and the number of significant figures in the
3529 times.
3531 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3533 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3534 a building problem on Windows NT with a FAT file system.
3537 Version 2.01 21-Oct-98
3538 ----------------------
3540 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3541 to character tables built by pcre_maketables() in the current locale. If NULL
3542 is passed, the default tables are used.
3545 Version 2.00 24-Sep-98
3546 ----------------------
3548 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3549 it any more.
3551 2. Allow quantification of (?>) groups, and make it work correctly.
3553 3. The first character computation wasn't working for (?>) groups.
3555 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3556 end of the subject) and add 5.005's \z, which really does match only at the
3557 very end of the subject.
3559 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3561 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3562 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3563 localized options. All options to pcre_study() were also removed.
3565 7. Add other new features from 5.005:
3567 $(?<= positive lookbehind
3568 $(?<! negative lookbehind
3569 (?imsx-imsx) added the unsetting capability
3570 such a setting is global if at outer level; local otherwise
3571 (?imsx-imsx:) non-capturing groups with option setting
3572 (?(cond)re|re) conditional pattern matching
3574 A backreference to itself in a repeated group matches the previous
3575 captured string.
3577 8. General tidying up of studying (both automatic and via "study")
3578 consequential on the addition of new assertions.
3580 9. As in 5.005, unlimited repeated groups that could match an empty substring
3581 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3582 runtime if any iteration does actually match an empty substring.
3584 10. Include the RunTest script in the distribution.
3586 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3587 discrepancies, some of which were old and were also with respect to 5.004. They
3588 have now been fixed.
3591 Version 1.09 28-Apr-98
3592 ----------------------
3594 1. A negated single character class followed by a quantifier with a minimum
3595 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3596 program crashes, or just wrong answers. This did not apply to negated classes
3597 containing more than one character, or to minima other than one.
3600 Version 1.08 27-Mar-98
3601 ----------------------
3603 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3605 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3606 latter must appear before anything that relies on it in the pattern.
3609 Version 1.07 16-Feb-98
3610 ----------------------
3612 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3613 repeat of a potentially empty string).
3616 Version 1.06 23-Jan-98
3617 ----------------------
3619 1. Added Markus Oberhumer's little patches for C++.
3621 2. Literal strings longer than 255 characters were broken.
3624 Version 1.05 23-Dec-97
3625 ----------------------
3627 1. Negated character classes containing more than one character were failing if
3628 PCRE_CASELESS was set at run time.
3631 Version 1.04 19-Dec-97
3632 ----------------------
3634 1. Corrected the man page, where some "const" qualifiers had been omitted.
3636 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3637 input syntax.
3639 3. Fixed memory leak which occurred when a regex with back references was
3640 matched with an offsets vector that wasn't big enough. The temporary memory
3641 that is used in this case wasn't being freed if the match failed.
3643 4. Tidied pcretest to ensure it frees memory that it gets.
3645 5. Temporary memory was being obtained in the case where the passed offsets
3646 vector was exactly big enough.
3648 6. Corrected definition of offsetof() from change 5 below.
3650 7. I had screwed up change 6 below and broken the rules for the use of
3651 setjmp(). Now fixed.
3654 Version 1.03 18-Dec-97
3655 ----------------------
3657 1. A erroneous regex with a missing opening parenthesis was correctly
3658 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3659 on some systems.
3661 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3662 it was reported that one broken compiler failed on the former because "code" is
3663 also an independent variable.
3665 3. The erroneous regex a[]b caused an array overrun reference.
3667 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3668 fail on data ending with that character. (It was going on too far, and checking
3669 the next character, typically a binary zero.) This was specific to the
3670 optimized code for single-character negative classes.
3672 5. Added a contributed patch from the TIN world which does the following:
3674 + Add an undef for memmove, in case the the system defines a macro for it.
3676 + Add a definition of offsetof(), in case there isn't one. (I don't know
3677 the reason behind this - offsetof() is part of the ANSI standard - but
3678 it does no harm).
3680 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3681 most of the places where whitespace preceded '#'. I have given up and
3682 allowed the remaining 2 cases to be at the margin.
3684 + Rename some variables in pcre to eliminate shadowing. This seems very
3685 pedantic, but does no harm, of course.
3687 6. Moved the call to setjmp() into its own function, to get rid of warnings
3688 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3690 7. Constructs such as \d{8,} were compiling into the equivalent of
3691 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3692 outcome, but in this particular case used more store than had been allocated,
3693 which caused the bug to be discovered because it threw up an internal error.
3695 8. The debugging code in both pcre and pcretest for outputting the compiled
3696 form of a regex was going wrong in the case of back references followed by
3697 curly-bracketed repeats.
3700 Version 1.02 12-Dec-97
3701 ----------------------
3703 1. Typos in pcre.3 and comments in the source fixed.
3705 2. Applied a contributed patch to get rid of places where it used to remove
3706 'const' from variables, and fixed some signed/unsigned and uninitialized
3707 variable warnings.
3709 3. Added the "runtest" target to Makefile.
3711 4. Set default compiler flag to -O2 rather than just -O.
3714 Version 1.01 19-Nov-97
3715 ----------------------
3717 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3718 like /([ab]*)*/, that is, for classes with more than one character in them.
3720 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3721 as /((?>a*))*/ (a PCRE_EXTRA facility).
3724 Version 1.00 18-Nov-97
3725 ----------------------
3727 1. Added compile-time macros to support systems such as SunOS4 which don't have
3728 memmove() or strerror() but have other things that can be used instead.
3730 2. Arranged that "make clean" removes the executables.
3733 Version 0.99 27-Oct-97
3734 ----------------------
3736 1. Fixed bug in code for optimizing classes with only one character. It was
3737 initializing a 32-byte map regardless, which could cause it to run off the end
3738 of the memory it had got.
3740 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3743 Version 0.98 22-Oct-97
3744 ----------------------
3746 1. Fixed bug in code for handling temporary memory usage when there are more
3747 back references than supplied space in the ovector. This could cause segfaults.
3750 Version 0.97 21-Oct-97
3751 ----------------------
3753 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3755 2. Optimized negated single characters not to use a bit map.
3757 3. Brought error texts together as macro definitions; clarified some of them;
3758 fixed one that was wrong - it said "range out of order" when it meant "invalid
3759 escape sequence".
3761 4. Changed some char * arguments to const char *.
3763 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3765 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3766 pcretest.
3769 Version 0.96 16-Oct-97
3770 ----------------------
3772 1. Added a simple "pgrep" utility to the distribution.
3774 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3775 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3776 where "ddd" means "one or more decimal digits".
3778 3. Fixed serious bug. If a pattern had a back reference, but the call to
3779 pcre_exec() didn't supply a large enough ovector to record the related
3780 identifying subpattern, the match always failed. PCRE now remembers the number
3781 of the largest back reference, and gets some temporary memory in which to save
3782 the offsets during matching if necessary, in order to ensure that
3783 backreferences always work.
3785 4. Increased the compatibility with Perl in a number of ways:
3787 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3788 to request this handling. The option can be set at compile or exec time.
3790 (b) $ matches before a terminating newline by default; an option
3791 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3792 mode). The option can be set at compile or exec time.
3794 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3795 the same as Perl's. If the decimal number it represents is less than 10
3796 or there aren't that many previous left capturing parentheses, an octal
3797 escape is read. Inside a character class, it's always an octal escape,
3798 even if it is a single digit.
3800 (d) An escaped but undefined alphabetic character is taken as a literal,
3801 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3802 escapes.
3804 (e) {0} is now permitted. (The previous item is removed from the compiled
3805 pattern).
3807 5. Changed all the names of code files so that the basic parts are no longer
3808 than 10 characters, and abolished the teeny "globals.c" file.
3810 6. Changed the handling of character classes; they are now done with a 32-byte
3811 bit map always.
3813 7. Added the -d and /D options to pcretest to make it possible to look at the
3814 internals of compilation without having to recompile pcre.
3817 Version 0.95 23-Sep-97
3818 ----------------------
3820 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3821 \x20 at the start of a run of normal characters. These were being treated as
3822 real characters, instead of the source characters being re-checked.
3825 Version 0.94 18-Sep-97
3826 ----------------------
3828 1. The functions are now thread-safe, with the caveat that the global variables
3829 containing pointers to malloc() and free() or alternative functions are the
3830 same for all threads.
3832 2. Get pcre_study() to generate a bitmap of initial characters for non-
3833 anchored patterns when this is possible, and use it if passed to pcre_exec().
3836 Version 0.93 15-Sep-97
3837 ----------------------
3839 1. /(b)|(:+)/ was computing an incorrect first character.
3841 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3842 but not actually doing anything yet.
3844 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3845 as Perl does (e.g. [-az] or [az-]).
3847 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3848 all possible positions.
3850 5. Split up into different modules to avoid including unneeded functions in a
3851 compiled binary. However, compile and exec are still in one module. The "study"
3852 function is split off.
3854 6. The character tables are now in a separate module whose source is generated
3855 by an auxiliary program - but can then be edited by hand if required. There are
3856 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3857 toupper() in the code.
3859 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3860 make them global. Abolish the function for setting them, as the caller can now
3861 set them directly.
3864 Version 0.92 11-Sep-97
3865 ----------------------
3867 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3868 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3870 2. Caseless matching was not working in character classes if the characters in
3871 the pattern were in upper case.
3873 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3875 4. Make PCRE_ANCHORED public and accept as a compile option.
3877 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3878 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3879 pass them.
3881 6. Give an error if bad option bits passed at compile or run time.
3883 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3884 pcretest to cause it to pass that flag.
3886 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3887 options, and the first character, if set.
3889 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3892 Version 0.91 10-Sep-97
3893 ----------------------
3895 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3896 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3898 2. PCRE was looping on encountering an indefinitely repeated back reference to
3899 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3900 Perl does - treats the match as successful.
3902 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5