ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 573 - (show annotations)
Fri Nov 19 19:29:44 2010 UTC (10 years, 11 months ago) by ph10
File size: 195520 byte(s)
Remove loops from GETCHAR etc. macros.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.11 10-Oct-2010
5 ------------------------
7 1. (*THEN) was not working properly if there were untried alternatives prior
8 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
9 backtracked to try for "b" instead of moving to the next alternative branch
10 at the same level (in this case, to look for "c"). The Perl documentation
11 is clear that when (*THEN) is backtracked onto, it goes to the "next
12 alternative in the innermost enclosing group".
14 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
15 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
16 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
17 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
18 (*THEN).
20 3. If \s appeared in a character class, it removed the VT character from
21 the class, even if it had been included by some previous item, for example
22 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
23 of \s, but is part of the POSIX "space" class.)
25 4. A partial match never returns an empty string (because you can always
26 match an empty string at the end of the subject); however the checking for
27 an empty string was starting at the "start of match" point. This has been
28 changed to the "earliest inspected character" point, because the returned
29 data for a partial match starts at this character. This means that, for
30 example, /(?<=abc)def/ gives a partial match for the subject "abc"
31 (previously it gave "no match").
33 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
34 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
35 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
36 has an implication that the given string is incomplete (because a partial
37 match is preferred over a full match). For this reason, these items now
38 give a partial match in this situation. [Aside: previously, the one case
39 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
40 match rather than a full match, which was wrong by the old rules, but is
41 now correct.]
43 6. There was a bug in the handling of #-introduced comments, recognized when
44 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
45 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
46 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
47 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
48 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
49 places in pcre_compile().
51 7. Related to (6) above, when pcre_compile() was skipping #-introduced
52 comments when looking ahead for named forward references to subpatterns,
53 the only newline sequence it recognized was NL. It now handles newlines
54 according to the set newline convention.
56 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
57 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
58 cater for a lack of strerror(). These oversights have been fixed.
60 9. Added --match-limit and --recursion-limit to pcregrep.
62 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
64 11. When the -o option was used, pcregrep was setting a return code of 1, even
65 when matches were found, and --line-buffered was not being honoured.
67 12. Added an optional parentheses number to the -o and --only-matching options
68 of pcregrep.
70 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
71 can match an empty string. The code to do it in pcretest and pcredemo
72 needed fixing:
74 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
75 only one byte after an empty string match just before CRLF (this case
76 just got forgotten; "any" and "anycrlf" were OK).
78 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
79 mode when an empty string match preceded an ASCII character followed by
80 a non-ASCII character. (The code for advancing by one character rather
81 than one byte was nonsense.)
83 (c) The pcredemo.c sample program did not have any code at all to handle
84 the cases when CRLF is a valid newline sequence.
86 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
87 as a starting offset was within the subject string. There is now a new
88 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
89 negative or greater than the length of the string. In order to test this,
90 pcretest is extended to allow the setting of negative starting offsets.
92 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
93 starting offset points to the beginning of a UTF-8 character was
94 unnecessarily clumsy. I tidied it up.
96 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
97 bad UTF-8 sequence and one that is incomplete.
99 17. Nobody had reported that the --include_dir option, which was added in
100 release 7.7 should have been called --include-dir (hyphen, not underscore)
101 for compatibility with GNU grep. I have changed it to --include-dir, but
102 left --include_dir as an undocumented synonym, and the same for
103 --exclude-dir, though that is not available in GNU grep, at least as of
104 release 2.5.4.
106 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
107 characters from a string of bytes) have been redefined so as not to use
108 loops, in order to improve performance in some environments. At the same
109 time, I abstracted some of the common code into auxiliary macros to save
110 repetition (this should not affect the compiled code).
113 Version 8.10 25-Jun-2010
114 ------------------------
116 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
117 THEN.
119 2. (*ACCEPT) was not working when inside an atomic group.
121 3. Inside a character class, \B is treated as a literal by default, but
122 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
123 causes the error). The code is unchanged, but I tidied the documentation.
125 4. Inside a character class, PCRE always treated \R and \X as literals,
126 whereas Perl faults them if its -w option is set. I have changed PCRE so
127 that it faults them when PCRE_EXTRA is set.
129 5. Added support for \N, which always matches any character other than
130 newline. (It is the same as "." when PCRE_DOTALL is not set.)
132 6. When compiling pcregrep with newer versions of gcc which may have
133 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
134 declared with attribute warn_unused_result" were given. Just casting the
135 result to (void) does not stop the warnings; a more elaborate fudge is
136 needed. I've used a macro to implement this.
138 7. Minor change to pcretest.c to avoid a compiler warning.
140 8. Added four artifical Unicode properties to help with an option to make
141 \s etc use properties (see next item). The new properties are: Xan
142 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
144 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
145 use Unicode properties. (*UCP) at the start of a pattern can be used to set
146 this option. Modified pcretest to add /W to test this facility. Added
147 REG_UCP to make it available via the POSIX interface.
149 10. Added --line-buffered to pcregrep.
151 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
152 studied, and the match started with a letter with a code point greater than
153 127 whose first byte was different to the first byte of the other case of
154 the letter, the other case of this starting letter was not recognized
155 (#976).
157 12. If a pattern that was studied started with a repeated Unicode property
158 test, for example, \p{Nd}+, there was the theoretical possibility of
159 setting up an incorrect bitmap of starting bytes, but fortunately it could
160 not have actually happened in practice until change 8 above was made (it
161 added property types that matched character-matching opcodes).
163 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
164 possible starting bytes for non-anchored patterns.
166 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
167 \R, and also a number of cases that involve Unicode properties, both
168 explicit and implicit when PCRE_UCP is set.
170 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
171 input, it could crash or give wrong results if characters with values
172 greater than 0xc0 were present in the subject string. (Detail: it assumed
173 UTF-8 input when processing these items.)
175 16. Added a lot of (int) casts to avoid compiler warnings in systems where
176 size_t is 64-bit (#991).
178 17. Added a check for running out of memory when PCRE is compiled with
179 --disable-stack-for-recursion (#990).
181 18. If the last data line in a file for pcretest does not have a newline on
182 the end, a newline was missing in the output.
184 19. The default pcre_chartables.c file recognizes only ASCII characters (values
185 less than 128) in its various bitmaps. However, there is a facility for
186 generating tables according to the current locale when PCRE is compiled. It
187 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
188 characters, are recognized by isspace() and therefore were getting set in
189 these tables, and indeed these tables seem to approximate to ISO 8859. This
190 caused a problem in UTF-8 mode when pcre_study() was used to create a list
191 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
192 which of course cannot start UTF-8 characters. I have changed the code so
193 that only real ASCII characters (less than 128) and the correct starting
194 bytes for UTF-8 encodings are set for characters greater than 127 when in
195 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
196 altogether.)
198 20. Added the /T option to pcretest so as to be able to run tests with non-
199 standard character tables, thus making it possible to include the tests
200 used for 19 above in the standard set of tests.
202 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
203 reference to a subpattern the other side of a comment that contains an
204 opening parenthesis caused either an internal compiling error, or a
205 reference to the wrong subpattern.
208 Version 8.02 19-Mar-2010
209 ------------------------
211 1. The Unicode data tables have been updated to Unicode 5.2.0.
213 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
214 configured.
216 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
217 original author of that file, following a query about its status.
219 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
220 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
222 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
223 quantifier applied to a forward-referencing subroutine call, could compile
224 incorrect code or give the error "internal error: previously-checked
225 referenced subpattern not found".
227 6. Both MS Visual Studio and Symbian OS have problems with initializing
228 variables to point to external functions. For these systems, therefore,
229 pcre_malloc etc. are now initialized to local functions that call the
230 relevant global functions.
232 7. There were two entries missing in the vectors called coptable and poptable
233 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
234 I've fixed the data, and added a kludgy way of testing at compile time that
235 the lengths are correct (equal to the number of opcodes).
237 8. Following on from 7, I added a similar kludge to check the length of the
238 eint vector in pcreposix.c.
240 9. Error texts for pcre_compile() are held as one long string to avoid too
241 much relocation at load time. To find a text, the string is searched,
242 counting zeros. There was no check for running off the end of the string,
243 which could happen if a new error number was added without updating the
244 string.
246 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
248 11. \K was not working if it appeared in an atomic group or in a group that
249 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
250 \K is "not well defined" if used in an assertion. PCRE now accepts it if
251 the assertion is positive, but not if it is negative.
253 12. Change 11 fortuitously reduced the size of the stack frame used in the
254 "match()" function of pcre_exec.c by one pointer. Forthcoming
255 implementation of support for (*MARK) will need an extra pointer on the
256 stack; I have reserved it now, so that the stack frame size does not
257 decrease.
259 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
260 item in branch that calls a recursion is a subroutine call - as in the
261 second branch in the above example - was incorrectly given the compile-
262 time error "recursive call could loop indefinitely" because pcre_compile()
263 was not correctly checking the subroutine for matching a non-empty string.
265 14. The checks for overrunning compiling workspace could trigger after an
266 overrun had occurred. This is a "should never occur" error, but it can be
267 triggered by pathological patterns such as hundreds of nested parentheses.
268 The checks now trigger 100 bytes before the end of the workspace.
270 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
273 Version 8.01 19-Jan-2010
274 ------------------------
276 1. If a pattern contained a conditional subpattern with only one branch (in
277 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
278 computed the wrong minimum data length (which is of course zero for such
279 subpatterns). This could cause incorrect "no match" results.
281 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
282 the pattern is reset in the first branch, pcre_compile() failed with
283 "internal error: code overflow at offset...". This happened only when
284 the reset was to the original external option setting. (An optimization
285 abstracts leading options settings into an external setting, which was the
286 cause of this.)
288 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
289 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
290 assertion pattern did not match (meaning that the assertion was true), it
291 was incorrectly treated as false if the SKIP had been reached during the
292 matching. This also applied to assertions used as conditions.
294 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
295 assertion subpattern, including such a pattern used as a condition,
296 unpredictable results occurred, instead of the error return
299 5. The C++ GlobalReplace function was not working like Perl for the special
300 situation when an empty string is matched. It now does the fancy magic
301 stuff that is necessary.
303 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
304 removed. (These were left over from very, very early versions of PCRE.)
306 7. Some cosmetic changes to the code to make life easier when compiling it
307 as part of something else:
309 (a) Change DEBUG to PCRE_DEBUG.
311 (b) In pcre_compile(), rename the member of the "branch_chain" structure
312 called "current" as "current_branch", to prevent a collision with the
313 Linux macro when compiled as a kernel module.
315 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
316 prevent a collision with the Linux macro when compiled as a kernel
317 module.
319 8. In pcre_compile() there are some checks for integer overflows that used to
320 cast potentially large values to (double). This has been changed to that
321 when building, a check for int64_t is made, and if it is found, it is used
322 instead, thus avoiding the use of floating point arithmetic. (There is no
323 other use of FP in PCRE.) If int64_t is not found, the fallback is to
324 double.
326 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
327 2005 (difference between two addresses compared to an unsigned value).
329 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
330 custom one, because of the following reported problem in Windows:
332 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
333 under Win32.
334 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
335 therefore missing the function definition.
336 - The compiler thus generates a "C" signature for the test function.
337 - The linker fails to find the "C" function.
338 - PCRE fails to configure if asked to do so against libbz2.
340 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
341 messages were output:
343 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
344 rerunning libtoolize, to keep the correct libtool macros in-tree.
345 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
347 I have done both of these things.
349 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
350 most of the time, it *can* run out if it is given a pattern that contains a
351 runaway infinite recursion. I updated the discussion in the pcrestack man
352 page.
354 13. Now that we have gone to the x.xx style of version numbers, the minor
355 version may start with zero. Using 08 or 09 is a bad idea because users
356 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
357 interpreted as invalid octal numbers. I've updated the previous comment in
358 configure.ac, and also added a check that gives an error if 08 or 09 are
359 used.
361 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
362 causing partial matching to fail when the end of the subject matched \W
363 in a UTF-8 pattern where \W was quantified with a minimum of 3.
365 15. There were some discrepancies between the declarations in pcre_internal.h
366 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
367 their definitions. The declarations used "const uschar *" and the
368 definitions used USPTR. Even though USPTR is normally defined as "const
369 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
370 reported that: "This difference in casting confuses some C++ compilers, for
371 example, SunCC recognizes above declarations as different functions and
372 generates broken code for hbpcre." I have changed the declarations to use
373 USPTR.
375 16. GNU libtool is named differently on some systems. The autogen.sh script now
376 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
377 (FreeBSD).
379 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
380 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
381 comment: "Figure out how to create a longlong from a string: strtoll and
382 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
383 instance, but it only takes 2 args instead of 3!"
385 18. A subtle bug concerned with back references has been fixed by a change of
386 specification, with a corresponding code fix. A pattern such as
387 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
388 refers, was giving matches when it shouldn't. For example, xa=xaaa would
389 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
390 same bug. Such groups have to be quantified to be useful, or contained
391 inside another quantified group. (If there's no repetition, the reference
392 can never match.) The problem arises because, having left the group and
393 moved on to the rest of the pattern, a later failure that backtracks into
394 the group uses the captured value from the final iteration of the group
395 rather than the correct earlier one. I have fixed this in PCRE by forcing
396 any group that contains a reference to itself to be an atomic group; that
397 is, there cannot be any backtracking into it once it has completed. This is
398 similar to recursive and subroutine calls.
401 Version 8.00 19-Oct-09
402 ----------------------
404 1. The table for translating pcre_compile() error codes into POSIX error codes
405 was out-of-date, and there was no check on the pcre_compile() error code
406 being within the table. This could lead to an OK return being given in
407 error.
409 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
410 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
411 in a Windows environment.
413 3. The pcregrep --count option prints the count for each file even when it is
414 zero, as does GNU grep. However, pcregrep was also printing all files when
415 --files-with-matches was added. Now, when both options are given, it prints
416 counts only for those files that have at least one match. (GNU grep just
417 prints the file name in this circumstance, but including the count seems
418 more useful - otherwise, why use --count?) Also ensured that the
419 combination -clh just lists non-zero counts, with no names.
421 4. The long form of the pcregrep -F option was incorrectly implemented as
422 --fixed_strings instead of --fixed-strings. This is an incompatible change,
423 but it seems right to fix it, and I didn't think it was worth preserving
424 the old behaviour.
426 5. The command line items --regex=pattern and --regexp=pattern were not
427 recognized by pcregrep, which required --regex pattern or --regexp pattern
428 (with a space rather than an '='). The man page documented the '=' forms,
429 which are compatible with GNU grep; these now work.
431 6. No libpcreposix.pc file was created for pkg-config; there was just
432 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
434 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
435 when UCP support is not needed, by modifying the Python script that
436 generates it from Unicode data files. This should not matter if the module
437 is correctly used as a library, but I received one complaint about 50K of
438 unwanted data. My guess is that the person linked everything into his
439 program rather than using a library. Anyway, it does no harm.
441 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
442 was a minimum greater than 1 for a wide character in a possessive
443 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
444 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
445 character. Chaos in the form of incorrect output or a compiling loop could
446 result.
448 9. The restrictions on what a pattern can contain when partial matching is
449 requested for pcre_exec() have been removed. All patterns can now be
450 partially matched by this function. In addition, if there are at least two
451 slots in the offset vector, the offset of the earliest inspected character
452 for the match and the offset of the end of the subject are set in them when
453 PCRE_ERROR_PARTIAL is returned.
455 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
456 synonymous with PCRE_PARTIAL, for backwards compatibility, and
457 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
458 and may be more useful for multi-segment matching.
460 11. Partial matching with pcre_exec() is now more intuitive. A partial match
461 used to be given if ever the end of the subject was reached; now it is
462 given only if matching could not proceed because another character was
463 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
464 string "Z", which now yields "no match" instead of "partial match". In the
465 case of pcre_dfa_exec(), "no match" is given if every matching path for the
466 final character ended with (*FAIL).
468 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
469 if the pattern had a "must contain" character that was already found in the
470 earlier partial match, unless partial matching was again requested. For
471 example, with the pattern /dog.(body)?/, the "must contain" character is
472 "g". If the first part-match was for the string "dog", restarting with
473 "sbody" failed. This bug has been fixed.
475 13. The string returned by pcre_dfa_exec() after a partial match has been
476 changed so that it starts at the first inspected character rather than the
477 first character of the match. This makes a difference only if the pattern
478 starts with a lookbehind assertion or \b or \B (\K is not supported by
479 pcre_dfa_exec()). It's an incompatible change, but it makes the two
480 matching functions compatible, and I think it's the right thing to do.
482 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
483 so that the demonstration program is easily available in environments where
484 PCRE has not been installed from source.
486 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
487 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
488 library.
490 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
491 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
492 is not the first non-POSIX option to be added. Clearly some people find
493 these options useful.
495 17. If a caller to the POSIX matching function regexec() passes a non-zero
496 value for nmatch with a NULL value for pmatch, the value of
497 nmatch is forced to zero.
499 18. RunGrepTest did not have a test for the availability of the -u option of
500 the diff command, as RunTest does. It now checks in the same way as
501 RunTest, and also checks for the -b option.
503 19. If an odd number of negated classes containing just a single character
504 interposed, within parentheses, between a forward reference to a named
505 subpattern and the definition of the subpattern, compilation crashed with
506 an internal error, complaining that it could not find the referenced
507 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
508 [The bug was that it was starting one character too far in when skipping
509 over the character class, thus treating the ] as data rather than
510 terminating the class. This meant it could skip too much.]
512 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
513 /g option in pcretest when the pattern contains \K, which makes it possible
514 to have an empty string match not at the start, even when the pattern is
515 anchored. Updated pcretest and pcredemo to use this option.
517 21. If the maximum number of capturing subpatterns in a recursion was greater
518 than the maximum at the outer level, the higher number was returned, but
519 with unset values at the outer level. The correct (outer level) value is
520 now given.
522 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
523 PCRE did not set those parentheses (unlike Perl). I have now found a way to
524 make it do so. The string so far is captured, making this feature
525 compatible with Perl.
527 23. The tests have been re-organized, adding tests 11 and 12, to make it
528 possible to check the Perl 5.10 features against Perl 5.10.
530 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
531 pattern matches a fixed length string. PCRE did not allow this; now it
532 does. Neither allows recursion.
534 25. I finally figured out how to implement a request to provide the minimum
535 length of subject string that was needed in order to match a given pattern.
536 (It was back references and recursion that I had previously got hung up
537 on.) This code has now been added to pcre_study(); it finds a lower bound
538 to the length of subject needed. It is not necessarily the greatest lower
539 bound, but using it to avoid searching strings that are too short does give
540 some useful speed-ups. The value is available to calling programs via
541 pcre_fullinfo().
543 26. While implementing 25, I discovered to my embarrassment that pcretest had
544 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
545 study optimizations had never been tested with that matching function.
546 Oops. What is worse, even when it was passed study data, there was a bug in
547 pcre_dfa_exec() that meant it never actually used it. Double oops. There
548 were also very few tests of studied patterns with pcre_dfa_exec().
550 27. If (?| is used to create subpatterns with duplicate numbers, they are now
551 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
552 on the other side of the coin, they are no longer allowed to have different
553 names, because these cannot be distinguished in PCRE, and this has caused
554 confusion. (This is a difference from Perl.)
556 28. When duplicate subpattern names are present (necessarily with different
557 numbers, as required by 27 above), and a test is made by name in a
558 conditional pattern, either for a subpattern having been matched, or for
559 recursion in such a pattern, all the associated numbered subpatterns are
560 tested, and the overall condition is true if the condition is true for any
561 one of them. This is the way Perl works, and is also more like the way
562 testing by number works.
565 Version 7.9 11-Apr-09
566 ---------------------
568 1. When building with support for bzlib/zlib (pcregrep) and/or readline
569 (pcretest), all targets were linked against these libraries. This included
570 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
571 libraries. This caused unwanted dependencies to be created. This problem
572 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
573 pcretest is linked with readline.
575 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
576 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
577 moved outside it again, because FALSE and TRUE are already defined in AIX,
578 but BOOL is not.
580 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
581 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
583 4. The pcregrep documentation said spaces were inserted as well as colons (or
584 hyphens) following file names and line numbers when outputting matching
585 lines. This is not true; no spaces are inserted. I have also clarified the
586 wording for the --colour (or --color) option.
588 5. In pcregrep, when --colour was used with -o, the list of matching strings
589 was not coloured; this is different to GNU grep, so I have changed it to be
590 the same.
592 6. When --colo(u)r was used in pcregrep, only the first matching substring in
593 each matching line was coloured. Now it goes on to look for further matches
594 of any of the test patterns, which is the same behaviour as GNU grep.
596 7. A pattern that could match an empty string could cause pcregrep to loop; it
597 doesn't make sense to accept an empty string match in pcregrep, so I have
598 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
599 seems to be how GNU grep behaves.
601 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
602 start or after a newline", because the conditional assertion was not being
603 correctly handled. The rule now is that both the assertion and what follows
604 in the first alternative must satisfy the test.
606 9. If auto-callout was enabled in a pattern with a conditional group whose
607 condition was an assertion, PCRE could crash during matching, both with
608 pcre_exec() and pcre_dfa_exec().
610 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
611 used for matching.
613 11. Unicode property support in character classes was not working for
614 characters (bytes) greater than 127 when not in UTF-8 mode.
616 12. Added the -M command line option to pcretest.
618 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
620 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
622 16. Added comments and documentation about mis-use of no_arg in the C++
623 wrapper.
625 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
626 from Martin Jerabek that uses macro names for all relevant character and
627 string constants.
629 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
630 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
631 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
632 these, but not everybody uses configure.
634 19. A conditional group that had only one branch was not being correctly
635 recognized as an item that could match an empty string. This meant that an
636 enclosing group might also not be so recognized, causing infinite looping
637 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
638 with the subject "ab", where knowledge that the repeated group can match
639 nothing is needed in order to break the loop.
641 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
642 exec(), but without supplying a callout function, matching went wrong.
644 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
645 leak if the size of the offset vector was greater than 30. When the vector
646 is smaller, the saved offsets during recursion go onto a local stack
647 vector, but for larger vectors malloc() is used. It was failing to free
648 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
649 error, in fact).
651 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
652 heapframe that is used only when UTF-8 support is enabled. This caused no
653 problem, but was untidy.
655 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
656 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
657 included within another project.
659 24. Steven Van Ingelgem's patches to add more options to the CMake support,
660 slightly modified by me:
662 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
663 not building pcregrep.
665 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
666 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
668 25. Forward references, both numeric and by name, in patterns that made use of
669 duplicate group numbers, could behave incorrectly or give incorrect errors,
670 because when scanning forward to find the reference group, PCRE was not
671 taking into account the duplicate group numbers. A pattern such as
672 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
674 26. Changed a few more instances of "const unsigned char *" to USPTR, making
675 the feature of a custom pointer more persuasive (as requested by a user).
677 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
678 pcretest.c, inside #ifndefs, because it seems they are sometimes already
679 pre-defined.
681 28. Added support for (*UTF8) at the start of a pattern.
683 29. Arrange for flags added by the "release type" setting in CMake to be shown
684 in the configuration summary.
687 Version 7.8 05-Sep-08
688 ---------------------
690 1. Replaced UCP searching code with optimized version as implemented for Ad
691 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
692 stage table and inline lookup instead of a function, giving speed ups of 2
693 to 5 times on some simple patterns that I tested. Permission was given to
694 distribute the MultiStage2.py script that generates the tables (it's not in
695 the tarball, but is in the Subversion repository).
697 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
698 scripts.
700 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
701 a group with a zero qualifier. The result of the study could be incorrect,
702 or the function might crash, depending on the pattern.
704 4. Caseless matching was not working for non-ASCII characters in back
705 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
706 It now works when Unicode Property Support is available.
708 5. In pcretest, an escape such as \x{de} in the data was always generating
709 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
710 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
711 truncation.
713 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
715 7. Added two (int) casts to pcregrep when printing the difference of two
716 pointers, in case they are 64-bit values.
718 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
719 test 2 if it fails.
721 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
722 and a #define of that name to empty if it is not externally set. This is to
723 allow users of MSVC to set it if necessary.
725 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
726 the convenience functions in the pcre_get.c source file.
728 11. An option change at the start of a pattern that had top-level alternatives
729 could cause overwriting and/or a crash. This command provoked a crash in
730 some environments:
732 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
734 This potential security problem was recorded as CVE-2008-2371.
736 12. For a pattern where the match had to start at the beginning or immediately
737 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
738 pcre_dfa_exec() could read past the end of the passed subject if there was
739 no match. To help with detecting such bugs (e.g. with valgrind), I modified
740 pcretest so that it places the subject at the end of its malloc-ed buffer.
742 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
743 exec() might read past the end of the data buffer in UTF-8 mode.
745 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
746 the data contained the byte 0x85 as part of a UTF-8 character within its
747 first line. This applied both to normal and DFA matching.
749 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
750 /^[^d]*?$/8 failed to match "abc".
752 16. Added a missing copyright notice to pcrecpp_internal.h.
754 17. Make it more clear in the documentation that values returned from
755 pcre_exec() in ovector are byte offsets, not character counts.
757 18. Tidied a few places to stop certain compilers from issuing warnings.
759 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
760 supplied by Stefan Weber. I made a further small update for 7.8 because
761 there is a change of source arrangements: the pcre_searchfuncs.c module is
762 replaced by pcre_ucd.c.
765 Version 7.7 07-May-08
766 ---------------------
768 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
769 a string to a long long, pretend we don't even have a long long." This is
770 done by checking for the strtoq, strtoll, and _strtoi64 functions.
772 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
773 pre-7.6 versions, which defined a global no_arg variable instead of putting
774 it in the RE class. (See also #8 below.)
776 3. Remove a line of dead code, identified by coverity and reported by Nuno
777 Lopes.
779 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
781 (1) The include/exclude patterns were being applied to the whole pathnames
782 of files, instead of just to the final components.
784 (2) If there was more than one level of directory, the subdirectories were
785 skipped unless they satisfied the include/exclude conditions. This is
786 inconsistent with GNU grep (and could even be seen as contrary to the
787 pcregrep specification - which I improved to make it absolutely clear).
788 The action now is always to scan all levels of directory, and just
789 apply the include/exclude patterns to regular files.
791 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
792 --exclude_dir in the tests to avoid scanning .svn directories.
794 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
795 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
796 doesn't support NULs in patterns.
798 7. Added some missing "const"s to declarations of static tables in
799 pcre_compile.c and pcre_dfa_exec.c.
801 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
802 caused by fix #2 above. (Subsequently also a second patch to fix the
803 first patch. And a third patch - this was a messy problem.)
805 9. Applied Craig's patch to remove the use of push_back().
807 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
808 matching function regexec().
810 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
811 which, however, unlike Perl's \g{...}, are subroutine calls, not back
812 references. PCRE supports relative numbers with this syntax (I don't think
813 Oniguruma does).
815 12. Previously, a group with a zero repeat such as (...){0} was completely
816 omitted from the compiled regex. However, this means that if the group
817 was called as a subroutine from elsewhere in the pattern, things went wrong
818 (an internal error was given). Such groups are now left in the compiled
819 pattern, with a new opcode that causes them to be skipped at execution
820 time.
822 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
823 to the way PCRE behaves:
825 (a) A lone ] character is dis-allowed (Perl treats it as data).
827 (b) A back reference to an unmatched subpattern matches an empty string
828 (Perl fails the current match path).
830 (c) A data ] in a character class must be notated as \] because if the
831 first data character in a class is ], it defines an empty class. (In
832 Perl it is not possible to have an empty class.) The empty class []
833 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
834 The negative empty class [^] matches any one character, independently
835 of the DOTALL setting.
837 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
838 non-existent subpattern following a character class starting with ']' and
839 containing () gave an internal compiling error instead of "reference to
840 non-existent subpattern". Fortunately, when the pattern did exist, the
841 compiled code was correct. (When scanning forwards to check for the
842 existencd of the subpattern, it was treating the data ']' as terminating
843 the class, so got the count wrong. When actually compiling, the reference
844 was subsequently set up correctly.)
846 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
847 it was being rejected as not supported by pcre_dfa_exec(), even though
848 other assertions are supported. I have made pcre_dfa_exec() support
849 (*FAIL).
851 16. The implementation of 13c above involved the invention of a new opcode,
852 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
853 cannot be changed at match time, I realized I could make a small
854 improvement to matching performance by compiling OP_ALLANY instead of
855 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
856 on the OP_ANY path.
858 17. Compiling pcretest on Windows with readline support failed without the
859 following two fixes: (1) Make the unistd.h include conditional on
860 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
862 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
863 ncurses library to be included for pcretest when ReadLine support is
864 requested, but also to allow for it to be overridden. This patch came from
865 Daniel Bergström.
867 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
868 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
869 any errors with the current Unicode tables. Thanks to Peter Kankowski for
870 spotting this.
873 Version 7.6 28-Jan-08
874 ---------------------
876 1. A character class containing a very large number of characters with
877 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
878 overflow.
880 2. Patch to cut out the "long long" test in pcrecpp_unittest when
881 HAVE_LONG_LONG is not defined.
883 3. Applied Christian Ehrlicher's patch to update the CMake build files to
884 bring them up to date and include new features. This patch includes:
886 - Fixed PH's badly added libz and libbz2 support.
887 - Fixed a problem with static linking.
888 - Added pcredemo. [But later removed - see 7 below.]
889 - Fixed dftables problem and added an option.
890 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
892 - Added readline support for pcretest.
893 - Added an listing of the option settings after cmake has run.
895 4. A user submitted a patch to Makefile that makes it easy to create
896 "pcre.dll" under mingw when using Configure/Make. I added stuff to
897 Makefile.am that cause it to include this special target, without
898 affecting anything else. Note that the same mingw target plus all
899 the other distribution libraries and programs are now supported
900 when configuring with CMake (see 6 below) instead of with
901 Configure/Make.
903 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
904 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
905 exported in the Windows port". It has not yet been confirmed that the patch
906 solves the problem, but it does no harm.
908 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
909 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
910 with CMake, and also correct the comment about stack recursion.
912 7. Remove the automatic building of pcredemo from the ./configure system and
913 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
914 of a program that users should build themselves after PCRE is installed, so
915 building it automatically is not really right. What is more, it gave
916 trouble in some build environments.
918 8. Further tidies to CMakeLists.txt from Sheri and Christian.
921 Version 7.5 10-Jan-08
922 ---------------------
924 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
925 values in parens when parsing an RE using the C++ wrapper."
927 2. Negative specials like \S did not work in character classes in UTF-8 mode.
928 Characters greater than 255 were excluded from the class instead of being
929 included.
931 3. The same bug as (2) above applied to negated POSIX classes such as
932 [:^space:].
934 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
935 defined or documented. It seems to have been a typo for PCRE_STATIC, so
936 I have changed it.
938 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
939 first named subpattern) and a construct such as (?&a) would reference the
940 first named subpattern whose name started with "a" (in other words, the
941 length check was missing). Both these problems are fixed. "Subpattern name
942 expected" is now given for (?&) (a zero-length name), and this patch also
943 makes it give the same error for \k'' (previously it complained that that
944 was a reference to a non-existent subpattern).
946 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
947 this is right because (?- can be followed by option settings as well as by
948 digits. I have, however, made the messages clearer.
950 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
951 than the number used in the conditional) now cause a compile-time error.
952 This is actually not compatible with Perl, which accepts such patterns, but
953 treats the conditional as always being FALSE (as PCRE used to), but it
954 seems to me that giving a diagnostic is better.
956 8. Change "alphameric" to the more common word "alphanumeric" in comments
957 and messages.
959 9. Fix two occurrences of "backslash" in comments that should have been
960 "backspace".
962 10. Remove two redundant lines of code that can never be obeyed (their function
963 was moved elsewhere).
965 11. The program that makes PCRE's Unicode character property table had a bug
966 which caused it to generate incorrect table entries for sequences of
967 characters that have the same character type, but are in different scripts.
968 It amalgamated them into a single range, with the script of the first of
969 them. In other words, some characters were in the wrong script. There were
970 thirteen such cases, affecting characters in the following ranges:
972 U+002b0 - U+002c1
973 U+0060c - U+0060d
974 U+0061e - U+00612
975 U+0064b - U+0065e
976 U+0074d - U+0076d
977 U+01800 - U+01805
978 U+01d00 - U+01d77
979 U+01d9b - U+01dbf
980 U+0200b - U+0200f
981 U+030fc - U+030fe
982 U+03260 - U+0327f
983 U+0fb46 - U+0fbb1
984 U+10450 - U+1049d
986 12. The -o option (show only the matching part of a line) for pcregrep was not
987 compatible with GNU grep in that, if there was more than one match in a
988 line, it showed only the first of them. It now behaves in the same way as
989 GNU grep.
991 13. If the -o and -v options were combined for pcregrep, it printed a blank
992 line for every non-matching line. GNU grep prints nothing, and pcregrep now
993 does the same. The return code can be used to tell if there were any
994 non-matching lines.
996 14. Added --file-offsets and --line-offsets to pcregrep.
998 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
999 infinitely looping recursion. The bug was that positive lookaheads were not
1000 being skipped when checking for a possible empty match (negative lookaheads
1001 and both kinds of lookbehind were skipped).
1003 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1004 inclusion of <windows.h> to before rather than after the definition of
1005 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1007 17. Specifying a possessive quantifier with a specific limit for a Unicode
1008 character property caused pcre_compile() to compile bad code, which led at
1009 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1010 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1011 caused the error; without that there was no problem.
1013 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1015 19. Added --enable-pcretest-libreadline.
1017 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1018 RE::GlobalReplace(). As a result, the number of replacements returned was
1019 double what it should be. I removed one of the increments, but Craig sent a
1020 later patch that removed the other one (the right fix) and added unit tests
1021 that check the return values (which was not done before).
1023 21. Several CMake things:
1025 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1026 the names libpcre and libpcreposix, not just pcre and pcreposix.
1028 (2) The above change means that pcretest and pcregrep are now correctly
1029 linked with the newly-built libraries, not previously installed ones.
1033 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1034 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1035 UTF-8 newline character). The key issue is that the pattern starts .*;
1036 this means that the match must be either at the beginning, or after a
1037 newline. The bug was in the code for advancing after a failed match and
1038 checking that the new position followed a newline. It was not taking
1039 account of UTF-8 characters correctly.
1041 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1042 character classes. PCRE was not treating the sequence [:...:] as a
1043 character class unless the ... were all letters. Perl, however, seems to
1044 allow any characters between [: and :], though of course it rejects as
1045 unknown any "names" that contain non-letters, because all the known class
1046 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1047 for example, whereas PCRE did not - it did not recognize a POSIX character
1048 class. This seemed a bit dangerous, so the code has been changed to be
1049 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1050 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1051 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1052 Perl does, and where it didn't before.
1054 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1055 Windows environments %n is disabled by default.
1058 Version 7.4 21-Sep-07
1059 ---------------------
1061 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1062 means that a class such as [\s] counted as "explicit reference to CR or
1063 LF". That isn't really right - the whole point of the change was to try to
1064 help when there was an actual mention of one of the two characters. So now
1065 the change happens only if \r or \n (or a literal CR or LF) character is
1066 encountered.
1068 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1069 of both had grown to the point where there were only 3 bits left.
1070 Fortunately, there was spare space in the data structure, and so I have
1071 moved the internal flags into a new 16-bit field to free up more option
1072 bits.
1074 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1075 but did not set the internal JCHANGED flag - either of these is enough to
1076 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1077 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1078 start sets both bits.
1080 4. Added options (at build time, compile time, exec time) to change \R from
1081 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1083 5. doc/pcresyntax.html was missing from the distribution.
1085 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1086 compatibility, even though it is no longer used.
1088 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1089 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1090 windows.h file is present (where different names are used). [This was
1091 reversed later after testing - see 16 below.]
1093 8. Changed all #include <config.h> to #include "config.h". There were also
1094 some further <pcre.h> cases that I changed to "pcre.h".
1096 9. When pcregrep was used with the --colour option, it missed the line ending
1097 sequence off the lines that it output.
1099 10. It was pointed out to me that arrays of string pointers cause lots of
1100 relocations when a shared library is dynamically loaded. A technique of
1101 using a single long string with a table of offsets can drastically reduce
1102 these. I have refactored PCRE in four places to do this. The result is
1103 dramatic:
1105 Originally: 290
1106 After changing UCP table: 187
1107 After changing error message table: 43
1108 After changing table of "verbs" 36
1109 After changing table of Posix names 22
1111 Thanks to the folks working on Gregex for glib for this insight.
1113 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1114 unicode-properties was also set.
1116 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1118 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1119 checked only for CRLF.
1121 14. Added casts to pcretest.c to avoid compiler warnings.
1123 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1125 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1126 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1127 entirely. This removes changes made in 7 above.
1129 17. The CMake files have been updated, and there is now more information about
1130 building with CMake in the NON-UNIX-USE document.
1133 Version 7.3 28-Aug-07
1134 ---------------------
1136 1. In the rejigging of the build system that eventually resulted in 7.1, the
1137 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1138 brackets there is not right, since it causes compilers to look for an
1139 installed pcre.h, not the version that is in the source that is being
1140 compiled (which of course may be different). I have changed it back to:
1142 #include "pcre.h"
1144 I have a vague recollection that the change was concerned with compiling in
1145 different directories, but in the new build system, that is taken care of
1146 by the VPATH setting the Makefile.
1148 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1149 when the subject happened to end in the byte 0x85 (e.g. if the last
1150 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1151 characters but of course it shouldn't be taken as a newline when it is part
1152 of another character. The bug was that, for an unlimited repeat of . in
1153 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1154 characters when looking for a newline.
1156 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1158 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1159 in debug output.
1161 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1162 long printing in the pcrecpp unittest when running under MinGW.
1164 6. ESC_K was left out of the EBCDIC table.
1166 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1167 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1168 limit also applies to "virtual nesting" when a pattern is recursive, and in
1169 this case 1000 isn't so big. I have been able to remove this limit at the
1170 expense of backing off one optimization in certain circumstances. Normally,
1171 when pcre_exec() would call its internal match() function recursively and
1172 immediately return the result unconditionally, it uses a "tail recursion"
1173 feature to save stack. However, when a subpattern that can match an empty
1174 string has an unlimited repetition quantifier, it no longer makes this
1175 optimization. That gives it a stack frame in which to save the data for
1176 checking that an empty string has been matched. Previously this was taken
1177 from the 1000-entry workspace that had been reserved. So now there is no
1178 explicit limit, but more stack is used.
1180 8. Applied Daniel's patches to solve problems with the import/export magic
1181 syntax that is required for Windows, and which was going wrong for the
1182 pcreposix and pcrecpp parts of the library. These were overlooked when this
1183 problem was solved for the main library.
1185 9. There were some crude static tests to avoid integer overflow when computing
1186 the size of patterns that contain repeated groups with explicit upper
1187 limits. As the maximum quantifier is 65535, the maximum group length was
1188 set at 30,000 so that the product of these two numbers did not overflow a
1189 32-bit integer. However, it turns out that people want to use groups that
1190 are longer than 30,000 bytes (though not repeat them that many times).
1191 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1192 made it possible to implement the integer overflow checks in a much more
1193 dynamic way, which I have now done. The artificial limitation on group
1194 length has been removed - we now have only the limit on the total length of
1195 the compiled pattern, which depends on the LINK_SIZE setting.
1197 10. Fixed a bug in the documentation for get/copy named substring when
1198 duplicate names are permitted. If none of the named substrings are set, the
1199 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1200 empty string.
1202 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1203 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1204 because the ] is interpreted as the first data character and the
1205 terminating ] is not found. PCRE has been made compatible with Perl in this
1206 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1207 cause memory overwriting.
1209 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1210 string has been matched (to stop an infinite loop). It was not recognizing
1211 a conditional subpattern that could match an empty string if that
1212 subpattern was within another subpattern. For example, it looped when
1213 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1214 condition was not nested. This bug has been fixed.
1216 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1217 past the start of the subject in the presence of bytes with the top bit
1218 set, for example "\x8aBCD".
1220 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1221 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1223 14. Optimized (?!) to (*FAIL).
1225 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1226 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1227 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1228 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1229 does: it's just the validity check that is more restrictive.
1231 16. Inserted checks for integer overflows during escape sequence (backslash)
1232 processing, and also fixed erroneous offset values for syntax errors during
1233 backslash processing.
1235 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1236 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1238 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1239 caused an overrun.
1241 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1242 something other than just ASCII characters) inside a group that had an
1243 unlimited repeat caused a loop at compile time (while checking to see
1244 whether the group could match an empty string).
1246 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1247 [\P{Any}] did so. (Error in the code for printing property names.)
1249 21. An orphan \E inside a character class could cause a crash.
1251 22. A repeated capturing bracket such as (A)? could cause a wild memory
1252 reference during compilation.
1254 23. There are several functions in pcre_compile() that scan along a compiled
1255 expression for various reasons (e.g. to see if it's fixed length for look
1256 behind). There were bugs in these functions when a repeated \p or \P was
1257 present in the pattern. These operators have additional parameters compared
1258 with \d, etc, and these were not being taken into account when moving along
1259 the compiled data. Specifically:
1261 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1262 length.
1264 (b) An item such as \pL+ within a repeated group could cause crashes or
1265 loops.
1267 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1268 "reference to non-existent subpattern" error.
1270 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1272 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1273 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1275 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1277 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1278 character were causing crashes (broken optimization).
1280 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1281 \p or \P) caused a compile-time loop.
1283 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1284 break. For example, the unstudied pattern [\r\n]A does not match the string
1285 "\r\nA" because change 7.0/46 below moves the current point on by two
1286 characters after failing to match at the start. However, the pattern \nA
1287 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1288 the same is true. There doesn't seem any very clean way out of this, but
1289 what I have chosen to do makes the common cases work: PCRE now takes note
1290 of whether there can be an explicit match for \r or \n anywhere in the
1291 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1292 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1293 pattern has explicit CR or LF references.
1295 29. Added (*CR) etc for changing newline setting at start of pattern.
1298 Version 7.2 19-Jun-07
1299 ---------------------
1301 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1302 which is apparently normally available under Windows.
1304 2. Re-jig the pcregrep tests with different newline settings in an attempt
1305 to make them independent of the local environment's newline setting.
1307 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1309 4. Some of the "internals" tests were previously cut out when the link size
1310 was not 2, because the output contained actual offsets. The recent new
1311 "Z" feature of pcretest means that these can be cut out, making the tests
1312 usable with all link sizes.
1314 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1315 stack recursion. This gives a massive performance boost under BSD, but just
1316 a small improvement under Linux. However, it saves one field in the frame
1317 in all cases.
1319 6. Added more features from the forthcoming Perl 5.10:
1321 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1322 recursion call. It refers to the nth most recently opened parentheses.
1324 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1325 to be opened parentheses.
1327 (c) Conditions that refer to capturing parentheses can be specified
1328 relatively, for example, (?(-2)... or (?(+3)...
1330 (d) \K resets the start of the current match so that everything before
1331 is not part of it.
1333 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1335 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1336 reference syntax.
1338 (g) (?| introduces a group in which the numbering of parentheses in each
1339 alternative starts with the same number.
1341 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1343 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1346 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1347 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1348 for detecting groups that can match an empty string.
1350 9. A pattern with a very large number of alternatives (more than several
1351 hundred) was running out of internal workspace during the pre-compile
1352 phase, where pcre_compile() figures out how much memory will be needed. A
1353 bit of new cunning has reduced the workspace needed for groups with
1354 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1355 workspace instead of running out of the 4096 that are available.
1357 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1359 11. Applied patch from Google to remove an optimization that didn't quite work.
1360 The report of the bug said:
1362 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1363 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1364 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1366 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1367 it matched the wrong number of bytes.
1370 Version 7.1 24-Apr-07
1371 ---------------------
1373 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1374 that is more "standard", making use of automake and other Autotools. There
1375 is some re-arrangement of the files and adjustment of comments consequent
1376 on this.
1378 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1379 for recursive directory scanning broke on some systems because the files
1380 are not scanned in any specific order and on different systems the order
1381 was different. A call to "sort" has been inserted into RunGrepTest for the
1382 approprate test as a short-term fix. In the longer term there may be an
1383 alternative.
1385 3. I had an email from Eric Raymond about problems translating some of PCRE's
1386 man pages to HTML (despite the fact that I distribute HTML pages, some
1387 people do their own conversions for various reasons). The problems
1388 concerned the use of low-level troff macros .br and .in. I have therefore
1389 removed all such uses from the man pages (some were redundant, some could
1390 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1391 HTML has been updated to handle .nf/.fi and to complain if it encounters
1392 .br or .in.
1394 4. Updated comments in configure.ac that get placed in config.h.in and also
1395 arranged for config.h to be included in the distribution, with the name
1396 config.h.generic, for the benefit of those who have to compile without
1397 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1399 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1400 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1401 makevp.bat for the current PCRE, using the additional files
1402 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1404 6. A Windows user reported a minor discrepancy with test 2, which turned out
1405 to be caused by a trailing space on an input line that had got lost in his
1406 copy. The trailing space was an accident, so I've just removed it.
1408 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1409 that is needed.
1411 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1412 as "const" (a) because they are and (b) because it helps the PHP
1413 maintainers who have recently made a script to detect big data structures
1414 in the php code that should be moved to the .rodata section. I remembered
1415 to update Builducptable as well, so it won't revert if ucptable.h is ever
1416 re-created.
1418 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1419 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1420 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1421 support is not required. This saves 1.5-2K of code, which is important in
1422 some applications.
1424 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1425 so as not to refer to the tables, even though these functions will never be
1426 called when UTF-8 support is disabled. Otherwise there are problems with a
1427 shared library.
1429 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1431 (a) It was defining its arguments as char * instead of void *.
1433 (b) It was assuming that all moves were upwards in memory; this was true
1434 a long time ago when I wrote it, but is no longer the case.
1436 The emulated memove() is provided for those environments that have neither
1437 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1438 is clearly not the case, as these two bugs were recently reported.
1440 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1441 and Detrail to create the HTML documentation, the .txt form of the man
1442 pages, and it removes trailing spaces from listed files. It also creates
1443 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1444 case, it wraps all the #defines with #ifndefs. This script should be run
1445 before "make dist".
1447 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1448 with Unicode property support.
1450 (a) For a maximizing quantifier, if the two different cases of the
1451 character were of different lengths in their UTF-8 codings (there are
1452 some cases like this - I found 11), and the matching function had to
1453 back up over a mixture of the two cases, it incorrectly assumed they
1454 were both the same length.
1456 (b) When PCRE was configured to use the heap rather than the stack for
1457 recursion during matching, it was not correctly preserving the data for
1458 the other case of a UTF-8 character when checking ahead for a match
1459 while processing a minimizing repeat. If the check also involved
1460 matching a wide character, but failed, corruption could cause an
1461 erroneous result when trying to check for a repeat of the original
1462 character.
1464 13. Some tidying changes to the testing mechanism:
1466 (a) The RunTest script now detects the internal link size and whether there
1467 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1468 values substituted by "configure". (The RunGrepTest script already did
1469 this for UTF-8.) The configure.ac script no longer substitutes the
1470 relevant variables.
1472 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1473 with length and offset values. This means that the output is different
1474 for different internal link sizes. Test 2 is skipped for link sizes
1475 other than 2 because of this, bypassing the problem. Unfortunately,
1476 there was also a test in test 3 (the locale tests) that used /B and
1477 failed for link sizes other than 2. Rather than cut the whole test out,
1478 I have added a new /Z option to pcretest that replaces the length and
1479 offset values with spaces. This is now used to make test 3 independent
1480 of link size. (Test 2 will be tidied up later.)
1482 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1483 segmentation fault instead of returning the appropriate error message.
1485 15. In multiline mode when the newline sequence was set to "any", the pattern
1486 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1487 This doesn't seem right; it now treats the CRLF combination as the line
1488 ending, and so does not match in that case. It's only a pattern such as ^$
1489 that would hit this one: something like ^ABC$ would have failed after \r
1490 and then tried again after \r\n.
1492 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1493 in an attempt to make files that differ only in their line terminators
1494 compare equal. This works on Linux.
1496 17. Under certain error circumstances pcregrep might try to free random memory
1497 as it exited. This is now fixed, thanks to valgrind.
1499 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1500 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1501 was because its rules for how to advance for /g after matching an empty
1502 string at the end of a line did not allow for this case. They now check for
1503 it specially.
1505 20. pcretest is supposed to handle patterns and data of any length, by
1506 extending its buffers when necessary. It was getting this wrong when the
1507 buffer for a data line had to be extended.
1509 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1510 CRLF as a newline sequence.
1512 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1513 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1514 I have nevertheless tidied it up.
1516 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1518 24. Added a man page for pcre-config.
1521 Version 7.0 19-Dec-06
1522 ---------------------
1524 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1525 moving to gcc 4.1.1.
1527 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1528 sys/time.h, which is documented as needed for this function. It doesn't
1529 seem to matter on Linux, but it showed up on some releases of OS X.
1531 3. It seems that there are systems where bytes whose values are greater than
1532 127 match isprint() in the "C" locale. The "C" locale should be the
1533 default when a C program starts up. In most systems, only ASCII printing
1534 characters match isprint(). This difference caused the output from pcretest
1535 to vary, making some of the tests fail. I have changed pcretest so that:
1537 (a) When it is outputting text in the compiled version of a pattern, bytes
1538 other than 32-126 are always shown as hex escapes.
1540 (b) When it is outputting text that is a matched part of a subject string,
1541 it does the same, unless a different locale has been set for the match
1542 (using the /L modifier). In this case, it uses isprint() to decide.
1544 4. Fixed a major bug that caused incorrect computation of the amount of memory
1545 required for a compiled pattern when options that changed within the
1546 pattern affected the logic of the preliminary scan that determines the
1547 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1548 that the computed length was too small. The symptoms of this bug were
1549 either the PCRE error "internal error: code overflow" from pcre_compile(),
1550 or a glibc crash with a message such as "pcretest: free(): invalid next
1551 size (fast)". Examples of patterns that provoked this bug (shown in
1552 pcretest format) are:
1554 /(?-x: )/x
1555 /(?x)(?-x: \s*#\s*)/
1556 /((?i)[\x{c0}])/8
1557 /(?i:[\x{c0}])/8
1559 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1560 is now done differently.
1562 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1563 wrapper classes; (b) implement a new function in the C++ scanner that is
1564 more efficient than the old way of doing things because it avoids levels of
1565 recursion in the regex matching; (c) add a paragraph to the documentation
1566 for the FullMatch() function.
1568 6. The escape sequence \n was being treated as whatever was defined as
1569 "newline". Not only was this contrary to the documentation, which states
1570 that \n is character 10 (hex 0A), but it also went horribly wrong when
1571 "newline" was defined as CRLF. This has been fixed.
1573 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1574 was being set to -1 for the "end of line" case (supposedly a value that no
1575 character can have). Though this value is never used (the check for end of
1576 line is "zero bytes in current character"), it caused compiler complaints.
1577 I've changed it to 0xffffffff.
1579 8. In pcre_version.c, the version string was being built by a sequence of
1580 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1581 string (as it is for production releases) called a macro with an empty
1582 argument. The C standard says the result of this is undefined. The gcc
1583 compiler treats it as an empty string (which was what was wanted) but it is
1584 reported that Visual C gives an error. The source has been hacked around to
1585 avoid this problem.
1587 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1588 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1589 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1590 of them did).
1592 10. Originally, pcretest opened its input and output without "b"; then I was
1593 told that "b" was needed in some environments, so it was added for release
1594 5.0 to both the input and output. (It makes no difference on Unix-like
1595 systems.) Later I was told that it is wrong for the input on Windows. I've
1596 now abstracted the modes into two macros, to make it easier to fiddle with
1597 them, and removed "b" from the input mode under Windows.
1599 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1601 12. Added -help and --help to pcretest as an official way of being reminded
1602 of the options.
1604 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1605 and pcrecpp.cc because they annoy compilers at high warning levels.
1607 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1609 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1610 scripts are not C programs :-) and which was not noticed because it works
1611 on Linux.
1613 16. pcretest is supposed to handle any length of pattern and data line (as one
1614 line or as a continued sequence of lines) by extending its input buffer if
1615 necessary. This feature was broken for very long pattern lines, leading to
1616 a string of junk being passed to pcre_compile() if the pattern was longer
1617 than about 50K.
1619 17. I have done a major re-factoring of the way pcre_compile() computes the
1620 amount of memory needed for a compiled pattern. Previously, there was code
1621 that made a preliminary scan of the pattern in order to do this. That was
1622 OK when PCRE was new, but as the facilities have expanded, it has become
1623 harder and harder to keep it in step with the real compile phase, and there
1624 have been a number of bugs (see for example, 4 above). I have now found a
1625 cunning way of running the real compile function in a "fake" mode that
1626 enables it to compute how much memory it would need, while actually only
1627 ever using a few hundred bytes of working memory and without too many
1628 tests of the mode. This should make future maintenance and development
1629 easier. A side effect of this work is that the limit of 200 on the nesting
1630 depth of parentheses has been removed (though this was never a serious
1631 limitation, I suspect). However, there is a downside: pcre_compile() now
1632 runs more slowly than before (30% or more, depending on the pattern). I
1633 hope this isn't a big issue. There is no effect on runtime performance.
1635 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1636 newline (only possible for the last line of a file) and it was a
1637 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1639 19. Added additional timing features to pcretest. (1) The -tm option now times
1640 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1641 separate command line item, by a number that specifies the number of
1642 repeats to use when timing. The default is 50000; this gives better
1643 precision, but takes uncomfortably long for very large patterns.
1645 20. Extended pcre_study() to be more clever in cases where a branch of a
1646 subpattern has no definite first character. For example, (a*|b*)[cd] would
1647 previously give no result from pcre_study(). Now it recognizes that the
1648 first character must be a, b, c, or d.
1650 21. There was an incorrect error "recursive call could loop indefinitely" if
1651 a subpattern (or the entire pattern) that was being tested for matching an
1652 empty string contained only one non-empty item after a nested subpattern.
1653 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1654 incorrectly, because the \d was being skipped in the check.
1656 22. The pcretest program now has a new pattern option /B and a command line
1657 option -b, which is equivalent to adding /B to every pattern. This causes
1658 it to show the compiled bytecode, without the additional information that
1659 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1660 is the same as /B/I).
1662 23. A new optimization is now able automatically to treat some sequences such
1663 as a*b as a*+b. More specifically, if something simple (such as a character
1664 or a simple class like \d) has an unlimited quantifier, and is followed by
1665 something that cannot possibly match the quantified thing, the quantifier
1666 is automatically "possessified".
1668 24. A recursive reference to a subpattern whose number was greater than 39
1669 went wrong under certain circumstances in UTF-8 mode. This bug could also
1670 have affected the operation of pcre_study().
1672 25. Realized that a little bit of performance could be had by replacing
1673 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1675 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1677 27. Possessive quantifiers such as a++ were previously implemented by turning
1678 them into atomic groups such as ($>a+). Now they have their own opcodes,
1679 which improves performance. This includes the automatically created ones
1680 from 23 above.
1682 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1683 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1684 the first matched character to be a colon. This applied both to named and
1685 numbered groups.
1687 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1689 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1690 building PCRE on Windows easier, so I have included it in the distribution.
1692 31. There is now a check in pcretest against a ridiculously large number being
1693 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1694 loop, the loop is abandoned.
1696 32. Forward references to subpatterns in conditions such as (?(2)...) where
1697 subpattern 2 is defined later cause pcre_compile() to search forwards in
1698 the pattern for the relevant set of parentheses. This search went wrong
1699 when there were unescaped parentheses in a character class, parentheses
1700 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1702 33. "Subroutine" calls and backreferences were previously restricted to
1703 referencing subpatterns earlier in the regex. This restriction has now
1704 been removed.
1706 34. Added a number of extra features that are going to be in Perl 5.10. On the
1707 whole, these are just syntactic alternatives for features that PCRE had
1708 previously implemented using the Python syntax or my own invention. The
1709 other formats are all retained for compatibility.
1711 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1712 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1713 also .NET compatible.
1715 (b) A recursion or subroutine call to a named group can now be defined as
1716 (?&name) as well as (?P>name).
1718 (c) A backreference to a named group can now be defined as \k<name> or
1719 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1720 5.10, are also .NET compatible.
1722 (d) A conditional reference to a named group can now use the syntax
1723 (?(<name>) or (?('name') as well as (?(name).
1725 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1726 groups (named and numbered) that are never evaluated inline, but can be
1727 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1728 is always false. There may be only one alternative in such a group.
1730 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1731 as the simple (?(R). The condition is true only if the most recent
1732 recursion is that of the given number or name. It does not search out
1733 through the entire recursion stack.
1735 (g) The escape \gN or \g{N} has been added, where N is a positive or
1736 negative number, specifying an absolute or relative reference.
1738 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1739 some "unreachable code" warnings.
1741 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1742 things, this adds five new scripts.
1744 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1745 There were also incompatibilities regarding the handling of \Q..\E inside
1746 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1747 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1749 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1750 matches an empty string, and forcibly breaks the loop. There were bugs in
1751 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1752 against aaaa the result was just "a" rather than "aaaa", for example. Two
1753 separate and independent bugs (that affected different cases) have been
1754 fixed.
1756 39. Refactored the code to abolish the use of different opcodes for small
1757 capturing bracket numbers. This is a tidy that I avoided doing when I
1758 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1759 The new approach is not only tidier, it makes it possible to reduce the
1760 memory needed to fix the previous bug (38).
1762 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1763 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1764 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1765 mode.
1767 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1768 report.
1770 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1771 copy construction and assignment in the C++ wrapper.
1773 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1774 couple of bugs that could have given wrong results in the "--newline=crlf"
1775 case.
1777 44. Added a number of casts and did some reorganization of signed/unsigned int
1778 variables following suggestions from Dair Grant. Also renamed the variable
1779 "this" as "item" because it is a C++ keyword.
1781 45. Arranged for dftables to add
1783 #include "pcre_internal.h"
1785 to pcre_chartables.c because without it, gcc 4.x may remove the array
1786 definition from the final binary if PCRE is built into a static library and
1787 dead code stripping is activated.
1789 46. For an unanchored pattern, if a match attempt fails at the start of a
1790 newline sequence, and the newline setting is CRLF or ANY, and the next two
1791 characters are CRLF, advance by two characters instead of one.
1794 Version 6.7 04-Jul-06
1795 ---------------------
1797 1. In order to handle tests when input lines are enormously long, pcretest has
1798 been re-factored so that it automatically extends its buffers when
1799 necessary. The code is crude, but this _is_ just a test program. The
1800 default size has been increased from 32K to 50K.
1802 2. The code in pcre_study() was using the value of the re argument before
1803 testing it for NULL. (Of course, in any sensible call of the function, it
1804 won't be NULL.)
1806 3. The memmove() emulation function in pcre_internal.h, which is used on
1807 systems that lack both memmove() and bcopy() - that is, hardly ever -
1808 was missing a "static" storage class specifier.
1810 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1811 containing an extended class (one that cannot be represented by a bitmap
1812 because it contains high-valued characters or Unicode property items, e.g.
1813 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1814 pattern, but PCRE should not loop if you do not (it no longer does).
1815 [Detail: two cases were found: (a) a repeated subpattern containing an
1816 extended class; (b) a recursive reference to a subpattern that followed a
1817 previous extended class. It wasn't skipping over the extended class
1818 correctly when UTF-8 mode was not set.]
1820 5. A negated single-character class was not being recognized as fixed-length
1821 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1822 compile error "lookbehind assertion is not fixed length".
1824 6. The RunPerlTest auxiliary script was showing an unexpected difference
1825 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1826 write a Perl script that can interpret lines of an input file either as
1827 byte characters or as UTF-8, which is what "perltest" was being required to
1828 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1829 can't do is switch easily at run time between having the "use utf8;" pragma
1830 or not. In the end, I fudged it by using the RunPerlTest script to insert
1831 "use utf8;" explicitly for the UTF-8 tests.
1833 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1834 the end of the subject string, contrary to the documentation and to what
1835 Perl does. This was true of both matching functions. Now it matches only at
1836 the start of the subject and immediately after *internal* newlines.
1838 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1839 a pointer to an int instead of a pointer to an unsigned long int. This
1840 caused problems on 64-bit systems.
1842 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1843 instance of the 'standard' template library not being so standard".
1845 10. There was no check on the number of named subpatterns nor the maximum
1846 length of a subpattern name. The product of these values is used to compute
1847 the size of the memory block for a compiled pattern. By supplying a very
1848 long subpattern name and a large number of named subpatterns, the size
1849 computation could be caused to overflow. This is now prevented by limiting
1850 the length of names to 32 characters, and the number of named subpatterns
1851 to 10,000.
1853 11. Subpatterns that are repeated with specific counts have to be replicated in
1854 the compiled pattern. The size of memory for this was computed from the
1855 length of the subpattern and the repeat count. The latter is limited to
1856 65535, but there was no limit on the former, meaning that integer overflow
1857 could in principle occur. The compiled length of a repeated subpattern is
1858 now limited to 30,000 bytes in order to prevent this.
1860 12. Added the optional facility to have named substrings with the same name.
1862 13. Added the ability to use a named substring as a condition, using the
1863 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1864 are numbers (not recommended). Forward references are permitted.
1866 14. Added forward references in named backreferences (if you see what I mean).
1868 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1869 pattern could run off the end of the subject. For example, the pattern
1870 "(?s)(.{1,5})"8 did this with the subject "ab".
1872 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1873 PCRE_CASELESS was set when matching characters that were quantified with ?
1874 or *.
1876 17. A character class other than a single negated character that had a minimum
1877 but no maximum quantifier - for example [ab]{6,} - was not handled
1878 correctly by pce_dfa_exec(). It would match only one character.
1880 18. A valid (though odd) pattern that looked like a POSIX character
1881 class but used an invalid character after [ (for example [[,abc,]]) caused
1882 pcre_compile() to give the error "Failed: internal error: code overflow" or
1883 in some cases to crash with a glibc free() error. This could even happen if
1884 the pattern terminated after [[ but there just happened to be a sequence of
1885 letters, a binary zero, and a closing ] in the memory that followed.
1887 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1888 over the years. Originally (before any Unicode support), just the bottom 8
1889 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1890 output from "man perlunicode" includes this:
1892 The regular expression compiler produces polymorphic opcodes. That
1893 is, the pattern adapts to the data and automatically switches to
1894 the Unicode character scheme when presented with Unicode data--or
1895 instead uses a traditional byte scheme when presented with byte
1896 data.
1898 Sadly, a wide octal escape does not cause a switch, and in a string with
1899 no other multibyte characters, these octal escapes are treated as before.
1900 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1901 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1902 Unicode string.
1904 I have not perpetrated such confusion in PCRE. Up till now, it took just
1905 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1906 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1907 translate to the appropriate multibyte character.
1909 29. Applied some refactoring to reduce the number of warnings from Microsoft
1910 and Borland compilers. This has included removing the fudge introduced
1911 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1912 a warning about an unused variable.
1914 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1915 characters since release 4.0, because Perl (from release 5.004) does not.
1916 [Or at least, is documented not to: some releases seem to be in conflict
1917 with the documentation.] However, when a pattern was studied with
1918 pcre_study() and all its branches started with \s, PCRE still included VT
1919 as a possible starting character. Of course, this did no harm; it just
1920 caused an unnecessary match attempt.
1922 22. Removed a now-redundant internal flag bit that recorded the fact that case
1923 dependency changed within the pattern. This was once needed for "required
1924 byte" processing, but is no longer used. This recovers a now-scarce options
1925 bit. Also moved the least significant internal flag bit to the most-
1926 significant bit of the word, which was not previously used (hangover from
1927 the days when it was an int rather than a uint) to free up another bit for
1928 the future.
1930 23. Added support for CRLF line endings as well as CR and LF. As well as the
1931 default being selectable at build time, it can now be changed at runtime
1932 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1933 specify that it is scanning data with non-default line endings.
1935 24. Changed the definition of CXXLINK to make it agree with the definition of
1936 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1938 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1939 recursions. This makes a big different to stack usage for some patterns.
1941 26. If a subpattern containing a named recursion or subroutine reference such
1942 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1943 the space required for the compiled pattern went wrong and gave too small a
1944 value. Depending on the environment, this could lead to "Failed: internal
1945 error: code overflow at offset 49" or "glibc detected double free or
1946 corruption" errors.
1948 27. Applied patches from Google (a) to support the new newline modes and (b) to
1949 advance over multibyte UTF-8 characters in GlobalReplace.
1951 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1952 difference for some implementation of PCRE in some Windows version.
1954 29. Added some extra testing facilities to pcretest:
1956 \q<number> in a data line sets the "match limit" value
1957 \Q<number> in a data line sets the "match recursion limt" value
1958 -S <number> sets the stack size, where <number> is in megabytes
1960 The -S option isn't available for Windows.
1963 Version 6.6 06-Feb-06
1964 ---------------------
1966 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1967 in pcreposix.h. I have copied the definition from pcre.h.
1969 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1970 because pcre.h is no longer a built file.
1972 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1973 not normally included in the compiled code.
1976 Version 6.5 01-Feb-06
1977 ---------------------
1979 1. When using the partial match feature with pcre_dfa_exec(), it was not
1980 anchoring the second and subsequent partial matches at the new starting
1981 point. This could lead to incorrect results. For example, with the pattern
1982 /1234/, partially matching against "123" and then "a4" gave a match.
1984 2. Changes to pcregrep:
1986 (a) All non-match returns from pcre_exec() were being treated as failures
1987 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1988 error message is output. Some extra information is given for the
1990 probably the only errors that are likely to be caused by users (by
1991 specifying a regex that has nested indefinite repeats, for instance).
1992 If there are more than 20 of these errors, pcregrep is abandoned.
1994 (b) A binary zero was treated as data while matching, but terminated the
1995 output line if it was written out. This has been fixed: binary zeroes
1996 are now no different to any other data bytes.
1998 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1999 used to set a locale for matching. The --locale=xxxx long option has
2000 been added (no short equivalent) to specify a locale explicitly on the
2001 pcregrep command, overriding the environment variables.
2003 (d) When -B was used with -n, some line numbers in the output were one less
2004 than they should have been.
2006 (e) Added the -o (--only-matching) option.
2008 (f) If -A or -C was used with -c (count only), some lines of context were
2009 accidentally printed for the final match.
2011 (g) Added the -H (--with-filename) option.
2013 (h) The combination of options -rh failed to suppress file names for files
2014 that were found from directory arguments.
2016 (i) Added the -D (--devices) and -d (--directories) options.
2018 (j) Added the -F (--fixed-strings) option.
2020 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2022 (l) Added the --colo(u)r option.
2024 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2025 is not present by default.
2027 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2028 items such as (?R) or (?1), when the recursion could match a number of
2029 alternatives. If it matched one of the alternatives, but subsequently,
2030 outside the recursion, there was a failure, the code tried to back up into
2031 the recursion. However, because of the way PCRE is implemented, this is not
2032 possible, and the result was an incorrect result from the match.
2034 In order to prevent this happening, the specification of recursion has
2035 been changed so that all such subpatterns are automatically treated as
2036 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2038 4. I had overlooked the fact that, in some locales, there are characters for
2039 which isalpha() is true but neither isupper() nor islower() are true. In
2040 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2041 and ordfeminine) are like this. This affected the treatment of \w and \W
2042 when they appeared in character classes, but not when they appeared outside
2043 a character class. The bit map for "word" characters is now created
2044 separately from the results of isalnum() instead of just taking it from the
2045 upper, lower, and digit maps. (Plus the underscore character, of course.)
2047 5. The above bug also affected the handling of POSIX character classes such as
2048 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2049 permanent tables. Instead, the bit maps for such a class were previously
2050 created as the appropriate unions of the upper, lower, and digit bitmaps.
2051 Now they are created by subtraction from the [[:word:]] class, which has
2052 its own bitmap.
2054 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2055 It is created by subtracting the vertical space characters (\x09, \x0a,
2056 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2057 subtraction was done in the overall bitmap for a character class, meaning
2058 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2059 be recognized. This bug has been fixed.
2061 7. Patches from the folks at Google:
2063 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2064 real life, but is still worth protecting against".
2066 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2067 regular expressions".
2069 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2070 have it.
2072 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2073 "configure" and the latter not, in order to fix a problem somebody had
2074 with compiling the Arg class on HP-UX.
2076 (e) Improve the error-handling of the C++ wrapper a little bit.
2078 (f) New tests for checking recursion limiting.
2080 8. The pcre_memmove() function, which is used only if the environment does not
2081 have a standard memmove() function (and is therefore rarely compiled),
2082 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2083 returning a result (though PCRE never actually uses the result).
2085 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2086 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2087 returned instead of calling malloc() with an overflowing number that would
2088 most likely cause subsequent chaos.
2090 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2092 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2093 with this option is matched, the nmatch and pmatch options of regexec() are
2094 ignored.
2096 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2097 provided in case anyone wants to the the POSIX interface with UTF-8
2098 strings.
2100 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2101 C++ linking (needed for some HP-UX environments).
2103 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2104 (unused parameter) and in the pcre_printint() function (omitted "default"
2105 switch label when the default is to do nothing).
2107 15. Added some code to make it possible, when PCRE is compiled as a C++
2108 library, to replace subject pointers for pcre_exec() with a smart pointer
2109 class, thus making it possible to process discontinuous strings.
2111 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2112 much the same function. They were added by different people who were trying
2113 to make PCRE easy to compile on non-Unix systems. It has been suggested
2114 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2115 for compiling on Windows systems. I have therefore replaced it with
2116 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2117 defaults to "extern" for C or "extern C" for C++, which works fine on
2118 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2119 SCOPE with something explicit in config.h. In addition:
2121 (a) pcreposix.h still had just "extern" instead of either of these macros;
2122 I have replaced it with PCRE_DATA_SCOPE.
2124 (b) Functions such as _pcre_xclass(), which are internal to the library,
2125 but external in the C sense, all had PCRE_EXPORT in their definitions.
2126 This is apparently wrong for the Windows case, so I have removed it.
2127 (It makes no difference on Unix-like systems.)
2129 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2130 of recursive calls to match(). This is different to MATCH_LIMIT because
2131 that limits the total number of calls to match(), not all of which increase
2132 the depth of recursion. Limiting the recursion depth limits the amount of
2133 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2134 when PCRE is compiled, and changed at run time. A patch from Google adds
2135 this functionality to the C++ interface.
2137 18. Changes to the handling of Unicode character properties:
2139 (a) Updated the table to Unicode 4.1.0.
2141 (b) Recognize characters that are not in the table as "Cn" (undefined).
2143 (c) I revised the way the table is implemented to a much improved format
2144 which includes recognition of ranges. It now supports the ranges that
2145 are defined in UnicodeData.txt, and it also amalgamates other
2146 characters into ranges. This has reduced the number of entries in the
2147 table from around 16,000 to around 3,000, thus reducing its size
2148 considerably. I realized I did not need to use a tree structure after
2149 all - a binary chop search is just as efficient. Having reduced the
2150 number of entries, I extended their size from 6 bytes to 8 bytes to
2151 allow for more data.
2153 (d) Added support for Unicode script names via properties such as \p{Han}.
2155 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2156 matching that character.
2158 20. When matching a repeated Unicode property with a minimum greater than zero,
2159 (for example \pL{2,}), PCRE could look past the end of the subject if it
2160 reached it while seeking the minimum number of characters. This could
2161 happen only if some of the characters were more than one byte long, because
2162 there is a check for at least the minimum number of bytes.
2164 21. Refactored the implementation of \p and \P so as to be more general, to
2165 allow for more different types of property in future. This has changed the
2166 compiled form incompatibly. Anybody with saved compiled patterns that use
2167 \p or \P will have to recompile them.
2169 22. Added "Any" and "L&" to the supported property types.
2171 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2172 but give a compile time error if the value is greater than 0xff.
2174 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2175 accidentally not being installed or uninstalled.
2177 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2178 made were to insert the current release number. This seemed silly, because
2179 it made things harder for people building PCRE on systems that don't run
2180 "configure". I have turned pcre.h into a distributed file, no longer built
2181 by "configure", with the version identification directly included. There is
2182 no longer a pcre.h.in file.
2184 However, this change necessitated a change to the pcre-config script as
2185 well. It is built from pcre-config.in, and one of the substitutions was the
2186 release number. I have updated configure.ac so that ./configure now finds
2187 the release number by grepping pcre.h.
2189 26. Added the ability to run the tests under valgrind.
2192 Version 6.4 05-Sep-05
2193 ---------------------
2195 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2196 "--" to be printed when multiple files were scanned, even when none of the
2197 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2198 consider it to be a bug, and have restored the previous behaviour.
2200 2. A couple of code tidies to get rid of compiler warnings.
2202 3. The pcretest program used to cheat by referring to symbols in the library
2203 whose names begin with _pcre_. These are internal symbols that are not
2204 really supposed to be visible externally, and in some environments it is
2205 possible to suppress them. The cheating is now confined to including
2206 certain files from the library's source, which is a bit cleaner.
2208 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2209 file's purpose clearer.
2211 5. Reorganized pcre_ucp_findchar().
2214 Version 6.3 15-Aug-05
2215 ---------------------
2217 1. The file libpcre.pc.in did not have general read permission in the tarball.
2219 2. There were some problems when building without C++ support:
2221 (a) If C++ support was not built, "make install" and "make test" still
2222 tried to test it.
2224 (b) There were problems when the value of CXX was explicitly set. Some
2225 changes have been made to try to fix these, and ...
2227 (c) --disable-cpp can now be used to explicitly disable C++ support.
2229 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2230 backslash in a target when C++ was disabled. This confuses some
2231 versions of "make", apparently. Using an intermediate variable solves
2232 this. (Same for CPP_LOBJ.)
2234 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2235 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2236 necessary on certain architectures.
2238 4. Added a setting of -export-symbols-regex to the link command to remove
2239 those symbols that are exported in the C sense, but actually are local
2240 within the library, and not documented. Their names all begin with
2241 "_pcre_". This is not a perfect job, because (a) we have to except some
2242 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2243 available (and never for static libraries). I have made a note to try to
2244 find a way round (a) in the future.
2247 Version 6.2 01-Aug-05
2248 ---------------------
2250 1. There was no test for integer overflow of quantifier values. A construction
2251 such as {1111111111111111} would give undefined results. What is worse, if
2252 a minimum quantifier for a parenthesized subpattern overflowed and became
2253 negative, the calculation of the memory size went wrong. This could have
2254 led to memory overwriting.
2256 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2258 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2259 operating environments where this matters.
2261 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2262 PCRE options from within the C++ wrapper.
2264 5. Named capturing subpatterns were not being correctly counted when a pattern
2265 was compiled. This caused two problems: (a) If there were more than 100
2266 such subpatterns, the calculation of the memory needed for the whole
2267 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2268 back references of the form \12, where the number was greater than 9, were
2269 not recognized as back references, even though there were sufficient
2270 previous subpatterns.
2272 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2273 versions of gcc, e.g. 2.95.4.
2276 Version 6.1 21-Jun-05
2277 ---------------------
2279 1. There was one reference to the variable "posix" in pcretest.c that was not
2280 surrounded by "#if !defined NOPOSIX".
2282 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2283 the cross-check on the old pcre_info() function, for the benefit of the
2284 cut-down version of PCRE that is currently imported into Exim.
2286 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2287 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2288 patterns that start (?i) but I don't think that matters. The use of (?i) is
2289 just an example; this all applies to the other options as well.
2291 4. Since libtool seems to echo the compile commands it is issuing, the output
2292 from "make" can be reduced a bit by putting "@" in front of each libtool
2293 compile command.
2295 5. Patch from the folks at Google for configure.in to be a bit more thorough
2296 in checking for a suitable C++ installation before trying to compile the
2297 C++ stuff. This should fix a reported problem when a compiler was present,
2298 but no suitable headers.
2300 6. The man pages all had just "PCRE" as their title. I have changed them to
2301 be the relevant file name. I have also arranged that these names are
2302 retained in the file doc/pcre.txt, which is a concatenation in text format
2303 of all the man pages except the little individual ones for each function.
2305 7. The NON-UNIX-USE file had not been updated for the different set of source
2306 files that come with release 6. I also added a few comments about the C++
2307 wrapper.
2310 Version 6.0 07-Jun-05
2311 ---------------------
2313 1. Some minor internal re-organization to help with my DFA experiments.
2315 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2316 didn't matter for the library itself when fully configured, but did matter
2317 when compiling without UCP support, or within Exim, where the ucp files are
2318 not imported.
2320 3. Refactoring of the library code to split up the various functions into
2321 different source modules. The addition of the new DFA matching code (see
2322 below) to a single monolithic source would have made it really too
2323 unwieldy, quite apart from causing all the code to be include in a
2324 statically linked application, when only some functions are used. This is
2325 relevant even without the DFA addition now that patterns can be compiled in
2326 one application and matched in another.
2328 The downside of splitting up is that there have to be some external
2329 functions and data tables that are used internally in different modules of
2330 the library but which are not part of the API. These have all had their
2331 names changed to start with "_pcre_" so that they are unlikely to clash
2332 with other external names.
2334 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2335 a different (DFA) algorithm. Although it is slower than the original
2336 function, it does have some advantages for certain types of matching
2337 problem.
2339 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2340 including restarting after a partial match.
2342 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2343 defined when compiling for Windows was sent to me. I have put it into the
2344 code, though I have no means of testing or verifying it.
2346 7. Added the pcre_refcount() auxiliary function.
2348 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2349 match before or at the first newline in the subject string. In pcretest,
2350 the /f option on a pattern can be used to set this.
2352 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2353 would behave wrongly. This has been present in PCRE since release 4.0.
2355 10. A number of changes to the pcregrep command:
2357 (a) Refactored how -x works; insert ^(...)$ instead of setting
2358 PCRE_ANCHORED and checking the length, in preparation for adding
2359 something similar for -w.
2361 (b) Added the -w (match as a word) option.
2363 (c) Refactored the way lines are read and buffered so as to have more
2364 than one at a time available.
2366 (d) Implemented a pcregrep test script.
2368 (e) Added the -M (multiline match) option. This allows patterns to match
2369 over several lines of the subject. The buffering ensures that at least
2370 8K, or the rest of the document (whichever is the shorter) is available
2371 for matching (and similarly the previous 8K for lookbehind assertions).
2373 (f) Changed the --help output so that it now says
2375 -w, --word-regex(p)
2377 instead of two lines, one with "regex" and the other with "regexp"
2378 because that confused at least one person since the short forms are the
2379 same. (This required a bit of code, as the output is generated
2380 automatically from a table. It wasn't just a text change.)
2382 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2383 option but starts with a hyphen. Could be a pattern or a path name
2384 starting with a hyphen, for instance.
2386 (h) "-" can be given as a file name to represent stdin.
2388 (i) When file names are being printed, "(standard input)" is used for
2389 the standard input, for compatibility with GNU grep. Previously
2390 "<stdin>" was used.
2392 (j) The option --label=xxx can be used to supply a name to be used for
2393 stdin when file names are being printed. There is no short form.
2395 (k) Re-factored the options decoding logic because we are going to add
2396 two more options that take data. Such options can now be given in four
2397 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2399 (l) Added the -A, -B, and -C options for requesting that lines of context
2400 around matches be printed.
2402 (m) Added the -L option to print the names of files that do not contain
2403 any matching lines, that is, the complement of -l.
2405 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2406 continue to scan other files.
2408 (o) The -s option was incorrectly implemented. For compatibility with other
2409 greps, it now suppresses the error message for a non-existent or non-
2410 accessible file (but not the return code). There is a new option called
2411 -q that suppresses the output of matching lines, which was what -s was
2412 previously doing.
2414 (p) Added --include and --exclude options to specify files for inclusion
2415 and exclusion when recursing.
2417 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2418 Hopefully, it now does.
2420 12. Missing cast in pcre_study().
2422 13. Added an "uninstall" target to the makefile.
2424 14. Replaced "extern" in the function prototypes in Makefile.in with
2425 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2426 world, but is set differently for Windows.
2428 15. Added a second compiling function called pcre_compile2(). The only
2429 difference is that it has an extra argument, which is a pointer to an
2430 integer error code. When there is a compile-time failure, this is set
2431 non-zero, in addition to the error test pointer being set to point to an
2432 error message. The new argument may be NULL if no error number is required
2433 (but then you may as well call pcre_compile(), which is now just a
2434 wrapper). This facility is provided because some applications need a
2435 numeric error indication, but it has also enabled me to tidy up the way
2436 compile-time errors are handled in the POSIX wrapper.
2438 16. Added VPATH=.libs to the makefile; this should help when building with one
2439 prefix path and installing with another. (Or so I'm told by someone who
2440 knows more about this stuff than I do.)
2442 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2443 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2444 match everything, including newlines. This is not POSIX-compatible, but
2445 somebody wanted the feature. From pcretest it can be activated by using
2446 both the P and the s flags.
2448 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2450 19. libpcre.pc was being incorrectly installed as executable.
2452 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2453 it now also looks for '\r' so that it will work unmodified on Windows.
2455 21. Added Google's contributed C++ wrapper to the distribution.
2457 22. Added some untidy missing memory free() calls in pcretest, to keep
2458 Electric Fence happy when testing.
2462 Version 5.0 13-Sep-04
2463 ---------------------
2465 1. Internal change: literal characters are no longer packed up into items
2466 containing multiple characters in a single byte-string. Each character
2467 is now matched using a separate opcode. However, there may be more than one
2468 byte in the character in UTF-8 mode.
2470 2. The pcre_callout_block structure has two new fields: pattern_position and
2471 next_item_length. These contain the offset in the pattern to the next match
2472 item, and its length, respectively.
2474 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2475 insertion of callouts before each pattern item. Added the /C option to
2476 pcretest to make use of this.
2478 4. On the advice of a Windows user, the lines
2480 #if defined(_WIN32) || defined(WIN32)
2481 _setmode( _fileno( stdout ), 0x8000 );
2482 #endif /* defined(_WIN32) || defined(WIN32) */
2484 have been added to the source of pcretest. This apparently does useful
2485 magic in relation to line terminators.
2487 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2488 for the benefit of those environments where the "b" makes a difference.
2490 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2491 to know about it. I have put a hack into configure.in that adds in code
2492 to set GCC=yes if CC=icc. This seems to end up at a point in the
2493 generated configure script that is early enough to affect the setting of
2494 compiler options, which is what is needed, but I have no means of testing
2495 whether it really works. (The user who reported this had patched the
2496 generated configure script, which of course I cannot do.)
2498 LATER: After change 22 below (new libtool files), the configure script
2499 seems to know about icc (and also ecc). Therefore, I have commented out
2500 this hack in configure.in.
2502 7. Added support for pkg-config (2 patches were sent in).
2504 8. Negated POSIX character classes that used a combination of internal tables
2505 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2506 [[:^ascii]]. Typically, they would match almost any characters. The other
2507 POSIX classes were not broken in this way.
2509 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2510 to find the match, as PCRE was deluded into thinking that the match had to
2511 start at the start point or following a newline. The same bug applied to
2512 patterns with negative forward assertions or any backward assertions
2513 preceding ".*" at the start, unless the pattern required a fixed first
2514 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2516 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2517 starting at the last subject character, bytes beyond the end of the subject
2518 string were read.
2520 11. Renamed the variable "class" as "classbits" to make life easier for C++
2521 users. (Previously there was a macro definition, but it apparently wasn't
2522 enough.)
2524 12. Added the new field "tables" to the extra data so that tables can be passed
2525 in at exec time, or the internal tables can be re-selected. This allows
2526 a compiled regex to be saved and re-used at a later time by a different
2527 program that might have everything at different addresses.
2529 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2530 -R library as well as a -L library.
2532 14. The debugging options of pcretest (-d on the command line or D on a
2533 pattern) showed incorrect output for anything following an extended class
2534 that contained multibyte characters and which was followed by a quantifier.
2536 15. Added optional support for general category Unicode character properties
2537 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2538 support. It adds about 90K to the size of the library. The meanings of the
2539 inbuilt class escapes such as \d and \s have NOT been changed.
2541 16. Updated pcredemo.c to include calls to free() to release the memory for the
2542 compiled pattern.
2544 17. The generated file chartables.c was being created in the source directory
2545 instead of in the building directory. This caused the build to fail if the
2546 source directory was different from the building directory, and was
2547 read-only.
2549 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2550 file. No doubt somebody will tell me if they don't make sense... Also added
2551 Dan Mooney's comments about building on OpenVMS.
2553 19. Added support for partial matching via the PCRE_PARTIAL option for
2554 pcre_exec() and the \P data escape in pcretest.
2556 20. Extended pcretest with 3 new pattern features:
2558 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2559 write the compiled pattern to the file whose name is "rest-of-line".
2560 This is a straight binary dump of the data, with the saved pointer to
2561 the character tables forced to be NULL. The study data, if any, is
2562 written too. After writing, pcretest reads a new pattern.
2564 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2565 compiled pattern from the given file. There must not be any
2566 occurrences of "<" in the file name (pretty unlikely); if there are,
2567 pcretest will instead treat the initial "<" as a pattern delimiter.
2568 After reading in the pattern, pcretest goes on to read data lines as
2569 usual.
2571 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2572 and 16-bit fields in a compiled pattern, to simulate a pattern that
2573 was compiled on a host of opposite endianness.
2575 21. The pcre-exec() function can now cope with patterns that were compiled on
2576 hosts of opposite endianness, with this restriction:
2578 As for any compiled expression that is saved and used later, the tables
2579 pointer field cannot be preserved; the extra_data field in the arguments
2580 to pcre_exec() should be used to pass in a tables address if a value
2581 other than the default internal tables were used at compile time.
2583 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2584 now diagnosed as an error. Previously, most of the time, a negative number
2585 would have been treated as zero, but if in addition "ovector" was passed as
2586 NULL, a crash could occur.
2588 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2589 new versions from the libtool 1.5 distribution (the last one is a copy of
2590 a file called libtool.m4). This seems to have fixed the need to patch
2591 "configure" to support Darwin 1.3 (which I used to do). However, I still
2592 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2593 workstation).
2595 24. Changed the PCRE licence to be the more standard "BSD" licence.
2598 Version 4.5 01-Dec-03
2599 ---------------------
2601 1. There has been some re-arrangement of the code for the match() function so
2602 that it can be compiled in a version that does not call itself recursively.
2603 Instead, it keeps those local variables that need separate instances for
2604 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2605 needs to "recurse". Keeping track of where control must go is done by means
2606 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2607 hide most of the details from the main code, and operates only if
2608 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2609 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2610 operating.
2612 To make it easier for callers to provide specially tailored get/free
2613 functions for this usage, two new functions, pcre_stack_malloc, and
2614 pcre_stack_free, are used. They are always called in strict stacking order,
2615 and the size of block requested is always the same.
2617 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2618 PCRE has been compiled to use the stack or the heap for recursion. The
2619 -C option of pcretest uses this to show which version is compiled.
2621 A new data escape \S, is added to pcretest; it causes the amounts of store
2622 obtained and freed by both kinds of malloc/free at match time to be added
2623 to the output.
2625 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2626 what's available on my current Linux desktop machine.
2628 3. When matching a UTF-8 string, the test for a valid string at the start has
2629 been extended. If start_offset is not zero, PCRE now checks that it points
2630 to a byte that is the start of a UTF-8 character. If not, it returns
2631 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2632 this is necessary because there may be backward assertions in the pattern.
2633 When matching the same subject several times, it may save resources to use
2634 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2636 4. The code for checking the validity of UTF-8 strings has been tightened so
2637 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2638 containing "overlong sequences".
2640 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2641 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2642 should have been "&", but it just so happened that all the cases this let
2643 through by mistake were picked up later in the function.
2645 6. I had used a variable called "isblank" - this is a C99 function, causing
2646 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2648 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2649 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2650 so the version is shown; (c) stop "make test" echoing "./RunTest".
2652 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2654 9. The prototype for memmove() for systems that don't have it was using
2655 size_t, but the inclusion of the header that defines size_t was later. I've
2656 moved the #includes for the C headers earlier to avoid this.
2658 10. Added some adjustments to the code to make it easier to compiler on certain
2659 special systems:
2661 (a) Some "const" qualifiers were missing.
2662 (b) Added the macro EXPORT before all exported functions; by default this
2663 is defined to be empty.
2664 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2665 that it reads its output file name as an argument instead of writing
2666 to the standard output and assuming this can be redirected.
2668 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2669 class containing characters with values greater than 255, PCRE compilation
2670 went into a loop.
2672 12. A recursive reference to a subpattern that was within another subpattern
2673 that had a minimum quantifier of zero caused PCRE to crash. For example,
2674 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2675 recursion. If the recursively-called subpattern itself had a zero repeat,
2676 that was OK.
2678 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2679 buffer into which it was copied (for escape processing) was still set at
2680 1024, so long lines caused crashes.
2682 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2683 "internal error: code overflow...". This applied to any character class
2684 that was followed by a possessive quantifier.
2686 15. Modified the Makefile to add libpcre.la as a prerequisite for
2687 libpcreposix.la because I was told this is needed for a parallel build to
2688 work.
2690 16. If a pattern that contained .* following optional items at the start was
2691 studied, the wrong optimizing data was generated, leading to matching
2692 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2693 matching string must start with a or b or c. The correct conclusion for
2694 this pattern is that a match can start with any character.
2697 Version 4.4 13-Aug-03
2698 ---------------------
2700 1. In UTF-8 mode, a character class containing characters with values between
2701 127 and 255 was not handled correctly if the compiled pattern was studied.
2702 In fixing this, I have also improved the studying algorithm for such
2703 classes (slightly).
2705 2. Three internal functions had redundant arguments passed to them. Removal
2706 might give a very teeny performance improvement.
2708 3. Documentation bug: the value of the capture_top field in a callout is *one
2709 more than* the number of the hightest numbered captured substring.
2711 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2712 in incorrectly linking with a previously installed version. They now link
2713 explicitly with libpcre.la.
2715 5. configure.in no longer needs to recognize Cygwin specially.
2717 6. A problem in pcre.in for Windows platforms is fixed.
2719 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2720 pcretest, it used to include the size of the study block as part of its
2721 output. Unfortunately, the structure contains a field that has a different
2722 size on different hardware architectures. This meant that the tests that
2723 showed this size failed. As the block is currently always of a fixed size,
2724 this information isn't actually particularly useful in pcretest output, so
2725 I have just removed it.
2727 8. Three pre-processor statements accidentally did not start in column 1.
2728 Sadly, there are *still* compilers around that complain, even though
2729 standard C has not required this for well over a decade. Sigh.
2731 9. In pcretest, the code for checking callouts passed small integers in the
2732 callout_data field, which is a void * field. However, some picky compilers
2733 complained about the casts involved for this on 64-bit systems. Now
2734 pcretest passes the address of the small integer instead, which should get
2735 rid of the warnings.
2737 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2738 both compile and run time, and gives an error if an invalid UTF-8 sequence
2739 is found. There is a option for disabling this check in cases where the
2740 string is known to be correct and/or the maximum performance is wanted.
2742 11. In response to a bug report, I changed one line in Makefile.in from
2744 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2745 to
2746 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2748 to look similar to other lines, but I have no way of telling whether this
2749 is the right thing to do, as I do not use Windows. No doubt I'll get told
2750 if it's wrong...
2753 Version 4.3 21-May-03
2754 ---------------------
2756 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2757 Makefile.
2759 2. Some refactoring to improve the quality of the code:
2761 (i) The utf8_table... variables are now declared "const".
2763 (ii) The code for \cx, which used the "case flipping" table to upper case
2764 lower case letters, now just substracts 32. This is ASCII-specific,
2765 but the whole concept of \cx is ASCII-specific, so it seems
2766 reasonable.
2768 (iii) PCRE was using its character types table to recognize decimal and
2769 hexadecimal digits in the pattern. This is silly, because it handles
2770 only 0-9, a-f, and A-F, but the character types table is locale-
2771 specific, which means strange things might happen. A private
2772 table is now used for this - though it costs 256 bytes, a table is
2773 much faster than multiple explicit tests. Of course, the standard
2774 character types table is still used for matching digits in subject
2775 strings against \d.
2777 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2778 ending in _t are). So I've renamed it as ESC_tee.
2780 3. The first argument for regexec() in the POSIX wrapper should have been
2781 defined as "const".
2783 4. Changed pcretest to use malloc() for its buffers so that they can be
2784 Electric Fenced for debugging.
2786 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2787 to read one or more bytes before the start of the subject string. Often this
2788 had no effect on PCRE's behaviour, but in some circumstances it could
2789 provoke a segmentation fault.
2791 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2792 to try to read one or more bytes before the start of the subject string.
2794 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2795 UTF-8 support could misbehave in various ways if the subject string
2796 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2797 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2798 back over UTF-8 characters.)
2801 Version 4.2 14-Apr-03
2802 ---------------------
2804 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2806 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2807 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2808 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2809 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2810 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2812 Note: automatic setting of the BUILD variables is not yet working
2813 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2814 compile-time but not at link-time
2815 [LINK]: use for linking executables only
2816 make different versions for Windows and non-Windows
2817 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2818 libraries
2819 [LINK_FOR_BUILD]: new variable
2820 [OBJEXT]: use throughout
2821 [EXEEXT]: use throughout
2822 <winshared>: new target
2823 <wininstall>: new target
2824 <dftables.o>: use native compiler
2825 <dftables>: use native linker
2826 <install>: handle Windows platform correctly
2827 <clean>: ditto
2828 <check>: ditto
2829 copy DLL to top builddir before testing
2831 As part of these changes, -no-undefined was removed again. This was reported
2832 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2833 in any case.
2835 3. Some tidies to get rid of compiler warnings:
2837 . In the match_data structure, match_limit was an unsigned long int, whereas
2838 match_call_count was an int. I've made them both unsigned long ints.
2840 . In pcretest the fact that a const uschar * doesn't automatically cast to
2841 a void * provoked a warning.
2843 . Turning on some more compiler warnings threw up some "shadow" variables
2844 and a few more missing casts.
2846 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2847 option, a class that contained a single character with a value between 128
2848 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2850 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2851 option, a class that contained several characters, but with at least one
2852 whose value was between 128 and 255 caused PCRE to crash.
2855 Version 4.1 12-Mar-03
2856 ---------------------
2858 1. Compiling with gcc -pedantic found a couple of places where casts were
2859 needed, and a string in dftables.c that was longer than standard compilers are
2860 required to support.
2862 2. Compiling with Sun's compiler found a few more places where the code could
2863 be tidied up in order to avoid warnings.
2865 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2866 first of these names is deprecated in the latest Autoconf in favour of the name
2867 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2868 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2869 analogy I have changed it to CFLAGS_FOR_BUILD.
2871 4. Added -no-undefined to the linking command in the Makefile, because this is
2872 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2873 linking step for the pcreposix library.
2875 5. PCRE was failing to diagnose the case of two named groups with the same
2876 name.
2878 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2879 literal character that is needed in the subject for a match, and scans along to
2880 ensure that it is present before embarking on the full matching process. This
2881 saves time in cases of nested unlimited repeats that are never going to match.
2882 Problem: the scan can take a lot of time if the subject is very long (e.g.
2883 megabytes), thus penalizing straightforward matches. It is now done only if the
2884 amount of subject to be scanned is less than 1000 bytes.
2886 7. A lesser problem with the same optimization is that it was recording the
2887 first character of an anchored pattern as "needed", thus provoking a search
2888 right along the subject, even when the first match of the pattern was going to
2889 fail. The "needed" character is now not set for anchored patterns, unless it
2890 follows something in the pattern that is of non-fixed length. Thus, it still
2891 fulfils its original purpose of finding quick non-matches in cases of nested
2892 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2895 Version 4.0 17-Feb-03
2896 ---------------------
2898 1. If a comment in an extended regex that started immediately after a meta-item
2899 extended to the end of string, PCRE compiled incorrect data. This could lead to
2900 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2902 2. Moved to autoconf 2.53 and libtool 1.4.2.
2904 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2905 the special perltest8 script is no longer needed - all the tests can be run
2906 from a single perltest script.
2908 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2909 by \s. It has now been removed in PCRE. This means it isn't recognized as
2910 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2911 class [:space:] *does* include VT, thereby creating a mess.
2913 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2914 space and tab.
2916 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2917 its new features into the main test script, reducing the number of scripts.
2919 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2920 were backward compatible, and made the (?i) apply to the whole pattern, as if
2921 /i were given. Now it behaves more logically, and applies the option setting
2922 only to what follows. PCRE has been changed to follow suit. However, if it
2923 finds options settings right at the start of the pattern, it extracts them into
2924 the global options, as before. Thus, they show up in the info data.
2926 8. Added support for the \Q...\E escape sequence. Characters in between are
2927 treated as literals. This is slightly different from Perl in that $ and @ are
2928 also handled as literals inside the quotes. In Perl, they will cause variable
2929 interpolation. Note the following examples:
2931 Pattern PCRE matches Perl matches
2933 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2934 \Qabc\$xyz\E abc\$xyz abc\$xyz
2935 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2937 For compatibility with Perl, \Q...\E sequences are recognized inside character
2938 classes as well as outside them.
2940 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2941 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2942 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2943 signed/unsigned warnings.
2945 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2946 option for pcretest, so I've replaced it by a simple function that does just
2947 that job.
2949 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2950 "pcregrep -".
2952 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2953 Java package. This provides some syntactic sugar for simple cases of what my
2954 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2955 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2956 item, you can use this simplified notation. Note that only makes sense with
2957 greedy quantifiers. Consequently, the use of the possessive quantifier forces
2958 greediness, whatever the setting of the PCRE_UNGREEDY option.
2960 13. A change of greediness default within a pattern was not taking effect at
2961 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2962 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2963 was abstracted outside.
2965 14. PCRE now supports the \G assertion. It is true when the current matching
2966 position is at the start point of the match. This differs from \A when the
2967 starting offset is non-zero. Used with the /g option of pcretest (or similar
2968 code), it works in the same way as it does for Perl's /g option. If all
2969 alternatives of a regex begin with \G, the expression is anchored to the start
2970 match position, and the "anchored" flag is set in the compiled expression.
2972 15. Some bugs concerning the handling of certain option changes within patterns
2973 have been fixed. These applied to options other than (?ims). For example,
2974 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2975 been the other way round. Some of this was related to change 7 above.
2977 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2978 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2979 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2980 POSIX classes only within a class (e.g. /[[:alpha:]]/).
2982 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2983 mode. Unlike ".", it always matches newline, whatever the setting of
2984 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2985 assertions. Perl allows it, but it doesn't (in general) work because it can't
2986 calculate the length of the lookbehind. At least, that's the case for Perl
2987 5.8.0 - I've been told they are going to document that it doesn't work in
2988 future.
2990 18. Added an error diagnosis for escapes that PCRE does not support: these are
2991 \L, \l, \N, \P, \p, \U, \u, and \X.
2993 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2994 reading past the end of the pattern in cases such as /[abcd/.
2996 20. PCRE was getting more memory than necessary for patterns with classes that
2997 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2999 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3000 compiling PCRE for use with Virtual Pascal.
3002 22. Small fix to the Makefile to make it work properly if the build is done
3003 outside the source tree.
3005 23. Added a new extension: a condition to go with recursion. If a conditional
3006 subpattern starts with (?(R) the "true" branch is used if recursion has
3007 happened, whereas the "false" branch is used only at the top level.
3009 24. When there was a very long string of literal characters (over 255 bytes
3010 without UTF support, over 250 bytes with UTF support), the computation of how
3011 much memory was required could be incorrect, leading to segfaults or other
3012 strange effects.
3014 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3015 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3016 there was a subsequent back reference to those brackets. This meant that, for
3017 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3018 possible to check for precisely this case. All we can do is abandon the
3019 optimization if .* occurs inside capturing brackets when there are any back
3020 references whatsoever. (See below for a better fix that came later.)
3022 26. The handling of the optimization for finding the first character of a
3023 non-anchored pattern, and for finding a character that is required later in the
3024 match were failing in some cases. This didn't break the matching; it just
3025 failed to optimize when it could. The way this is done has been re-implemented.
3027 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3029 28. Added a new feature that provides some of the functionality that Perl
3030 provides with (?{...}). The facility is termed a "callout". The way it is done
3031 in PCRE is for the caller to provide an optional function, by setting
3032 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3033 global variable. By default it is unset, which disables all calling out. To get
3034 the function called, the regex must include (?C) at appropriate points. This
3035 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3036 This provides a means of identifying different callout points. When PCRE
3037 reaches such a point in the regex, if pcre_callout has been set, the external
3038 function is called. It is provided with data in a structure called
3039 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3040 matching continues; if it returns a non-zero value, the match at the current
3041 point fails. However, backtracking will occur if possible. [This was changed
3042 later and other features added - see item 49 below.]
3044 29. pcretest is upgraded to test the callout functionality. It provides a
3045 callout function that displays information. By default, it shows the start of
3046 the match and the current position in the text. There are some new data escapes
3047 to vary what happens:
3049 \C+ in addition, show current contents of captured substrings
3050 \C- do not supply a callout function
3051 \C!n return 1 when callout number n is reached
3052 \C!n!m return 1 when callout number n is reached for the mth time
3054 30. If pcregrep was called with the -l option and just a single file name, it
3055 output "<stdin>" if a match was found, instead of the file name.
3057 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3058 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3059 pcre_exec(). This saves a malloc/free per call. The default value of
3060 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3061 when configuring.
3063 32. The default maximum size of a compiled pattern is 64K. There have been a
3064 few cases of people hitting this limit. The code now uses macros to handle the
3065 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3066 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3067 configuring. Tests 2 and 5 work only with 2-byte links because they output
3068 debugging information about compiled patterns.
3070 33. Internal code re-arrangements:
3072 (a) Moved the debugging function for printing out a compiled regex into
3073 its own source file (printint.c) and used #include to pull it into
3074 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3075 separate copies.
3077 (b) Defined the list of op-code names for debugging as a macro in
3078 internal.h so that it is next to the definition of the opcodes.
3080 (c) Defined a table of op-code lengths for simpler skipping along compiled
3081 code. This is again a macro in internal.h so that it is next to the
3082 definition of the opcodes.
3084 34. Added support for recursive calls to individual subpatterns, along the
3085 lines of Robin Houston's patch (but implemented somewhat differently).
3087 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3088 allow it to read and process whole directories in Win32. This code was
3089 contributed by Lionel Fourquaux; it has not been tested by me.
3091 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3092 used to name a group. Names consist of alphanumerics and underscores, and must
3093 be unique. Back references use the syntax (?P=name) and recursive calls use
3094 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3095 numbers. The function pcre_fullinfo() can be used after compilation to extract
3096 a name/number map. There are three relevant calls:
3098 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3099 PCRE_INFO_NAMECOUNT yields the number of entries
3100 PCRE_INFO_NAMETABLE yields a pointer to the map.
3102 The map is a vector of fixed-size entries. The size of each entry depends on
3103 the length of the longest name used. The first two bytes of each entry are the
3104 group number, most significant byte first. There follows the corresponding
3105 name, zero terminated. The names are in alphabetical order.
3107 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3108 case instead of 255. Making it the same both with and without UTF-8 support
3109 means that the same test output works with both.
3111 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3112 calling malloc() with a zero argument.
3114 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3115 optimization. I've improved things by keeping a bitmap of backreferences with
3116 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3117 fact referenced, the optimization can be applied. It is unlikely that a
3118 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3119 the match to follow \n) will appear inside brackets with a number greater than
3120 31, but if it does, any back reference > 31 suppresses the optimization.
3122 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3123 of disabling numbered capturing parentheses. Any opening parenthesis that is
3124 not followed by ? behaves as if it were followed by ?: but named parentheses
3125 can still be used for capturing (and they will acquire numbers in the usual
3126 way).
3128 41. Redesigned the return codes from the match() function into yes/no/error so
3129 that errors can be passed back from deep inside the nested calls. A malloc
3130 failure while inside a recursive subpattern call now causes the
3131 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3133 42. It is now possible to set a limit on the number of times the match()
3134 function is called in a call to pcre_exec(). This facility makes it possible to
3135 limit the amount of recursion and backtracking, though not in a directly
3136 obvious way, because the match() function is used in a number of different
3137 circumstances. The count starts from zero for each position in the subject
3138 string (for non-anchored patterns). The default limit is, for compatibility, a
3139 large number, namely 10 000 000. You can change this in two ways:
3141 (a) When configuring PCRE before making, you can use --with-match-limit=n
3142 to set a default value for the compiled library.
3144 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3145 a different value is set. See 45 below.
3147 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3149 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3150 of things that can be changed at compile time. The first argument specifies
3151 what is wanted and the second points to where the information is to be placed.
3152 The current list of available information is:
3156 The output is an integer that is set to one if UTF-8 support is available;
3157 otherwise it is set to zero.
3161 The output is an integer that it set to the value of the code that is used for
3162 newline. It is either LF (10) or CR (13).
3166 The output is an integer that contains the number of bytes used for internal
3167 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3171 The output is an integer that contains the threshold above which the POSIX
3172 interface uses malloc() for output vectors. See item 31 above.
3176 The output is an unsigned integer that contains the default limit of the number
3177 of match() calls in a pcre_exec() execution. See 42 above.
3179 44. pcretest has been upgraded by the addition of the -C option. This causes it
3180 to extract all the available output from the new pcre_config() function, and to
3181 output it. The program then exits immediately.
3183 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3184 order to support additional features. One way would have been to define
3185 pcre_exec2() (for example) with extra arguments, but this would not have been
3186 extensible, and would also have required all calls to the original function to
3187 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3188 is used for passing in "extra" data from pcre_study().
3190 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3191 contains the following fields:
3193 flags a bitmap indicating which of the following fields are set
3194 study_data opaque data from pcre_study()
3195 match_limit a way of specifying a limit on match() calls for a specific
3196 call to pcre_exec()
3197 callout_data data for callouts (see 49 below)
3199 The flag bits are also defined in pcre.h, and are
3205 The pcre_study() function now returns one of these new pcre_extra blocks, with
3206 the actual study data pointed to by the study_data field, and the
3207 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3208 before. That is, this change is entirely upwards-compatible and requires no
3209 change to existing code.
3211 If you want to pass in additional data to pcre_exec(), you can either place it
3212 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3213 block.
3215 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3216 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3217 times with different match limits, until it finds the minimum value needed for
3218 pcre_exec() to complete. The value is then output. This can be instructive; for
3219 most simple matches the number is quite small, but for pathological cases it
3220 gets very large very quickly.
3222 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3223 returns the size of the data block pointed to by the study_data field in a
3224 pcre_extra block, that is, the value that was passed as the argument to
3225 pcre_malloc() when PCRE was getting memory in which to place the information
3226 created by pcre_study(). The fourth argument should point to a size_t variable.
3227 pcretest has been extended so that this information is shown after a successful
3228 pcre_study() call when information about the compiled regex is being displayed.
3230 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3231 because what follows is always an absolute path. (Later: it turns out that this
3232 is more than cosmetic for MinGW, because it doesn't like empty path
3233 components.)
3235 49. Some changes have been made to the callout feature (see 28 above):
3237 (i) A callout function now has three choices for what it returns:
3239 0 => success, carry on matching
3240 > 0 => failure at this point, but backtrack if possible
3241 < 0 => serious error, return this value from pcre_exec()
3243 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3244 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3245 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3246 use by callout functions. It will never be used by PCRE itself.
3248 (ii) The pcre_extra structure (see 45 above) has a void * field called
3249 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3250 pcre_callout_block structure has a field of the same name. The contents of
3251 the field passed in the pcre_extra structure are passed to the callout
3252 function in the corresponding field in the callout block. This makes it
3253 easier to use the same callout-containing regex from multiple threads. For
3254 testing, the pcretest program has a new data escape
3256 \C*n pass the number n (may be negative) as callout_data
3258 If the callout function in pcretest receives a non-zero value as
3259 callout_data, it returns that value.
3261 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3262 there were some redundant $(CFLAGS) in commands that are now specified as
3263 $(LINK), which already includes $(CFLAGS).
3265 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3266 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3267 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3268 one-byte characters throughout. Note that case-insensitive matching applies
3269 only to characters whose values are less than 256. PCRE doesn't support the
3270 notion of cases for higher-valued characters.
3272 (i) A character class whose characters are all within 0-255 is handled as
3273 a bit map, and the map is inverted for negative classes. Previously, a
3274 character > 255 always failed to match such a class; however it should
3275 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3277 (ii) A negated character class with a single character < 255 is coded as
3278 "not this character" (OP_NOT). This wasn't working properly when the test
3279 character was multibyte, either singly or repeated.
3281 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3282 mode, for example: \x{100}{2,3}.
3284 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3285 singly or repeated) now correctly test multibyte characters. However,
3286 PCRE doesn't recognize any characters with values greater than 255 as
3287 digits, spaces, or word characters. Such characters always match \D, \S,
3288 and \W, and never match \d, \s, or \w.
3290 (v) Classes may now contain characters and character ranges with values
3291 greater than 255. For example: [ab\x{100}-\x{400}].
3293 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3294 PCRE in UTF-8 mode.
3296 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3297 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3298 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3299 value.)
3301 53. The single man page has become too large. I have therefore split it up into
3302 a number of separate man pages. These also give rise to individual HTML pages;
3303 these are now put in a separate directory, and there is an index.html page that
3304 lists them all. Some hyperlinking between the pages has been installed.
3306 54. Added convenience functions for handling named capturing parentheses.
3308 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3309 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3310 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3311 are faulted.
3313 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3314 calling configure. These values are used when compiling the dftables.c program
3315 which is run to generate the source of the default character tables. They
3316 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3317 you will need to set these values.
3319 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3322 Version 3.9 02-Jan-02
3323 ---------------------
3325 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3327 2. If --disable-static was given, the building process failed when trying to
3328 build pcretest and pcregrep. (For some reason it was using libtool to compile
3329 them, which is not right, as they aren't part of the library.)
3332 Version 3.8 18-Dec-01
3333 ---------------------
3335 1. The experimental UTF-8 code was completely screwed up. It was packing the
3336 bytes in the wrong order. How dumb can you get?
3339 Version 3.7 29-Oct-01
3340 ---------------------
3342 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3343 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3344 this didn't happen under Solaris 8, where I normally test things.
3346 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3347 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3348 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3349 this fix an hour or so after the initial 3.7 release.)
3352 Version 3.6 23-Oct-01
3353 ---------------------
3355 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3356 offsets passed as NULL with zero offset count.
3358 2. The config.guess and config.sub files had not been updated when I moved to
3359 the latest autoconf.
3362 Version 3.5 15-Aug-01
3363 ---------------------
3365 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3366 had been forgotten.
3368 2. By using declared but undefined structures, we can avoid using "void"
3369 definitions in pcre.h while keeping the internal definitions of the structures
3370 private.
3372 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3373 user point of view, this means that both static and shared libraries are built
3374 by default, but this can be individually controlled. More of the work of
3375 handling this static/shared cases is now inside libtool instead of PCRE's make
3376 file.
3378 4. The pcretest utility is now installed along with pcregrep because it is
3379 useful for users (to test regexs) and by doing this, it automatically gets
3380 relinked by libtool. The documentation has been turned into a man page, so
3381 there are now .1, .txt, and .html versions in /doc.
3383 5. Upgrades to pcregrep:
3384 (i) Added long-form option names like gnu grep.
3385 (ii) Added --help to list all options with an explanatory phrase.
3386 (iii) Added -r, --recursive to recurse into sub-directories.
3387 (iv) Added -f, --file to read patterns from a file.
3389 6. pcre_exec() was referring to its "code" argument before testing that
3390 argument for NULL (and giving an error if it was NULL).
3392 7. Upgraded Makefile.in to allow for compiling in a different directory from
3393 the source directory.
3395 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3396 options bits, the pointer it was passed was to an int instead of to an unsigned
3397 long int. This mattered only on 64-bit systems.
3399 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3400 generated) instead of pcre.in, which it its source. Also made the same change
3401 in several of the .c files.
3403 10. A new release of gcc defines printf() as a macro, which broke pcretest
3404 because it had an ifdef in the middle of a string argument for printf(). Fixed
3405 by using separate calls to printf().
3407 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3408 script, to force use of CR or LF instead of \n in the source. On non-Unix
3409 systems, the value can be set in config.h.
3411 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3412 absolute limit. Changed the text of the error message to make this clear, and
3413 likewise updated the man page.
3415 13. The limit of 99 on the number of capturing subpatterns has been removed.
3416 The new limit is 65535, which I hope will not be a "real" limit.
3419 Version 3.4 22-Aug-00
3420 ---------------------
3422 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3424 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3427 Version 3.3 01-Aug-00
3428 ---------------------
3430 1. If an octal character was given, but the value was greater than \377, it
3431 was not getting masked to the least significant bits, as documented. This could
3432 lead to crashes in some systems.
3434 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3435 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3437 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3438 These just pass their arguments on to (pcre_free)(), but they are provided
3439 because some uses of PCRE bind it to non-C systems that can call its functions,
3440 but cannot call free() or pcre_free() directly.
3442 4. Add "make test" as a synonym for "make check". Corrected some comments in
3443 the Makefile.
3445 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3446 Makefile.
3448 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3449 command called pgrep for grepping around the active processes.
3451 7. Added the beginnings of support for UTF-8 character strings.
3453 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3454 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3455 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3456 out for the ar command.)
3459 Version 3.2 12-May-00
3460 ---------------------
3462 This is purely a bug fixing release.
3464 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3465 of ZA. This was just one example of several cases that could provoke this bug,
3466 which was introduced by change 9 of version 2.00. The code for breaking
3467 infinite loops after an iteration that matches an empty string was't working
3468 correctly.
3470 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3471 when matched against abbab (for example). After matching an empty string, it
3472 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3473 caused it to match further down the string than it should.
3475 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3476 was there because it doesn't seem to be needed, and it causes trouble on some
3477 systems, as it is not a Standard C header. It has been removed.
3479 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3480 were reported on the Macintosh. The changes were from
3482 while ((c = *(++ptr)) != 0 && c != '\n');
3483 to
3484 while ((c = *(++ptr)) != 0 && c != '\n') ;
3486 Totally extraordinary, but if that's what it takes...
3488 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3489 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3490 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3491 assumes the way PCRE uses memmove() (always moving upwards).
3493 6. PCRE is being used in one environment where strchr() is not available. There
3494 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3495 faster code anyway.
3498 Version 3.1 09-Feb-00
3499 ---------------------
3501 The only change in this release is the fixing of some bugs in Makefile.in for
3502 the "install" target:
3504 (1) It was failing to install pcreposix.h.
3506 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3509 Version 3.0 01-Feb-00
3510 ---------------------
3512 1. Add support for the /+ modifier to perltest (to output $` like it does in
3513 pcretest).
3515 2. Add support for the /g modifier to perltest.
3517 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3518 matches null strings.
3520 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3521 pattern. Perl treats empty patterns specially - it reuses the most recent
3522 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3523 effect.
3525 5. The POSIX interface was broken in that it was just handing over the POSIX
3526 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3527 required a bigger vector, with some working space on the end. This means that
3528 the POSIX wrapper now has to get and free some memory, and copy the results.
3530 6. Added some simple autoconf support, placing the test data and the
3531 documentation in separate directories, re-organizing some of the
3532 information files, and making it build pcre-config (a GNU standard). Also added
3533 libtool support for building PCRE as a shared library, which is now the
3534 default.
3536 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3537 09 are not valid octal constants. Single digits will be used for minor values
3538 less than 10.
3540 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3541 existing programs that set these in the POSIX interface can use PCRE without
3542 modification.
3544 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3545 return all that pcre_info() returns, plus additional data. The pcre_info()
3546 function is retained for compatibility, but is considered to be obsolete.
3548 10. Added experimental recursion feature (?R) to handle one common case that
3549 Perl 5.6 will be able to do with (?p{...}).
3551 11. Added support for POSIX character classes like [:alpha:], which Perl is
3552 adopting.
3555 Version 2.08 31-Aug-99
3556 ----------------------
3558 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3559 trying to match at the startoffset position, but instead was moving forward to
3560 the next newline as if a previous match had failed.
3562 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3563 and could get into a loop if a null string was matched other than at the start
3564 of the subject.
3566 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3567 be distinguished at compile time, and for completeness also added PCRE_DATE.
3569 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3570 in GnuWin32 environments.
3573 Version 2.07 29-Jul-99
3574 ----------------------
3576 1. The documentation is now supplied in plain text form and HTML as well as in
3577 the form of man page sources.
3579 2. C++ compilers don't like assigning (void *) values to other pointer types.
3580 In particular this affects malloc(). Although there is no problem in Standard
3581 C, I've put in casts to keep C++ compilers happy.
3583 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3584 should be (const char *).
3586 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3587 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3588 However, I haven't made this a standard facility. The documentation doesn't
3589 mention it, and the Makefile doesn't support it.
3591 5. The Makefile now contains an "install" target, with editable destinations at
3592 the top of the file. The pcretest program is not installed.
3594 6. pgrep -V now gives the PCRE version number and date.
3596 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3597 causing the entire string to be ignored, instead of just the last character.
3599 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3600 non-matching string, it can take a very, very long time, even for strings of
3601 quite modest length, because of the nested recursion. PCRE now does better in
3602 some of these cases. It does this by remembering the last required literal
3603 character in the pattern, and pre-searching the subject to ensure it is present
3604 before running the real match. In other words, it applies a heuristic to detect
3605 some types of certain failure quickly, and in the above example, if presented
3606 with a string that has no trailing " it gives "no match" very quickly.
3608 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3609 other alternatives are tried instead.
3612 Version 2.06 09-Jun-99
3613 ----------------------
3615 1. Change pcretest's output for amount of store used to show just the code
3616 space, because the remainder (the data block) varies in size between 32-bit and
3617 64-bit systems.
3619 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3620 start matching at. This allows lookbehinds to work when searching for multiple
3621 occurrences in a string.
3623 3. Added additional options to pcretest for testing multiple occurrences:
3625 /+ outputs the rest of the string that follows a match
3626 /g loops for multiple occurrences, using the new startoffset argument
3627 /G loops for multiple occurrences by passing an incremented pointer
3629 4. PCRE wasn't doing the "first character" optimization for patterns starting
3630 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3631 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3632 the letter 'x'. On long subject strings, this gives a significant speed-up.
3635 Version 2.05 21-Apr-99
3636 ----------------------
3638 1. Changed the type of magic_number from int to long int so that it works
3639 properly on 16-bit systems.
3641 2. Fixed a bug which caused patterns starting with .* not to work correctly
3642 when the subject string contained newline characters. PCRE was assuming
3643 anchoring for such patterns in all cases, which is not correct because .* will
3644 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3645 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3646 must be retried after every newline in the subject.
3649 Version 2.04 18-Feb-99
3650 ----------------------
3652 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3653 computation of the store needed to hold the pattern was incorrect (too large).
3654 If such patterns were nested a few deep, this could multiply and become a real
3655 problem.
3657 2. Added /M option to pcretest to show the memory requirement of a specific
3658 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3660 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3661 compiled in such a way that the backtracking after subsequent failure was
3662 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3663 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3666 Version 2.03 02-Feb-99
3667 ----------------------
3669 1. Fixed typo and small mistake in man page.
3671 2. Added 4th condition (GPL supersedes if conflict) and created separate
3672 LICENCE file containing the conditions.
3674 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3675 Perl, that is the internal \ allows the delimiter to be included in the
3676 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3677 the final delimiter, add \ to the end of the pattern (to test the error).
3679 4. Added the convenience functions for extracting substrings after a successful
3680 match. Updated pcretest to make it able to test these functions.
3683 Version 2.02 14-Jan-99
3684 ----------------------
3686 1. Initialized the working variables associated with each extraction so that
3687 their saving and restoring doesn't refer to uninitialized store.
3689 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3690 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3691 fix the problem.
3693 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3694 calls, and wasn't printing the correct value for compiling calls. Increased the
3695 default value of LOOPREPEAT, and the number of significant figures in the
3696 times.
3698 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3700 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3701 a building problem on Windows NT with a FAT file system.
3704 Version 2.01 21-Oct-98
3705 ----------------------
3707 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3708 to character tables built by pcre_maketables() in the current locale. If NULL
3709 is passed, the default tables are used.
3712 Version 2.00 24-Sep-98
3713 ----------------------
3715 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3716 it any more.
3718 2. Allow quantification of (?>) groups, and make it work correctly.
3720 3. The first character computation wasn't working for (?>) groups.
3722 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3723 end of the subject) and add 5.005's \z, which really does match only at the
3724 very end of the subject.
3726 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3728 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3729 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3730 localized options. All options to pcre_study() were also removed.
3732 7. Add other new features from 5.005:
3734 $(?<= positive lookbehind
3735 $(?<! negative lookbehind
3736 (?imsx-imsx) added the unsetting capability
3737 such a setting is global if at outer level; local otherwise
3738 (?imsx-imsx:) non-capturing groups with option setting
3739 (?(cond)re|re) conditional pattern matching
3741 A backreference to itself in a repeated group matches the previous
3742 captured string.
3744 8. General tidying up of studying (both automatic and via "study")
3745 consequential on the addition of new assertions.
3747 9. As in 5.005, unlimited repeated groups that could match an empty substring
3748 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3749 runtime if any iteration does actually match an empty substring.
3751 10. Include the RunTest script in the distribution.
3753 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3754 discrepancies, some of which were old and were also with respect to 5.004. They
3755 have now been fixed.
3758 Version 1.09 28-Apr-98
3759 ----------------------
3761 1. A negated single character class followed by a quantifier with a minimum
3762 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3763 program crashes, or just wrong answers. This did not apply to negated classes
3764 containing more than one character, or to minima other than one.
3767 Version 1.08 27-Mar-98
3768 ----------------------
3770 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3772 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3773 latter must appear before anything that relies on it in the pattern.
3776 Version 1.07 16-Feb-98
3777 ----------------------
3779 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3780 repeat of a potentially empty string).
3783 Version 1.06 23-Jan-98
3784 ----------------------
3786 1. Added Markus Oberhumer's little patches for C++.
3788 2. Literal strings longer than 255 characters were broken.
3791 Version 1.05 23-Dec-97
3792 ----------------------
3794 1. Negated character classes containing more than one character were failing if
3795 PCRE_CASELESS was set at run time.
3798 Version 1.04 19-Dec-97
3799 ----------------------
3801 1. Corrected the man page, where some "const" qualifiers had been omitted.
3803 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3804 input syntax.
3806 3. Fixed memory leak which occurred when a regex with back references was
3807 matched with an offsets vector that wasn't big enough. The temporary memory
3808 that is used in this case wasn't being freed if the match failed.
3810 4. Tidied pcretest to ensure it frees memory that it gets.
3812 5. Temporary memory was being obtained in the case where the passed offsets
3813 vector was exactly big enough.
3815 6. Corrected definition of offsetof() from change 5 below.
3817 7. I had screwed up change 6 below and broken the rules for the use of
3818 setjmp(). Now fixed.
3821 Version 1.03 18-Dec-97
3822 ----------------------
3824 1. A erroneous regex with a missing opening parenthesis was correctly
3825 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3826 on some systems.
3828 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3829 it was reported that one broken compiler failed on the former because "code" is
3830 also an independent variable.
3832 3. The erroneous regex a[]b caused an array overrun reference.
3834 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3835 fail on data ending with that character. (It was going on too far, and checking
3836 the next character, typically a binary zero.) This was specific to the
3837 optimized code for single-character negative classes.
3839 5. Added a contributed patch from the TIN world which does the following:
3841 + Add an undef for memmove, in case the the system defines a macro for it.
3843 + Add a definition of offsetof(), in case there isn't one. (I don't know
3844 the reason behind this - offsetof() is part of the ANSI standard - but
3845 it does no harm).
3847 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3848 most of the places where whitespace preceded '#'. I have given up and
3849 allowed the remaining 2 cases to be at the margin.
3851 + Rename some variables in pcre to eliminate shadowing. This seems very
3852 pedantic, but does no harm, of course.
3854 6. Moved the call to setjmp() into its own function, to get rid of warnings
3855 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3857 7. Constructs such as \d{8,} were compiling into the equivalent of
3858 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3859 outcome, but in this particular case used more store than had been allocated,
3860 which caused the bug to be discovered because it threw up an internal error.
3862 8. The debugging code in both pcre and pcretest for outputting the compiled
3863 form of a regex was going wrong in the case of back references followed by
3864 curly-bracketed repeats.
3867 Version 1.02 12-Dec-97
3868 ----------------------
3870 1. Typos in pcre.3 and comments in the source fixed.
3872 2. Applied a contributed patch to get rid of places where it used to remove
3873 'const' from variables, and fixed some signed/unsigned and uninitialized
3874 variable warnings.
3876 3. Added the "runtest" target to Makefile.
3878 4. Set default compiler flag to -O2 rather than just -O.
3881 Version 1.01 19-Nov-97
3882 ----------------------
3884 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3885 like /([ab]*)*/, that is, for classes with more than one character in them.
3887 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3888 as /((?>a*))*/ (a PCRE_EXTRA facility).
3891 Version 1.00 18-Nov-97
3892 ----------------------
3894 1. Added compile-time macros to support systems such as SunOS4 which don't have
3895 memmove() or strerror() but have other things that can be used instead.
3897 2. Arranged that "make clean" removes the executables.
3900 Version 0.99 27-Oct-97
3901 ----------------------
3903 1. Fixed bug in code for optimizing classes with only one character. It was
3904 initializing a 32-byte map regardless, which could cause it to run off the end
3905 of the memory it had got.
3907 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3910 Version 0.98 22-Oct-97
3911 ----------------------
3913 1. Fixed bug in code for handling temporary memory usage when there are more
3914 back references than supplied space in the ovector. This could cause segfaults.
3917 Version 0.97 21-Oct-97
3918 ----------------------
3920 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3922 2. Optimized negated single characters not to use a bit map.
3924 3. Brought error texts together as macro definitions; clarified some of them;
3925 fixed one that was wrong - it said "range out of order" when it meant "invalid
3926 escape sequence".
3928 4. Changed some char * arguments to const char *.
3930 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3932 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3933 pcretest.
3936 Version 0.96 16-Oct-97
3937 ----------------------
3939 1. Added a simple "pgrep" utility to the distribution.
3941 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3942 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3943 where "ddd" means "one or more decimal digits".
3945 3. Fixed serious bug. If a pattern had a back reference, but the call to
3946 pcre_exec() didn't supply a large enough ovector to record the related
3947 identifying subpattern, the match always failed. PCRE now remembers the number
3948 of the largest back reference, and gets some temporary memory in which to save
3949 the offsets during matching if necessary, in order to ensure that
3950 backreferences always work.
3952 4. Increased the compatibility with Perl in a number of ways:
3954 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3955 to request this handling. The option can be set at compile or exec time.
3957 (b) $ matches before a terminating newline by default; an option
3958 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3959 mode). The option can be set at compile or exec time.
3961 (c) The handling of \ followed by a digit other than 0 is now supposed to be
3962 the same as Perl's. If the decimal number it represents is less than 10
3963 or there aren't that many previous left capturing parentheses, an octal
3964 escape is read. Inside a character class, it's always an octal escape,
3965 even if it is a single digit.
3967 (d) An escaped but undefined alphabetic character is taken as a literal,
3968 unless PCRE_EXTRA is set. Currently this just reserves the remaining
3969 escapes.
3971 (e) {0} is now permitted. (The previous item is removed from the compiled
3972 pattern).
3974 5. Changed all the names of code files so that the basic parts are no longer
3975 than 10 characters, and abolished the teeny "globals.c" file.
3977 6. Changed the handling of character classes; they are now done with a 32-byte
3978 bit map always.
3980 7. Added the -d and /D options to pcretest to make it possible to look at the
3981 internals of compilation without having to recompile pcre.
3984 Version 0.95 23-Sep-97
3985 ----------------------
3987 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3988 \x20 at the start of a run of normal characters. These were being treated as
3989 real characters, instead of the source characters being re-checked.
3992 Version 0.94 18-Sep-97
3993 ----------------------
3995 1. The functions are now thread-safe, with the caveat that the global variables
3996 containing pointers to malloc() and free() or alternative functions are the
3997 same for all threads.
3999 2. Get pcre_study() to generate a bitmap of initial characters for non-
4000 anchored patterns when this is possible, and use it if passed to pcre_exec().
4003 Version 0.93 15-Sep-97
4004 ----------------------
4006 1. /(b)|(:+)/ was computing an incorrect first character.
4008 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4009 but not actually doing anything yet.
4011 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4012 as Perl does (e.g. [-az] or [az-]).
4014 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4015 all possible positions.
4017 5. Split up into different modules to avoid including unneeded functions in a
4018 compiled binary. However, compile and exec are still in one module. The "study"
4019 function is split off.
4021 6. The character tables are now in a separate module whose source is generated
4022 by an auxiliary program - but can then be edited by hand if required. There are
4023 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4024 toupper() in the code.
4026 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4027 make them global. Abolish the function for setting them, as the caller can now
4028 set them directly.
4031 Version 0.92 11-Sep-97
4032 ----------------------
4034 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4035 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4037 2. Caseless matching was not working in character classes if the characters in
4038 the pattern were in upper case.
4040 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4042 4. Make PCRE_ANCHORED public and accept as a compile option.
4044 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4045 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4046 pass them.
4048 6. Give an error if bad option bits passed at compile or run time.
4050 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4051 pcretest to cause it to pass that flag.
4053 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4054 options, and the first character, if set.
4056 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4059 Version 0.91 10-Sep-97
4060 ----------------------
4062 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4063 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4065 2. PCRE was looping on encountering an indefinitely repeated back reference to
4066 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4067 Perl does - treats the match as successful.
4069 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5