ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 513 - (show annotations)
Mon May 3 11:13:37 2010 UTC (11 years, 2 months ago) by ph10
File size: 185361 byte(s)
Make \R and \X in a character class behave more like Perl
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.
21 Version 8.02 19-Mar-2010
22 ------------------------
24 1. The Unicode data tables have been updated to Unicode 5.2.0.
26 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
27 configured.
29 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
30 original author of that file, following a query about its status.
32 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
33 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
35 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
36 quantifier applied to a forward-referencing subroutine call, could compile
37 incorrect code or give the error "internal error: previously-checked
38 referenced subpattern not found".
40 6. Both MS Visual Studio and Symbian OS have problems with initializing
41 variables to point to external functions. For these systems, therefore,
42 pcre_malloc etc. are now initialized to local functions that call the
43 relevant global functions.
45 7. There were two entries missing in the vectors called coptable and poptable
46 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
47 I've fixed the data, and added a kludgy way of testing at compile time that
48 the lengths are correct (equal to the number of opcodes).
50 8. Following on from 7, I added a similar kludge to check the length of the
51 eint vector in pcreposix.c.
53 9. Error texts for pcre_compile() are held as one long string to avoid too
54 much relocation at load time. To find a text, the string is searched,
55 counting zeros. There was no check for running off the end of the string,
56 which could happen if a new error number was added without updating the
57 string.
59 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
61 11. \K was not working if it appeared in an atomic group or in a group that
62 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
63 \K is "not well defined" if used in an assertion. PCRE now accepts it if
64 the assertion is positive, but not if it is negative.
66 12. Change 11 fortuitously reduced the size of the stack frame used in the
67 "match()" function of pcre_exec.c by one pointer. Forthcoming
68 implementation of support for (*MARK) will need an extra pointer on the
69 stack; I have reserved it now, so that the stack frame size does not
70 decrease.
72 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
73 item in branch that calls a recursion is a subroutine call - as in the
74 second branch in the above example - was incorrectly given the compile-
75 time error "recursive call could loop indefinitely" because pcre_compile()
76 was not correctly checking the subroutine for matching a non-empty string.
78 14. The checks for overrunning compiling workspace could trigger after an
79 overrun had occurred. This is a "should never occur" error, but it can be
80 triggered by pathological patterns such as hundreds of nested parentheses.
81 The checks now trigger 100 bytes before the end of the workspace.
83 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
86 Version 8.01 19-Jan-2010
87 ------------------------
89 1. If a pattern contained a conditional subpattern with only one branch (in
90 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
91 computed the wrong minimum data length (which is of course zero for such
92 subpatterns). This could cause incorrect "no match" results.
94 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
95 the pattern is reset in the first branch, pcre_compile() failed with
96 "internal error: code overflow at offset...". This happened only when
97 the reset was to the original external option setting. (An optimization
98 abstracts leading options settings into an external setting, which was the
99 cause of this.)
101 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
102 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
103 assertion pattern did not match (meaning that the assertion was true), it
104 was incorrectly treated as false if the SKIP had been reached during the
105 matching. This also applied to assertions used as conditions.
107 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
108 assertion subpattern, including such a pattern used as a condition,
109 unpredictable results occurred, instead of the error return
112 5. The C++ GlobalReplace function was not working like Perl for the special
113 situation when an empty string is matched. It now does the fancy magic
114 stuff that is necessary.
116 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
117 removed. (These were left over from very, very early versions of PCRE.)
119 7. Some cosmetic changes to the code to make life easier when compiling it
120 as part of something else:
122 (a) Change DEBUG to PCRE_DEBUG.
124 (b) In pcre_compile(), rename the member of the "branch_chain" structure
125 called "current" as "current_branch", to prevent a collision with the
126 Linux macro when compiled as a kernel module.
128 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
129 prevent a collision with the Linux macro when compiled as a kernel
130 module.
132 8. In pcre_compile() there are some checks for integer overflows that used to
133 cast potentially large values to (double). This has been changed to that
134 when building, a check for int64_t is made, and if it is found, it is used
135 instead, thus avoiding the use of floating point arithmetic. (There is no
136 other use of FP in PCRE.) If int64_t is not found, the fallback is to
137 double.
139 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
140 2005 (difference between two addresses compared to an unsigned value).
142 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
143 custom one, because of the following reported problem in Windows:
145 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
146 under Win32.
147 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
148 therefore missing the function definition.
149 - The compiler thus generates a "C" signature for the test function.
150 - The linker fails to find the "C" function.
151 - PCRE fails to configure if asked to do so against libbz2.
153 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
154 messages were output:
156 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
157 rerunning libtoolize, to keep the correct libtool macros in-tree.
158 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
160 I have done both of these things.
162 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
163 most of the time, it *can* run out if it is given a pattern that contains a
164 runaway infinite recursion. I updated the discussion in the pcrestack man
165 page.
167 13. Now that we have gone to the x.xx style of version numbers, the minor
168 version may start with zero. Using 08 or 09 is a bad idea because users
169 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
170 interpreted as invalid octal numbers. I've updated the previous comment in
171 configure.ac, and also added a check that gives an error if 08 or 09 are
172 used.
174 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
175 causing partial matching to fail when the end of the subject matched \W
176 in a UTF-8 pattern where \W was quantified with a minimum of 3.
178 15. There were some discrepancies between the declarations in pcre_internal.h
179 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
180 their definitions. The declarations used "const uschar *" and the
181 definitions used USPTR. Even though USPTR is normally defined as "const
182 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
183 reported that: "This difference in casting confuses some C++ compilers, for
184 example, SunCC recognizes above declarations as different functions and
185 generates broken code for hbpcre." I have changed the declarations to use
186 USPTR.
188 16. GNU libtool is named differently on some systems. The autogen.sh script now
189 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
190 (FreeBSD).
192 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
193 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
194 comment: "Figure out how to create a longlong from a string: strtoll and
195 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
196 instance, but it only takes 2 args instead of 3!"
198 18. A subtle bug concerned with back references has been fixed by a change of
199 specification, with a corresponding code fix. A pattern such as
200 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
201 refers, was giving matches when it shouldn't. For example, xa=xaaa would
202 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
203 same bug. Such groups have to be quantified to be useful, or contained
204 inside another quantified group. (If there's no repetition, the reference
205 can never match.) The problem arises because, having left the group and
206 moved on to the rest of the pattern, a later failure that backtracks into
207 the group uses the captured value from the final iteration of the group
208 rather than the correct earlier one. I have fixed this in PCRE by forcing
209 any group that contains a reference to itself to be an atomic group; that
210 is, there cannot be any backtracking into it once it has completed. This is
211 similar to recursive and subroutine calls.
214 Version 8.00 19-Oct-09
215 ----------------------
217 1. The table for translating pcre_compile() error codes into POSIX error codes
218 was out-of-date, and there was no check on the pcre_compile() error code
219 being within the table. This could lead to an OK return being given in
220 error.
222 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
223 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
224 in a Windows environment.
226 3. The pcregrep --count option prints the count for each file even when it is
227 zero, as does GNU grep. However, pcregrep was also printing all files when
228 --files-with-matches was added. Now, when both options are given, it prints
229 counts only for those files that have at least one match. (GNU grep just
230 prints the file name in this circumstance, but including the count seems
231 more useful - otherwise, why use --count?) Also ensured that the
232 combination -clh just lists non-zero counts, with no names.
234 4. The long form of the pcregrep -F option was incorrectly implemented as
235 --fixed_strings instead of --fixed-strings. This is an incompatible change,
236 but it seems right to fix it, and I didn't think it was worth preserving
237 the old behaviour.
239 5. The command line items --regex=pattern and --regexp=pattern were not
240 recognized by pcregrep, which required --regex pattern or --regexp pattern
241 (with a space rather than an '='). The man page documented the '=' forms,
242 which are compatible with GNU grep; these now work.
244 6. No libpcreposix.pc file was created for pkg-config; there was just
245 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
247 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
248 when UCP support is not needed, by modifying the Python script that
249 generates it from Unicode data files. This should not matter if the module
250 is correctly used as a library, but I received one complaint about 50K of
251 unwanted data. My guess is that the person linked everything into his
252 program rather than using a library. Anyway, it does no harm.
254 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
255 was a minimum greater than 1 for a wide character in a possessive
256 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
257 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
258 character. Chaos in the form of incorrect output or a compiling loop could
259 result.
261 9. The restrictions on what a pattern can contain when partial matching is
262 requested for pcre_exec() have been removed. All patterns can now be
263 partially matched by this function. In addition, if there are at least two
264 slots in the offset vector, the offset of the earliest inspected character
265 for the match and the offset of the end of the subject are set in them when
266 PCRE_ERROR_PARTIAL is returned.
268 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
269 synonymous with PCRE_PARTIAL, for backwards compatibility, and
270 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
271 and may be more useful for multi-segment matching.
273 11. Partial matching with pcre_exec() is now more intuitive. A partial match
274 used to be given if ever the end of the subject was reached; now it is
275 given only if matching could not proceed because another character was
276 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
277 string "Z", which now yields "no match" instead of "partial match". In the
278 case of pcre_dfa_exec(), "no match" is given if every matching path for the
279 final character ended with (*FAIL).
281 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
282 if the pattern had a "must contain" character that was already found in the
283 earlier partial match, unless partial matching was again requested. For
284 example, with the pattern /dog.(body)?/, the "must contain" character is
285 "g". If the first part-match was for the string "dog", restarting with
286 "sbody" failed. This bug has been fixed.
288 13. The string returned by pcre_dfa_exec() after a partial match has been
289 changed so that it starts at the first inspected character rather than the
290 first character of the match. This makes a difference only if the pattern
291 starts with a lookbehind assertion or \b or \B (\K is not supported by
292 pcre_dfa_exec()). It's an incompatible change, but it makes the two
293 matching functions compatible, and I think it's the right thing to do.
295 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
296 so that the demonstration program is easily available in environments where
297 PCRE has not been installed from source.
299 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
300 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
301 library.
303 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
304 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
305 is not the first non-POSIX option to be added. Clearly some people find
306 these options useful.
308 17. If a caller to the POSIX matching function regexec() passes a non-zero
309 value for nmatch with a NULL value for pmatch, the value of
310 nmatch is forced to zero.
312 18. RunGrepTest did not have a test for the availability of the -u option of
313 the diff command, as RunTest does. It now checks in the same way as
314 RunTest, and also checks for the -b option.
316 19. If an odd number of negated classes containing just a single character
317 interposed, within parentheses, between a forward reference to a named
318 subpattern and the definition of the subpattern, compilation crashed with
319 an internal error, complaining that it could not find the referenced
320 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
321 [The bug was that it was starting one character too far in when skipping
322 over the character class, thus treating the ] as data rather than
323 terminating the class. This meant it could skip too much.]
325 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
326 /g option in pcretest when the pattern contains \K, which makes it possible
327 to have an empty string match not at the start, even when the pattern is
328 anchored. Updated pcretest and pcredemo to use this option.
330 21. If the maximum number of capturing subpatterns in a recursion was greater
331 than the maximum at the outer level, the higher number was returned, but
332 with unset values at the outer level. The correct (outer level) value is
333 now given.
335 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
336 PCRE did not set those parentheses (unlike Perl). I have now found a way to
337 make it do so. The string so far is captured, making this feature
338 compatible with Perl.
340 23. The tests have been re-organized, adding tests 11 and 12, to make it
341 possible to check the Perl 5.10 features against Perl 5.10.
343 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
344 pattern matches a fixed length string. PCRE did not allow this; now it
345 does. Neither allows recursion.
347 25. I finally figured out how to implement a request to provide the minimum
348 length of subject string that was needed in order to match a given pattern.
349 (It was back references and recursion that I had previously got hung up
350 on.) This code has now been added to pcre_study(); it finds a lower bound
351 to the length of subject needed. It is not necessarily the greatest lower
352 bound, but using it to avoid searching strings that are too short does give
353 some useful speed-ups. The value is available to calling programs via
354 pcre_fullinfo().
356 26. While implementing 25, I discovered to my embarrassment that pcretest had
357 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
358 study optimizations had never been tested with that matching function.
359 Oops. What is worse, even when it was passed study data, there was a bug in
360 pcre_dfa_exec() that meant it never actually used it. Double oops. There
361 were also very few tests of studied patterns with pcre_dfa_exec().
363 27. If (?| is used to create subpatterns with duplicate numbers, they are now
364 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
365 on the other side of the coin, they are no longer allowed to have different
366 names, because these cannot be distinguished in PCRE, and this has caused
367 confusion. (This is a difference from Perl.)
369 28. When duplicate subpattern names are present (necessarily with different
370 numbers, as required by 27 above), and a test is made by name in a
371 conditional pattern, either for a subpattern having been matched, or for
372 recursion in such a pattern, all the associated numbered subpatterns are
373 tested, and the overall condition is true if the condition is true for any
374 one of them. This is the way Perl works, and is also more like the way
375 testing by number works.
378 Version 7.9 11-Apr-09
379 ---------------------
381 1. When building with support for bzlib/zlib (pcregrep) and/or readline
382 (pcretest), all targets were linked against these libraries. This included
383 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
384 libraries. This caused unwanted dependencies to be created. This problem
385 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
386 pcretest is linked with readline.
388 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
389 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
390 moved outside it again, because FALSE and TRUE are already defined in AIX,
391 but BOOL is not.
393 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
394 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
396 4. The pcregrep documentation said spaces were inserted as well as colons (or
397 hyphens) following file names and line numbers when outputting matching
398 lines. This is not true; no spaces are inserted. I have also clarified the
399 wording for the --colour (or --color) option.
401 5. In pcregrep, when --colour was used with -o, the list of matching strings
402 was not coloured; this is different to GNU grep, so I have changed it to be
403 the same.
405 6. When --colo(u)r was used in pcregrep, only the first matching substring in
406 each matching line was coloured. Now it goes on to look for further matches
407 of any of the test patterns, which is the same behaviour as GNU grep.
409 7. A pattern that could match an empty string could cause pcregrep to loop; it
410 doesn't make sense to accept an empty string match in pcregrep, so I have
411 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
412 seems to be how GNU grep behaves.
414 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
415 start or after a newline", because the conditional assertion was not being
416 correctly handled. The rule now is that both the assertion and what follows
417 in the first alternative must satisfy the test.
419 9. If auto-callout was enabled in a pattern with a conditional group whose
420 condition was an assertion, PCRE could crash during matching, both with
421 pcre_exec() and pcre_dfa_exec().
423 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
424 used for matching.
426 11. Unicode property support in character classes was not working for
427 characters (bytes) greater than 127 when not in UTF-8 mode.
429 12. Added the -M command line option to pcretest.
431 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
433 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
435 16. Added comments and documentation about mis-use of no_arg in the C++
436 wrapper.
438 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
439 from Martin Jerabek that uses macro names for all relevant character and
440 string constants.
442 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
443 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
444 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
445 these, but not everybody uses configure.
447 19. A conditional group that had only one branch was not being correctly
448 recognized as an item that could match an empty string. This meant that an
449 enclosing group might also not be so recognized, causing infinite looping
450 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
451 with the subject "ab", where knowledge that the repeated group can match
452 nothing is needed in order to break the loop.
454 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
455 exec(), but without supplying a callout function, matching went wrong.
457 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
458 leak if the size of the offset vector was greater than 30. When the vector
459 is smaller, the saved offsets during recursion go onto a local stack
460 vector, but for larger vectors malloc() is used. It was failing to free
461 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
462 error, in fact).
464 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
465 heapframe that is used only when UTF-8 support is enabled. This caused no
466 problem, but was untidy.
468 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
469 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
470 included within another project.
472 24. Steven Van Ingelgem's patches to add more options to the CMake support,
473 slightly modified by me:
475 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
476 not building pcregrep.
478 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
479 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
481 25. Forward references, both numeric and by name, in patterns that made use of
482 duplicate group numbers, could behave incorrectly or give incorrect errors,
483 because when scanning forward to find the reference group, PCRE was not
484 taking into account the duplicate group numbers. A pattern such as
485 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
487 26. Changed a few more instances of "const unsigned char *" to USPTR, making
488 the feature of a custom pointer more persuasive (as requested by a user).
490 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
491 pcretest.c, inside #ifndefs, because it seems they are sometimes already
492 pre-defined.
494 28. Added support for (*UTF8) at the start of a pattern.
496 29. Arrange for flags added by the "release type" setting in CMake to be shown
497 in the configuration summary.
500 Version 7.8 05-Sep-08
501 ---------------------
503 1. Replaced UCP searching code with optimized version as implemented for Ad
504 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
505 stage table and inline lookup instead of a function, giving speed ups of 2
506 to 5 times on some simple patterns that I tested. Permission was given to
507 distribute the MultiStage2.py script that generates the tables (it's not in
508 the tarball, but is in the Subversion repository).
510 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
511 scripts.
513 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
514 a group with a zero qualifier. The result of the study could be incorrect,
515 or the function might crash, depending on the pattern.
517 4. Caseless matching was not working for non-ASCII characters in back
518 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
519 It now works when Unicode Property Support is available.
521 5. In pcretest, an escape such as \x{de} in the data was always generating
522 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
523 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
524 truncation.
526 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
528 7. Added two (int) casts to pcregrep when printing the difference of two
529 pointers, in case they are 64-bit values.
531 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
532 test 2 if it fails.
534 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
535 and a #define of that name to empty if it is not externally set. This is to
536 allow users of MSVC to set it if necessary.
538 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
539 the convenience functions in the pcre_get.c source file.
541 11. An option change at the start of a pattern that had top-level alternatives
542 could cause overwriting and/or a crash. This command provoked a crash in
543 some environments:
545 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
547 This potential security problem was recorded as CVE-2008-2371.
549 12. For a pattern where the match had to start at the beginning or immediately
550 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
551 pcre_dfa_exec() could read past the end of the passed subject if there was
552 no match. To help with detecting such bugs (e.g. with valgrind), I modified
553 pcretest so that it places the subject at the end of its malloc-ed buffer.
555 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
556 exec() might read past the end of the data buffer in UTF-8 mode.
558 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
559 the data contained the byte 0x85 as part of a UTF-8 character within its
560 first line. This applied both to normal and DFA matching.
562 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
563 /^[^d]*?$/8 failed to match "abc".
565 16. Added a missing copyright notice to pcrecpp_internal.h.
567 17. Make it more clear in the documentation that values returned from
568 pcre_exec() in ovector are byte offsets, not character counts.
570 18. Tidied a few places to stop certain compilers from issuing warnings.
572 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
573 supplied by Stefan Weber. I made a further small update for 7.8 because
574 there is a change of source arrangements: the pcre_searchfuncs.c module is
575 replaced by pcre_ucd.c.
578 Version 7.7 07-May-08
579 ---------------------
581 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
582 a string to a long long, pretend we don't even have a long long." This is
583 done by checking for the strtoq, strtoll, and _strtoi64 functions.
585 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
586 pre-7.6 versions, which defined a global no_arg variable instead of putting
587 it in the RE class. (See also #8 below.)
589 3. Remove a line of dead code, identified by coverity and reported by Nuno
590 Lopes.
592 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
594 (1) The include/exclude patterns were being applied to the whole pathnames
595 of files, instead of just to the final components.
597 (2) If there was more than one level of directory, the subdirectories were
598 skipped unless they satisfied the include/exclude conditions. This is
599 inconsistent with GNU grep (and could even be seen as contrary to the
600 pcregrep specification - which I improved to make it absolutely clear).
601 The action now is always to scan all levels of directory, and just
602 apply the include/exclude patterns to regular files.
604 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
605 --exclude_dir in the tests to avoid scanning .svn directories.
607 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
608 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
609 doesn't support NULs in patterns.
611 7. Added some missing "const"s to declarations of static tables in
612 pcre_compile.c and pcre_dfa_exec.c.
614 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
615 caused by fix #2 above. (Subsequently also a second patch to fix the
616 first patch. And a third patch - this was a messy problem.)
618 9. Applied Craig's patch to remove the use of push_back().
620 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
621 matching function regexec().
623 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
624 which, however, unlike Perl's \g{...}, are subroutine calls, not back
625 references. PCRE supports relative numbers with this syntax (I don't think
626 Oniguruma does).
628 12. Previously, a group with a zero repeat such as (...){0} was completely
629 omitted from the compiled regex. However, this means that if the group
630 was called as a subroutine from elsewhere in the pattern, things went wrong
631 (an internal error was given). Such groups are now left in the compiled
632 pattern, with a new opcode that causes them to be skipped at execution
633 time.
635 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
636 to the way PCRE behaves:
638 (a) A lone ] character is dis-allowed (Perl treats it as data).
640 (b) A back reference to an unmatched subpattern matches an empty string
641 (Perl fails the current match path).
643 (c) A data ] in a character class must be notated as \] because if the
644 first data character in a class is ], it defines an empty class. (In
645 Perl it is not possible to have an empty class.) The empty class []
646 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
647 The negative empty class [^] matches any one character, independently
648 of the DOTALL setting.
650 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
651 non-existent subpattern following a character class starting with ']' and
652 containing () gave an internal compiling error instead of "reference to
653 non-existent subpattern". Fortunately, when the pattern did exist, the
654 compiled code was correct. (When scanning forwards to check for the
655 existencd of the subpattern, it was treating the data ']' as terminating
656 the class, so got the count wrong. When actually compiling, the reference
657 was subsequently set up correctly.)
659 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
660 it was being rejected as not supported by pcre_dfa_exec(), even though
661 other assertions are supported. I have made pcre_dfa_exec() support
662 (*FAIL).
664 16. The implementation of 13c above involved the invention of a new opcode,
665 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
666 cannot be changed at match time, I realized I could make a small
667 improvement to matching performance by compiling OP_ALLANY instead of
668 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
669 on the OP_ANY path.
671 17. Compiling pcretest on Windows with readline support failed without the
672 following two fixes: (1) Make the unistd.h include conditional on
673 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
675 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
676 ncurses library to be included for pcretest when ReadLine support is
677 requested, but also to allow for it to be overridden. This patch came from
678 Daniel Bergström.
680 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
681 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
682 any errors with the current Unicode tables. Thanks to Peter Kankowski for
683 spotting this.
686 Version 7.6 28-Jan-08
687 ---------------------
689 1. A character class containing a very large number of characters with
690 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
691 overflow.
693 2. Patch to cut out the "long long" test in pcrecpp_unittest when
694 HAVE_LONG_LONG is not defined.
696 3. Applied Christian Ehrlicher's patch to update the CMake build files to
697 bring them up to date and include new features. This patch includes:
699 - Fixed PH's badly added libz and libbz2 support.
700 - Fixed a problem with static linking.
701 - Added pcredemo. [But later removed - see 7 below.]
702 - Fixed dftables problem and added an option.
703 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
705 - Added readline support for pcretest.
706 - Added an listing of the option settings after cmake has run.
708 4. A user submitted a patch to Makefile that makes it easy to create
709 "pcre.dll" under mingw when using Configure/Make. I added stuff to
710 Makefile.am that cause it to include this special target, without
711 affecting anything else. Note that the same mingw target plus all
712 the other distribution libraries and programs are now supported
713 when configuring with CMake (see 6 below) instead of with
714 Configure/Make.
716 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
717 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
718 exported in the Windows port". It has not yet been confirmed that the patch
719 solves the problem, but it does no harm.
721 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
722 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
723 with CMake, and also correct the comment about stack recursion.
725 7. Remove the automatic building of pcredemo from the ./configure system and
726 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
727 of a program that users should build themselves after PCRE is installed, so
728 building it automatically is not really right. What is more, it gave
729 trouble in some build environments.
731 8. Further tidies to CMakeLists.txt from Sheri and Christian.
734 Version 7.5 10-Jan-08
735 ---------------------
737 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
738 values in parens when parsing an RE using the C++ wrapper."
740 2. Negative specials like \S did not work in character classes in UTF-8 mode.
741 Characters greater than 255 were excluded from the class instead of being
742 included.
744 3. The same bug as (2) above applied to negated POSIX classes such as
745 [:^space:].
747 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
748 defined or documented. It seems to have been a typo for PCRE_STATIC, so
749 I have changed it.
751 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
752 first named subpattern) and a construct such as (?&a) would reference the
753 first named subpattern whose name started with "a" (in other words, the
754 length check was missing). Both these problems are fixed. "Subpattern name
755 expected" is now given for (?&) (a zero-length name), and this patch also
756 makes it give the same error for \k'' (previously it complained that that
757 was a reference to a non-existent subpattern).
759 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
760 this is right because (?- can be followed by option settings as well as by
761 digits. I have, however, made the messages clearer.
763 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
764 than the number used in the conditional) now cause a compile-time error.
765 This is actually not compatible with Perl, which accepts such patterns, but
766 treats the conditional as always being FALSE (as PCRE used to), but it
767 seems to me that giving a diagnostic is better.
769 8. Change "alphameric" to the more common word "alphanumeric" in comments
770 and messages.
772 9. Fix two occurrences of "backslash" in comments that should have been
773 "backspace".
775 10. Remove two redundant lines of code that can never be obeyed (their function
776 was moved elsewhere).
778 11. The program that makes PCRE's Unicode character property table had a bug
779 which caused it to generate incorrect table entries for sequences of
780 characters that have the same character type, but are in different scripts.
781 It amalgamated them into a single range, with the script of the first of
782 them. In other words, some characters were in the wrong script. There were
783 thirteen such cases, affecting characters in the following ranges:
785 U+002b0 - U+002c1
786 U+0060c - U+0060d
787 U+0061e - U+00612
788 U+0064b - U+0065e
789 U+0074d - U+0076d
790 U+01800 - U+01805
791 U+01d00 - U+01d77
792 U+01d9b - U+01dbf
793 U+0200b - U+0200f
794 U+030fc - U+030fe
795 U+03260 - U+0327f
796 U+0fb46 - U+0fbb1
797 U+10450 - U+1049d
799 12. The -o option (show only the matching part of a line) for pcregrep was not
800 compatible with GNU grep in that, if there was more than one match in a
801 line, it showed only the first of them. It now behaves in the same way as
802 GNU grep.
804 13. If the -o and -v options were combined for pcregrep, it printed a blank
805 line for every non-matching line. GNU grep prints nothing, and pcregrep now
806 does the same. The return code can be used to tell if there were any
807 non-matching lines.
809 14. Added --file-offsets and --line-offsets to pcregrep.
811 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
812 infinitely looping recursion. The bug was that positive lookaheads were not
813 being skipped when checking for a possible empty match (negative lookaheads
814 and both kinds of lookbehind were skipped).
816 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
817 inclusion of <windows.h> to before rather than after the definition of
818 INVALID_FILE_ATTRIBUTES (patch from David Byron).
820 17. Specifying a possessive quantifier with a specific limit for a Unicode
821 character property caused pcre_compile() to compile bad code, which led at
822 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
823 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
824 caused the error; without that there was no problem.
826 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
828 19. Added --enable-pcretest-libreadline.
830 20. In pcrecpp.cc, the variable 'count' was incremented twice in
831 RE::GlobalReplace(). As a result, the number of replacements returned was
832 double what it should be. I removed one of the increments, but Craig sent a
833 later patch that removed the other one (the right fix) and added unit tests
834 that check the return values (which was not done before).
836 21. Several CMake things:
838 (1) Arranged that, when cmake is used on Unix, the libraries end up with
839 the names libpcre and libpcreposix, not just pcre and pcreposix.
841 (2) The above change means that pcretest and pcregrep are now correctly
842 linked with the newly-built libraries, not previously installed ones.
846 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
847 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
848 UTF-8 newline character). The key issue is that the pattern starts .*;
849 this means that the match must be either at the beginning, or after a
850 newline. The bug was in the code for advancing after a failed match and
851 checking that the new position followed a newline. It was not taking
852 account of UTF-8 characters correctly.
854 23. PCRE was behaving differently from Perl in the way it recognized POSIX
855 character classes. PCRE was not treating the sequence [:...:] as a
856 character class unless the ... were all letters. Perl, however, seems to
857 allow any characters between [: and :], though of course it rejects as
858 unknown any "names" that contain non-letters, because all the known class
859 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
860 for example, whereas PCRE did not - it did not recognize a POSIX character
861 class. This seemed a bit dangerous, so the code has been changed to be
862 closer to Perl. The behaviour is not identical to Perl, because PCRE will
863 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
864 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
865 Perl does, and where it didn't before.
867 24. Rewrite so as to remove the single use of %n from pcregrep because in some
868 Windows environments %n is disabled by default.
871 Version 7.4 21-Sep-07
872 ---------------------
874 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
875 means that a class such as [\s] counted as "explicit reference to CR or
876 LF". That isn't really right - the whole point of the change was to try to
877 help when there was an actual mention of one of the two characters. So now
878 the change happens only if \r or \n (or a literal CR or LF) character is
879 encountered.
881 2. The 32-bit options word was also used for 6 internal flags, but the numbers
882 of both had grown to the point where there were only 3 bits left.
883 Fortunately, there was spare space in the data structure, and so I have
884 moved the internal flags into a new 16-bit field to free up more option
885 bits.
887 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
888 but did not set the internal JCHANGED flag - either of these is enough to
889 control the way the "get" function works - but the PCRE_INFO_JCHANGED
890 facility is supposed to tell if (?J) was ever used, so now (?J) at the
891 start sets both bits.
893 4. Added options (at build time, compile time, exec time) to change \R from
894 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
896 5. doc/pcresyntax.html was missing from the distribution.
898 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
899 compatibility, even though it is no longer used.
901 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
902 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
903 windows.h file is present (where different names are used). [This was
904 reversed later after testing - see 16 below.]
906 8. Changed all #include <config.h> to #include "config.h". There were also
907 some further <pcre.h> cases that I changed to "pcre.h".
909 9. When pcregrep was used with the --colour option, it missed the line ending
910 sequence off the lines that it output.
912 10. It was pointed out to me that arrays of string pointers cause lots of
913 relocations when a shared library is dynamically loaded. A technique of
914 using a single long string with a table of offsets can drastically reduce
915 these. I have refactored PCRE in four places to do this. The result is
916 dramatic:
918 Originally: 290
919 After changing UCP table: 187
920 After changing error message table: 43
921 After changing table of "verbs" 36
922 After changing table of Posix names 22
924 Thanks to the folks working on Gregex for glib for this insight.
926 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
927 unicode-properties was also set.
929 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
931 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
932 checked only for CRLF.
934 14. Added casts to pcretest.c to avoid compiler warnings.
936 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
938 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
939 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
940 entirely. This removes changes made in 7 above.
942 17. The CMake files have been updated, and there is now more information about
943 building with CMake in the NON-UNIX-USE document.
946 Version 7.3 28-Aug-07
947 ---------------------
949 1. In the rejigging of the build system that eventually resulted in 7.1, the
950 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
951 brackets there is not right, since it causes compilers to look for an
952 installed pcre.h, not the version that is in the source that is being
953 compiled (which of course may be different). I have changed it back to:
955 #include "pcre.h"
957 I have a vague recollection that the change was concerned with compiling in
958 different directories, but in the new build system, that is taken care of
959 by the VPATH setting the Makefile.
961 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
962 when the subject happened to end in the byte 0x85 (e.g. if the last
963 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
964 characters but of course it shouldn't be taken as a newline when it is part
965 of another character. The bug was that, for an unlimited repeat of . in
966 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
967 characters when looking for a newline.
969 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
971 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
972 in debug output.
974 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
975 long printing in the pcrecpp unittest when running under MinGW.
977 6. ESC_K was left out of the EBCDIC table.
979 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
980 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
981 limit also applies to "virtual nesting" when a pattern is recursive, and in
982 this case 1000 isn't so big. I have been able to remove this limit at the
983 expense of backing off one optimization in certain circumstances. Normally,
984 when pcre_exec() would call its internal match() function recursively and
985 immediately return the result unconditionally, it uses a "tail recursion"
986 feature to save stack. However, when a subpattern that can match an empty
987 string has an unlimited repetition quantifier, it no longer makes this
988 optimization. That gives it a stack frame in which to save the data for
989 checking that an empty string has been matched. Previously this was taken
990 from the 1000-entry workspace that had been reserved. So now there is no
991 explicit limit, but more stack is used.
993 8. Applied Daniel's patches to solve problems with the import/export magic
994 syntax that is required for Windows, and which was going wrong for the
995 pcreposix and pcrecpp parts of the library. These were overlooked when this
996 problem was solved for the main library.
998 9. There were some crude static tests to avoid integer overflow when computing
999 the size of patterns that contain repeated groups with explicit upper
1000 limits. As the maximum quantifier is 65535, the maximum group length was
1001 set at 30,000 so that the product of these two numbers did not overflow a
1002 32-bit integer. However, it turns out that people want to use groups that
1003 are longer than 30,000 bytes (though not repeat them that many times).
1004 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1005 made it possible to implement the integer overflow checks in a much more
1006 dynamic way, which I have now done. The artificial limitation on group
1007 length has been removed - we now have only the limit on the total length of
1008 the compiled pattern, which depends on the LINK_SIZE setting.
1010 10. Fixed a bug in the documentation for get/copy named substring when
1011 duplicate names are permitted. If none of the named substrings are set, the
1012 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1013 empty string.
1015 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1016 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1017 because the ] is interpreted as the first data character and the
1018 terminating ] is not found. PCRE has been made compatible with Perl in this
1019 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1020 cause memory overwriting.
1022 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1023 string has been matched (to stop an infinite loop). It was not recognizing
1024 a conditional subpattern that could match an empty string if that
1025 subpattern was within another subpattern. For example, it looped when
1026 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1027 condition was not nested. This bug has been fixed.
1029 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1030 past the start of the subject in the presence of bytes with the top bit
1031 set, for example "\x8aBCD".
1033 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1034 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1036 14. Optimized (?!) to (*FAIL).
1038 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1039 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1040 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1041 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1042 does: it's just the validity check that is more restrictive.
1044 16. Inserted checks for integer overflows during escape sequence (backslash)
1045 processing, and also fixed erroneous offset values for syntax errors during
1046 backslash processing.
1048 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1049 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1051 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1052 caused an overrun.
1054 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1055 something other than just ASCII characters) inside a group that had an
1056 unlimited repeat caused a loop at compile time (while checking to see
1057 whether the group could match an empty string).
1059 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1060 [\P{Any}] did so. (Error in the code for printing property names.)
1062 21. An orphan \E inside a character class could cause a crash.
1064 22. A repeated capturing bracket such as (A)? could cause a wild memory
1065 reference during compilation.
1067 23. There are several functions in pcre_compile() that scan along a compiled
1068 expression for various reasons (e.g. to see if it's fixed length for look
1069 behind). There were bugs in these functions when a repeated \p or \P was
1070 present in the pattern. These operators have additional parameters compared
1071 with \d, etc, and these were not being taken into account when moving along
1072 the compiled data. Specifically:
1074 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1075 length.
1077 (b) An item such as \pL+ within a repeated group could cause crashes or
1078 loops.
1080 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1081 "reference to non-existent subpattern" error.
1083 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1085 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1086 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1088 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1090 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1091 character were causing crashes (broken optimization).
1093 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1094 \p or \P) caused a compile-time loop.
1096 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1097 break. For example, the unstudied pattern [\r\n]A does not match the string
1098 "\r\nA" because change 7.0/46 below moves the current point on by two
1099 characters after failing to match at the start. However, the pattern \nA
1100 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1101 the same is true. There doesn't seem any very clean way out of this, but
1102 what I have chosen to do makes the common cases work: PCRE now takes note
1103 of whether there can be an explicit match for \r or \n anywhere in the
1104 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1105 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1106 pattern has explicit CR or LF references.
1108 29. Added (*CR) etc for changing newline setting at start of pattern.
1111 Version 7.2 19-Jun-07
1112 ---------------------
1114 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1115 which is apparently normally available under Windows.
1117 2. Re-jig the pcregrep tests with different newline settings in an attempt
1118 to make them independent of the local environment's newline setting.
1120 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1122 4. Some of the "internals" tests were previously cut out when the link size
1123 was not 2, because the output contained actual offsets. The recent new
1124 "Z" feature of pcretest means that these can be cut out, making the tests
1125 usable with all link sizes.
1127 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1128 stack recursion. This gives a massive performance boost under BSD, but just
1129 a small improvement under Linux. However, it saves one field in the frame
1130 in all cases.
1132 6. Added more features from the forthcoming Perl 5.10:
1134 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1135 recursion call. It refers to the nth most recently opened parentheses.
1137 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1138 to be opened parentheses.
1140 (c) Conditions that refer to capturing parentheses can be specified
1141 relatively, for example, (?(-2)... or (?(+3)...
1143 (d) \K resets the start of the current match so that everything before
1144 is not part of it.
1146 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1148 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1149 reference syntax.
1151 (g) (?| introduces a group in which the numbering of parentheses in each
1152 alternative starts with the same number.
1154 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1156 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1159 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1160 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1161 for detecting groups that can match an empty string.
1163 9. A pattern with a very large number of alternatives (more than several
1164 hundred) was running out of internal workspace during the pre-compile
1165 phase, where pcre_compile() figures out how much memory will be needed. A
1166 bit of new cunning has reduced the workspace needed for groups with
1167 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1168 workspace instead of running out of the 4096 that are available.
1170 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1172 11. Applied patch from Google to remove an optimization that didn't quite work.
1173 The report of the bug said:
1175 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1176 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1177 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1179 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1180 it matched the wrong number of bytes.
1183 Version 7.1 24-Apr-07
1184 ---------------------
1186 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1187 that is more "standard", making use of automake and other Autotools. There
1188 is some re-arrangement of the files and adjustment of comments consequent
1189 on this.
1191 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1192 for recursive directory scanning broke on some systems because the files
1193 are not scanned in any specific order and on different systems the order
1194 was different. A call to "sort" has been inserted into RunGrepTest for the
1195 approprate test as a short-term fix. In the longer term there may be an
1196 alternative.
1198 3. I had an email from Eric Raymond about problems translating some of PCRE's
1199 man pages to HTML (despite the fact that I distribute HTML pages, some
1200 people do their own conversions for various reasons). The problems
1201 concerned the use of low-level troff macros .br and .in. I have therefore
1202 removed all such uses from the man pages (some were redundant, some could
1203 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1204 HTML has been updated to handle .nf/.fi and to complain if it encounters
1205 .br or .in.
1207 4. Updated comments in configure.ac that get placed in config.h.in and also
1208 arranged for config.h to be included in the distribution, with the name
1209 config.h.generic, for the benefit of those who have to compile without
1210 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1212 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1213 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1214 makevp.bat for the current PCRE, using the additional files
1215 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1217 6. A Windows user reported a minor discrepancy with test 2, which turned out
1218 to be caused by a trailing space on an input line that had got lost in his
1219 copy. The trailing space was an accident, so I've just removed it.
1221 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1222 that is needed.
1224 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1225 as "const" (a) because they are and (b) because it helps the PHP
1226 maintainers who have recently made a script to detect big data structures
1227 in the php code that should be moved to the .rodata section. I remembered
1228 to update Builducptable as well, so it won't revert if ucptable.h is ever
1229 re-created.
1231 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1232 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1233 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1234 support is not required. This saves 1.5-2K of code, which is important in
1235 some applications.
1237 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1238 so as not to refer to the tables, even though these functions will never be
1239 called when UTF-8 support is disabled. Otherwise there are problems with a
1240 shared library.
1242 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1244 (a) It was defining its arguments as char * instead of void *.
1246 (b) It was assuming that all moves were upwards in memory; this was true
1247 a long time ago when I wrote it, but is no longer the case.
1249 The emulated memove() is provided for those environments that have neither
1250 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1251 is clearly not the case, as these two bugs were recently reported.
1253 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1254 and Detrail to create the HTML documentation, the .txt form of the man
1255 pages, and it removes trailing spaces from listed files. It also creates
1256 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1257 case, it wraps all the #defines with #ifndefs. This script should be run
1258 before "make dist".
1260 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1261 with Unicode property support.
1263 (a) For a maximizing quantifier, if the two different cases of the
1264 character were of different lengths in their UTF-8 codings (there are
1265 some cases like this - I found 11), and the matching function had to
1266 back up over a mixture of the two cases, it incorrectly assumed they
1267 were both the same length.
1269 (b) When PCRE was configured to use the heap rather than the stack for
1270 recursion during matching, it was not correctly preserving the data for
1271 the other case of a UTF-8 character when checking ahead for a match
1272 while processing a minimizing repeat. If the check also involved
1273 matching a wide character, but failed, corruption could cause an
1274 erroneous result when trying to check for a repeat of the original
1275 character.
1277 13. Some tidying changes to the testing mechanism:
1279 (a) The RunTest script now detects the internal link size and whether there
1280 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1281 values substituted by "configure". (The RunGrepTest script already did
1282 this for UTF-8.) The configure.ac script no longer substitutes the
1283 relevant variables.
1285 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1286 with length and offset values. This means that the output is different
1287 for different internal link sizes. Test 2 is skipped for link sizes
1288 other than 2 because of this, bypassing the problem. Unfortunately,
1289 there was also a test in test 3 (the locale tests) that used /B and
1290 failed for link sizes other than 2. Rather than cut the whole test out,
1291 I have added a new /Z option to pcretest that replaces the length and
1292 offset values with spaces. This is now used to make test 3 independent
1293 of link size. (Test 2 will be tidied up later.)
1295 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1296 segmentation fault instead of returning the appropriate error message.
1298 15. In multiline mode when the newline sequence was set to "any", the pattern
1299 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1300 This doesn't seem right; it now treats the CRLF combination as the line
1301 ending, and so does not match in that case. It's only a pattern such as ^$
1302 that would hit this one: something like ^ABC$ would have failed after \r
1303 and then tried again after \r\n.
1305 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1306 in an attempt to make files that differ only in their line terminators
1307 compare equal. This works on Linux.
1309 17. Under certain error circumstances pcregrep might try to free random memory
1310 as it exited. This is now fixed, thanks to valgrind.
1312 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1313 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1314 was because its rules for how to advance for /g after matching an empty
1315 string at the end of a line did not allow for this case. They now check for
1316 it specially.
1318 20. pcretest is supposed to handle patterns and data of any length, by
1319 extending its buffers when necessary. It was getting this wrong when the
1320 buffer for a data line had to be extended.
1322 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1323 CRLF as a newline sequence.
1325 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1326 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1327 I have nevertheless tidied it up.
1329 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1331 24. Added a man page for pcre-config.
1334 Version 7.0 19-Dec-06
1335 ---------------------
1337 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1338 moving to gcc 4.1.1.
1340 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1341 sys/time.h, which is documented as needed for this function. It doesn't
1342 seem to matter on Linux, but it showed up on some releases of OS X.
1344 3. It seems that there are systems where bytes whose values are greater than
1345 127 match isprint() in the "C" locale. The "C" locale should be the
1346 default when a C program starts up. In most systems, only ASCII printing
1347 characters match isprint(). This difference caused the output from pcretest
1348 to vary, making some of the tests fail. I have changed pcretest so that:
1350 (a) When it is outputting text in the compiled version of a pattern, bytes
1351 other than 32-126 are always shown as hex escapes.
1353 (b) When it is outputting text that is a matched part of a subject string,
1354 it does the same, unless a different locale has been set for the match
1355 (using the /L modifier). In this case, it uses isprint() to decide.
1357 4. Fixed a major bug that caused incorrect computation of the amount of memory
1358 required for a compiled pattern when options that changed within the
1359 pattern affected the logic of the preliminary scan that determines the
1360 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1361 that the computed length was too small. The symptoms of this bug were
1362 either the PCRE error "internal error: code overflow" from pcre_compile(),
1363 or a glibc crash with a message such as "pcretest: free(): invalid next
1364 size (fast)". Examples of patterns that provoked this bug (shown in
1365 pcretest format) are:
1367 /(?-x: )/x
1368 /(?x)(?-x: \s*#\s*)/
1369 /((?i)[\x{c0}])/8
1370 /(?i:[\x{c0}])/8
1372 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1373 is now done differently.
1375 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1376 wrapper classes; (b) implement a new function in the C++ scanner that is
1377 more efficient than the old way of doing things because it avoids levels of
1378 recursion in the regex matching; (c) add a paragraph to the documentation
1379 for the FullMatch() function.
1381 6. The escape sequence \n was being treated as whatever was defined as
1382 "newline". Not only was this contrary to the documentation, which states
1383 that \n is character 10 (hex 0A), but it also went horribly wrong when
1384 "newline" was defined as CRLF. This has been fixed.
1386 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1387 was being set to -1 for the "end of line" case (supposedly a value that no
1388 character can have). Though this value is never used (the check for end of
1389 line is "zero bytes in current character"), it caused compiler complaints.
1390 I've changed it to 0xffffffff.
1392 8. In pcre_version.c, the version string was being built by a sequence of
1393 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1394 string (as it is for production releases) called a macro with an empty
1395 argument. The C standard says the result of this is undefined. The gcc
1396 compiler treats it as an empty string (which was what was wanted) but it is
1397 reported that Visual C gives an error. The source has been hacked around to
1398 avoid this problem.
1400 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1401 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1402 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1403 of them did).
1405 10. Originally, pcretest opened its input and output without "b"; then I was
1406 told that "b" was needed in some environments, so it was added for release
1407 5.0 to both the input and output. (It makes no difference on Unix-like
1408 systems.) Later I was told that it is wrong for the input on Windows. I've
1409 now abstracted the modes into two macros, to make it easier to fiddle with
1410 them, and removed "b" from the input mode under Windows.
1412 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1414 12. Added -help and --help to pcretest as an official way of being reminded
1415 of the options.
1417 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1418 and pcrecpp.cc because they annoy compilers at high warning levels.
1420 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1422 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1423 scripts are not C programs :-) and which was not noticed because it works
1424 on Linux.
1426 16. pcretest is supposed to handle any length of pattern and data line (as one
1427 line or as a continued sequence of lines) by extending its input buffer if
1428 necessary. This feature was broken for very long pattern lines, leading to
1429 a string of junk being passed to pcre_compile() if the pattern was longer
1430 than about 50K.
1432 17. I have done a major re-factoring of the way pcre_compile() computes the
1433 amount of memory needed for a compiled pattern. Previously, there was code
1434 that made a preliminary scan of the pattern in order to do this. That was
1435 OK when PCRE was new, but as the facilities have expanded, it has become
1436 harder and harder to keep it in step with the real compile phase, and there
1437 have been a number of bugs (see for example, 4 above). I have now found a
1438 cunning way of running the real compile function in a "fake" mode that
1439 enables it to compute how much memory it would need, while actually only
1440 ever using a few hundred bytes of working memory and without too many
1441 tests of the mode. This should make future maintenance and development
1442 easier. A side effect of this work is that the limit of 200 on the nesting
1443 depth of parentheses has been removed (though this was never a serious
1444 limitation, I suspect). However, there is a downside: pcre_compile() now
1445 runs more slowly than before (30% or more, depending on the pattern). I
1446 hope this isn't a big issue. There is no effect on runtime performance.
1448 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1449 newline (only possible for the last line of a file) and it was a
1450 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1452 19. Added additional timing features to pcretest. (1) The -tm option now times
1453 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1454 separate command line item, by a number that specifies the number of
1455 repeats to use when timing. The default is 50000; this gives better
1456 precision, but takes uncomfortably long for very large patterns.
1458 20. Extended pcre_study() to be more clever in cases where a branch of a
1459 subpattern has no definite first character. For example, (a*|b*)[cd] would
1460 previously give no result from pcre_study(). Now it recognizes that the
1461 first character must be a, b, c, or d.
1463 21. There was an incorrect error "recursive call could loop indefinitely" if
1464 a subpattern (or the entire pattern) that was being tested for matching an
1465 empty string contained only one non-empty item after a nested subpattern.
1466 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1467 incorrectly, because the \d was being skipped in the check.
1469 22. The pcretest program now has a new pattern option /B and a command line
1470 option -b, which is equivalent to adding /B to every pattern. This causes
1471 it to show the compiled bytecode, without the additional information that
1472 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1473 is the same as /B/I).
1475 23. A new optimization is now able automatically to treat some sequences such
1476 as a*b as a*+b. More specifically, if something simple (such as a character
1477 or a simple class like \d) has an unlimited quantifier, and is followed by
1478 something that cannot possibly match the quantified thing, the quantifier
1479 is automatically "possessified".
1481 24. A recursive reference to a subpattern whose number was greater than 39
1482 went wrong under certain circumstances in UTF-8 mode. This bug could also
1483 have affected the operation of pcre_study().
1485 25. Realized that a little bit of performance could be had by replacing
1486 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1488 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1490 27. Possessive quantifiers such as a++ were previously implemented by turning
1491 them into atomic groups such as ($>a+). Now they have their own opcodes,
1492 which improves performance. This includes the automatically created ones
1493 from 23 above.
1495 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1496 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1497 the first matched character to be a colon. This applied both to named and
1498 numbered groups.
1500 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1502 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1503 building PCRE on Windows easier, so I have included it in the distribution.
1505 31. There is now a check in pcretest against a ridiculously large number being
1506 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1507 loop, the loop is abandoned.
1509 32. Forward references to subpatterns in conditions such as (?(2)...) where
1510 subpattern 2 is defined later cause pcre_compile() to search forwards in
1511 the pattern for the relevant set of parentheses. This search went wrong
1512 when there were unescaped parentheses in a character class, parentheses
1513 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1515 33. "Subroutine" calls and backreferences were previously restricted to
1516 referencing subpatterns earlier in the regex. This restriction has now
1517 been removed.
1519 34. Added a number of extra features that are going to be in Perl 5.10. On the
1520 whole, these are just syntactic alternatives for features that PCRE had
1521 previously implemented using the Python syntax or my own invention. The
1522 other formats are all retained for compatibility.
1524 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1525 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1526 also .NET compatible.
1528 (b) A recursion or subroutine call to a named group can now be defined as
1529 (?&name) as well as (?P>name).
1531 (c) A backreference to a named group can now be defined as \k<name> or
1532 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1533 5.10, are also .NET compatible.
1535 (d) A conditional reference to a named group can now use the syntax
1536 (?(<name>) or (?('name') as well as (?(name).
1538 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1539 groups (named and numbered) that are never evaluated inline, but can be
1540 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1541 is always false. There may be only one alternative in such a group.
1543 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1544 as the simple (?(R). The condition is true only if the most recent
1545 recursion is that of the given number or name. It does not search out
1546 through the entire recursion stack.
1548 (g) The escape \gN or \g{N} has been added, where N is a positive or
1549 negative number, specifying an absolute or relative reference.
1551 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1552 some "unreachable code" warnings.
1554 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1555 things, this adds five new scripts.
1557 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1558 There were also incompatibilities regarding the handling of \Q..\E inside
1559 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1560 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1562 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1563 matches an empty string, and forcibly breaks the loop. There were bugs in
1564 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1565 against aaaa the result was just "a" rather than "aaaa", for example. Two
1566 separate and independent bugs (that affected different cases) have been
1567 fixed.
1569 39. Refactored the code to abolish the use of different opcodes for small
1570 capturing bracket numbers. This is a tidy that I avoided doing when I
1571 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1572 The new approach is not only tidier, it makes it possible to reduce the
1573 memory needed to fix the previous bug (38).
1575 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1576 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1577 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1578 mode.
1580 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1581 report.
1583 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1584 copy construction and assignment in the C++ wrapper.
1586 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1587 couple of bugs that could have given wrong results in the "--newline=crlf"
1588 case.
1590 44. Added a number of casts and did some reorganization of signed/unsigned int
1591 variables following suggestions from Dair Grant. Also renamed the variable
1592 "this" as "item" because it is a C++ keyword.
1594 45. Arranged for dftables to add
1596 #include "pcre_internal.h"
1598 to pcre_chartables.c because without it, gcc 4.x may remove the array
1599 definition from the final binary if PCRE is built into a static library and
1600 dead code stripping is activated.
1602 46. For an unanchored pattern, if a match attempt fails at the start of a
1603 newline sequence, and the newline setting is CRLF or ANY, and the next two
1604 characters are CRLF, advance by two characters instead of one.
1607 Version 6.7 04-Jul-06
1608 ---------------------
1610 1. In order to handle tests when input lines are enormously long, pcretest has
1611 been re-factored so that it automatically extends its buffers when
1612 necessary. The code is crude, but this _is_ just a test program. The
1613 default size has been increased from 32K to 50K.
1615 2. The code in pcre_study() was using the value of the re argument before
1616 testing it for NULL. (Of course, in any sensible call of the function, it
1617 won't be NULL.)
1619 3. The memmove() emulation function in pcre_internal.h, which is used on
1620 systems that lack both memmove() and bcopy() - that is, hardly ever -
1621 was missing a "static" storage class specifier.
1623 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1624 containing an extended class (one that cannot be represented by a bitmap
1625 because it contains high-valued characters or Unicode property items, e.g.
1626 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1627 pattern, but PCRE should not loop if you do not (it no longer does).
1628 [Detail: two cases were found: (a) a repeated subpattern containing an
1629 extended class; (b) a recursive reference to a subpattern that followed a
1630 previous extended class. It wasn't skipping over the extended class
1631 correctly when UTF-8 mode was not set.]
1633 5. A negated single-character class was not being recognized as fixed-length
1634 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1635 compile error "lookbehind assertion is not fixed length".
1637 6. The RunPerlTest auxiliary script was showing an unexpected difference
1638 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1639 write a Perl script that can interpret lines of an input file either as
1640 byte characters or as UTF-8, which is what "perltest" was being required to
1641 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1642 can't do is switch easily at run time between having the "use utf8;" pragma
1643 or not. In the end, I fudged it by using the RunPerlTest script to insert
1644 "use utf8;" explicitly for the UTF-8 tests.
1646 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1647 the end of the subject string, contrary to the documentation and to what
1648 Perl does. This was true of both matching functions. Now it matches only at
1649 the start of the subject and immediately after *internal* newlines.
1651 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1652 a pointer to an int instead of a pointer to an unsigned long int. This
1653 caused problems on 64-bit systems.
1655 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1656 instance of the 'standard' template library not being so standard".
1658 10. There was no check on the number of named subpatterns nor the maximum
1659 length of a subpattern name. The product of these values is used to compute
1660 the size of the memory block for a compiled pattern. By supplying a very
1661 long subpattern name and a large number of named subpatterns, the size
1662 computation could be caused to overflow. This is now prevented by limiting
1663 the length of names to 32 characters, and the number of named subpatterns
1664 to 10,000.
1666 11. Subpatterns that are repeated with specific counts have to be replicated in
1667 the compiled pattern. The size of memory for this was computed from the
1668 length of the subpattern and the repeat count. The latter is limited to
1669 65535, but there was no limit on the former, meaning that integer overflow
1670 could in principle occur. The compiled length of a repeated subpattern is
1671 now limited to 30,000 bytes in order to prevent this.
1673 12. Added the optional facility to have named substrings with the same name.
1675 13. Added the ability to use a named substring as a condition, using the
1676 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1677 are numbers (not recommended). Forward references are permitted.
1679 14. Added forward references in named backreferences (if you see what I mean).
1681 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1682 pattern could run off the end of the subject. For example, the pattern
1683 "(?s)(.{1,5})"8 did this with the subject "ab".
1685 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1686 PCRE_CASELESS was set when matching characters that were quantified with ?
1687 or *.
1689 17. A character class other than a single negated character that had a minimum
1690 but no maximum quantifier - for example [ab]{6,} - was not handled
1691 correctly by pce_dfa_exec(). It would match only one character.
1693 18. A valid (though odd) pattern that looked like a POSIX character
1694 class but used an invalid character after [ (for example [[,abc,]]) caused
1695 pcre_compile() to give the error "Failed: internal error: code overflow" or
1696 in some cases to crash with a glibc free() error. This could even happen if
1697 the pattern terminated after [[ but there just happened to be a sequence of
1698 letters, a binary zero, and a closing ] in the memory that followed.
1700 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1701 over the years. Originally (before any Unicode support), just the bottom 8
1702 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1703 output from "man perlunicode" includes this:
1705 The regular expression compiler produces polymorphic opcodes. That
1706 is, the pattern adapts to the data and automatically switches to
1707 the Unicode character scheme when presented with Unicode data--or
1708 instead uses a traditional byte scheme when presented with byte
1709 data.
1711 Sadly, a wide octal escape does not cause a switch, and in a string with
1712 no other multibyte characters, these octal escapes are treated as before.
1713 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1714 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1715 Unicode string.
1717 I have not perpetrated such confusion in PCRE. Up till now, it took just
1718 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1719 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1720 translate to the appropriate multibyte character.
1722 29. Applied some refactoring to reduce the number of warnings from Microsoft
1723 and Borland compilers. This has included removing the fudge introduced
1724 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1725 a warning about an unused variable.
1727 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1728 characters since release 4.0, because Perl (from release 5.004) does not.
1729 [Or at least, is documented not to: some releases seem to be in conflict
1730 with the documentation.] However, when a pattern was studied with
1731 pcre_study() and all its branches started with \s, PCRE still included VT
1732 as a possible starting character. Of course, this did no harm; it just
1733 caused an unnecessary match attempt.
1735 22. Removed a now-redundant internal flag bit that recorded the fact that case
1736 dependency changed within the pattern. This was once needed for "required
1737 byte" processing, but is no longer used. This recovers a now-scarce options
1738 bit. Also moved the least significant internal flag bit to the most-
1739 significant bit of the word, which was not previously used (hangover from
1740 the days when it was an int rather than a uint) to free up another bit for
1741 the future.
1743 23. Added support for CRLF line endings as well as CR and LF. As well as the
1744 default being selectable at build time, it can now be changed at runtime
1745 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1746 specify that it is scanning data with non-default line endings.
1748 24. Changed the definition of CXXLINK to make it agree with the definition of
1749 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1751 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1752 recursions. This makes a big different to stack usage for some patterns.
1754 26. If a subpattern containing a named recursion or subroutine reference such
1755 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1756 the space required for the compiled pattern went wrong and gave too small a
1757 value. Depending on the environment, this could lead to "Failed: internal
1758 error: code overflow at offset 49" or "glibc detected double free or
1759 corruption" errors.
1761 27. Applied patches from Google (a) to support the new newline modes and (b) to
1762 advance over multibyte UTF-8 characters in GlobalReplace.
1764 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1765 difference for some implementation of PCRE in some Windows version.
1767 29. Added some extra testing facilities to pcretest:
1769 \q<number> in a data line sets the "match limit" value
1770 \Q<number> in a data line sets the "match recursion limt" value
1771 -S <number> sets the stack size, where <number> is in megabytes
1773 The -S option isn't available for Windows.
1776 Version 6.6 06-Feb-06
1777 ---------------------
1779 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1780 in pcreposix.h. I have copied the definition from pcre.h.
1782 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1783 because pcre.h is no longer a built file.
1785 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1786 not normally included in the compiled code.
1789 Version 6.5 01-Feb-06
1790 ---------------------
1792 1. When using the partial match feature with pcre_dfa_exec(), it was not
1793 anchoring the second and subsequent partial matches at the new starting
1794 point. This could lead to incorrect results. For example, with the pattern
1795 /1234/, partially matching against "123" and then "a4" gave a match.
1797 2. Changes to pcregrep:
1799 (a) All non-match returns from pcre_exec() were being treated as failures
1800 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1801 error message is output. Some extra information is given for the
1803 probably the only errors that are likely to be caused by users (by
1804 specifying a regex that has nested indefinite repeats, for instance).
1805 If there are more than 20 of these errors, pcregrep is abandoned.
1807 (b) A binary zero was treated as data while matching, but terminated the
1808 output line if it was written out. This has been fixed: binary zeroes
1809 are now no different to any other data bytes.
1811 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1812 used to set a locale for matching. The --locale=xxxx long option has
1813 been added (no short equivalent) to specify a locale explicitly on the
1814 pcregrep command, overriding the environment variables.
1816 (d) When -B was used with -n, some line numbers in the output were one less
1817 than they should have been.
1819 (e) Added the -o (--only-matching) option.
1821 (f) If -A or -C was used with -c (count only), some lines of context were
1822 accidentally printed for the final match.
1824 (g) Added the -H (--with-filename) option.
1826 (h) The combination of options -rh failed to suppress file names for files
1827 that were found from directory arguments.
1829 (i) Added the -D (--devices) and -d (--directories) options.
1831 (j) Added the -F (--fixed-strings) option.
1833 (k) Allow "-" to be used as a file name for -f as well as for a data file.
1835 (l) Added the --colo(u)r option.
1837 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1838 is not present by default.
1840 3. A nasty bug was discovered in the handling of recursive patterns, that is,
1841 items such as (?R) or (?1), when the recursion could match a number of
1842 alternatives. If it matched one of the alternatives, but subsequently,
1843 outside the recursion, there was a failure, the code tried to back up into
1844 the recursion. However, because of the way PCRE is implemented, this is not
1845 possible, and the result was an incorrect result from the match.
1847 In order to prevent this happening, the specification of recursion has
1848 been changed so that all such subpatterns are automatically treated as
1849 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1851 4. I had overlooked the fact that, in some locales, there are characters for
1852 which isalpha() is true but neither isupper() nor islower() are true. In
1853 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1854 and ordfeminine) are like this. This affected the treatment of \w and \W
1855 when they appeared in character classes, but not when they appeared outside
1856 a character class. The bit map for "word" characters is now created
1857 separately from the results of isalnum() instead of just taking it from the
1858 upper, lower, and digit maps. (Plus the underscore character, of course.)
1860 5. The above bug also affected the handling of POSIX character classes such as
1861 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1862 permanent tables. Instead, the bit maps for such a class were previously
1863 created as the appropriate unions of the upper, lower, and digit bitmaps.
1864 Now they are created by subtraction from the [[:word:]] class, which has
1865 its own bitmap.
1867 6. The [[:blank:]] character class matches horizontal, but not vertical space.
1868 It is created by subtracting the vertical space characters (\x09, \x0a,
1869 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1870 subtraction was done in the overall bitmap for a character class, meaning
1871 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1872 be recognized. This bug has been fixed.
1874 7. Patches from the folks at Google:
1876 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1877 real life, but is still worth protecting against".
1879 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1880 regular expressions".
1882 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1883 have it.
1885 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1886 "configure" and the latter not, in order to fix a problem somebody had
1887 with compiling the Arg class on HP-UX.
1889 (e) Improve the error-handling of the C++ wrapper a little bit.
1891 (f) New tests for checking recursion limiting.
1893 8. The pcre_memmove() function, which is used only if the environment does not
1894 have a standard memmove() function (and is therefore rarely compiled),
1895 contained two bugs: (a) use of int instead of size_t, and (b) it was not
1896 returning a result (though PCRE never actually uses the result).
1898 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1899 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1900 returned instead of calling malloc() with an overflowing number that would
1901 most likely cause subsequent chaos.
1903 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1905 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1906 with this option is matched, the nmatch and pmatch options of regexec() are
1907 ignored.
1909 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1910 provided in case anyone wants to the the POSIX interface with UTF-8
1911 strings.
1913 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1914 C++ linking (needed for some HP-UX environments).
1916 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1917 (unused parameter) and in the pcre_printint() function (omitted "default"
1918 switch label when the default is to do nothing).
1920 15. Added some code to make it possible, when PCRE is compiled as a C++
1921 library, to replace subject pointers for pcre_exec() with a smart pointer
1922 class, thus making it possible to process discontinuous strings.
1924 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1925 much the same function. They were added by different people who were trying
1926 to make PCRE easy to compile on non-Unix systems. It has been suggested
1927 that PCRE_EXPORT be abolished now that there is more automatic apparatus
1928 for compiling on Windows systems. I have therefore replaced it with
1929 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1930 defaults to "extern" for C or "extern C" for C++, which works fine on
1931 Unix-like systems. It is now possible to override the value of PCRE_DATA_
1932 SCOPE with something explicit in config.h. In addition:
1934 (a) pcreposix.h still had just "extern" instead of either of these macros;
1935 I have replaced it with PCRE_DATA_SCOPE.
1937 (b) Functions such as _pcre_xclass(), which are internal to the library,
1938 but external in the C sense, all had PCRE_EXPORT in their definitions.
1939 This is apparently wrong for the Windows case, so I have removed it.
1940 (It makes no difference on Unix-like systems.)
1942 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1943 of recursive calls to match(). This is different to MATCH_LIMIT because
1944 that limits the total number of calls to match(), not all of which increase
1945 the depth of recursion. Limiting the recursion depth limits the amount of
1946 stack (or heap if NO_RECURSE is set) that is used. The default can be set
1947 when PCRE is compiled, and changed at run time. A patch from Google adds
1948 this functionality to the C++ interface.
1950 18. Changes to the handling of Unicode character properties:
1952 (a) Updated the table to Unicode 4.1.0.
1954 (b) Recognize characters that are not in the table as "Cn" (undefined).
1956 (c) I revised the way the table is implemented to a much improved format
1957 which includes recognition of ranges. It now supports the ranges that
1958 are defined in UnicodeData.txt, and it also amalgamates other
1959 characters into ranges. This has reduced the number of entries in the
1960 table from around 16,000 to around 3,000, thus reducing its size
1961 considerably. I realized I did not need to use a tree structure after
1962 all - a binary chop search is just as efficient. Having reduced the
1963 number of entries, I extended their size from 6 bytes to 8 bytes to
1964 allow for more data.
1966 (d) Added support for Unicode script names via properties such as \p{Han}.
1968 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1969 matching that character.
1971 20. When matching a repeated Unicode property with a minimum greater than zero,
1972 (for example \pL{2,}), PCRE could look past the end of the subject if it
1973 reached it while seeking the minimum number of characters. This could
1974 happen only if some of the characters were more than one byte long, because
1975 there is a check for at least the minimum number of bytes.
1977 21. Refactored the implementation of \p and \P so as to be more general, to
1978 allow for more different types of property in future. This has changed the
1979 compiled form incompatibly. Anybody with saved compiled patterns that use
1980 \p or \P will have to recompile them.
1982 22. Added "Any" and "L&" to the supported property types.
1984 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1985 but give a compile time error if the value is greater than 0xff.
1987 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1988 accidentally not being installed or uninstalled.
1990 25. The pcre.h file was built from pcre.h.in, but the only changes that were
1991 made were to insert the current release number. This seemed silly, because
1992 it made things harder for people building PCRE on systems that don't run
1993 "configure". I have turned pcre.h into a distributed file, no longer built
1994 by "configure", with the version identification directly included. There is
1995 no longer a pcre.h.in file.
1997 However, this change necessitated a change to the pcre-config script as
1998 well. It is built from pcre-config.in, and one of the substitutions was the
1999 release number. I have updated configure.ac so that ./configure now finds
2000 the release number by grepping pcre.h.
2002 26. Added the ability to run the tests under valgrind.
2005 Version 6.4 05-Sep-05
2006 ---------------------
2008 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2009 "--" to be printed when multiple files were scanned, even when none of the
2010 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2011 consider it to be a bug, and have restored the previous behaviour.
2013 2. A couple of code tidies to get rid of compiler warnings.
2015 3. The pcretest program used to cheat by referring to symbols in the library
2016 whose names begin with _pcre_. These are internal symbols that are not
2017 really supposed to be visible externally, and in some environments it is
2018 possible to suppress them. The cheating is now confined to including
2019 certain files from the library's source, which is a bit cleaner.
2021 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2022 file's purpose clearer.
2024 5. Reorganized pcre_ucp_findchar().
2027 Version 6.3 15-Aug-05
2028 ---------------------
2030 1. The file libpcre.pc.in did not have general read permission in the tarball.
2032 2. There were some problems when building without C++ support:
2034 (a) If C++ support was not built, "make install" and "make test" still
2035 tried to test it.
2037 (b) There were problems when the value of CXX was explicitly set. Some
2038 changes have been made to try to fix these, and ...
2040 (c) --disable-cpp can now be used to explicitly disable C++ support.
2042 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2043 backslash in a target when C++ was disabled. This confuses some
2044 versions of "make", apparently. Using an intermediate variable solves
2045 this. (Same for CPP_LOBJ.)
2047 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2048 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2049 necessary on certain architectures.
2051 4. Added a setting of -export-symbols-regex to the link command to remove
2052 those symbols that are exported in the C sense, but actually are local
2053 within the library, and not documented. Their names all begin with
2054 "_pcre_". This is not a perfect job, because (a) we have to except some
2055 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2056 available (and never for static libraries). I have made a note to try to
2057 find a way round (a) in the future.
2060 Version 6.2 01-Aug-05
2061 ---------------------
2063 1. There was no test for integer overflow of quantifier values. A construction
2064 such as {1111111111111111} would give undefined results. What is worse, if
2065 a minimum quantifier for a parenthesized subpattern overflowed and became
2066 negative, the calculation of the memory size went wrong. This could have
2067 led to memory overwriting.
2069 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2071 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2072 operating environments where this matters.
2074 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2075 PCRE options from within the C++ wrapper.
2077 5. Named capturing subpatterns were not being correctly counted when a pattern
2078 was compiled. This caused two problems: (a) If there were more than 100
2079 such subpatterns, the calculation of the memory needed for the whole
2080 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2081 back references of the form \12, where the number was greater than 9, were
2082 not recognized as back references, even though there were sufficient
2083 previous subpatterns.
2085 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2086 versions of gcc, e.g. 2.95.4.
2089 Version 6.1 21-Jun-05
2090 ---------------------
2092 1. There was one reference to the variable "posix" in pcretest.c that was not
2093 surrounded by "#if !defined NOPOSIX".
2095 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2096 the cross-check on the old pcre_info() function, for the benefit of the
2097 cut-down version of PCRE that is currently imported into Exim.
2099 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2100 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2101 patterns that start (?i) but I don't think that matters. The use of (?i) is
2102 just an example; this all applies to the other options as well.
2104 4. Since libtool seems to echo the compile commands it is issuing, the output
2105 from "make" can be reduced a bit by putting "@" in front of each libtool
2106 compile command.
2108 5. Patch from the folks at Google for configure.in to be a bit more thorough
2109 in checking for a suitable C++ installation before trying to compile the
2110 C++ stuff. This should fix a reported problem when a compiler was present,
2111 but no suitable headers.
2113 6. The man pages all had just "PCRE" as their title. I have changed them to
2114 be the relevant file name. I have also arranged that these names are
2115 retained in the file doc/pcre.txt, which is a concatenation in text format
2116 of all the man pages except the little individual ones for each function.
2118 7. The NON-UNIX-USE file had not been updated for the different set of source
2119 files that come with release 6. I also added a few comments about the C++
2120 wrapper.
2123 Version 6.0 07-Jun-05
2124 ---------------------
2126 1. Some minor internal re-organization to help with my DFA experiments.
2128 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2129 didn't matter for the library itself when fully configured, but did matter
2130 when compiling without UCP support, or within Exim, where the ucp files are
2131 not imported.
2133 3. Refactoring of the library code to split up the various functions into
2134 different source modules. The addition of the new DFA matching code (see
2135 below) to a single monolithic source would have made it really too
2136 unwieldy, quite apart from causing all the code to be include in a
2137 statically linked application, when only some functions are used. This is
2138 relevant even without the DFA addition now that patterns can be compiled in
2139 one application and matched in another.
2141 The downside of splitting up is that there have to be some external
2142 functions and data tables that are used internally in different modules of
2143 the library but which are not part of the API. These have all had their
2144 names changed to start with "_pcre_" so that they are unlikely to clash
2145 with other external names.
2147 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2148 a different (DFA) algorithm. Although it is slower than the original
2149 function, it does have some advantages for certain types of matching
2150 problem.
2152 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2153 including restarting after a partial match.
2155 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2156 defined when compiling for Windows was sent to me. I have put it into the
2157 code, though I have no means of testing or verifying it.
2159 7. Added the pcre_refcount() auxiliary function.
2161 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2162 match before or at the first newline in the subject string. In pcretest,
2163 the /f option on a pattern can be used to set this.
2165 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2166 would behave wrongly. This has been present in PCRE since release 4.0.
2168 10. A number of changes to the pcregrep command:
2170 (a) Refactored how -x works; insert ^(...)$ instead of setting
2171 PCRE_ANCHORED and checking the length, in preparation for adding
2172 something similar for -w.
2174 (b) Added the -w (match as a word) option.
2176 (c) Refactored the way lines are read and buffered so as to have more
2177 than one at a time available.
2179 (d) Implemented a pcregrep test script.
2181 (e) Added the -M (multiline match) option. This allows patterns to match
2182 over several lines of the subject. The buffering ensures that at least
2183 8K, or the rest of the document (whichever is the shorter) is available
2184 for matching (and similarly the previous 8K for lookbehind assertions).
2186 (f) Changed the --help output so that it now says
2188 -w, --word-regex(p)
2190 instead of two lines, one with "regex" and the other with "regexp"
2191 because that confused at least one person since the short forms are the
2192 same. (This required a bit of code, as the output is generated
2193 automatically from a table. It wasn't just a text change.)
2195 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2196 option but starts with a hyphen. Could be a pattern or a path name
2197 starting with a hyphen, for instance.
2199 (h) "-" can be given as a file name to represent stdin.
2201 (i) When file names are being printed, "(standard input)" is used for
2202 the standard input, for compatibility with GNU grep. Previously
2203 "<stdin>" was used.
2205 (j) The option --label=xxx can be used to supply a name to be used for
2206 stdin when file names are being printed. There is no short form.
2208 (k) Re-factored the options decoding logic because we are going to add
2209 two more options that take data. Such options can now be given in four
2210 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2212 (l) Added the -A, -B, and -C options for requesting that lines of context
2213 around matches be printed.
2215 (m) Added the -L option to print the names of files that do not contain
2216 any matching lines, that is, the complement of -l.
2218 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2219 continue to scan other files.
2221 (o) The -s option was incorrectly implemented. For compatibility with other
2222 greps, it now suppresses the error message for a non-existent or non-
2223 accessible file (but not the return code). There is a new option called
2224 -q that suppresses the output of matching lines, which was what -s was
2225 previously doing.
2227 (p) Added --include and --exclude options to specify files for inclusion
2228 and exclusion when recursing.
2230 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2231 Hopefully, it now does.
2233 12. Missing cast in pcre_study().
2235 13. Added an "uninstall" target to the makefile.
2237 14. Replaced "extern" in the function prototypes in Makefile.in with
2238 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2239 world, but is set differently for Windows.
2241 15. Added a second compiling function called pcre_compile2(). The only
2242 difference is that it has an extra argument, which is a pointer to an
2243 integer error code. When there is a compile-time failure, this is set
2244 non-zero, in addition to the error test pointer being set to point to an
2245 error message. The new argument may be NULL if no error number is required
2246 (but then you may as well call pcre_compile(), which is now just a
2247 wrapper). This facility is provided because some applications need a
2248 numeric error indication, but it has also enabled me to tidy up the way
2249 compile-time errors are handled in the POSIX wrapper.
2251 16. Added VPATH=.libs to the makefile; this should help when building with one
2252 prefix path and installing with another. (Or so I'm told by someone who
2253 knows more about this stuff than I do.)
2255 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2256 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2257 match everything, including newlines. This is not POSIX-compatible, but
2258 somebody wanted the feature. From pcretest it can be activated by using
2259 both the P and the s flags.
2261 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2263 19. libpcre.pc was being incorrectly installed as executable.
2265 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2266 it now also looks for '\r' so that it will work unmodified on Windows.
2268 21. Added Google's contributed C++ wrapper to the distribution.
2270 22. Added some untidy missing memory free() calls in pcretest, to keep
2271 Electric Fence happy when testing.
2275 Version 5.0 13-Sep-04
2276 ---------------------
2278 1. Internal change: literal characters are no longer packed up into items
2279 containing multiple characters in a single byte-string. Each character
2280 is now matched using a separate opcode. However, there may be more than one
2281 byte in the character in UTF-8 mode.
2283 2. The pcre_callout_block structure has two new fields: pattern_position and
2284 next_item_length. These contain the offset in the pattern to the next match
2285 item, and its length, respectively.
2287 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2288 insertion of callouts before each pattern item. Added the /C option to
2289 pcretest to make use of this.
2291 4. On the advice of a Windows user, the lines
2293 #if defined(_WIN32) || defined(WIN32)
2294 _setmode( _fileno( stdout ), 0x8000 );
2295 #endif /* defined(_WIN32) || defined(WIN32) */
2297 have been added to the source of pcretest. This apparently does useful
2298 magic in relation to line terminators.
2300 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2301 for the benefit of those environments where the "b" makes a difference.
2303 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2304 to know about it. I have put a hack into configure.in that adds in code
2305 to set GCC=yes if CC=icc. This seems to end up at a point in the
2306 generated configure script that is early enough to affect the setting of
2307 compiler options, which is what is needed, but I have no means of testing
2308 whether it really works. (The user who reported this had patched the
2309 generated configure script, which of course I cannot do.)
2311 LATER: After change 22 below (new libtool files), the configure script
2312 seems to know about icc (and also ecc). Therefore, I have commented out
2313 this hack in configure.in.
2315 7. Added support for pkg-config (2 patches were sent in).
2317 8. Negated POSIX character classes that used a combination of internal tables
2318 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2319 [[:^ascii]]. Typically, they would match almost any characters. The other
2320 POSIX classes were not broken in this way.
2322 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2323 to find the match, as PCRE was deluded into thinking that the match had to
2324 start at the start point or following a newline. The same bug applied to
2325 patterns with negative forward assertions or any backward assertions
2326 preceding ".*" at the start, unless the pattern required a fixed first
2327 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2329 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2330 starting at the last subject character, bytes beyond the end of the subject
2331 string were read.
2333 11. Renamed the variable "class" as "classbits" to make life easier for C++
2334 users. (Previously there was a macro definition, but it apparently wasn't
2335 enough.)
2337 12. Added the new field "tables" to the extra data so that tables can be passed
2338 in at exec time, or the internal tables can be re-selected. This allows
2339 a compiled regex to be saved and re-used at a later time by a different
2340 program that might have everything at different addresses.
2342 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2343 -R library as well as a -L library.
2345 14. The debugging options of pcretest (-d on the command line or D on a
2346 pattern) showed incorrect output for anything following an extended class
2347 that contained multibyte characters and which was followed by a quantifier.
2349 15. Added optional support for general category Unicode character properties
2350 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2351 support. It adds about 90K to the size of the library. The meanings of the
2352 inbuilt class escapes such as \d and \s have NOT been changed.
2354 16. Updated pcredemo.c to include calls to free() to release the memory for the
2355 compiled pattern.
2357 17. The generated file chartables.c was being created in the source directory
2358 instead of in the building directory. This caused the build to fail if the
2359 source directory was different from the building directory, and was
2360 read-only.
2362 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2363 file. No doubt somebody will tell me if they don't make sense... Also added
2364 Dan Mooney's comments about building on OpenVMS.
2366 19. Added support for partial matching via the PCRE_PARTIAL option for
2367 pcre_exec() and the \P data escape in pcretest.
2369 20. Extended pcretest with 3 new pattern features:
2371 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2372 write the compiled pattern to the file whose name is "rest-of-line".
2373 This is a straight binary dump of the data, with the saved pointer to
2374 the character tables forced to be NULL. The study data, if any, is
2375 written too. After writing, pcretest reads a new pattern.
2377 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2378 compiled pattern from the given file. There must not be any
2379 occurrences of "<" in the file name (pretty unlikely); if there are,
2380 pcretest will instead treat the initial "<" as a pattern delimiter.
2381 After reading in the pattern, pcretest goes on to read data lines as
2382 usual.
2384 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2385 and 16-bit fields in a compiled pattern, to simulate a pattern that
2386 was compiled on a host of opposite endianness.
2388 21. The pcre-exec() function can now cope with patterns that were compiled on
2389 hosts of opposite endianness, with this restriction:
2391 As for any compiled expression that is saved and used later, the tables
2392 pointer field cannot be preserved; the extra_data field in the arguments
2393 to pcre_exec() should be used to pass in a tables address if a value
2394 other than the default internal tables were used at compile time.
2396 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2397 now diagnosed as an error. Previously, most of the time, a negative number
2398 would have been treated as zero, but if in addition "ovector" was passed as
2399 NULL, a crash could occur.
2401 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2402 new versions from the libtool 1.5 distribution (the last one is a copy of
2403 a file called libtool.m4). This seems to have fixed the need to patch
2404 "configure" to support Darwin 1.3 (which I used to do). However, I still
2405 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2406 workstation).
2408 24. Changed the PCRE licence to be the more standard "BSD" licence.
2411 Version 4.5 01-Dec-03
2412 ---------------------
2414 1. There has been some re-arrangement of the code for the match() function so
2415 that it can be compiled in a version that does not call itself recursively.
2416 Instead, it keeps those local variables that need separate instances for
2417 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2418 needs to "recurse". Keeping track of where control must go is done by means
2419 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2420 hide most of the details from the main code, and operates only if
2421 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2422 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2423 operating.
2425 To make it easier for callers to provide specially tailored get/free
2426 functions for this usage, two new functions, pcre_stack_malloc, and
2427 pcre_stack_free, are used. They are always called in strict stacking order,
2428 and the size of block requested is always the same.
2430 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2431 PCRE has been compiled to use the stack or the heap for recursion. The
2432 -C option of pcretest uses this to show which version is compiled.
2434 A new data escape \S, is added to pcretest; it causes the amounts of store
2435 obtained and freed by both kinds of malloc/free at match time to be added
2436 to the output.
2438 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2439 what's available on my current Linux desktop machine.
2441 3. When matching a UTF-8 string, the test for a valid string at the start has
2442 been extended. If start_offset is not zero, PCRE now checks that it points
2443 to a byte that is the start of a UTF-8 character. If not, it returns
2444 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2445 this is necessary because there may be backward assertions in the pattern.
2446 When matching the same subject several times, it may save resources to use
2447 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2449 4. The code for checking the validity of UTF-8 strings has been tightened so
2450 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2451 containing "overlong sequences".
2453 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2454 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2455 should have been "&", but it just so happened that all the cases this let
2456 through by mistake were picked up later in the function.
2458 6. I had used a variable called "isblank" - this is a C99 function, causing
2459 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2461 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2462 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2463 so the version is shown; (c) stop "make test" echoing "./RunTest".
2465 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2467 9. The prototype for memmove() for systems that don't have it was using
2468 size_t, but the inclusion of the header that defines size_t was later. I've
2469 moved the #includes for the C headers earlier to avoid this.
2471 10. Added some adjustments to the code to make it easier to compiler on certain
2472 special systems:
2474 (a) Some "const" qualifiers were missing.
2475 (b) Added the macro EXPORT before all exported functions; by default this
2476 is defined to be empty.
2477 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2478 that it reads its output file name as an argument instead of writing
2479 to the standard output and assuming this can be redirected.
2481 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2482 class containing characters with values greater than 255, PCRE compilation
2483 went into a loop.
2485 12. A recursive reference to a subpattern that was within another subpattern
2486 that had a minimum quantifier of zero caused PCRE to crash. For example,
2487 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2488 recursion. If the recursively-called subpattern itself had a zero repeat,
2489 that was OK.
2491 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2492 buffer into which it was copied (for escape processing) was still set at
2493 1024, so long lines caused crashes.
2495 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2496 "internal error: code overflow...". This applied to any character class
2497 that was followed by a possessive quantifier.
2499 15. Modified the Makefile to add libpcre.la as a prerequisite for
2500 libpcreposix.la because I was told this is needed for a parallel build to
2501 work.
2503 16. If a pattern that contained .* following optional items at the start was
2504 studied, the wrong optimizing data was generated, leading to matching
2505 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2506 matching string must start with a or b or c. The correct conclusion for
2507 this pattern is that a match can start with any character.
2510 Version 4.4 13-Aug-03
2511 ---------------------
2513 1. In UTF-8 mode, a character class containing characters with values between
2514 127 and 255 was not handled correctly if the compiled pattern was studied.
2515 In fixing this, I have also improved the studying algorithm for such
2516 classes (slightly).
2518 2. Three internal functions had redundant arguments passed to them. Removal
2519 might give a very teeny performance improvement.
2521 3. Documentation bug: the value of the capture_top field in a callout is *one
2522 more than* the number of the hightest numbered captured substring.
2524 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2525 in incorrectly linking with a previously installed version. They now link
2526 explicitly with libpcre.la.
2528 5. configure.in no longer needs to recognize Cygwin specially.
2530 6. A problem in pcre.in for Windows platforms is fixed.
2532 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2533 pcretest, it used to include the size of the study block as part of its
2534 output. Unfortunately, the structure contains a field that has a different
2535 size on different hardware architectures. This meant that the tests that
2536 showed this size failed. As the block is currently always of a fixed size,
2537 this information isn't actually particularly useful in pcretest output, so
2538 I have just removed it.
2540 8. Three pre-processor statements accidentally did not start in column 1.
2541 Sadly, there are *still* compilers around that complain, even though
2542 standard C has not required this for well over a decade. Sigh.
2544 9. In pcretest, the code for checking callouts passed small integers in the
2545 callout_data field, which is a void * field. However, some picky compilers
2546 complained about the casts involved for this on 64-bit systems. Now
2547 pcretest passes the address of the small integer instead, which should get
2548 rid of the warnings.
2550 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2551 both compile and run time, and gives an error if an invalid UTF-8 sequence
2552 is found. There is a option for disabling this check in cases where the
2553 string is known to be correct and/or the maximum performance is wanted.
2555 11. In response to a bug report, I changed one line in Makefile.in from
2557 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2558 to
2559 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2561 to look similar to other lines, but I have no way of telling whether this
2562 is the right thing to do, as I do not use Windows. No doubt I'll get told
2563 if it's wrong...
2566 Version 4.3 21-May-03
2567 ---------------------
2569 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2570 Makefile.
2572 2. Some refactoring to improve the quality of the code:
2574 (i) The utf8_table... variables are now declared "const".
2576 (ii) The code for \cx, which used the "case flipping" table to upper case
2577 lower case letters, now just substracts 32. This is ASCII-specific,
2578 but the whole concept of \cx is ASCII-specific, so it seems
2579 reasonable.
2581 (iii) PCRE was using its character types table to recognize decimal and
2582 hexadecimal digits in the pattern. This is silly, because it handles
2583 only 0-9, a-f, and A-F, but the character types table is locale-
2584 specific, which means strange things might happen. A private
2585 table is now used for this - though it costs 256 bytes, a table is
2586 much faster than multiple explicit tests. Of course, the standard
2587 character types table is still used for matching digits in subject
2588 strings against \d.
2590 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2591 ending in _t are). So I've renamed it as ESC_tee.
2593 3. The first argument for regexec() in the POSIX wrapper should have been
2594 defined as "const".
2596 4. Changed pcretest to use malloc() for its buffers so that they can be
2597 Electric Fenced for debugging.
2599 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2600 to read one or more bytes before the start of the subject string. Often this
2601 had no effect on PCRE's behaviour, but in some circumstances it could
2602 provoke a segmentation fault.
2604 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2605 to try to read one or more bytes before the start of the subject string.
2607 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2608 UTF-8 support could misbehave in various ways if the subject string
2609 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2610 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2611 back over UTF-8 characters.)
2614 Version 4.2 14-Apr-03
2615 ---------------------
2617 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2619 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2620 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2621 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2622 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2623 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2625 Note: automatic setting of the BUILD variables is not yet working
2626 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2627 compile-time but not at link-time
2628 [LINK]: use for linking executables only
2629 make different versions for Windows and non-Windows
2630 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2631 libraries
2632 [LINK_FOR_BUILD]: new variable
2633 [OBJEXT]: use throughout
2634 [EXEEXT]: use throughout
2635 <winshared>: new target
2636 <wininstall>: new target
2637 <dftables.o>: use native compiler
2638 <dftables>: use native linker
2639 <install>: handle Windows platform correctly
2640 <clean>: ditto
2641 <check>: ditto
2642 copy DLL to top builddir before testing
2644 As part of these changes, -no-undefined was removed again. This was reported
2645 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2646 in any case.
2648 3. Some tidies to get rid of compiler warnings:
2650 . In the match_data structure, match_limit was an unsigned long int, whereas
2651 match_call_count was an int. I've made them both unsigned long ints.
2653 . In pcretest the fact that a const uschar * doesn't automatically cast to
2654 a void * provoked a warning.
2656 . Turning on some more compiler warnings threw up some "shadow" variables
2657 and a few more missing casts.
2659 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2660 option, a class that contained a single character with a value between 128
2661 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2663 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2664 option, a class that contained several characters, but with at least one
2665 whose value was between 128 and 255 caused PCRE to crash.
2668 Version 4.1 12-Mar-03
2669 ---------------------
2671 1. Compiling with gcc -pedantic found a couple of places where casts were
2672 needed, and a string in dftables.c that was longer than standard compilers are
2673 required to support.
2675 2. Compiling with Sun's compiler found a few more places where the code could
2676 be tidied up in order to avoid warnings.
2678 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2679 first of these names is deprecated in the latest Autoconf in favour of the name
2680 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2681 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2682 analogy I have changed it to CFLAGS_FOR_BUILD.
2684 4. Added -no-undefined to the linking command in the Makefile, because this is
2685 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2686 linking step for the pcreposix library.
2688 5. PCRE was failing to diagnose the case of two named groups with the same
2689 name.
2691 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2692 literal character that is needed in the subject for a match, and scans along to
2693 ensure that it is present before embarking on the full matching process. This
2694 saves time in cases of nested unlimited repeats that are never going to match.
2695 Problem: the scan can take a lot of time if the subject is very long (e.g.
2696 megabytes), thus penalizing straightforward matches. It is now done only if the
2697 amount of subject to be scanned is less than 1000 bytes.
2699 7. A lesser problem with the same optimization is that it was recording the
2700 first character of an anchored pattern as "needed", thus provoking a search
2701 right along the subject, even when the first match of the pattern was going to
2702 fail. The "needed" character is now not set for anchored patterns, unless it
2703 follows something in the pattern that is of non-fixed length. Thus, it still
2704 fulfils its original purpose of finding quick non-matches in cases of nested
2705 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2708 Version 4.0 17-Feb-03
2709 ---------------------
2711 1. If a comment in an extended regex that started immediately after a meta-item
2712 extended to the end of string, PCRE compiled incorrect data. This could lead to
2713 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2715 2. Moved to autoconf 2.53 and libtool 1.4.2.
2717 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2718 the special perltest8 script is no longer needed - all the tests can be run
2719 from a single perltest script.
2721 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2722 by \s. It has now been removed in PCRE. This means it isn't recognized as
2723 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2724 class [:space:] *does* include VT, thereby creating a mess.
2726 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2727 space and tab.
2729 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2730 its new features into the main test script, reducing the number of scripts.
2732 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2733 were backward compatible, and made the (?i) apply to the whole pattern, as if
2734 /i were given. Now it behaves more logically, and applies the option setting
2735 only to what follows. PCRE has been changed to follow suit. However, if it
2736 finds options settings right at the start of the pattern, it extracts them into
2737 the global options, as before. Thus, they show up in the info data.
2739 8. Added support for the \Q...\E escape sequence. Characters in between are
2740 treated as literals. This is slightly different from Perl in that $ and @ are
2741 also handled as literals inside the quotes. In Perl, they will cause variable
2742 interpolation. Note the following examples:
2744 Pattern PCRE matches Perl matches
2746 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2747 \Qabc\$xyz\E abc\$xyz abc\$xyz
2748 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2750 For compatibility with Perl, \Q...\E sequences are recognized inside character
2751 classes as well as outside them.
2753 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2754 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2755 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2756 signed/unsigned warnings.
2758 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2759 option for pcretest, so I've replaced it by a simple function that does just
2760 that job.
2762 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2763 "pcregrep -".
2765 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2766 Java package. This provides some syntactic sugar for simple cases of what my
2767 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2768 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2769 item, you can use this simplified notation. Note that only makes sense with
2770 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2771 greediness, whatever the setting of the PCRE_UNGREEDY option.
2773 13. A change of greediness default within a pattern was not taking effect at
2774 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2775 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2776 was abstracted outside.
2778 14. PCRE now supports the \G assertion. It is true when the current matching
2779 position is at the start point of the match. This differs from \A when the
2780 starting offset is non-zero. Used with the /g option of pcretest (or similar
2781 code), it works in the same way as it does for Perl's /g option. If all
2782 alternatives of a regex begin with \G, the expression is anchored to the start
2783 match position, and the "anchored" flag is set in the compiled expression.
2785 15. Some bugs concerning the handling of certain option changes within patterns
2786 have been fixed. These applied to options other than (?ims). For example,
2787 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2788 been the other way round. Some of this was related to change 7 above.
2790 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2791 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2792 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2793 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2795 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2796 mode. Unlike ".", it always matches newline, whatever the setting of
2797 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2798 assertions. Perl allows it, but it doesn't (in general) work because it can't
2799 calculate the length of the lookbehind. At least, that's the case for Perl
2800 5.8.0 - I've been told they are going to document that it doesn't work in
2801 future.
2803 18. Added an error diagnosis for escapes that PCRE does not support: these are
2804 \L, \l, \N, \P, \p, \U, \u, and \X.
2806 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2807 reading past the end of the pattern in cases such as /[abcd/.
2809 20. PCRE was getting more memory than necessary for patterns with classes that
2810 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2812 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2813 compiling PCRE for use with Virtual Pascal.
2815 22. Small fix to the Makefile to make it work properly if the build is done
2816 outside the source tree.
2818 23. Added a new extension: a condition to go with recursion. If a conditional
2819 subpattern starts with (?(R) the "true" branch is used if recursion has
2820 happened, whereas the "false" branch is used only at the top level.
2822 24. When there was a very long string of literal characters (over 255 bytes
2823 without UTF support, over 250 bytes with UTF support), the computation of how
2824 much memory was required could be incorrect, leading to segfaults or other
2825 strange effects.
2827 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2828 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2829 there was a subsequent back reference to those brackets. This meant that, for
2830 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2831 possible to check for precisely this case. All we can do is abandon the
2832 optimization if .* occurs inside capturing brackets when there are any back
2833 references whatsoever. (See below for a better fix that came later.)
2835 26. The handling of the optimization for finding the first character of a
2836 non-anchored pattern, and for finding a character that is required later in the
2837 match were failing in some cases. This didn't break the matching; it just
2838 failed to optimize when it could. The way this is done has been re-implemented.
2840 27. Fixed typo in error message for invalid (?R item (it said "(?p").
2842 28. Added a new feature that provides some of the functionality that Perl
2843 provides with (?{...}). The facility is termed a "callout". The way it is done
2844 in PCRE is for the caller to provide an optional function, by setting
2845 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2846 global variable. By default it is unset, which disables all calling out. To get
2847 the function called, the regex must include (?C) at appropriate points. This
2848 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2849 This provides a means of identifying different callout points. When PCRE
2850 reaches such a point in the regex, if pcre_callout has been set, the external
2851 function is called. It is provided with data in a structure called
2852 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2853 matching continues; if it returns a non-zero value, the match at the current
2854 point fails. However, backtracking will occur if possible. [This was changed
2855 later and other features added - see item 49 below.]
2857 29. pcretest is upgraded to test the callout functionality. It provides a
2858 callout function that displays information. By default, it shows the start of
2859 the match and the current position in the text. There are some new data escapes
2860 to vary what happens:
2862 \C+ in addition, show current contents of captured substrings
2863 \C- do not supply a callout function
2864 \C!n return 1 when callout number n is reached
2865 \C!n!m return 1 when callout number n is reached for the mth time
2867 30. If pcregrep was called with the -l option and just a single file name, it
2868 output "<stdin>" if a match was found, instead of the file name.
2870 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2871 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2872 pcre_exec(). This saves a malloc/free per call. The default value of
2873 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2874 when configuring.
2876 32. The default maximum size of a compiled pattern is 64K. There have been a
2877 few cases of people hitting this limit. The code now uses macros to handle the
2878 storing of links as offsets within the compiled pattern. It defaults to 2-byte
2879 links, but this can be changed to 3 or 4 bytes by --with-link-size when
2880 configuring. Tests 2 and 5 work only with 2-byte links because they output
2881 debugging information about compiled patterns.
2883 33. Internal code re-arrangements:
2885 (a) Moved the debugging function for printing out a compiled regex into
2886 its own source file (printint.c) and used #include to pull it into
2887 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2888 separate copies.
2890 (b) Defined the list of op-code names for debugging as a macro in
2891 internal.h so that it is next to the definition of the opcodes.
2893 (c) Defined a table of op-code lengths for simpler skipping along compiled
2894 code. This is again a macro in internal.h so that it is next to the
2895 definition of the opcodes.
2897 34. Added support for recursive calls to individual subpatterns, along the
2898 lines of Robin Houston's patch (but implemented somewhat differently).
2900 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2901 allow it to read and process whole directories in Win32. This code was
2902 contributed by Lionel Fourquaux; it has not been tested by me.
2904 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2905 used to name a group. Names consist of alphanumerics and underscores, and must
2906 be unique. Back references use the syntax (?P=name) and recursive calls use
2907 (?P>name) which is a PCRE extension to the Python extension. Groups still have
2908 numbers. The function pcre_fullinfo() can be used after compilation to extract
2909 a name/number map. There are three relevant calls:
2911 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2912 PCRE_INFO_NAMECOUNT yields the number of entries
2913 PCRE_INFO_NAMETABLE yields a pointer to the map.
2915 The map is a vector of fixed-size entries. The size of each entry depends on
2916 the length of the longest name used. The first two bytes of each entry are the
2917 group number, most significant byte first. There follows the corresponding
2918 name, zero terminated. The names are in alphabetical order.
2920 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2921 case instead of 255. Making it the same both with and without UTF-8 support
2922 means that the same test output works with both.
2924 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2925 calling malloc() with a zero argument.
2927 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2928 optimization. I've improved things by keeping a bitmap of backreferences with
2929 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2930 fact referenced, the optimization can be applied. It is unlikely that a
2931 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2932 the match to follow \n) will appear inside brackets with a number greater than
2933 31, but if it does, any back reference > 31 suppresses the optimization.
2935 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2936 of disabling numbered capturing parentheses. Any opening parenthesis that is
2937 not followed by ? behaves as if it were followed by ?: but named parentheses
2938 can still be used for capturing (and they will acquire numbers in the usual
2939 way).
2941 41. Redesigned the return codes from the match() function into yes/no/error so
2942 that errors can be passed back from deep inside the nested calls. A malloc
2943 failure while inside a recursive subpattern call now causes the
2944 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2946 42. It is now possible to set a limit on the number of times the match()
2947 function is called in a call to pcre_exec(). This facility makes it possible to
2948 limit the amount of recursion and backtracking, though not in a directly
2949 obvious way, because the match() function is used in a number of different
2950 circumstances. The count starts from zero for each position in the subject
2951 string (for non-anchored patterns). The default limit is, for compatibility, a
2952 large number, namely 10 000 000. You can change this in two ways:
2954 (a) When configuring PCRE before making, you can use --with-match-limit=n
2955 to set a default value for the compiled library.
2957 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2958 a different value is set. See 45 below.
2960 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2962 43. Added a new function pcre_config(int, void *) to enable run-time extraction
2963 of things that can be changed at compile time. The first argument specifies
2964 what is wanted and the second points to where the information is to be placed.
2965 The current list of available information is:
2969 The output is an integer that is set to one if UTF-8 support is available;
2970 otherwise it is set to zero.
2974 The output is an integer that it set to the value of the code that is used for
2975 newline. It is either LF (10) or CR (13).
2979 The output is an integer that contains the number of bytes used for internal
2980 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2984 The output is an integer that contains the threshold above which the POSIX
2985 interface uses malloc() for output vectors. See item 31 above.
2989 The output is an unsigned integer that contains the default limit of the number
2990 of match() calls in a pcre_exec() execution. See 42 above.
2992 44. pcretest has been upgraded by the addition of the -C option. This causes it
2993 to extract all the available output from the new pcre_config() function, and to
2994 output it. The program then exits immediately.
2996 45. A need has arisen to pass over additional data with calls to pcre_exec() in
2997 order to support additional features. One way would have been to define
2998 pcre_exec2() (for example) with extra arguments, but this would not have been
2999 extensible, and would also have required all calls to the original function to
3000 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3001 is used for passing in "extra" data from pcre_study().
3003 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3004 contains the following fields:
3006 flags a bitmap indicating which of the following fields are set
3007 study_data opaque data from pcre_study()
3008 match_limit a way of specifying a limit on match() calls for a specific
3009 call to pcre_exec()
3010 callout_data data for callouts (see 49 below)
3012 The flag bits are also defined in pcre.h, and are
3018 The pcre_study() function now returns one of these new pcre_extra blocks, with
3019 the actual study data pointed to by the study_data field, and the
3020 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3021 before. That is, this change is entirely upwards-compatible and requires no
3022 change to existing code.
3024 If you want to pass in additional data to pcre_exec(), you can either place it
3025 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3026 block.
3028 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3029 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3030 times with different match limits, until it finds the minimum value needed for
3031 pcre_exec() to complete. The value is then output. This can be instructive; for
3032 most simple matches the number is quite small, but for pathological cases it
3033 gets very large very quickly.
3035 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3036 returns the size of the data block pointed to by the study_data field in a
3037 pcre_extra block, that is, the value that was passed as the argument to
3038 pcre_malloc() when PCRE was getting memory in which to place the information
3039 created by pcre_study(). The fourth argument should point to a size_t variable.
3040 pcretest has been extended so that this information is shown after a successful
3041 pcre_study() call when information about the compiled regex is being displayed.
3043 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3044 because what follows is always an absolute path. (Later: it turns out that this
3045 is more than cosmetic for MinGW, because it doesn't like empty path
3046 components.)
3048 49. Some changes have been made to the callout feature (see 28 above):
3050 (i) A callout function now has three choices for what it returns:
3052 0 => success, carry on matching
3053 > 0 => failure at this point, but backtrack if possible
3054 < 0 => serious error, return this value from pcre_exec()
3056 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3057 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3058 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3059 use by callout functions. It will never be used by PCRE itself.
3061 (ii) The pcre_extra structure (see 45 above) has a void * field called
3062 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3063 pcre_callout_block structure has a field of the same name. The contents of
3064 the field passed in the pcre_extra structure are passed to the callout
3065 function in the corresponding field in the callout block. This makes it
3066 easier to use the same callout-containing regex from multiple threads. For
3067 testing, the pcretest program has a new data escape
3069 \C*n pass the number n (may be negative) as callout_data
3071 If the callout function in pcretest receives a non-zero value as
3072 callout_data, it returns that value.
3074 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3075 there were some redundant $(CFLAGS) in commands that are now specified as
3076 $(LINK), which already includes $(CFLAGS).
3078 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3079 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3080 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3081 one-byte characters throughout. Note that case-insensitive matching applies
3082 only to characters whose values are less than 256. PCRE doesn't support the
3083 notion of cases for higher-valued characters.
3085 (i) A character class whose characters are all within 0-255 is handled as
3086 a bit map, and the map is inverted for negative classes. Previously, a
3087 character > 255 always failed to match such a class; however it should
3088 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3090 (ii) A negated character class with a single character < 255 is coded as
3091 "not this character" (OP_NOT). This wasn't working properly when the test
3092 character was multibyte, either singly or repeated.
3094 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3095 mode, for example: \x{100}{2,3}.
3097 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3098 singly or repeated) now correctly test multibyte characters. However,
3099 PCRE doesn't recognize any characters with values greater than 255 as
3100 digits, spaces, or word characters. Such characters always match \D, \S,
3101 and \W, and never match \d, \s, or \w.
3103 (v) Classes may now contain characters and character ranges with values
3104 greater than 255. For example: [ab\x{100}-\x{400}].
3106 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3107 PCRE in UTF-8 mode.
3109 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3110 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3111 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3112 value.)
3114 53. The single man page has become too large. I have therefore split it up into
3115 a number of separate man pages. These also give rise to individual HTML pages;
3116 these are now put in a separate directory, and there is an index.html page that
3117 lists them all. Some hyperlinking between the pages has been installed.
3119 54. Added convenience functions for handling named capturing parentheses.
3121 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3122 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3123 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3124 are faulted.
3126 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3127 calling configure. These values are used when compiling the dftables.c program
3128 which is run to generate the source of the default character tables. They
3129 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3130 you will need to set these values.
3132 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3135 Version 3.9 02-Jan-02
3136 ---------------------
3138 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3140 2. If --disable-static was given, the building process failed when trying to
3141 build pcretest and pcregrep. (For some reason it was using libtool to compile
3142 them, which is not right, as they aren't part of the library.)
3145 Version 3.8 18-Dec-01
3146 ---------------------
3148 1. The experimental UTF-8 code was completely screwed up. It was packing the
3149 bytes in the wrong order. How dumb can you get?
3152 Version 3.7 29-Oct-01
3153 ---------------------
3155 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3156 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3157 this didn't happen under Solaris 8, where I normally test things.
3159 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3160 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3161 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3162 this fix an hour or so after the initial 3.7 release.)
3165 Version 3.6 23-Oct-01
3166 ---------------------
3168 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3169 offsets passed as NULL with zero offset count.
3171 2. The config.guess and config.sub files had not been updated when I moved to
3172 the latest autoconf.
3175 Version 3.5 15-Aug-01
3176 ---------------------
3178 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3179 had been forgotten.
3181 2. By using declared but undefined structures, we can avoid using "void"
3182 definitions in pcre.h while keeping the internal definitions of the structures
3183 private.
3185 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3186 user point of view, this means that both static and shared libraries are built
3187 by default, but this can be individually controlled. More of the work of
3188 handling this static/shared cases is now inside libtool instead of PCRE's make
3189 file.
3191 4. The pcretest utility is now installed along with pcregrep because it is
3192 useful for users (to test regexs) and by doing this, it automatically gets
3193 relinked by libtool. The documentation has been turned into a man page, so
3194 there are now .1, .txt, and .html versions in /doc.
3196 5. Upgrades to pcregrep:
3197 (i) Added long-form option names like gnu grep.
3198 (ii) Added --help to list all options with an explanatory phrase.
3199 (iii) Added -r, --recursive to recurse into sub-directories.
3200 (iv) Added -f, --file to read patterns from a file.
3202 6. pcre_exec() was referring to its "code" argument before testing that
3203 argument for NULL (and giving an error if it was NULL).
3205 7. Upgraded Makefile.in to allow for compiling in a different directory from
3206 the source directory.
3208 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3209 options bits, the pointer it was passed was to an int instead of to an unsigned
3210 long int. This mattered only on 64-bit systems.
3212 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3213 generated) instead of pcre.in, which it its source. Also made the same change
3214 in several of the .c files.
3216 10. A new release of gcc defines printf() as a macro, which broke pcretest
3217 because it had an ifdef in the middle of a string argument for printf(). Fixed
3218 by using separate calls to printf().
3220 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3221 script, to force use of CR or LF instead of \n in the source. On non-Unix
3222 systems, the value can be set in config.h.
3224 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3225 absolute limit. Changed the text of the error message to make this clear, and
3226 likewise updated the man page.
3228 13. The limit of 99 on the number of capturing subpatterns has been removed.
3229 The new limit is 65535, which I hope will not be a "real" limit.
3232 Version 3.4 22-Aug-00
3233 ---------------------
3235 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3237 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3240 Version 3.3 01-Aug-00
3241 ---------------------
3243 1. If an octal character was given, but the value was greater than \377, it
3244 was not getting masked to the least significant bits, as documented. This could
3245 lead to crashes in some systems.
3247 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3248 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3250 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3251 These just pass their arguments on to (pcre_free)(), but they are provided
3252 because some uses of PCRE bind it to non-C systems that can call its functions,
3253 but cannot call free() or pcre_free() directly.
3255 4. Add "make test" as a synonym for "make check". Corrected some comments in
3256 the Makefile.
3258 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3259 Makefile.
3261 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3262 command called pgrep for grepping around the active processes.
3264 7. Added the beginnings of support for UTF-8 character strings.
3266 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3267 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3268 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3269 out for the ar command.)
3272 Version 3.2 12-May-00
3273 ---------------------
3275 This is purely a bug fixing release.
3277 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3278 of ZA. This was just one example of several cases that could provoke this bug,
3279 which was introduced by change 9 of version 2.00. The code for breaking
3280 infinite loops after an iteration that matches an empty string was't working
3281 correctly.
3283 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3284 when matched against abbab (for example). After matching an empty string, it
3285 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3286 caused it to match further down the string than it should.
3288 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3289 was there because it doesn't seem to be needed, and it causes trouble on some
3290 systems, as it is not a Standard C header. It has been removed.
3292 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3293 were reported on the Macintosh. The changes were from
3295 while ((c = *(++ptr)) != 0 && c != '\n');
3296 to
3297 while ((c = *(++ptr)) != 0 && c != '\n') ;
3299 Totally extraordinary, but if that's what it takes...
3301 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3302 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3303 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3304 assumes the way PCRE uses memmove() (always moving upwards).
3306 6. PCRE is being used in one environment where strchr() is not available. There
3307 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3308 faster code anyway.
3311 Version 3.1 09-Feb-00
3312 ---------------------
3314 The only change in this release is the fixing of some bugs in Makefile.in for
3315 the "install" target:
3317 (1) It was failing to install pcreposix.h.
3319 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3322 Version 3.0 01-Feb-00
3323 ---------------------
3325 1. Add support for the /+ modifier to perltest (to output $` like it does in
3326 pcretest).
3328 2. Add support for the /g modifier to perltest.
3330 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3331 matches null strings.
3333 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3334 pattern. Perl treats empty patterns specially - it reuses the most recent
3335 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3336 effect.
3338 5. The POSIX interface was broken in that it was just handing over the POSIX
3339 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3340 required a bigger vector, with some working space on the end. This means that
3341 the POSIX wrapper now has to get and free some memory, and copy the results.
3343 6. Added some simple autoconf support, placing the test data and the
3344 documentation in separate directories, re-organizing some of the
3345 information files, and making it build pcre-config (a GNU standard). Also added
3346 libtool support for building PCRE as a shared library, which is now the
3347 default.
3349 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3350 09 are not valid octal constants. Single digits will be used for minor values
3351 less than 10.
3353 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3354 existing programs that set these in the POSIX interface can use PCRE without
3355 modification.
3357 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3358 return all that pcre_info() returns, plus additional data. The pcre_info()
3359 function is retained for compatibility, but is considered to be obsolete.
3361 10. Added experimental recursion feature (?R) to handle one common case that
3362 Perl 5.6 will be able to do with (?p{...}).
3364 11. Added support for POSIX character classes like [:alpha:], which Perl is
3365 adopting.
3368 Version 2.08 31-Aug-99
3369 ----------------------
3371 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3372 trying to match at the startoffset position, but instead was moving forward to
3373 the next newline as if a previous match had failed.
3375 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3376 and could get into a loop if a null string was matched other than at the start
3377 of the subject.
3379 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3380 be distinguished at compile time, and for completeness also added PCRE_DATE.
3382 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3383 in GnuWin32 environments.
3386 Version 2.07 29-Jul-99
3387 ----------------------
3389 1. The documentation is now supplied in plain text form and HTML as well as in
3390 the form of man page sources.
3392 2. C++ compilers don't like assigning (void *) values to other pointer types.
3393 In particular this affects malloc(). Although there is no problem in Standard
3394 C, I've put in casts to keep C++ compilers happy.
3396 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3397 should be (const char *).
3399 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3400 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3401 However, I haven't made this a standard facility. The documentation doesn't
3402 mention it, and the Makefile doesn't support it.
3404 5. The Makefile now contains an "install" target, with editable destinations at
3405 the top of the file. The pcretest program is not installed.
3407 6. pgrep -V now gives the PCRE version number and date.
3409 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3410 causing the entire string to be ignored, instead of just the last character.
3412 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3413 non-matching string, it can take a very, very long time, even for strings of
3414 quite modest length, because of the nested recursion. PCRE now does better in
3415 some of these cases. It does this by remembering the last required literal
3416 character in the pattern, and pre-searching the subject to ensure it is present
3417 before running the real match. In other words, it applies a heuristic to detect
3418 some types of certain failure quickly, and in the above example, if presented
3419 with a string that has no trailing " it gives "no match" very quickly.
3421 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3422 other alternatives are tried instead.
3425 Version 2.06 09-Jun-99
3426 ----------------------
3428 1. Change pcretest's output for amount of store used to show just the code
3429 space, because the remainder (the data block) varies in size between 32-bit and
3430 64-bit systems.
3432 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3433 start matching at. This allows lookbehinds to work when searching for multiple
3434 occurrences in a string.
3436 3. Added additional options to pcretest for testing multiple occurrences:
3438 /+ outputs the rest of the string that follows a match
3439 /g loops for multiple occurrences, using the new startoffset argument
3440 /G loops for multiple occurrences by passing an incremented pointer
3442 4. PCRE wasn't doing the "first character" optimization for patterns starting
3443 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3444 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3445 the letter 'x'. On long subject strings, this gives a significant speed-up.
3448 Version 2.05 21-Apr-99
3449 ----------------------
3451 1. Changed the type of magic_number from int to long int so that it works
3452 properly on 16-bit systems.
3454 2. Fixed a bug which caused patterns starting with .* not to work correctly
3455 when the subject string contained newline characters. PCRE was assuming
3456 anchoring for such patterns in all cases, which is not correct because .* will
3457 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3458 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3459 must be retried after every newline in the subject.
3462 Version 2.04 18-Feb-99
3463 ----------------------
3465 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3466 computation of the store needed to hold the pattern was incorrect (too large).
3467 If such patterns were nested a few deep, this could multiply and become a real
3468 problem.
3470 2. Added /M option to pcretest to show the memory requirement of a specific
3471 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3473 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3474 compiled in such a way that the backtracking after subsequent failure was
3475 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3476 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3479 Version 2.03 02-Feb-99
3480 ----------------------
3482 1. Fixed typo and small mistake in man page.
3484 2. Added 4th condition (GPL supersedes if conflict) and created separate
3485 LICENCE file containing the conditions.
3487 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3488 Perl, that is the internal \ allows the delimiter to be included in the
3489 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3490 the final delimiter, add \ to the end of the pattern (to test the error).
3492 4. Added the convenience functions for extracting substrings after a successful
3493 match. Updated pcretest to make it able to test these functions.
3496 Version 2.02 14-Jan-99
3497 ----------------------
3499 1. Initialized the working variables associated with each extraction so that
3500 their saving and restoring doesn't refer to uninitialized store.
3502 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3503 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3504 fix the problem.
3506 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3507 calls, and wasn't printing the correct value for compiling calls. Increased the
3508 default value of LOOPREPEAT, and the number of significant figures in the
3509 times.
3511 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3513 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3514 a building problem on Windows NT with a FAT file system.
3517 Version 2.01 21-Oct-98
3518 ----------------------
3520 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3521 to character tables built by pcre_maketables() in the current locale. If NULL
3522 is passed, the default tables are used.
3525 Version 2.00 24-Sep-98
3526 ----------------------
3528 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3529 it any more.
3531 2. Allow quantification of (?>) groups, and make it work correctly.
3533 3. The first character computation wasn't working for (?>) groups.
3535 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3536 end of the subject) and add 5.005's \z, which really does match only at the
3537 very end of the subject.
3539 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3541 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3542 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3543 localized options. All options to pcre_study() were also removed.
3545 7. Add other new features from 5.005:
3547 $(?<= positive lookbehind
3548 $(?<! negative lookbehind
3549 (?imsx-imsx) added the unsetting capability
3550 such a setting is global if at outer level; local otherwise
3551 (?imsx-imsx:) non-capturing groups with option setting
3552 (?(cond)re|re) conditional pattern matching
3554 A backreference to itself in a repeated group matches the previous
3555 captured string.
3557 8. General tidying up of studying (both automatic and via "study")
3558 consequential on the addition of new assertions.
3560 9. As in 5.005, unlimited repeated groups that could match an empty substring
3561 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3562 runtime if any iteration does actually match an empty substring.
3564 10. Include the RunTest script in the distribution.
3566 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3567 discrepancies, some of which were old and were also with respect to 5.004. They
3568 have now been fixed.
3571 Version 1.09 28-Apr-98
3572 ----------------------
3574 1. A negated single character class followed by a quantifier with a minimum
3575 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3576 program crashes, or just wrong answers. This did not apply to negated classes
3577 containing more than one character, or to minima other than one.
3580 Version 1.08 27-Mar-98
3581 ----------------------
3583 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3585 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3586 latter must appear before anything that relies on it in the pattern.
3589 Version 1.07 16-Feb-98
3590 ----------------------
3592 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3593 repeat of a potentially empty string).
3596 Version 1.06 23-Jan-98
3597 ----------------------
3599 1. Added Markus Oberhumer's little patches for C++.
3601 2. Literal strings longer than 255 characters were broken.
3604 Version 1.05 23-Dec-97
3605 ----------------------
3607 1. Negated character classes containing more than one character were failing if
3608 PCRE_CASELESS was set at run time.
3611 Version 1.04 19-Dec-97
3612 ----------------------
3614 1. Corrected the man page, where some "const" qualifiers had been omitted.
3616 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3617 input syntax.
3619 3. Fixed memory leak which occurred when a regex with back references was
3620 matched with an offsets vector that wasn't big enough. The temporary memory
3621 that is used in this case wasn't being freed if the match failed.
3623 4. Tidied pcretest to ensure it frees memory that it gets.
3625 5. Temporary memory was being obtained in the case where the passed offsets
3626 vector was exactly big enough.
3628 6. Corrected definition of offsetof() from change 5 below.
3630 7. I had screwed up change 6 below and broken the rules for the use of
3631 setjmp(). Now fixed.
3634 Version 1.03 18-Dec-97
3635 ----------------------
3637 1. A erroneous regex with a missing opening parenthesis was correctly
3638 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3639 on some systems.
3641 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3642 it was reported that one broken compiler failed on the former because "code" is
3643 also an independent variable.
3645 3. The erroneous regex a[]b caused an array overrun reference.
3647 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3648 fail on data ending with that character. (It was going on too far, and checking
3649 the next character, typically a binary zero.) This was specific to the
3650 optimized code for single-character negative classes.
3652 5. Added a contributed patch from the TIN world which does the following:
3654 + Add an undef for memmove, in case the the system defines a macro for it.
3656 + Add a definition of offsetof(), in case there isn't one. (I don't know
3657 the reason behind this - offsetof() is part of the ANSI standard - but
3658 it does no harm).
3660 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3661 most of the places where whitespace preceded '#'. I have given up and
3662 allowed the remaining 2 cases to be at the margin.
3664 + Rename some variables in pcre to eliminate shadowing. This seems very
3665 pedantic, but does no harm, of course.
3667 6. Moved the call to setjmp() into its own function, to get rid of warnings
3668 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3670 7. Constructs such as \d{8,} were compiling into the equivalent of
3671 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3672 outcome, but in this particular case used more store than had been allocated,
3673 which caused the bug to be discovered because it threw up an internal error.
3675 8. The debugging code in both pcre and pcretest for outputting the compiled
3676 form of a regex was going wrong in the case of back references followed by
3677 curly-bracketed repeats.
3680 Version 1.02 12-Dec-97
3681 ----------------------
3683 1. Typos in pcre.3 and comments in the source fixed.
3685 2. Applied a contributed patch to get rid of places where it used to remove
3686 'const' from variables, and fixed some signed/unsigned and uninitialized
3687 variable warnings.
3689 3. Added the "runtest" target to Makefile.
3691 4. Set default compiler flag to -O2 rather than just -O.
3694 Version 1.01 19-Nov-97
3695 ----------------------
3697 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3698 like /([ab]*)*/, that is, for classes with more than one character in them.
3700 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3701 as /((?>a*))*/ (a PCRE_EXTRA facility).
3704 Version 1.00 18-Nov-97
3705 ----------------------
3707 1. Added compile-time macros to support systems such as SunOS4 which don't have
3708 memmove() or strerror() but have other things that can be used instead.
3710 2. Arranged that "make clean" removes the executables.
3713 Version 0.99 27-Oct-97
3714 ----------------------
3716 1. Fixed bug in code for optimizing classes with only one character. It was
3717 initializing a 32-byte map regardless, which could cause it to run off the end
3718 of the memory it had got.
3720 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3723 Version 0.98 22-Oct-97
3724 ----------------------
3726 1. Fixed bug in code for handling temporary memory usage when there are more
3727 back references than supplied space in the ovector. This could cause segfaults.
3730 Version 0.97 21-Oct-97
3731 ----------------------
3733 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3735 2. Optimized negated single characters not to use a bit map.
3737 3. Brought error texts together as macro definitions; clarified some of them;
3738 fixed one that was wrong - it said "range out of order" when it meant "invalid
3739 escape sequence".
3741 4. Changed some char * arguments to const char *.
3743 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3745 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3746 pcretest.
3749 Version 0.96 16-Oct-97
3750 ----------------------
3752 1. Added a simple "pgrep" utility to the distribution.
3754 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3755 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3756 where "ddd" means "one or more decimal digits".
3758 3. Fixed serious bug. If a pattern had a back reference, but the call to
3759 pcre_exec() didn't supply a large enough ovector to record the related
3760 identifying subpattern, the match always failed. PCRE now remembers the number
3761 of the largest back reference, and gets some temporary memory in which to save
3762 the offsets during matching if necessary, in order to ensure that
3763 backreferences always work.
3765 4. Increased the compatibility with Perl in a number of ways:
3767 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3768 to request this handling. The option can be set at compile or exec time.
3770 (b) $ matches before a terminating newline by default; an option
3771 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3772 mode). The option can be set at compile or exec time.
3774 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3775 the same as Perl's. If the decimal number it represents is less than 10
3776 or there aren't that many previous left capturing parentheses, an octal
3777 escape is read. Inside a character class, it's always an octal escape,
3778 even if it is a single digit.
3780 (d) An escaped but undefined alphabetic character is taken as a literal,
3781 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3782 escapes.
3784 (e) {0} is now permitted. (The previous item is removed from the compiled
3785 pattern).
3787 5. Changed all the names of code files so that the basic parts are no longer
3788 than 10 characters, and abolished the teeny "globals.c" file.
3790 6. Changed the handling of character classes; they are now done with a 32-byte
3791 bit map always.
3793 7. Added the -d and /D options to pcretest to make it possible to look at the
3794 internals of compilation without having to recompile pcre.
3797 Version 0.95 23-Sep-97
3798 ----------------------
3800 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3801 \x20 at the start of a run of normal characters. These were being treated as
3802 real characters, instead of the source characters being re-checked.
3805 Version 0.94 18-Sep-97
3806 ----------------------
3808 1. The functions are now thread-safe, with the caveat that the global variables
3809 containing pointers to malloc() and free() or alternative functions are the
3810 same for all threads.
3812 2. Get pcre_study() to generate a bitmap of initial characters for non-
3813 anchored patterns when this is possible, and use it if passed to pcre_exec().
3816 Version 0.93 15-Sep-97
3817 ----------------------
3819 1. /(b)|(:+)/ was computing an incorrect first character.
3821 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3822 but not actually doing anything yet.
3824 3. Treat "-" characters in classes that cannot be part of ranges as literals,
3825 as Perl does (e.g. [-az] or [az-]).
3827 4. Set the anchored flag if a branch starts with .* or .*? because that tests
3828 all possible positions.
3830 5. Split up into different modules to avoid including unneeded functions in a
3831 compiled binary. However, compile and exec are still in one module. The "study"
3832 function is split off.
3834 6. The character tables are now in a separate module whose source is generated
3835 by an auxiliary program - but can then be edited by hand if required. There are
3836 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3837 toupper() in the code.
3839 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3840 make them global. Abolish the function for setting them, as the caller can now
3841 set them directly.
3844 Version 0.92 11-Sep-97
3845 ----------------------
3847 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3848 (e.g. /a{1,3}/) was broken (I mis-optimized it).
3850 2. Caseless matching was not working in character classes if the characters in
3851 the pattern were in upper case.
3853 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3855 4. Make PCRE_ANCHORED public and accept as a compile option.
3857 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3858 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3859 pass them.
3861 6. Give an error if bad option bits passed at compile or run time.
3863 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
3864 pcretest to cause it to pass that flag.
3866 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
3867 options, and the first character, if set.
3869 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
3872 Version 0.91 10-Sep-97
3873 ----------------------
3875 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
3876 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
3878 2. PCRE was looping on encountering an indefinitely repeated back reference to
3879 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
3880 Perl does - treats the match as successful.
3882 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5