ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 586 - (show annotations)
Wed Jan 12 17:36:47 2011 UTC (4 years, 8 months ago) by ph10
File size: 197627 byte(s)
Error occurred while calculating annotation data.
Move definition of pcregrep_exit() above its first reference; this applies only 
to Windows.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.12 12-Jan-2011
5 ------------------------
7 1. Fixed some typos in the markup of the man pages, and wrote a script that
8 checks for such things as part of the documentation building process.
10 2. On a big-endian 64-bit system, pcregrep did not correctly process the
11 --match-limit and --recursion-limit options (added for 8.11). In
12 particular, this made one of the standard tests crash. (The integer value
13 went into the wrong half of a long int.)
15 3. If the --colour option was given to pcregrep with -v (invert match), it
16 did strange things, either producing crazy output, or crashing. It should,
17 of course, ignore a request for colour when reporting lines that do not
18 match.
20 4. If pcregrep was compiled under Windows, there was a reference to the
21 function pcregrep_exit() before it was defined. I am assuming this was
22 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
23 reported by a user. I've moved the definition above the reference.
26 Version 8.11 10-Dec-2010
27 ------------------------
29 1. (*THEN) was not working properly if there were untried alternatives prior
30 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
31 backtracked to try for "b" instead of moving to the next alternative branch
32 at the same level (in this case, to look for "c"). The Perl documentation
33 is clear that when (*THEN) is backtracked onto, it goes to the "next
34 alternative in the innermost enclosing group".
36 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
37 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
38 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
39 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
40 (*THEN).
42 3. If \s appeared in a character class, it removed the VT character from
43 the class, even if it had been included by some previous item, for example
44 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
45 of \s, but is part of the POSIX "space" class.)
47 4. A partial match never returns an empty string (because you can always
48 match an empty string at the end of the subject); however the checking for
49 an empty string was starting at the "start of match" point. This has been
50 changed to the "earliest inspected character" point, because the returned
51 data for a partial match starts at this character. This means that, for
52 example, /(?<=abc)def/ gives a partial match for the subject "abc"
53 (previously it gave "no match").
55 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
56 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
57 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
58 has an implication that the given string is incomplete (because a partial
59 match is preferred over a full match). For this reason, these items now
60 give a partial match in this situation. [Aside: previously, the one case
61 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
62 match rather than a full match, which was wrong by the old rules, but is
63 now correct.]
65 6. There was a bug in the handling of #-introduced comments, recognized when
66 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
67 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
68 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
69 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
70 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
71 places in pcre_compile().
73 7. Related to (6) above, when pcre_compile() was skipping #-introduced
74 comments when looking ahead for named forward references to subpatterns,
75 the only newline sequence it recognized was NL. It now handles newlines
76 according to the set newline convention.
78 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
79 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
80 cater for a lack of strerror(). These oversights have been fixed.
82 9. Added --match-limit and --recursion-limit to pcregrep.
84 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
86 11. When the -o option was used, pcregrep was setting a return code of 1, even
87 when matches were found, and --line-buffered was not being honoured.
89 12. Added an optional parentheses number to the -o and --only-matching options
90 of pcregrep.
92 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
93 can match an empty string. The code to do it in pcretest and pcredemo
94 needed fixing:
96 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
97 only one byte after an empty string match just before CRLF (this case
98 just got forgotten; "any" and "anycrlf" were OK).
100 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
101 mode when an empty string match preceded an ASCII character followed by
102 a non-ASCII character. (The code for advancing by one character rather
103 than one byte was nonsense.)
105 (c) The pcredemo.c sample program did not have any code at all to handle
106 the cases when CRLF is a valid newline sequence.
108 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
109 as a starting offset was within the subject string. There is now a new
110 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
111 negative or greater than the length of the string. In order to test this,
112 pcretest is extended to allow the setting of negative starting offsets.
114 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
115 starting offset points to the beginning of a UTF-8 character was
116 unnecessarily clumsy. I tidied it up.
118 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
119 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
121 17. Nobody had reported that the --include_dir option, which was added in
122 release 7.7 should have been called --include-dir (hyphen, not underscore)
123 for compatibility with GNU grep. I have changed it to --include-dir, but
124 left --include_dir as an undocumented synonym, and the same for
125 --exclude-dir, though that is not available in GNU grep, at least as of
126 release 2.5.4.
128 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
129 characters from a string of bytes) have been redefined so as not to use
130 loops, in order to improve performance in some environments. At the same
131 time, I abstracted some of the common code into auxiliary macros to save
132 repetition (this should not affect the compiled code).
134 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
135 compile-time error is now given if \c is not followed by an ASCII
136 character, that is, a byte less than 128. (In EBCDIC mode, the code is
137 different, and any byte value is allowed.)
139 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
140 START_OPTIMIZE option, which is now allowed at compile time - but just
141 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
142 to pcregrep and other applications that have no direct access to PCRE
143 options. The new /Y option in pcretest sets this option when calling
144 pcre_compile().
146 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
147 back references. Groups containing recursive back references were forced to
148 be atomic by that change, but in the case of named groups, the amount of
149 memory required was incorrectly computed, leading to "Failed: internal
150 error: code overflow". This has been fixed.
152 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
153 pcretest.c, to avoid build problems in some Borland environments.
156 Version 8.10 25-Jun-2010
157 ------------------------
159 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
160 THEN.
162 2. (*ACCEPT) was not working when inside an atomic group.
164 3. Inside a character class, \B is treated as a literal by default, but
165 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
166 causes the error). The code is unchanged, but I tidied the documentation.
168 4. Inside a character class, PCRE always treated \R and \X as literals,
169 whereas Perl faults them if its -w option is set. I have changed PCRE so
170 that it faults them when PCRE_EXTRA is set.
172 5. Added support for \N, which always matches any character other than
173 newline. (It is the same as "." when PCRE_DOTALL is not set.)
175 6. When compiling pcregrep with newer versions of gcc which may have
176 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
177 declared with attribute warn_unused_result" were given. Just casting the
178 result to (void) does not stop the warnings; a more elaborate fudge is
179 needed. I've used a macro to implement this.
181 7. Minor change to pcretest.c to avoid a compiler warning.
183 8. Added four artifical Unicode properties to help with an option to make
184 \s etc use properties (see next item). The new properties are: Xan
185 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
187 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
188 use Unicode properties. (*UCP) at the start of a pattern can be used to set
189 this option. Modified pcretest to add /W to test this facility. Added
190 REG_UCP to make it available via the POSIX interface.
192 10. Added --line-buffered to pcregrep.
194 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
195 studied, and the match started with a letter with a code point greater than
196 127 whose first byte was different to the first byte of the other case of
197 the letter, the other case of this starting letter was not recognized
198 (#976).
200 12. If a pattern that was studied started with a repeated Unicode property
201 test, for example, \p{Nd}+, there was the theoretical possibility of
202 setting up an incorrect bitmap of starting bytes, but fortunately it could
203 not have actually happened in practice until change 8 above was made (it
204 added property types that matched character-matching opcodes).
206 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
207 possible starting bytes for non-anchored patterns.
209 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
210 \R, and also a number of cases that involve Unicode properties, both
211 explicit and implicit when PCRE_UCP is set.
213 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
214 input, it could crash or give wrong results if characters with values
215 greater than 0xc0 were present in the subject string. (Detail: it assumed
216 UTF-8 input when processing these items.)
218 16. Added a lot of (int) casts to avoid compiler warnings in systems where
219 size_t is 64-bit (#991).
221 17. Added a check for running out of memory when PCRE is compiled with
222 --disable-stack-for-recursion (#990).
224 18. If the last data line in a file for pcretest does not have a newline on
225 the end, a newline was missing in the output.
227 19. The default pcre_chartables.c file recognizes only ASCII characters (values
228 less than 128) in its various bitmaps. However, there is a facility for
229 generating tables according to the current locale when PCRE is compiled. It
230 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
231 characters, are recognized by isspace() and therefore were getting set in
232 these tables, and indeed these tables seem to approximate to ISO 8859. This
233 caused a problem in UTF-8 mode when pcre_study() was used to create a list
234 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
235 which of course cannot start UTF-8 characters. I have changed the code so
236 that only real ASCII characters (less than 128) and the correct starting
237 bytes for UTF-8 encodings are set for characters greater than 127 when in
238 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
239 altogether.)
241 20. Added the /T option to pcretest so as to be able to run tests with non-
242 standard character tables, thus making it possible to include the tests
243 used for 19 above in the standard set of tests.
245 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
246 reference to a subpattern the other side of a comment that contains an
247 opening parenthesis caused either an internal compiling error, or a
248 reference to the wrong subpattern.
251 Version 8.02 19-Mar-2010
252 ------------------------
254 1. The Unicode data tables have been updated to Unicode 5.2.0.
256 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
257 configured.
259 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
260 original author of that file, following a query about its status.
262 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
263 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
265 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
266 quantifier applied to a forward-referencing subroutine call, could compile
267 incorrect code or give the error "internal error: previously-checked
268 referenced subpattern not found".
270 6. Both MS Visual Studio and Symbian OS have problems with initializing
271 variables to point to external functions. For these systems, therefore,
272 pcre_malloc etc. are now initialized to local functions that call the
273 relevant global functions.
275 7. There were two entries missing in the vectors called coptable and poptable
276 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
277 I've fixed the data, and added a kludgy way of testing at compile time that
278 the lengths are correct (equal to the number of opcodes).
280 8. Following on from 7, I added a similar kludge to check the length of the
281 eint vector in pcreposix.c.
283 9. Error texts for pcre_compile() are held as one long string to avoid too
284 much relocation at load time. To find a text, the string is searched,
285 counting zeros. There was no check for running off the end of the string,
286 which could happen if a new error number was added without updating the
287 string.
289 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
291 11. \K was not working if it appeared in an atomic group or in a group that
292 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
293 \K is "not well defined" if used in an assertion. PCRE now accepts it if
294 the assertion is positive, but not if it is negative.
296 12. Change 11 fortuitously reduced the size of the stack frame used in the
297 "match()" function of pcre_exec.c by one pointer. Forthcoming
298 implementation of support for (*MARK) will need an extra pointer on the
299 stack; I have reserved it now, so that the stack frame size does not
300 decrease.
302 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
303 item in branch that calls a recursion is a subroutine call - as in the
304 second branch in the above example - was incorrectly given the compile-
305 time error "recursive call could loop indefinitely" because pcre_compile()
306 was not correctly checking the subroutine for matching a non-empty string.
308 14. The checks for overrunning compiling workspace could trigger after an
309 overrun had occurred. This is a "should never occur" error, but it can be
310 triggered by pathological patterns such as hundreds of nested parentheses.
311 The checks now trigger 100 bytes before the end of the workspace.
313 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
316 Version 8.01 19-Jan-2010
317 ------------------------
319 1. If a pattern contained a conditional subpattern with only one branch (in
320 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
321 computed the wrong minimum data length (which is of course zero for such
322 subpatterns). This could cause incorrect "no match" results.
324 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
325 the pattern is reset in the first branch, pcre_compile() failed with
326 "internal error: code overflow at offset...". This happened only when
327 the reset was to the original external option setting. (An optimization
328 abstracts leading options settings into an external setting, which was the
329 cause of this.)
331 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
332 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
333 assertion pattern did not match (meaning that the assertion was true), it
334 was incorrectly treated as false if the SKIP had been reached during the
335 matching. This also applied to assertions used as conditions.
337 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
338 assertion subpattern, including such a pattern used as a condition,
339 unpredictable results occurred, instead of the error return
342 5. The C++ GlobalReplace function was not working like Perl for the special
343 situation when an empty string is matched. It now does the fancy magic
344 stuff that is necessary.
346 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
347 removed. (These were left over from very, very early versions of PCRE.)
349 7. Some cosmetic changes to the code to make life easier when compiling it
350 as part of something else:
352 (a) Change DEBUG to PCRE_DEBUG.
354 (b) In pcre_compile(), rename the member of the "branch_chain" structure
355 called "current" as "current_branch", to prevent a collision with the
356 Linux macro when compiled as a kernel module.
358 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
359 prevent a collision with the Linux macro when compiled as a kernel
360 module.
362 8. In pcre_compile() there are some checks for integer overflows that used to
363 cast potentially large values to (double). This has been changed to that
364 when building, a check for int64_t is made, and if it is found, it is used
365 instead, thus avoiding the use of floating point arithmetic. (There is no
366 other use of FP in PCRE.) If int64_t is not found, the fallback is to
367 double.
369 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
370 2005 (difference between two addresses compared to an unsigned value).
372 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
373 custom one, because of the following reported problem in Windows:
375 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
376 under Win32.
377 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
378 therefore missing the function definition.
379 - The compiler thus generates a "C" signature for the test function.
380 - The linker fails to find the "C" function.
381 - PCRE fails to configure if asked to do so against libbz2.
383 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
384 messages were output:
386 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
387 rerunning libtoolize, to keep the correct libtool macros in-tree.
388 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
390 I have done both of these things.
392 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
393 most of the time, it *can* run out if it is given a pattern that contains a
394 runaway infinite recursion. I updated the discussion in the pcrestack man
395 page.
397 13. Now that we have gone to the x.xx style of version numbers, the minor
398 version may start with zero. Using 08 or 09 is a bad idea because users
399 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
400 interpreted as invalid octal numbers. I've updated the previous comment in
401 configure.ac, and also added a check that gives an error if 08 or 09 are
402 used.
404 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
405 causing partial matching to fail when the end of the subject matched \W
406 in a UTF-8 pattern where \W was quantified with a minimum of 3.
408 15. There were some discrepancies between the declarations in pcre_internal.h
409 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
410 their definitions. The declarations used "const uschar *" and the
411 definitions used USPTR. Even though USPTR is normally defined as "const
412 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
413 reported that: "This difference in casting confuses some C++ compilers, for
414 example, SunCC recognizes above declarations as different functions and
415 generates broken code for hbpcre." I have changed the declarations to use
416 USPTR.
418 16. GNU libtool is named differently on some systems. The autogen.sh script now
419 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
420 (FreeBSD).
422 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
423 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
424 comment: "Figure out how to create a longlong from a string: strtoll and
425 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
426 instance, but it only takes 2 args instead of 3!"
428 18. A subtle bug concerned with back references has been fixed by a change of
429 specification, with a corresponding code fix. A pattern such as
430 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
431 refers, was giving matches when it shouldn't. For example, xa=xaaa would
432 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
433 same bug. Such groups have to be quantified to be useful, or contained
434 inside another quantified group. (If there's no repetition, the reference
435 can never match.) The problem arises because, having left the group and
436 moved on to the rest of the pattern, a later failure that backtracks into
437 the group uses the captured value from the final iteration of the group
438 rather than the correct earlier one. I have fixed this in PCRE by forcing
439 any group that contains a reference to itself to be an atomic group; that
440 is, there cannot be any backtracking into it once it has completed. This is
441 similar to recursive and subroutine calls.
444 Version 8.00 19-Oct-09
445 ----------------------
447 1. The table for translating pcre_compile() error codes into POSIX error codes
448 was out-of-date, and there was no check on the pcre_compile() error code
449 being within the table. This could lead to an OK return being given in
450 error.
452 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
453 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
454 in a Windows environment.
456 3. The pcregrep --count option prints the count for each file even when it is
457 zero, as does GNU grep. However, pcregrep was also printing all files when
458 --files-with-matches was added. Now, when both options are given, it prints
459 counts only for those files that have at least one match. (GNU grep just
460 prints the file name in this circumstance, but including the count seems
461 more useful - otherwise, why use --count?) Also ensured that the
462 combination -clh just lists non-zero counts, with no names.
464 4. The long form of the pcregrep -F option was incorrectly implemented as
465 --fixed_strings instead of --fixed-strings. This is an incompatible change,
466 but it seems right to fix it, and I didn't think it was worth preserving
467 the old behaviour.
469 5. The command line items --regex=pattern and --regexp=pattern were not
470 recognized by pcregrep, which required --regex pattern or --regexp pattern
471 (with a space rather than an '='). The man page documented the '=' forms,
472 which are compatible with GNU grep; these now work.
474 6. No libpcreposix.pc file was created for pkg-config; there was just
475 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
477 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
478 when UCP support is not needed, by modifying the Python script that
479 generates it from Unicode data files. This should not matter if the module
480 is correctly used as a library, but I received one complaint about 50K of
481 unwanted data. My guess is that the person linked everything into his
482 program rather than using a library. Anyway, it does no harm.
484 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
485 was a minimum greater than 1 for a wide character in a possessive
486 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
487 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
488 character. Chaos in the form of incorrect output or a compiling loop could
489 result.
491 9. The restrictions on what a pattern can contain when partial matching is
492 requested for pcre_exec() have been removed. All patterns can now be
493 partially matched by this function. In addition, if there are at least two
494 slots in the offset vector, the offset of the earliest inspected character
495 for the match and the offset of the end of the subject are set in them when
496 PCRE_ERROR_PARTIAL is returned.
498 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
499 synonymous with PCRE_PARTIAL, for backwards compatibility, and
500 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
501 and may be more useful for multi-segment matching.
503 11. Partial matching with pcre_exec() is now more intuitive. A partial match
504 used to be given if ever the end of the subject was reached; now it is
505 given only if matching could not proceed because another character was
506 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
507 string "Z", which now yields "no match" instead of "partial match". In the
508 case of pcre_dfa_exec(), "no match" is given if every matching path for the
509 final character ended with (*FAIL).
511 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
512 if the pattern had a "must contain" character that was already found in the
513 earlier partial match, unless partial matching was again requested. For
514 example, with the pattern /dog.(body)?/, the "must contain" character is
515 "g". If the first part-match was for the string "dog", restarting with
516 "sbody" failed. This bug has been fixed.
518 13. The string returned by pcre_dfa_exec() after a partial match has been
519 changed so that it starts at the first inspected character rather than the
520 first character of the match. This makes a difference only if the pattern
521 starts with a lookbehind assertion or \b or \B (\K is not supported by
522 pcre_dfa_exec()). It's an incompatible change, but it makes the two
523 matching functions compatible, and I think it's the right thing to do.
525 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
526 so that the demonstration program is easily available in environments where
527 PCRE has not been installed from source.
529 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
530 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
531 library.
533 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
534 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
535 is not the first non-POSIX option to be added. Clearly some people find
536 these options useful.
538 17. If a caller to the POSIX matching function regexec() passes a non-zero
539 value for nmatch with a NULL value for pmatch, the value of
540 nmatch is forced to zero.
542 18. RunGrepTest did not have a test for the availability of the -u option of
543 the diff command, as RunTest does. It now checks in the same way as
544 RunTest, and also checks for the -b option.
546 19. If an odd number of negated classes containing just a single character
547 interposed, within parentheses, between a forward reference to a named
548 subpattern and the definition of the subpattern, compilation crashed with
549 an internal error, complaining that it could not find the referenced
550 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
551 [The bug was that it was starting one character too far in when skipping
552 over the character class, thus treating the ] as data rather than
553 terminating the class. This meant it could skip too much.]
555 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
556 /g option in pcretest when the pattern contains \K, which makes it possible
557 to have an empty string match not at the start, even when the pattern is
558 anchored. Updated pcretest and pcredemo to use this option.
560 21. If the maximum number of capturing subpatterns in a recursion was greater
561 than the maximum at the outer level, the higher number was returned, but
562 with unset values at the outer level. The correct (outer level) value is
563 now given.
565 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
566 PCRE did not set those parentheses (unlike Perl). I have now found a way to
567 make it do so. The string so far is captured, making this feature
568 compatible with Perl.
570 23. The tests have been re-organized, adding tests 11 and 12, to make it
571 possible to check the Perl 5.10 features against Perl 5.10.
573 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
574 pattern matches a fixed length string. PCRE did not allow this; now it
575 does. Neither allows recursion.
577 25. I finally figured out how to implement a request to provide the minimum
578 length of subject string that was needed in order to match a given pattern.
579 (It was back references and recursion that I had previously got hung up
580 on.) This code has now been added to pcre_study(); it finds a lower bound
581 to the length of subject needed. It is not necessarily the greatest lower
582 bound, but using it to avoid searching strings that are too short does give
583 some useful speed-ups. The value is available to calling programs via
584 pcre_fullinfo().
586 26. While implementing 25, I discovered to my embarrassment that pcretest had
587 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
588 study optimizations had never been tested with that matching function.
589 Oops. What is worse, even when it was passed study data, there was a bug in
590 pcre_dfa_exec() that meant it never actually used it. Double oops. There
591 were also very few tests of studied patterns with pcre_dfa_exec().
593 27. If (?| is used to create subpatterns with duplicate numbers, they are now
594 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
595 on the other side of the coin, they are no longer allowed to have different
596 names, because these cannot be distinguished in PCRE, and this has caused
597 confusion. (This is a difference from Perl.)
599 28. When duplicate subpattern names are present (necessarily with different
600 numbers, as required by 27 above), and a test is made by name in a
601 conditional pattern, either for a subpattern having been matched, or for
602 recursion in such a pattern, all the associated numbered subpatterns are
603 tested, and the overall condition is true if the condition is true for any
604 one of them. This is the way Perl works, and is also more like the way
605 testing by number works.
608 Version 7.9 11-Apr-09
609 ---------------------
611 1. When building with support for bzlib/zlib (pcregrep) and/or readline
612 (pcretest), all targets were linked against these libraries. This included
613 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
614 libraries. This caused unwanted dependencies to be created. This problem
615 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
616 pcretest is linked with readline.
618 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
619 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
620 moved outside it again, because FALSE and TRUE are already defined in AIX,
621 but BOOL is not.
623 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
624 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
626 4. The pcregrep documentation said spaces were inserted as well as colons (or
627 hyphens) following file names and line numbers when outputting matching
628 lines. This is not true; no spaces are inserted. I have also clarified the
629 wording for the --colour (or --color) option.
631 5. In pcregrep, when --colour was used with -o, the list of matching strings
632 was not coloured; this is different to GNU grep, so I have changed it to be
633 the same.
635 6. When --colo(u)r was used in pcregrep, only the first matching substring in
636 each matching line was coloured. Now it goes on to look for further matches
637 of any of the test patterns, which is the same behaviour as GNU grep.
639 7. A pattern that could match an empty string could cause pcregrep to loop; it
640 doesn't make sense to accept an empty string match in pcregrep, so I have
641 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
642 seems to be how GNU grep behaves.
644 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
645 start or after a newline", because the conditional assertion was not being
646 correctly handled. The rule now is that both the assertion and what follows
647 in the first alternative must satisfy the test.
649 9. If auto-callout was enabled in a pattern with a conditional group whose
650 condition was an assertion, PCRE could crash during matching, both with
651 pcre_exec() and pcre_dfa_exec().
653 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
654 used for matching.
656 11. Unicode property support in character classes was not working for
657 characters (bytes) greater than 127 when not in UTF-8 mode.
659 12. Added the -M command line option to pcretest.
661 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
663 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
665 16. Added comments and documentation about mis-use of no_arg in the C++
666 wrapper.
668 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
669 from Martin Jerabek that uses macro names for all relevant character and
670 string constants.
672 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
673 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
674 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
675 these, but not everybody uses configure.
677 19. A conditional group that had only one branch was not being correctly
678 recognized as an item that could match an empty string. This meant that an
679 enclosing group might also not be so recognized, causing infinite looping
680 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
681 with the subject "ab", where knowledge that the repeated group can match
682 nothing is needed in order to break the loop.
684 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
685 exec(), but without supplying a callout function, matching went wrong.
687 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
688 leak if the size of the offset vector was greater than 30. When the vector
689 is smaller, the saved offsets during recursion go onto a local stack
690 vector, but for larger vectors malloc() is used. It was failing to free
691 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
692 error, in fact).
694 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
695 heapframe that is used only when UTF-8 support is enabled. This caused no
696 problem, but was untidy.
698 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
699 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
700 included within another project.
702 24. Steven Van Ingelgem's patches to add more options to the CMake support,
703 slightly modified by me:
705 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
706 not building pcregrep.
708 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
709 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
711 25. Forward references, both numeric and by name, in patterns that made use of
712 duplicate group numbers, could behave incorrectly or give incorrect errors,
713 because when scanning forward to find the reference group, PCRE was not
714 taking into account the duplicate group numbers. A pattern such as
715 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
717 26. Changed a few more instances of "const unsigned char *" to USPTR, making
718 the feature of a custom pointer more persuasive (as requested by a user).
720 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
721 pcretest.c, inside #ifndefs, because it seems they are sometimes already
722 pre-defined.
724 28. Added support for (*UTF8) at the start of a pattern.
726 29. Arrange for flags added by the "release type" setting in CMake to be shown
727 in the configuration summary.
730 Version 7.8 05-Sep-08
731 ---------------------
733 1. Replaced UCP searching code with optimized version as implemented for Ad
734 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
735 stage table and inline lookup instead of a function, giving speed ups of 2
736 to 5 times on some simple patterns that I tested. Permission was given to
737 distribute the MultiStage2.py script that generates the tables (it's not in
738 the tarball, but is in the Subversion repository).
740 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
741 scripts.
743 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
744 a group with a zero qualifier. The result of the study could be incorrect,
745 or the function might crash, depending on the pattern.
747 4. Caseless matching was not working for non-ASCII characters in back
748 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
749 It now works when Unicode Property Support is available.
751 5. In pcretest, an escape such as \x{de} in the data was always generating
752 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
753 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
754 truncation.
756 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
758 7. Added two (int) casts to pcregrep when printing the difference of two
759 pointers, in case they are 64-bit values.
761 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
762 test 2 if it fails.
764 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
765 and a #define of that name to empty if it is not externally set. This is to
766 allow users of MSVC to set it if necessary.
768 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
769 the convenience functions in the pcre_get.c source file.
771 11. An option change at the start of a pattern that had top-level alternatives
772 could cause overwriting and/or a crash. This command provoked a crash in
773 some environments:
775 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
777 This potential security problem was recorded as CVE-2008-2371.
779 12. For a pattern where the match had to start at the beginning or immediately
780 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
781 pcre_dfa_exec() could read past the end of the passed subject if there was
782 no match. To help with detecting such bugs (e.g. with valgrind), I modified
783 pcretest so that it places the subject at the end of its malloc-ed buffer.
785 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
786 exec() might read past the end of the data buffer in UTF-8 mode.
788 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
789 the data contained the byte 0x85 as part of a UTF-8 character within its
790 first line. This applied both to normal and DFA matching.
792 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
793 /^[^d]*?$/8 failed to match "abc".
795 16. Added a missing copyright notice to pcrecpp_internal.h.
797 17. Make it more clear in the documentation that values returned from
798 pcre_exec() in ovector are byte offsets, not character counts.
800 18. Tidied a few places to stop certain compilers from issuing warnings.
802 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
803 supplied by Stefan Weber. I made a further small update for 7.8 because
804 there is a change of source arrangements: the pcre_searchfuncs.c module is
805 replaced by pcre_ucd.c.
808 Version 7.7 07-May-08
809 ---------------------
811 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
812 a string to a long long, pretend we don't even have a long long." This is
813 done by checking for the strtoq, strtoll, and _strtoi64 functions.
815 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
816 pre-7.6 versions, which defined a global no_arg variable instead of putting
817 it in the RE class. (See also #8 below.)
819 3. Remove a line of dead code, identified by coverity and reported by Nuno
820 Lopes.
822 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
824 (1) The include/exclude patterns were being applied to the whole pathnames
825 of files, instead of just to the final components.
827 (2) If there was more than one level of directory, the subdirectories were
828 skipped unless they satisfied the include/exclude conditions. This is
829 inconsistent with GNU grep (and could even be seen as contrary to the
830 pcregrep specification - which I improved to make it absolutely clear).
831 The action now is always to scan all levels of directory, and just
832 apply the include/exclude patterns to regular files.
834 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
835 --exclude_dir in the tests to avoid scanning .svn directories.
837 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
838 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
839 doesn't support NULs in patterns.
841 7. Added some missing "const"s to declarations of static tables in
842 pcre_compile.c and pcre_dfa_exec.c.
844 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
845 caused by fix #2 above. (Subsequently also a second patch to fix the
846 first patch. And a third patch - this was a messy problem.)
848 9. Applied Craig's patch to remove the use of push_back().
850 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
851 matching function regexec().
853 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
854 which, however, unlike Perl's \g{...}, are subroutine calls, not back
855 references. PCRE supports relative numbers with this syntax (I don't think
856 Oniguruma does).
858 12. Previously, a group with a zero repeat such as (...){0} was completely
859 omitted from the compiled regex. However, this means that if the group
860 was called as a subroutine from elsewhere in the pattern, things went wrong
861 (an internal error was given). Such groups are now left in the compiled
862 pattern, with a new opcode that causes them to be skipped at execution
863 time.
865 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
866 to the way PCRE behaves:
868 (a) A lone ] character is dis-allowed (Perl treats it as data).
870 (b) A back reference to an unmatched subpattern matches an empty string
871 (Perl fails the current match path).
873 (c) A data ] in a character class must be notated as \] because if the
874 first data character in a class is ], it defines an empty class. (In
875 Perl it is not possible to have an empty class.) The empty class []
876 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
877 The negative empty class [^] matches any one character, independently
878 of the DOTALL setting.
880 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
881 non-existent subpattern following a character class starting with ']' and
882 containing () gave an internal compiling error instead of "reference to
883 non-existent subpattern". Fortunately, when the pattern did exist, the
884 compiled code was correct. (When scanning forwards to check for the
885 existencd of the subpattern, it was treating the data ']' as terminating
886 the class, so got the count wrong. When actually compiling, the reference
887 was subsequently set up correctly.)
889 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
890 it was being rejected as not supported by pcre_dfa_exec(), even though
891 other assertions are supported. I have made pcre_dfa_exec() support
892 (*FAIL).
894 16. The implementation of 13c above involved the invention of a new opcode,
895 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
896 cannot be changed at match time, I realized I could make a small
897 improvement to matching performance by compiling OP_ALLANY instead of
898 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
899 on the OP_ANY path.
901 17. Compiling pcretest on Windows with readline support failed without the
902 following two fixes: (1) Make the unistd.h include conditional on
903 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
905 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
906 ncurses library to be included for pcretest when ReadLine support is
907 requested, but also to allow for it to be overridden. This patch came from
908 Daniel Bergström.
910 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
911 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
912 any errors with the current Unicode tables. Thanks to Peter Kankowski for
913 spotting this.
916 Version 7.6 28-Jan-08
917 ---------------------
919 1. A character class containing a very large number of characters with
920 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
921 overflow.
923 2. Patch to cut out the "long long" test in pcrecpp_unittest when
924 HAVE_LONG_LONG is not defined.
926 3. Applied Christian Ehrlicher's patch to update the CMake build files to
927 bring them up to date and include new features. This patch includes:
929 - Fixed PH's badly added libz and libbz2 support.
930 - Fixed a problem with static linking.
931 - Added pcredemo. [But later removed - see 7 below.]
932 - Fixed dftables problem and added an option.
933 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
935 - Added readline support for pcretest.
936 - Added an listing of the option settings after cmake has run.
938 4. A user submitted a patch to Makefile that makes it easy to create
939 "pcre.dll" under mingw when using Configure/Make. I added stuff to
940 Makefile.am that cause it to include this special target, without
941 affecting anything else. Note that the same mingw target plus all
942 the other distribution libraries and programs are now supported
943 when configuring with CMake (see 6 below) instead of with
944 Configure/Make.
946 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
947 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
948 exported in the Windows port". It has not yet been confirmed that the patch
949 solves the problem, but it does no harm.
951 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
952 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
953 with CMake, and also correct the comment about stack recursion.
955 7. Remove the automatic building of pcredemo from the ./configure system and
956 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
957 of a program that users should build themselves after PCRE is installed, so
958 building it automatically is not really right. What is more, it gave
959 trouble in some build environments.
961 8. Further tidies to CMakeLists.txt from Sheri and Christian.
964 Version 7.5 10-Jan-08
965 ---------------------
967 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
968 values in parens when parsing an RE using the C++ wrapper."
970 2. Negative specials like \S did not work in character classes in UTF-8 mode.
971 Characters greater than 255 were excluded from the class instead of being
972 included.
974 3. The same bug as (2) above applied to negated POSIX classes such as
975 [:^space:].
977 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
978 defined or documented. It seems to have been a typo for PCRE_STATIC, so
979 I have changed it.
981 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
982 first named subpattern) and a construct such as (?&a) would reference the
983 first named subpattern whose name started with "a" (in other words, the
984 length check was missing). Both these problems are fixed. "Subpattern name
985 expected" is now given for (?&) (a zero-length name), and this patch also
986 makes it give the same error for \k'' (previously it complained that that
987 was a reference to a non-existent subpattern).
989 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
990 this is right because (?- can be followed by option settings as well as by
991 digits. I have, however, made the messages clearer.
993 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
994 than the number used in the conditional) now cause a compile-time error.
995 This is actually not compatible with Perl, which accepts such patterns, but
996 treats the conditional as always being FALSE (as PCRE used to), but it
997 seems to me that giving a diagnostic is better.
999 8. Change "alphameric" to the more common word "alphanumeric" in comments
1000 and messages.
1002 9. Fix two occurrences of "backslash" in comments that should have been
1003 "backspace".
1005 10. Remove two redundant lines of code that can never be obeyed (their function
1006 was moved elsewhere).
1008 11. The program that makes PCRE's Unicode character property table had a bug
1009 which caused it to generate incorrect table entries for sequences of
1010 characters that have the same character type, but are in different scripts.
1011 It amalgamated them into a single range, with the script of the first of
1012 them. In other words, some characters were in the wrong script. There were
1013 thirteen such cases, affecting characters in the following ranges:
1015 U+002b0 - U+002c1
1016 U+0060c - U+0060d
1017 U+0061e - U+00612
1018 U+0064b - U+0065e
1019 U+0074d - U+0076d
1020 U+01800 - U+01805
1021 U+01d00 - U+01d77
1022 U+01d9b - U+01dbf
1023 U+0200b - U+0200f
1024 U+030fc - U+030fe
1025 U+03260 - U+0327f
1026 U+0fb46 - U+0fbb1
1027 U+10450 - U+1049d
1029 12. The -o option (show only the matching part of a line) for pcregrep was not
1030 compatible with GNU grep in that, if there was more than one match in a
1031 line, it showed only the first of them. It now behaves in the same way as
1032 GNU grep.
1034 13. If the -o and -v options were combined for pcregrep, it printed a blank
1035 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1036 does the same. The return code can be used to tell if there were any
1037 non-matching lines.
1039 14. Added --file-offsets and --line-offsets to pcregrep.
1041 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1042 infinitely looping recursion. The bug was that positive lookaheads were not
1043 being skipped when checking for a possible empty match (negative lookaheads
1044 and both kinds of lookbehind were skipped).
1046 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1047 inclusion of <windows.h> to before rather than after the definition of
1048 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1050 17. Specifying a possessive quantifier with a specific limit for a Unicode
1051 character property caused pcre_compile() to compile bad code, which led at
1052 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1053 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1054 caused the error; without that there was no problem.
1056 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1058 19. Added --enable-pcretest-libreadline.
1060 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1061 RE::GlobalReplace(). As a result, the number of replacements returned was
1062 double what it should be. I removed one of the increments, but Craig sent a
1063 later patch that removed the other one (the right fix) and added unit tests
1064 that check the return values (which was not done before).
1066 21. Several CMake things:
1068 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1069 the names libpcre and libpcreposix, not just pcre and pcreposix.
1071 (2) The above change means that pcretest and pcregrep are now correctly
1072 linked with the newly-built libraries, not previously installed ones.
1076 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1077 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1078 UTF-8 newline character). The key issue is that the pattern starts .*;
1079 this means that the match must be either at the beginning, or after a
1080 newline. The bug was in the code for advancing after a failed match and
1081 checking that the new position followed a newline. It was not taking
1082 account of UTF-8 characters correctly.
1084 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1085 character classes. PCRE was not treating the sequence [:...:] as a
1086 character class unless the ... were all letters. Perl, however, seems to
1087 allow any characters between [: and :], though of course it rejects as
1088 unknown any "names" that contain non-letters, because all the known class
1089 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1090 for example, whereas PCRE did not - it did not recognize a POSIX character
1091 class. This seemed a bit dangerous, so the code has been changed to be
1092 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1093 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1094 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1095 Perl does, and where it didn't before.
1097 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1098 Windows environments %n is disabled by default.
1101 Version 7.4 21-Sep-07
1102 ---------------------
1104 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1105 means that a class such as [\s] counted as "explicit reference to CR or
1106 LF". That isn't really right - the whole point of the change was to try to
1107 help when there was an actual mention of one of the two characters. So now
1108 the change happens only if \r or \n (or a literal CR or LF) character is
1109 encountered.
1111 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1112 of both had grown to the point where there were only 3 bits left.
1113 Fortunately, there was spare space in the data structure, and so I have
1114 moved the internal flags into a new 16-bit field to free up more option
1115 bits.
1117 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1118 but did not set the internal JCHANGED flag - either of these is enough to
1119 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1120 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1121 start sets both bits.
1123 4. Added options (at build time, compile time, exec time) to change \R from
1124 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1126 5. doc/pcresyntax.html was missing from the distribution.
1128 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1129 compatibility, even though it is no longer used.
1131 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1132 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1133 windows.h file is present (where different names are used). [This was
1134 reversed later after testing - see 16 below.]
1136 8. Changed all #include <config.h> to #include "config.h". There were also
1137 some further <pcre.h> cases that I changed to "pcre.h".
1139 9. When pcregrep was used with the --colour option, it missed the line ending
1140 sequence off the lines that it output.
1142 10. It was pointed out to me that arrays of string pointers cause lots of
1143 relocations when a shared library is dynamically loaded. A technique of
1144 using a single long string with a table of offsets can drastically reduce
1145 these. I have refactored PCRE in four places to do this. The result is
1146 dramatic:
1148 Originally: 290
1149 After changing UCP table: 187
1150 After changing error message table: 43
1151 After changing table of "verbs" 36
1152 After changing table of Posix names 22
1154 Thanks to the folks working on Gregex for glib for this insight.
1156 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1157 unicode-properties was also set.
1159 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1161 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1162 checked only for CRLF.
1164 14. Added casts to pcretest.c to avoid compiler warnings.
1166 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1168 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1169 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1170 entirely. This removes changes made in 7 above.
1172 17. The CMake files have been updated, and there is now more information about
1173 building with CMake in the NON-UNIX-USE document.
1176 Version 7.3 28-Aug-07
1177 ---------------------
1179 1. In the rejigging of the build system that eventually resulted in 7.1, the
1180 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1181 brackets there is not right, since it causes compilers to look for an
1182 installed pcre.h, not the version that is in the source that is being
1183 compiled (which of course may be different). I have changed it back to:
1185 #include "pcre.h"
1187 I have a vague recollection that the change was concerned with compiling in
1188 different directories, but in the new build system, that is taken care of
1189 by the VPATH setting the Makefile.
1191 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1192 when the subject happened to end in the byte 0x85 (e.g. if the last
1193 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1194 characters but of course it shouldn't be taken as a newline when it is part
1195 of another character. The bug was that, for an unlimited repeat of . in
1196 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1197 characters when looking for a newline.
1199 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1201 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1202 in debug output.
1204 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1205 long printing in the pcrecpp unittest when running under MinGW.
1207 6. ESC_K was left out of the EBCDIC table.
1209 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1210 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1211 limit also applies to "virtual nesting" when a pattern is recursive, and in
1212 this case 1000 isn't so big. I have been able to remove this limit at the
1213 expense of backing off one optimization in certain circumstances. Normally,
1214 when pcre_exec() would call its internal match() function recursively and
1215 immediately return the result unconditionally, it uses a "tail recursion"
1216 feature to save stack. However, when a subpattern that can match an empty
1217 string has an unlimited repetition quantifier, it no longer makes this
1218 optimization. That gives it a stack frame in which to save the data for
1219 checking that an empty string has been matched. Previously this was taken
1220 from the 1000-entry workspace that had been reserved. So now there is no
1221 explicit limit, but more stack is used.
1223 8. Applied Daniel's patches to solve problems with the import/export magic
1224 syntax that is required for Windows, and which was going wrong for the
1225 pcreposix and pcrecpp parts of the library. These were overlooked when this
1226 problem was solved for the main library.
1228 9. There were some crude static tests to avoid integer overflow when computing
1229 the size of patterns that contain repeated groups with explicit upper
1230 limits. As the maximum quantifier is 65535, the maximum group length was
1231 set at 30,000 so that the product of these two numbers did not overflow a
1232 32-bit integer. However, it turns out that people want to use groups that
1233 are longer than 30,000 bytes (though not repeat them that many times).
1234 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1235 made it possible to implement the integer overflow checks in a much more
1236 dynamic way, which I have now done. The artificial limitation on group
1237 length has been removed - we now have only the limit on the total length of
1238 the compiled pattern, which depends on the LINK_SIZE setting.
1240 10. Fixed a bug in the documentation for get/copy named substring when
1241 duplicate names are permitted. If none of the named substrings are set, the
1242 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1243 empty string.
1245 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1246 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1247 because the ] is interpreted as the first data character and the
1248 terminating ] is not found. PCRE has been made compatible with Perl in this
1249 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1250 cause memory overwriting.
1252 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1253 string has been matched (to stop an infinite loop). It was not recognizing
1254 a conditional subpattern that could match an empty string if that
1255 subpattern was within another subpattern. For example, it looped when
1256 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1257 condition was not nested. This bug has been fixed.
1259 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1260 past the start of the subject in the presence of bytes with the top bit
1261 set, for example "\x8aBCD".
1263 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1264 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1266 14. Optimized (?!) to (*FAIL).
1268 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1269 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1270 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1271 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1272 does: it's just the validity check that is more restrictive.
1274 16. Inserted checks for integer overflows during escape sequence (backslash)
1275 processing, and also fixed erroneous offset values for syntax errors during
1276 backslash processing.
1278 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1279 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1281 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1282 caused an overrun.
1284 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1285 something other than just ASCII characters) inside a group that had an
1286 unlimited repeat caused a loop at compile time (while checking to see
1287 whether the group could match an empty string).
1289 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1290 [\P{Any}] did so. (Error in the code for printing property names.)
1292 21. An orphan \E inside a character class could cause a crash.
1294 22. A repeated capturing bracket such as (A)? could cause a wild memory
1295 reference during compilation.
1297 23. There are several functions in pcre_compile() that scan along a compiled
1298 expression for various reasons (e.g. to see if it's fixed length for look
1299 behind). There were bugs in these functions when a repeated \p or \P was
1300 present in the pattern. These operators have additional parameters compared
1301 with \d, etc, and these were not being taken into account when moving along
1302 the compiled data. Specifically:
1304 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1305 length.
1307 (b) An item such as \pL+ within a repeated group could cause crashes or
1308 loops.
1310 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1311 "reference to non-existent subpattern" error.
1313 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1315 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1316 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1318 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1320 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1321 character were causing crashes (broken optimization).
1323 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1324 \p or \P) caused a compile-time loop.
1326 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1327 break. For example, the unstudied pattern [\r\n]A does not match the string
1328 "\r\nA" because change 7.0/46 below moves the current point on by two
1329 characters after failing to match at the start. However, the pattern \nA
1330 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1331 the same is true. There doesn't seem any very clean way out of this, but
1332 what I have chosen to do makes the common cases work: PCRE now takes note
1333 of whether there can be an explicit match for \r or \n anywhere in the
1334 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1335 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1336 pattern has explicit CR or LF references.
1338 29. Added (*CR) etc for changing newline setting at start of pattern.
1341 Version 7.2 19-Jun-07
1342 ---------------------
1344 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1345 which is apparently normally available under Windows.
1347 2. Re-jig the pcregrep tests with different newline settings in an attempt
1348 to make them independent of the local environment's newline setting.
1350 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1352 4. Some of the "internals" tests were previously cut out when the link size
1353 was not 2, because the output contained actual offsets. The recent new
1354 "Z" feature of pcretest means that these can be cut out, making the tests
1355 usable with all link sizes.
1357 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1358 stack recursion. This gives a massive performance boost under BSD, but just
1359 a small improvement under Linux. However, it saves one field in the frame
1360 in all cases.
1362 6. Added more features from the forthcoming Perl 5.10:
1364 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1365 recursion call. It refers to the nth most recently opened parentheses.
1367 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1368 to be opened parentheses.
1370 (c) Conditions that refer to capturing parentheses can be specified
1371 relatively, for example, (?(-2)... or (?(+3)...
1373 (d) \K resets the start of the current match so that everything before
1374 is not part of it.
1376 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1378 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1379 reference syntax.
1381 (g) (?| introduces a group in which the numbering of parentheses in each
1382 alternative starts with the same number.
1384 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1386 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1389 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1390 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1391 for detecting groups that can match an empty string.
1393 9. A pattern with a very large number of alternatives (more than several
1394 hundred) was running out of internal workspace during the pre-compile
1395 phase, where pcre_compile() figures out how much memory will be needed. A
1396 bit of new cunning has reduced the workspace needed for groups with
1397 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1398 workspace instead of running out of the 4096 that are available.
1400 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1402 11. Applied patch from Google to remove an optimization that didn't quite work.
1403 The report of the bug said:
1405 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1406 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1407 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1409 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1410 it matched the wrong number of bytes.
1413 Version 7.1 24-Apr-07
1414 ---------------------
1416 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1417 that is more "standard", making use of automake and other Autotools. There
1418 is some re-arrangement of the files and adjustment of comments consequent
1419 on this.
1421 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1422 for recursive directory scanning broke on some systems because the files
1423 are not scanned in any specific order and on different systems the order
1424 was different. A call to "sort" has been inserted into RunGrepTest for the
1425 approprate test as a short-term fix. In the longer term there may be an
1426 alternative.
1428 3. I had an email from Eric Raymond about problems translating some of PCRE's
1429 man pages to HTML (despite the fact that I distribute HTML pages, some
1430 people do their own conversions for various reasons). The problems
1431 concerned the use of low-level troff macros .br and .in. I have therefore
1432 removed all such uses from the man pages (some were redundant, some could
1433 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1434 HTML has been updated to handle .nf/.fi and to complain if it encounters
1435 .br or .in.
1437 4. Updated comments in configure.ac that get placed in config.h.in and also
1438 arranged for config.h to be included in the distribution, with the name
1439 config.h.generic, for the benefit of those who have to compile without
1440 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1442 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1443 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1444 makevp.bat for the current PCRE, using the additional files
1445 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1447 6. A Windows user reported a minor discrepancy with test 2, which turned out
1448 to be caused by a trailing space on an input line that had got lost in his
1449 copy. The trailing space was an accident, so I've just removed it.
1451 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1452 that is needed.
1454 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1455 as "const" (a) because they are and (b) because it helps the PHP
1456 maintainers who have recently made a script to detect big data structures
1457 in the php code that should be moved to the .rodata section. I remembered
1458 to update Builducptable as well, so it won't revert if ucptable.h is ever
1459 re-created.
1461 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1462 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1463 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1464 support is not required. This saves 1.5-2K of code, which is important in
1465 some applications.
1467 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1468 so as not to refer to the tables, even though these functions will never be
1469 called when UTF-8 support is disabled. Otherwise there are problems with a
1470 shared library.
1472 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1474 (a) It was defining its arguments as char * instead of void *.
1476 (b) It was assuming that all moves were upwards in memory; this was true
1477 a long time ago when I wrote it, but is no longer the case.
1479 The emulated memove() is provided for those environments that have neither
1480 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1481 is clearly not the case, as these two bugs were recently reported.
1483 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1484 and Detrail to create the HTML documentation, the .txt form of the man
1485 pages, and it removes trailing spaces from listed files. It also creates
1486 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1487 case, it wraps all the #defines with #ifndefs. This script should be run
1488 before "make dist".
1490 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1491 with Unicode property support.
1493 (a) For a maximizing quantifier, if the two different cases of the
1494 character were of different lengths in their UTF-8 codings (there are
1495 some cases like this - I found 11), and the matching function had to
1496 back up over a mixture of the two cases, it incorrectly assumed they
1497 were both the same length.
1499 (b) When PCRE was configured to use the heap rather than the stack for
1500 recursion during matching, it was not correctly preserving the data for
1501 the other case of a UTF-8 character when checking ahead for a match
1502 while processing a minimizing repeat. If the check also involved
1503 matching a wide character, but failed, corruption could cause an
1504 erroneous result when trying to check for a repeat of the original
1505 character.
1507 13. Some tidying changes to the testing mechanism:
1509 (a) The RunTest script now detects the internal link size and whether there
1510 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1511 values substituted by "configure". (The RunGrepTest script already did
1512 this for UTF-8.) The configure.ac script no longer substitutes the
1513 relevant variables.
1515 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1516 with length and offset values. This means that the output is different
1517 for different internal link sizes. Test 2 is skipped for link sizes
1518 other than 2 because of this, bypassing the problem. Unfortunately,
1519 there was also a test in test 3 (the locale tests) that used /B and
1520 failed for link sizes other than 2. Rather than cut the whole test out,
1521 I have added a new /Z option to pcretest that replaces the length and
1522 offset values with spaces. This is now used to make test 3 independent
1523 of link size. (Test 2 will be tidied up later.)
1525 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1526 segmentation fault instead of returning the appropriate error message.
1528 15. In multiline mode when the newline sequence was set to "any", the pattern
1529 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1530 This doesn't seem right; it now treats the CRLF combination as the line
1531 ending, and so does not match in that case. It's only a pattern such as ^$
1532 that would hit this one: something like ^ABC$ would have failed after \r
1533 and then tried again after \r\n.
1535 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1536 in an attempt to make files that differ only in their line terminators
1537 compare equal. This works on Linux.
1539 17. Under certain error circumstances pcregrep might try to free random memory
1540 as it exited. This is now fixed, thanks to valgrind.
1542 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1543 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1544 was because its rules for how to advance for /g after matching an empty
1545 string at the end of a line did not allow for this case. They now check for
1546 it specially.
1548 20. pcretest is supposed to handle patterns and data of any length, by
1549 extending its buffers when necessary. It was getting this wrong when the
1550 buffer for a data line had to be extended.
1552 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1553 CRLF as a newline sequence.
1555 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1556 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1557 I have nevertheless tidied it up.
1559 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1561 24. Added a man page for pcre-config.
1564 Version 7.0 19-Dec-06
1565 ---------------------
1567 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1568 moving to gcc 4.1.1.
1570 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1571 sys/time.h, which is documented as needed for this function. It doesn't
1572 seem to matter on Linux, but it showed up on some releases of OS X.
1574 3. It seems that there are systems where bytes whose values are greater than
1575 127 match isprint() in the "C" locale. The "C" locale should be the
1576 default when a C program starts up. In most systems, only ASCII printing
1577 characters match isprint(). This difference caused the output from pcretest
1578 to vary, making some of the tests fail. I have changed pcretest so that:
1580 (a) When it is outputting text in the compiled version of a pattern, bytes
1581 other than 32-126 are always shown as hex escapes.
1583 (b) When it is outputting text that is a matched part of a subject string,
1584 it does the same, unless a different locale has been set for the match
1585 (using the /L modifier). In this case, it uses isprint() to decide.
1587 4. Fixed a major bug that caused incorrect computation of the amount of memory
1588 required for a compiled pattern when options that changed within the
1589 pattern affected the logic of the preliminary scan that determines the
1590 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1591 that the computed length was too small. The symptoms of this bug were
1592 either the PCRE error "internal error: code overflow" from pcre_compile(),
1593 or a glibc crash with a message such as "pcretest: free(): invalid next
1594 size (fast)". Examples of patterns that provoked this bug (shown in
1595 pcretest format) are:
1597 /(?-x: )/x
1598 /(?x)(?-x: \s*#\s*)/
1599 /((?i)[\x{c0}])/8
1600 /(?i:[\x{c0}])/8
1602 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1603 is now done differently.
1605 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1606 wrapper classes; (b) implement a new function in the C++ scanner that is
1607 more efficient than the old way of doing things because it avoids levels of
1608 recursion in the regex matching; (c) add a paragraph to the documentation
1609 for the FullMatch() function.
1611 6. The escape sequence \n was being treated as whatever was defined as
1612 "newline". Not only was this contrary to the documentation, which states
1613 that \n is character 10 (hex 0A), but it also went horribly wrong when
1614 "newline" was defined as CRLF. This has been fixed.
1616 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1617 was being set to -1 for the "end of line" case (supposedly a value that no
1618 character can have). Though this value is never used (the check for end of
1619 line is "zero bytes in current character"), it caused compiler complaints.
1620 I've changed it to 0xffffffff.
1622 8. In pcre_version.c, the version string was being built by a sequence of
1623 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1624 string (as it is for production releases) called a macro with an empty
1625 argument. The C standard says the result of this is undefined. The gcc
1626 compiler treats it as an empty string (which was what was wanted) but it is
1627 reported that Visual C gives an error. The source has been hacked around to
1628 avoid this problem.
1630 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1631 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1632 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1633 of them did).
1635 10. Originally, pcretest opened its input and output without "b"; then I was
1636 told that "b" was needed in some environments, so it was added for release
1637 5.0 to both the input and output. (It makes no difference on Unix-like
1638 systems.) Later I was told that it is wrong for the input on Windows. I've
1639 now abstracted the modes into two macros, to make it easier to fiddle with
1640 them, and removed "b" from the input mode under Windows.
1642 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1644 12. Added -help and --help to pcretest as an official way of being reminded
1645 of the options.
1647 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1648 and pcrecpp.cc because they annoy compilers at high warning levels.
1650 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1652 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1653 scripts are not C programs :-) and which was not noticed because it works
1654 on Linux.
1656 16. pcretest is supposed to handle any length of pattern and data line (as one
1657 line or as a continued sequence of lines) by extending its input buffer if
1658 necessary. This feature was broken for very long pattern lines, leading to
1659 a string of junk being passed to pcre_compile() if the pattern was longer
1660 than about 50K.
1662 17. I have done a major re-factoring of the way pcre_compile() computes the
1663 amount of memory needed for a compiled pattern. Previously, there was code
1664 that made a preliminary scan of the pattern in order to do this. That was
1665 OK when PCRE was new, but as the facilities have expanded, it has become
1666 harder and harder to keep it in step with the real compile phase, and there
1667 have been a number of bugs (see for example, 4 above). I have now found a
1668 cunning way of running the real compile function in a "fake" mode that
1669 enables it to compute how much memory it would need, while actually only
1670 ever using a few hundred bytes of working memory and without too many
1671 tests of the mode. This should make future maintenance and development
1672 easier. A side effect of this work is that the limit of 200 on the nesting
1673 depth of parentheses has been removed (though this was never a serious
1674 limitation, I suspect). However, there is a downside: pcre_compile() now
1675 runs more slowly than before (30% or more, depending on the pattern). I
1676 hope this isn't a big issue. There is no effect on runtime performance.
1678 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1679 newline (only possible for the last line of a file) and it was a
1680 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1682 19. Added additional timing features to pcretest. (1) The -tm option now times
1683 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1684 separate command line item, by a number that specifies the number of
1685 repeats to use when timing. The default is 50000; this gives better
1686 precision, but takes uncomfortably long for very large patterns.
1688 20. Extended pcre_study() to be more clever in cases where a branch of a
1689 subpattern has no definite first character. For example, (a*|b*)[cd] would
1690 previously give no result from pcre_study(). Now it recognizes that the
1691 first character must be a, b, c, or d.
1693 21. There was an incorrect error "recursive call could loop indefinitely" if
1694 a subpattern (or the entire pattern) that was being tested for matching an
1695 empty string contained only one non-empty item after a nested subpattern.
1696 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1697 incorrectly, because the \d was being skipped in the check.
1699 22. The pcretest program now has a new pattern option /B and a command line
1700 option -b, which is equivalent to adding /B to every pattern. This causes
1701 it to show the compiled bytecode, without the additional information that
1702 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1703 is the same as /B/I).
1705 23. A new optimization is now able automatically to treat some sequences such
1706 as a*b as a*+b. More specifically, if something simple (such as a character
1707 or a simple class like \d) has an unlimited quantifier, and is followed by
1708 something that cannot possibly match the quantified thing, the quantifier
1709 is automatically "possessified".
1711 24. A recursive reference to a subpattern whose number was greater than 39
1712 went wrong under certain circumstances in UTF-8 mode. This bug could also
1713 have affected the operation of pcre_study().
1715 25. Realized that a little bit of performance could be had by replacing
1716 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1718 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1720 27. Possessive quantifiers such as a++ were previously implemented by turning
1721 them into atomic groups such as ($>a+). Now they have their own opcodes,
1722 which improves performance. This includes the automatically created ones
1723 from 23 above.
1725 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1726 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1727 the first matched character to be a colon. This applied both to named and
1728 numbered groups.
1730 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1732 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1733 building PCRE on Windows easier, so I have included it in the distribution.
1735 31. There is now a check in pcretest against a ridiculously large number being
1736 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1737 loop, the loop is abandoned.
1739 32. Forward references to subpatterns in conditions such as (?(2)...) where
1740 subpattern 2 is defined later cause pcre_compile() to search forwards in
1741 the pattern for the relevant set of parentheses. This search went wrong
1742 when there were unescaped parentheses in a character class, parentheses
1743 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1745 33. "Subroutine" calls and backreferences were previously restricted to
1746 referencing subpatterns earlier in the regex. This restriction has now
1747 been removed.
1749 34. Added a number of extra features that are going to be in Perl 5.10. On the
1750 whole, these are just syntactic alternatives for features that PCRE had
1751 previously implemented using the Python syntax or my own invention. The
1752 other formats are all retained for compatibility.
1754 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1755 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1756 also .NET compatible.
1758 (b) A recursion or subroutine call to a named group can now be defined as
1759 (?&name) as well as (?P>name).
1761 (c) A backreference to a named group can now be defined as \k<name> or
1762 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1763 5.10, are also .NET compatible.
1765 (d) A conditional reference to a named group can now use the syntax
1766 (?(<name>) or (?('name') as well as (?(name).
1768 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1769 groups (named and numbered) that are never evaluated inline, but can be
1770 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1771 is always false. There may be only one alternative in such a group.
1773 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1774 as the simple (?(R). The condition is true only if the most recent
1775 recursion is that of the given number or name. It does not search out
1776 through the entire recursion stack.
1778 (g) The escape \gN or \g{N} has been added, where N is a positive or
1779 negative number, specifying an absolute or relative reference.
1781 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1782 some "unreachable code" warnings.
1784 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1785 things, this adds five new scripts.
1787 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1788 There were also incompatibilities regarding the handling of \Q..\E inside
1789 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1790 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1792 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1793 matches an empty string, and forcibly breaks the loop. There were bugs in
1794 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1795 against aaaa the result was just "a" rather than "aaaa", for example. Two
1796 separate and independent bugs (that affected different cases) have been
1797 fixed.
1799 39. Refactored the code to abolish the use of different opcodes for small
1800 capturing bracket numbers. This is a tidy that I avoided doing when I
1801 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1802 The new approach is not only tidier, it makes it possible to reduce the
1803 memory needed to fix the previous bug (38).
1805 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1806 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1807 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1808 mode.
1810 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1811 report.
1813 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1814 copy construction and assignment in the C++ wrapper.
1816 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1817 couple of bugs that could have given wrong results in the "--newline=crlf"
1818 case.
1820 44. Added a number of casts and did some reorganization of signed/unsigned int
1821 variables following suggestions from Dair Grant. Also renamed the variable
1822 "this" as "item" because it is a C++ keyword.
1824 45. Arranged for dftables to add
1826 #include "pcre_internal.h"
1828 to pcre_chartables.c because without it, gcc 4.x may remove the array
1829 definition from the final binary if PCRE is built into a static library and
1830 dead code stripping is activated.
1832 46. For an unanchored pattern, if a match attempt fails at the start of a
1833 newline sequence, and the newline setting is CRLF or ANY, and the next two
1834 characters are CRLF, advance by two characters instead of one.
1837 Version 6.7 04-Jul-06
1838 ---------------------
1840 1. In order to handle tests when input lines are enormously long, pcretest has
1841 been re-factored so that it automatically extends its buffers when
1842 necessary. The code is crude, but this _is_ just a test program. The
1843 default size has been increased from 32K to 50K.
1845 2. The code in pcre_study() was using the value of the re argument before
1846 testing it for NULL. (Of course, in any sensible call of the function, it
1847 won't be NULL.)
1849 3. The memmove() emulation function in pcre_internal.h, which is used on
1850 systems that lack both memmove() and bcopy() - that is, hardly ever -
1851 was missing a "static" storage class specifier.
1853 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1854 containing an extended class (one that cannot be represented by a bitmap
1855 because it contains high-valued characters or Unicode property items, e.g.
1856 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1857 pattern, but PCRE should not loop if you do not (it no longer does).
1858 [Detail: two cases were found: (a) a repeated subpattern containing an
1859 extended class; (b) a recursive reference to a subpattern that followed a
1860 previous extended class. It wasn't skipping over the extended class
1861 correctly when UTF-8 mode was not set.]
1863 5. A negated single-character class was not being recognized as fixed-length
1864 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1865 compile error "lookbehind assertion is not fixed length".
1867 6. The RunPerlTest auxiliary script was showing an unexpected difference
1868 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1869 write a Perl script that can interpret lines of an input file either as
1870 byte characters or as UTF-8, which is what "perltest" was being required to
1871 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1872 can't do is switch easily at run time between having the "use utf8;" pragma
1873 or not. In the end, I fudged it by using the RunPerlTest script to insert
1874 "use utf8;" explicitly for the UTF-8 tests.
1876 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1877 the end of the subject string, contrary to the documentation and to what
1878 Perl does. This was true of both matching functions. Now it matches only at
1879 the start of the subject and immediately after *internal* newlines.
1881 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1882 a pointer to an int instead of a pointer to an unsigned long int. This
1883 caused problems on 64-bit systems.
1885 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1886 instance of the 'standard' template library not being so standard".
1888 10. There was no check on the number of named subpatterns nor the maximum
1889 length of a subpattern name. The product of these values is used to compute
1890 the size of the memory block for a compiled pattern. By supplying a very
1891 long subpattern name and a large number of named subpatterns, the size
1892 computation could be caused to overflow. This is now prevented by limiting
1893 the length of names to 32 characters, and the number of named subpatterns
1894 to 10,000.
1896 11. Subpatterns that are repeated with specific counts have to be replicated in
1897 the compiled pattern. The size of memory for this was computed from the
1898 length of the subpattern and the repeat count. The latter is limited to
1899 65535, but there was no limit on the former, meaning that integer overflow
1900 could in principle occur. The compiled length of a repeated subpattern is
1901 now limited to 30,000 bytes in order to prevent this.
1903 12. Added the optional facility to have named substrings with the same name.
1905 13. Added the ability to use a named substring as a condition, using the
1906 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1907 are numbers (not recommended). Forward references are permitted.
1909 14. Added forward references in named backreferences (if you see what I mean).
1911 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1912 pattern could run off the end of the subject. For example, the pattern
1913 "(?s)(.{1,5})"8 did this with the subject "ab".
1915 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1916 PCRE_CASELESS was set when matching characters that were quantified with ?
1917 or *.
1919 17. A character class other than a single negated character that had a minimum
1920 but no maximum quantifier - for example [ab]{6,} - was not handled
1921 correctly by pce_dfa_exec(). It would match only one character.
1923 18. A valid (though odd) pattern that looked like a POSIX character
1924 class but used an invalid character after [ (for example [[,abc,]]) caused
1925 pcre_compile() to give the error "Failed: internal error: code overflow" or
1926 in some cases to crash with a glibc free() error. This could even happen if
1927 the pattern terminated after [[ but there just happened to be a sequence of
1928 letters, a binary zero, and a closing ] in the memory that followed.
1930 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1931 over the years. Originally (before any Unicode support), just the bottom 8
1932 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1933 output from "man perlunicode" includes this:
1935 The regular expression compiler produces polymorphic opcodes. That
1936 is, the pattern adapts to the data and automatically switches to
1937 the Unicode character scheme when presented with Unicode data--or
1938 instead uses a traditional byte scheme when presented with byte
1939 data.
1941 Sadly, a wide octal escape does not cause a switch, and in a string with
1942 no other multibyte characters, these octal escapes are treated as before.
1943 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1944 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1945 Unicode string.
1947 I have not perpetrated such confusion in PCRE. Up till now, it took just
1948 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1949 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1950 translate to the appropriate multibyte character.
1952 29. Applied some refactoring to reduce the number of warnings from Microsoft
1953 and Borland compilers. This has included removing the fudge introduced
1954 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1955 a warning about an unused variable.
1957 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1958 characters since release 4.0, because Perl (from release 5.004) does not.
1959 [Or at least, is documented not to: some releases seem to be in conflict
1960 with the documentation.] However, when a pattern was studied with
1961 pcre_study() and all its branches started with \s, PCRE still included VT
1962 as a possible starting character. Of course, this did no harm; it just
1963 caused an unnecessary match attempt.
1965 22. Removed a now-redundant internal flag bit that recorded the fact that case
1966 dependency changed within the pattern. This was once needed for "required
1967 byte" processing, but is no longer used. This recovers a now-scarce options
1968 bit. Also moved the least significant internal flag bit to the most-
1969 significant bit of the word, which was not previously used (hangover from
1970 the days when it was an int rather than a uint) to free up another bit for
1971 the future.
1973 23. Added support for CRLF line endings as well as CR and LF. As well as the
1974 default being selectable at build time, it can now be changed at runtime
1975 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1976 specify that it is scanning data with non-default line endings.
1978 24. Changed the definition of CXXLINK to make it agree with the definition of
1979 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1981 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1982 recursions. This makes a big different to stack usage for some patterns.
1984 26. If a subpattern containing a named recursion or subroutine reference such
1985 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1986 the space required for the compiled pattern went wrong and gave too small a
1987 value. Depending on the environment, this could lead to "Failed: internal
1988 error: code overflow at offset 49" or "glibc detected double free or
1989 corruption" errors.
1991 27. Applied patches from Google (a) to support the new newline modes and (b) to
1992 advance over multibyte UTF-8 characters in GlobalReplace.
1994 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1995 difference for some implementation of PCRE in some Windows version.
1997 29. Added some extra testing facilities to pcretest:
1999 \q<number> in a data line sets the "match limit" value
2000 \Q<number> in a data line sets the "match recursion limt" value
2001 -S <number> sets the stack size, where <number> is in megabytes
2003 The -S option isn't available for Windows.
2006 Version 6.6 06-Feb-06
2007 ---------------------
2009 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2010 in pcreposix.h. I have copied the definition from pcre.h.
2012 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2013 because pcre.h is no longer a built file.
2015 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2016 not normally included in the compiled code.
2019 Version 6.5 01-Feb-06
2020 ---------------------
2022 1. When using the partial match feature with pcre_dfa_exec(), it was not
2023 anchoring the second and subsequent partial matches at the new starting
2024 point. This could lead to incorrect results. For example, with the pattern
2025 /1234/, partially matching against "123" and then "a4" gave a match.
2027 2. Changes to pcregrep:
2029 (a) All non-match returns from pcre_exec() were being treated as failures
2030 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2031 error message is output. Some extra information is given for the
2033 probably the only errors that are likely to be caused by users (by
2034 specifying a regex that has nested indefinite repeats, for instance).
2035 If there are more than 20 of these errors, pcregrep is abandoned.
2037 (b) A binary zero was treated as data while matching, but terminated the
2038 output line if it was written out. This has been fixed: binary zeroes
2039 are now no different to any other data bytes.
2041 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2042 used to set a locale for matching. The --locale=xxxx long option has
2043 been added (no short equivalent) to specify a locale explicitly on the
2044 pcregrep command, overriding the environment variables.
2046 (d) When -B was used with -n, some line numbers in the output were one less
2047 than they should have been.
2049 (e) Added the -o (--only-matching) option.
2051 (f) If -A or -C was used with -c (count only), some lines of context were
2052 accidentally printed for the final match.
2054 (g) Added the -H (--with-filename) option.
2056 (h) The combination of options -rh failed to suppress file names for files
2057 that were found from directory arguments.
2059 (i) Added the -D (--devices) and -d (--directories) options.
2061 (j) Added the -F (--fixed-strings) option.
2063 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2065 (l) Added the --colo(u)r option.
2067 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2068 is not present by default.
2070 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2071 items such as (?R) or (?1), when the recursion could match a number of
2072 alternatives. If it matched one of the alternatives, but subsequently,
2073 outside the recursion, there was a failure, the code tried to back up into
2074 the recursion. However, because of the way PCRE is implemented, this is not
2075 possible, and the result was an incorrect result from the match.
2077 In order to prevent this happening, the specification of recursion has
2078 been changed so that all such subpatterns are automatically treated as
2079 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2081 4. I had overlooked the fact that, in some locales, there are characters for
2082 which isalpha() is true but neither isupper() nor islower() are true. In
2083 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2084 and ordfeminine) are like this. This affected the treatment of \w and \W
2085 when they appeared in character classes, but not when they appeared outside
2086 a character class. The bit map for "word" characters is now created
2087 separately from the results of isalnum() instead of just taking it from the
2088 upper, lower, and digit maps. (Plus the underscore character, of course.)
2090 5. The above bug also affected the handling of POSIX character classes such as
2091 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2092 permanent tables. Instead, the bit maps for such a class were previously
2093 created as the appropriate unions of the upper, lower, and digit bitmaps.
2094 Now they are created by subtraction from the [[:word:]] class, which has
2095 its own bitmap.
2097 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2098 It is created by subtracting the vertical space characters (\x09, \x0a,
2099 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2100 subtraction was done in the overall bitmap for a character class, meaning
2101 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2102 be recognized. This bug has been fixed.
2104 7. Patches from the folks at Google:
2106 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2107 real life, but is still worth protecting against".
2109 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2110 regular expressions".
2112 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2113 have it.
2115 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2116 "configure" and the latter not, in order to fix a problem somebody had
2117 with compiling the Arg class on HP-UX.
2119 (e) Improve the error-handling of the C++ wrapper a little bit.
2121 (f) New tests for checking recursion limiting.
2123 8. The pcre_memmove() function, which is used only if the environment does not
2124 have a standard memmove() function (and is therefore rarely compiled),
2125 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2126 returning a result (though PCRE never actually uses the result).
2128 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2129 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2130 returned instead of calling malloc() with an overflowing number that would
2131 most likely cause subsequent chaos.
2133 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2135 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2136 with this option is matched, the nmatch and pmatch options of regexec() are
2137 ignored.
2139 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2140 provided in case anyone wants to the the POSIX interface with UTF-8
2141 strings.
2143 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2144 C++ linking (needed for some HP-UX environments).
2146 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2147 (unused parameter) and in the pcre_printint() function (omitted "default"
2148 switch label when the default is to do nothing).
2150 15. Added some code to make it possible, when PCRE is compiled as a C++
2151 library, to replace subject pointers for pcre_exec() with a smart pointer
2152 class, thus making it possible to process discontinuous strings.
2154 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2155 much the same function. They were added by different people who were trying
2156 to make PCRE easy to compile on non-Unix systems. It has been suggested
2157 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2158 for compiling on Windows systems. I have therefore replaced it with
2159 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2160 defaults to "extern" for C or "extern C" for C++, which works fine on
2161 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2162 SCOPE with something explicit in config.h. In addition:
2164 (a) pcreposix.h still had just "extern" instead of either of these macros;
2165 I have replaced it with PCRE_DATA_SCOPE.
2167 (b) Functions such as _pcre_xclass(), which are internal to the library,
2168 but external in the C sense, all had PCRE_EXPORT in their definitions.
2169 This is apparently wrong for the Windows case, so I have removed it.
2170 (It makes no difference on Unix-like systems.)
2172 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2173 of recursive calls to match(). This is different to MATCH_LIMIT because
2174 that limits the total number of calls to match(), not all of which increase
2175 the depth of recursion. Limiting the recursion depth limits the amount of
2176 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2177 when PCRE is compiled, and changed at run time. A patch from Google adds
2178 this functionality to the C++ interface.
2180 18. Changes to the handling of Unicode character properties:
2182 (a) Updated the table to Unicode 4.1.0.
2184 (b) Recognize characters that are not in the table as "Cn" (undefined).
2186 (c) I revised the way the table is implemented to a much improved format
2187 which includes recognition of ranges. It now supports the ranges that
2188 are defined in UnicodeData.txt, and it also amalgamates other
2189 characters into ranges. This has reduced the number of entries in the
2190 table from around 16,000 to around 3,000, thus reducing its size
2191 considerably. I realized I did not need to use a tree structure after
2192 all - a binary chop search is just as efficient. Having reduced the
2193 number of entries, I extended their size from 6 bytes to 8 bytes to
2194 allow for more data.
2196 (d) Added support for Unicode script names via properties such as \p{Han}.
2198 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2199 matching that character.
2201 20. When matching a repeated Unicode property with a minimum greater than zero,
2202 (for example \pL{2,}), PCRE could look past the end of the subject if it
2203 reached it while seeking the minimum number of characters. This could
2204 happen only if some of the characters were more than one byte long, because
2205 there is a check for at least the minimum number of bytes.
2207 21. Refactored the implementation of \p and \P so as to be more general, to
2208 allow for more different types of property in future. This has changed the
2209 compiled form incompatibly. Anybody with saved compiled patterns that use
2210 \p or \P will have to recompile them.
2212 22. Added "Any" and "L&" to the supported property types.
2214 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2215 but give a compile time error if the value is greater than 0xff.
2217 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2218 accidentally not being installed or uninstalled.
2220 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2221 made were to insert the current release number. This seemed silly, because
2222 it made things harder for people building PCRE on systems that don't run
2223 "configure". I have turned pcre.h into a distributed file, no longer built
2224 by "configure", with the version identification directly included. There is
2225 no longer a pcre.h.in file.
2227 However, this change necessitated a change to the pcre-config script as
2228 well. It is built from pcre-config.in, and one of the substitutions was the
2229 release number. I have updated configure.ac so that ./configure now finds
2230 the release number by grepping pcre.h.
2232 26. Added the ability to run the tests under valgrind.
2235 Version 6.4 05-Sep-05
2236 ---------------------
2238 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2239 "--" to be printed when multiple files were scanned, even when none of the
2240 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2241 consider it to be a bug, and have restored the previous behaviour.
2243 2. A couple of code tidies to get rid of compiler warnings.
2245 3. The pcretest program used to cheat by referring to symbols in the library
2246 whose names begin with _pcre_. These are internal symbols that are not
2247 really supposed to be visible externally, and in some environments it is
2248 possible to suppress them. The cheating is now confined to including
2249 certain files from the library's source, which is a bit cleaner.
2251 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2252 file's purpose clearer.
2254 5. Reorganized pcre_ucp_findchar().
2257 Version 6.3 15-Aug-05
2258 ---------------------
2260 1. The file libpcre.pc.in did not have general read permission in the tarball.
2262 2. There were some problems when building without C++ support:
2264 (a) If C++ support was not built, "make install" and "make test" still
2265 tried to test it.
2267 (b) There were problems when the value of CXX was explicitly set. Some
2268 changes have been made to try to fix these, and ...
2270 (c) --disable-cpp can now be used to explicitly disable C++ support.
2272 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2273 backslash in a target when C++ was disabled. This confuses some
2274 versions of "make", apparently. Using an intermediate variable solves
2275 this. (Same for CPP_LOBJ.)
2277 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2278 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2279 necessary on certain architectures.
2281 4. Added a setting of -export-symbols-regex to the link command to remove
2282 those symbols that are exported in the C sense, but actually are local
2283 within the library, and not documented. Their names all begin with
2284 "_pcre_". This is not a perfect job, because (a) we have to except some
2285 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2286 available (and never for static libraries). I have made a note to try to
2287 find a way round (a) in the future.
2290 Version 6.2 01-Aug-05
2291 ---------------------
2293 1. There was no test for integer overflow of quantifier values. A construction
2294 such as {1111111111111111} would give undefined results. What is worse, if
2295 a minimum quantifier for a parenthesized subpattern overflowed and became
2296 negative, the calculation of the memory size went wrong. This could have
2297 led to memory overwriting.
2299 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2301 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2302 operating environments where this matters.
2304 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2305 PCRE options from within the C++ wrapper.
2307 5. Named capturing subpatterns were not being correctly counted when a pattern
2308 was compiled. This caused two problems: (a) If there were more than 100
2309 such subpatterns, the calculation of the memory needed for the whole
2310 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2311 back references of the form \12, where the number was greater than 9, were
2312 not recognized as back references, even though there were sufficient
2313 previous subpatterns.
2315 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2316 versions of gcc, e.g. 2.95.4.
2319 Version 6.1 21-Jun-05
2320 ---------------------
2322 1. There was one reference to the variable "posix" in pcretest.c that was not
2323 surrounded by "#if !defined NOPOSIX".
2325 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2326 the cross-check on the old pcre_info() function, for the benefit of the
2327 cut-down version of PCRE that is currently imported into Exim.
2329 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2330 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2331 patterns that start (?i) but I don't think that matters. The use of (?i) is
2332 just an example; this all applies to the other options as well.
2334 4. Since libtool seems to echo the compile commands it is issuing, the output
2335 from "make" can be reduced a bit by putting "@" in front of each libtool
2336 compile command.
2338 5. Patch from the folks at Google for configure.in to be a bit more thorough
2339 in checking for a suitable C++ installation before trying to compile the
2340 C++ stuff. This should fix a reported problem when a compiler was present,
2341 but no suitable headers.
2343 6. The man pages all had just "PCRE" as their title. I have changed them to
2344 be the relevant file name. I have also arranged that these names are
2345 retained in the file doc/pcre.txt, which is a concatenation in text format
2346 of all the man pages except the little individual ones for each function.
2348 7. The NON-UNIX-USE file had not been updated for the different set of source
2349 files that come with release 6. I also added a few comments about the C++
2350 wrapper.
2353 Version 6.0 07-Jun-05
2354 ---------------------
2356 1. Some minor internal re-organization to help with my DFA experiments.
2358 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2359 didn't matter for the library itself when fully configured, but did matter
2360 when compiling without UCP support, or within Exim, where the ucp files are
2361 not imported.
2363 3. Refactoring of the library code to split up the various functions into
2364 different source modules. The addition of the new DFA matching code (see
2365 below) to a single monolithic source would have made it really too
2366 unwieldy, quite apart from causing all the code to be include in a
2367 statically linked application, when only some functions are used. This is
2368 relevant even without the DFA addition now that patterns can be compiled in
2369 one application and matched in another.
2371 The downside of splitting up is that there have to be some external
2372 functions and data tables that are used internally in different modules of
2373 the library but which are not part of the API. These have all had their
2374 names changed to start with "_pcre_" so that they are unlikely to clash
2375 with other external names.
2377 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2378 a different (DFA) algorithm. Although it is slower than the original
2379 function, it does have some advantages for certain types of matching
2380 problem.
2382 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2383 including restarting after a partial match.
2385 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2386 defined when compiling for Windows was sent to me. I have put it into the
2387 code, though I have no means of testing or verifying it.
2389 7. Added the pcre_refcount() auxiliary function.
2391 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2392 match before or at the first newline in the subject string. In pcretest,
2393 the /f option on a pattern can be used to set this.
2395 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2396 would behave wrongly. This has been present in PCRE since release 4.0.
2398 10. A number of changes to the pcregrep command:
2400 (a) Refactored how -x works; insert ^(...)$ instead of setting
2401 PCRE_ANCHORED and checking the length, in preparation for adding
2402 something similar for -w.
2404 (b) Added the -w (match as a word) option.
2406 (c) Refactored the way lines are read and buffered so as to have more
2407 than one at a time available.
2409 (d) Implemented a pcregrep test script.
2411 (e) Added the -M (multiline match) option. This allows patterns to match
2412 over several lines of the subject. The buffering ensures that at least
2413 8K, or the rest of the document (whichever is the shorter) is available
2414 for matching (and similarly the previous 8K for lookbehind assertions).
2416 (f) Changed the --help output so that it now says
2418 -w, --word-regex(p)
2420 instead of two lines, one with "regex" and the other with "regexp"
2421 because that confused at least one person since the short forms are the
2422 same. (This required a bit of code, as the output is generated
2423 automatically from a table. It wasn't just a text change.)
2425 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2426 option but starts with a hyphen. Could be a pattern or a path name
2427 starting with a hyphen, for instance.
2429 (h) "-" can be given as a file name to represent stdin.
2431 (i) When file names are being printed, "(standard input)" is used for
2432 the standard input, for compatibility with GNU grep. Previously
2433 "<stdin>" was used.
2435 (j) The option --label=xxx can be used to supply a name to be used for
2436 stdin when file names are being printed. There is no short form.
2438 (k) Re-factored the options decoding logic because we are going to add
2439 two more options that take data. Such options can now be given in four
2440 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2442 (l) Added the -A, -B, and -C options for requesting that lines of context
2443 around matches be printed.
2445 (m) Added the -L option to print the names of files that do not contain
2446 any matching lines, that is, the complement of -l.
2448 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2449 continue to scan other files.
2451 (o) The -s option was incorrectly implemented. For compatibility with other
2452 greps, it now suppresses the error message for a non-existent or non-
2453 accessible file (but not the return code). There is a new option called
2454 -q that suppresses the output of matching lines, which was what -s was
2455 previously doing.
2457 (p) Added --include and --exclude options to specify files for inclusion
2458 and exclusion when recursing.
2460 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2461 Hopefully, it now does.
2463 12. Missing cast in pcre_study().
2465 13. Added an "uninstall" target to the makefile.
2467 14. Replaced "extern" in the function prototypes in Makefile.in with
2468 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2469 world, but is set differently for Windows.
2471 15. Added a second compiling function called pcre_compile2(). The only
2472 difference is that it has an extra argument, which is a pointer to an
2473 integer error code. When there is a compile-time failure, this is set
2474 non-zero, in addition to the error test pointer being set to point to an
2475 error message. The new argument may be NULL if no error number is required
2476 (but then you may as well call pcre_compile(), which is now just a
2477 wrapper). This facility is provided because some applications need a
2478 numeric error indication, but it has also enabled me to tidy up the way
2479 compile-time errors are handled in the POSIX wrapper.
2481 16. Added VPATH=.libs to the makefile; this should help when building with one
2482 prefix path and installing with another. (Or so I'm told by someone who
2483 knows more about this stuff than I do.)
2485 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2486 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2487 match everything, including newlines. This is not POSIX-compatible, but
2488 somebody wanted the feature. From pcretest it can be activated by using
2489 both the P and the s flags.
2491 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2493 19. libpcre.pc was being incorrectly installed as executable.
2495 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2496 it now also looks for '\r' so that it will work unmodified on Windows.
2498 21. Added Google's contributed C++ wrapper to the distribution.
2500 22. Added some untidy missing memory free() calls in pcretest, to keep
2501 Electric Fence happy when testing.
2505 Version 5.0 13-Sep-04
2506 ---------------------
2508 1. Internal change: literal characters are no longer packed up into items
2509 containing multiple characters in a single byte-string. Each character
2510 is now matched using a separate opcode. However, there may be more than one
2511 byte in the character in UTF-8 mode.
2513 2. The pcre_callout_block structure has two new fields: pattern_position and
2514 next_item_length. These contain the offset in the pattern to the next match
2515 item, and its length, respectively.
2517 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2518 insertion of callouts before each pattern item. Added the /C option to
2519 pcretest to make use of this.
2521 4. On the advice of a Windows user, the lines
2523 #if defined(_WIN32) || defined(WIN32)
2524 _setmode( _fileno( stdout ), 0x8000 );
2525 #endif /* defined(_WIN32) || defined(WIN32) */
2527 have been added to the source of pcretest. This apparently does useful
2528 magic in relation to line terminators.
2530 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2531 for the benefit of those environments where the "b" makes a difference.
2533 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2534 to know about it. I have put a hack into configure.in that adds in code
2535 to set GCC=yes if CC=icc. This seems to end up at a point in the
2536 generated configure script that is early enough to affect the setting of
2537 compiler options, which is what is needed, but I have no means of testing
2538 whether it really works. (The user who reported this had patched the
2539 generated configure script, which of course I cannot do.)
2541 LATER: After change 22 below (new libtool files), the configure script
2542 seems to know about icc (and also ecc). Therefore, I have commented out
2543 this hack in configure.in.
2545 7. Added support for pkg-config (2 patches were sent in).
2547 8. Negated POSIX character classes that used a combination of internal tables
2548 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2549 [[:^ascii]]. Typically, they would match almost any characters. The other
2550 POSIX classes were not broken in this way.
2552 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2553 to find the match, as PCRE was deluded into thinking that the match had to
2554 start at the start point or following a newline. The same bug applied to
2555 patterns with negative forward assertions or any backward assertions
2556 preceding ".*" at the start, unless the pattern required a fixed first
2557 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2559 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2560 starting at the last subject character, bytes beyond the end of the subject
2561 string were read.
2563 11. Renamed the variable "class" as "classbits" to make life easier for C++
2564 users. (Previously there was a macro definition, but it apparently wasn't
2565 enough.)
2567 12. Added the new field "tables" to the extra data so that tables can be passed
2568 in at exec time, or the internal tables can be re-selected. This allows
2569 a compiled regex to be saved and re-used at a later time by a different
2570 program that might have everything at different addresses.
2572 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2573 -R library as well as a -L library.
2575 14. The debugging options of pcretest (-d on the command line or D on a
2576 pattern) showed incorrect output for anything following an extended class
2577 that contained multibyte characters and which was followed by a quantifier.
2579 15. Added optional support for general category Unicode character properties
2580 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2581 support. It adds about 90K to the size of the library. The meanings of the
2582 inbuilt class escapes such as \d and \s have NOT been changed.
2584 16. Updated pcredemo.c to include calls to free() to release the memory for the
2585 compiled pattern.
2587 17. The generated file chartables.c was being created in the source directory
2588 instead of in the building directory. This caused the build to fail if the
2589 source directory was different from the building directory, and was
2590 read-only.
2592 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2593 file. No doubt somebody will tell me if they don't make sense... Also added
2594 Dan Mooney's comments about building on OpenVMS.
2596 19. Added support for partial matching via the PCRE_PARTIAL option for
2597 pcre_exec() and the \P data escape in pcretest.
2599 20. Extended pcretest with 3 new pattern features:
2601 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2602 write the compiled pattern to the file whose name is "rest-of-line".
2603 This is a straight binary dump of the data, with the saved pointer to
2604 the character tables forced to be NULL. The study data, if any, is
2605 written too. After writing, pcretest reads a new pattern.
2607 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2608 compiled pattern from the given file. There must not be any
2609 occurrences of "<" in the file name (pretty unlikely); if there are,
2610 pcretest will instead treat the initial "<" as a pattern delimiter.
2611 After reading in the pattern, pcretest goes on to read data lines as
2612 usual.
2614 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2615 and 16-bit fields in a compiled pattern, to simulate a pattern that
2616 was compiled on a host of opposite endianness.
2618 21. The pcre-exec() function can now cope with patterns that were compiled on
2619 hosts of opposite endianness, with this restriction:
2621 As for any compiled expression that is saved and used later, the tables
2622 pointer field cannot be preserved; the extra_data field in the arguments
2623 to pcre_exec() should be used to pass in a tables address if a value
2624 other than the default internal tables were used at compile time.
2626 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2627 now diagnosed as an error. Previously, most of the time, a negative number
2628 would have been treated as zero, but if in addition "ovector" was passed as
2629 NULL, a crash could occur.
2631 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2632 new versions from the libtool 1.5 distribution (the last one is a copy of
2633 a file called libtool.m4). This seems to have fixed the need to patch
2634 "configure" to support Darwin 1.3 (which I used to do). However, I still
2635 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2636 workstation).
2638 24. Changed the PCRE licence to be the more standard "BSD" licence.
2641 Version 4.5 01-Dec-03
2642 ---------------------
2644 1. There has been some re-arrangement of the code for the match() function so
2645 that it can be compiled in a version that does not call itself recursively.
2646 Instead, it keeps those local variables that need separate instances for
2647 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2648 needs to "recurse". Keeping track of where control must go is done by means
2649 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2650 hide most of the details from the main code, and operates only if
2651 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2652 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2653 operating.
2655 To make it easier for callers to provide specially tailored get/free
2656 functions for this usage, two new functions, pcre_stack_malloc, and
2657 pcre_stack_free, are used. They are always called in strict stacking order,
2658 and the size of block requested is always the same.
2660 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2661 PCRE has been compiled to use the stack or the heap for recursion. The
2662 -C option of pcretest uses this to show which version is compiled.
2664 A new data escape \S, is added to pcretest; it causes the amounts of store
2665 obtained and freed by both kinds of malloc/free at match time to be added
2666 to the output.
2668 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2669 what's available on my current Linux desktop machine.
2671 3. When matching a UTF-8 string, the test for a valid string at the start has
2672 been extended. If start_offset is not zero, PCRE now checks that it points
2673 to a byte that is the start of a UTF-8 character. If not, it returns
2674 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2675 this is necessary because there may be backward assertions in the pattern.
2676 When matching the same subject several times, it may save resources to use
2677 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2679 4. The code for checking the validity of UTF-8 strings has been tightened so
2680 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2681 containing "overlong sequences".
2683 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2684 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2685 should have been "&", but it just so happened that all the cases this let
2686 through by mistake were picked up later in the function.
2688 6. I had used a variable called "isblank" - this is a C99 function, causing
2689 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2691 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2692 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2693 so the version is shown; (c) stop "make test" echoing "./RunTest".
2695 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2697 9. The prototype for memmove() for systems that don't have it was using
2698 size_t, but the inclusion of the header that defines size_t was later. I've
2699 moved the #includes for the C headers earlier to avoid this.
2701 10. Added some adjustments to the code to make it easier to compiler on certain
2702 special systems:
2704 (a) Some "const" qualifiers were missing.
2705 (b) Added the macro EXPORT before all exported functions; by default this
2706 is defined to be empty.
2707 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2708 that it reads its output file name as an argument instead of writing
2709 to the standard output and assuming this can be redirected.
2711 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2712 class containing characters with values greater than 255, PCRE compilation
2713 went into a loop.
2715 12. A recursive reference to a subpattern that was within another subpattern
2716 that had a minimum quantifier of zero caused PCRE to crash. For example,
2717 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2718 recursion. If the recursively-called subpattern itself had a zero repeat,
2719 that was OK.
2721 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2722 buffer into which it was copied (for escape processing) was still set at
2723 1024, so long lines caused crashes.
2725 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2726 "internal error: code overflow...". This applied to any character class
2727 that was followed by a possessive quantifier.
2729 15. Modified the Makefile to add libpcre.la as a prerequisite for
2730 libpcreposix.la because I was told this is needed for a parallel build to
2731 work.
2733 16. If a pattern that contained .* following optional items at the start was
2734 studied, the wrong optimizing data was generated, leading to matching
2735 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2736 matching string must start with a or b or c. The correct conclusion for
2737 this pattern is that a match can start with any character.
2740 Version 4.4 13-Aug-03
2741 ---------------------
2743 1. In UTF-8 mode, a character class containing characters with values between
2744 127 and 255 was not handled correctly if the compiled pattern was studied.
2745 In fixing this, I have also improved the studying algorithm for such
2746 classes (slightly).
2748 2. Three internal functions had redundant arguments passed to them. Removal
2749 might give a very teeny performance improvement.
2751 3. Documentation bug: the value of the capture_top field in a callout is *one
2752 more than* the number of the hightest numbered captured substring.
2754 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2755 in incorrectly linking with a previously installed version. They now link
2756 explicitly with libpcre.la.
2758 5. configure.in no longer needs to recognize Cygwin specially.
2760 6. A problem in pcre.in for Windows platforms is fixed.
2762 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2763 pcretest, it used to include the size of the study block as part of its
2764 output. Unfortunately, the structure contains a field that has a different
2765 size on different hardware architectures. This meant that the tests that
2766 showed this size failed. As the block is currently always of a fixed size,
2767 this information isn't actually particularly useful in pcretest output, so
2768 I have just removed it.
2770 8. Three pre-processor statements accidentally did not start in column 1.
2771 Sadly, there are *still* compilers around that complain, even though
2772 standard C has not required this for well over a decade. Sigh.
2774 9. In pcretest, the code for checking callouts passed small integers in the
2775 callout_data field, which is a void * field. However, some picky compilers
2776 complained about the casts involved for this on 64-bit systems. Now
2777 pcretest passes the address of the small integer instead, which should get
2778 rid of the warnings.
2780 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2781 both compile and run time, and gives an error if an invalid UTF-8 sequence
2782 is found. There is a option for disabling this check in cases where the
2783 string is known to be correct and/or the maximum performance is wanted.
2785 11. In response to a bug report, I changed one line in Makefile.in from
2787 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2788 to
2789 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2791 to look similar to other lines, but I have no way of telling whether this
2792 is the right thing to do, as I do not use Windows. No doubt I'll get told
2793 if it's wrong...
2796 Version 4.3 21-May-03
2797 ---------------------
2799 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2800 Makefile.
2802 2. Some refactoring to improve the quality of the code:
2804 (i) The utf8_table... variables are now declared "const".
2806 (ii) The code for \cx, which used the "case flipping" table to upper case
2807 lower case letters, now just substracts 32. This is ASCII-specific,
2808 but the whole concept of \cx is ASCII-specific, so it seems
2809 reasonable.
2811 (iii) PCRE was using its character types table to recognize decimal and
2812 hexadecimal digits in the pattern. This is silly, because it handles
2813 only 0-9, a-f, and A-F, but the character types table is locale-
2814 specific, which means strange things might happen. A private
2815 table is now used for this - though it costs 256 bytes, a table is
2816 much faster than multiple explicit tests. Of course, the standard
2817 character types table is still used for matching digits in subject
2818 strings against \d.
2820 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2821 ending in _t are). So I've renamed it as ESC_tee.
2823 3. The first argument for regexec() in the POSIX wrapper should have been
2824 defined as "const".
2826 4. Changed pcretest to use malloc() for its buffers so that they can be
2827 Electric Fenced for debugging.
2829 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2830 to read one or more bytes before the start of the subject string. Often this
2831 had no effect on PCRE's behaviour, but in some circumstances it could
2832 provoke a segmentation fault.
2834 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2835 to try to read one or more bytes before the start of the subject string.
2837 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2838 UTF-8 support could misbehave in various ways if the subject string
2839 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2840 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2841 back over UTF-8 characters.)
2844 Version 4.2 14-Apr-03
2845 ---------------------
2847 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2849 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2850 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2851 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2852 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2853 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2855 Note: automatic setting of the BUILD variables is not yet working
2856 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2857 compile-time but not at link-time
2858 [LINK]: use for linking executables only
2859 make different versions for Windows and non-Windows
2860 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2861 libraries
2862 [LINK_FOR_BUILD]: new variable
2863 [OBJEXT]: use throughout
2864 [EXEEXT]: use throughout
2865 <winshared>: new target
2866 <wininstall>: new target
2867 <dftables.o>: use native compiler
2868 <dftables>: use native linker
2869 <install>: handle Windows platform correctly
2870 <clean>: ditto
2871 <check>: ditto
2872 copy DLL to top builddir before testing
2874 As part of these changes, -no-undefined was removed again. This was reported
2875 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2876 in any case.
2878 3. Some tidies to get rid of compiler warnings:
2880 . In the match_data structure, match_limit was an unsigned long int, whereas
2881 match_call_count was an int. I've made them both unsigned long ints.
2883 . In pcretest the fact that a const uschar * doesn't automatically cast to
2884 a void * provoked a warning.
2886 . Turning on some more compiler warnings threw up some "shadow" variables
2887 and a few more missing casts.
2889 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2890 option, a class that contained a single character with a value between 128
2891 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2893 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2894 option, a class that contained several characters, but with at least one
2895 whose value was between 128 and 255 caused PCRE to crash.
2898 Version 4.1 12-Mar-03
2899 ---------------------
2901 1. Compiling with gcc -pedantic found a couple of places where casts were
2902 needed, and a string in dftables.c that was longer than standard compilers are
2903 required to support.
2905 2. Compiling with Sun's compiler found a few more places where the code could
2906 be tidied up in order to avoid warnings.
2908 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2909 first of these names is deprecated in the latest Autoconf in favour of the name
2910 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2911 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2912 analogy I have changed it to CFLAGS_FOR_BUILD.
2914 4. Added -no-undefined to the linking command in the Makefile, because this is
2915 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2916 linking step for the pcreposix library.
2918 5. PCRE was failing to diagnose the case of two named groups with the same
2919 name.
2921 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2922 literal character that is needed in the subject for a match, and scans along to
2923 ensure that it is present before embarking on the full matching process. This
2924 saves time in cases of nested unlimited repeats that are never going to match.
2925 Problem: the scan can take a lot of time if the subject is very long (e.g.
2926 megabytes), thus penalizing straightforward matches. It is now done only if the
2927 amount of subject to be scanned is less than 1000 bytes.
2929 7. A lesser problem with the same optimization is that it was recording the
2930 first character of an anchored pattern as "needed", thus provoking a search
2931 right along the subject, even when the first match of the pattern was going to
2932 fail. The "needed" character is now not set for anchored patterns, unless it
2933 follows something in the pattern that is of non-fixed length. Thus, it still
2934 fulfils its original purpose of finding quick non-matches in cases of nested
2935 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2938 Version 4.0 17-Feb-03
2939 ---------------------
2941 1. If a comment in an extended regex that started immediately after a meta-item
2942 extended to the end of string, PCRE compiled incorrect data. This could lead to
2943 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2945 2. Moved to autoconf 2.53 and libtool 1.4.2.
2947 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2948 the special perltest8 script is no longer needed - all the tests can be run
2949 from a single perltest script.
2951 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2952 by \s. It has now been removed in PCRE. This means it isn't recognized as
2953 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2954 class [:space:] *does* include VT, thereby creating a mess.
2956 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2957 space and tab.
2959 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2960 its new features into the main test script, reducing the number of scripts.
2962 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2963 were backward compatible, and made the (?i) apply to the whole pattern, as if
2964 /i were given. Now it behaves more logically, and applies the option setting
2965 only to what follows. PCRE has been changed to follow suit. However, if it
2966 finds options settings right at the start of the pattern, it extracts them into
2967 the global options, as before. Thus, they show up in the info data.
2969 8. Added support for the \Q...\E escape sequence. Characters in between are
2970 treated as literals. This is slightly different from Perl in that $ and @ are
2971 also handled as literals inside the quotes. In Perl, they will cause variable
2972 interpolation. Note the following examples:
2974 Pattern PCRE matches Perl matches
2976 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2977 \Qabc\$xyz\E abc\$xyz abc\$xyz
2978 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2980 For compatibility with Perl, \Q...\E sequences are recognized inside character
2981 classes as well as outside them.
2983 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2984 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2985 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2986 signed/unsigned warnings.
2988 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2989 option for pcretest, so I've replaced it by a simple function that does just
2990 that job.
2992 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2993 "pcregrep -".
2995 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2996 Java package. This provides some syntactic sugar for simple cases of what my
2997 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2998 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2999 item, you can use this simplified notation. Note that only makes sense with
3000 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3001 greediness, whatever the setting of the PCRE_UNGREEDY option.
3003 13. A change of greediness default within a pattern was not taking effect at
3004 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3005 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3006 was abstracted outside.
3008 14. PCRE now supports the \G assertion. It is true when the current matching
3009 position is at the start point of the match. This differs from \A when the
3010 starting offset is non-zero. Used with the /g option of pcretest (or similar
3011 code), it works in the same way as it does for Perl's /g option. If all
3012 alternatives of a regex begin with \G, the expression is anchored to the start
3013 match position, and the "anchored" flag is set in the compiled expression.
3015 15. Some bugs concerning the handling of certain option changes within patterns
3016 have been fixed. These applied to options other than (?ims). For example,
3017 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3018 been the other way round. Some of this was related to change 7 above.
3020 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3021 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3022 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3023 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3025 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3026 mode. Unlike ".", it always matches newline, whatever the setting of
3027 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3028 assertions. Perl allows it, but it doesn't (in general) work because it can't
3029 calculate the length of the lookbehind. At least, that's the case for Perl
3030 5.8.0 - I've been told they are going to document that it doesn't work in
3031 future.
3033 18. Added an error diagnosis for escapes that PCRE does not support: these are
3034 \L, \l, \N, \P, \p, \U, \u, and \X.
3036 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3037 reading past the end of the pattern in cases such as /[abcd/.
3039 20. PCRE was getting more memory than necessary for patterns with classes that
3040 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3042 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3043 compiling PCRE for use with Virtual Pascal.
3045 22. Small fix to the Makefile to make it work properly if the build is done
3046 outside the source tree.
3048 23. Added a new extension: a condition to go with recursion. If a conditional
3049 subpattern starts with (?(R) the "true" branch is used if recursion has
3050 happened, whereas the "false" branch is used only at the top level.
3052 24. When there was a very long string of literal characters (over 255 bytes
3053 without UTF support, over 250 bytes with UTF support), the computation of how
3054 much memory was required could be incorrect, leading to segfaults or other
3055 strange effects.
3057 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3058 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3059 there was a subsequent back reference to those brackets. This meant that, for
3060 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3061 possible to check for precisely this case. All we can do is abandon the
3062 optimization if .* occurs inside capturing brackets when there are any back
3063 references whatsoever. (See below for a better fix that came later.)
3065 26. The handling of the optimization for finding the first character of a
3066 non-anchored pattern, and for finding a character that is required later in the
3067 match were failing in some cases. This didn't break the matching; it just
3068 failed to optimize when it could. The way this is done has been re-implemented.
3070 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3072 28. Added a new feature that provides some of the functionality that Perl
3073 provides with (?{...}). The facility is termed a "callout". The way it is done
3074 in PCRE is for the caller to provide an optional function, by setting
3075 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3076 global variable. By default it is unset, which disables all calling out. To get
3077 the function called, the regex must include (?C) at appropriate points. This
3078 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3079 This provides a means of identifying different callout points. When PCRE
3080 reaches such a point in the regex, if pcre_callout has been set, the external
3081 function is called. It is provided with data in a structure called
3082 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3083 matching continues; if it returns a non-zero value, the match at the current
3084 point fails. However, backtracking will occur if possible. [This was changed
3085 later and other features added - see item 49 below.]
3087 29. pcretest is upgraded to test the callout functionality. It provides a
3088 callout function that displays information. By default, it shows the start of
3089 the match and the current position in the text. There are some new data escapes
3090 to vary what happens:
3092 \C+ in addition, show current contents of captured substrings
3093 \C- do not supply a callout function
3094 \C!n return 1 when callout number n is reached
3095 \C!n!m return 1 when callout number n is reached for the mth time
3097 30. If pcregrep was called with the -l option and just a single file name, it
3098 output "<stdin>" if a match was found, instead of the file name.
3100 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3101 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3102 pcre_exec(). This saves a malloc/free per call. The default value of
3103 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3104 when configuring.
3106 32. The default maximum size of a compiled pattern is 64K. There have been a
3107 few cases of people hitting this limit. The code now uses macros to handle the
3108 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3109 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3110 configuring. Tests 2 and 5 work only with 2-byte links because they output
3111 debugging information about compiled patterns.
3113 33. Internal code re-arrangements:
3115 (a) Moved the debugging function for printing out a compiled regex into
3116 its own source file (printint.c) and used #include to pull it into
3117 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3118 separate copies.
3120 (b) Defined the list of op-code names for debugging as a macro in
3121 internal.h so that it is next to the definition of the opcodes.
3123 (c) Defined a table of op-code lengths for simpler skipping along compiled
3124 code. This is again a macro in internal.h so that it is next to the
3125 definition of the opcodes.
3127 34. Added support for recursive calls to individual subpatterns, along the
3128 lines of Robin Houston's patch (but implemented somewhat differently).
3130 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3131 allow it to read and process whole directories in Win32. This code was
3132 contributed by Lionel Fourquaux; it has not been tested by me.
3134 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3135 used to name a group. Names consist of alphanumerics and underscores, and must
3136 be unique. Back references use the syntax (?P=name) and recursive calls use
3137 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3138 numbers. The function pcre_fullinfo() can be used after compilation to extract
3139 a name/number map. There are three relevant calls:
3141 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3142 PCRE_INFO_NAMECOUNT yields the number of entries
3143 PCRE_INFO_NAMETABLE yields a pointer to the map.
3145 The map is a vector of fixed-size entries. The size of each entry depends on
3146 the length of the longest name used. The first two bytes of each entry are the
3147 group number, most significant byte first. There follows the corresponding
3148 name, zero terminated. The names are in alphabetical order.
3150 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3151 case instead of 255. Making it the same both with and without UTF-8 support
3152 means that the same test output works with both.
3154 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3155 calling malloc() with a zero argument.
3157 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3158 optimization. I've improved things by keeping a bitmap of backreferences with
3159 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3160 fact referenced, the optimization can be applied. It is unlikely that a
3161 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3162 the match to follow \n) will appear inside brackets with a number greater than
3163 31, but if it does, any back reference > 31 suppresses the optimization.
3165 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3166 of disabling numbered capturing parentheses. Any opening parenthesis that is
3167 not followed by ? behaves as if it were followed by ?: but named parentheses
3168 can still be used for capturing (and they will acquire numbers in the usual
3169 way).
3171 41. Redesigned the return codes from the match() function into yes/no/error so
3172 that errors can be passed back from deep inside the nested calls. A malloc
3173 failure while inside a recursive subpattern call now causes the
3174 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3176 42. It is now possible to set a limit on the number of times the match()
3177 function is called in a call to pcre_exec(). This facility makes it possible to
3178 limit the amount of recursion and backtracking, though not in a directly
3179 obvious way, because the match() function is used in a number of different
3180 circumstances. The count starts from zero for each position in the subject
3181 string (for non-anchored patterns). The default limit is, for compatibility, a
3182 large number, namely 10 000 000. You can change this in two ways:
3184 (a) When configuring PCRE before making, you can use --with-match-limit=n
3185 to set a default value for the compiled library.
3187 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3188 a different value is set. See 45 below.
3190 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3192 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3193 of things that can be changed at compile time. The first argument specifies
3194 what is wanted and the second points to where the information is to be placed.
3195 The current list of available information is:
3199 The output is an integer that is set to one if UTF-8 support is available;
3200 otherwise it is set to zero.
3204 The output is an integer that it set to the value of the code that is used for
3205 newline. It is either LF (10) or CR (13).
3209 The output is an integer that contains the number of bytes used for internal
3210 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3214 The output is an integer that contains the threshold above which the POSIX
3215 interface uses malloc() for output vectors. See item 31 above.
3219 The output is an unsigned integer that contains the default limit of the number
3220 of match() calls in a pcre_exec() execution. See 42 above.
3222 44. pcretest has been upgraded by the addition of the -C option. This causes it
3223 to extract all the available output from the new pcre_config() function, and to
3224 output it. The program then exits immediately.
3226 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3227 order to support additional features. One way would have been to define
3228 pcre_exec2() (for example) with extra arguments, but this would not have been
3229 extensible, and would also have required all calls to the original function to
3230 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3231 is used for passing in "extra" data from pcre_study().
3233 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3234 contains the following fields:
3236 flags a bitmap indicating which of the following fields are set
3237 study_data opaque data from pcre_study()
3238 match_limit a way of specifying a limit on match() calls for a specific
3239 call to pcre_exec()
3240 callout_data data for callouts (see 49 below)
3242 The flag bits are also defined in pcre.h, and are
3248 The pcre_study() function now returns one of these new pcre_extra blocks, with
3249 the actual study data pointed to by the study_data field, and the
3250 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3251 before. That is, this change is entirely upwards-compatible and requires no
3252 change to existing code.
3254 If you want to pass in additional data to pcre_exec(), you can either place it
3255 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3256 block.
3258 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3259 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3260 times with different match limits, until it finds the minimum value needed for
3261 pcre_exec() to complete. The value is then output. This can be instructive; for
3262 most simple matches the number is quite small, but for pathological cases it
3263 gets very large very quickly.
3265 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3266 returns the size of the data block pointed to by the study_data field in a
3267 pcre_extra block, that is, the value that was passed as the argument to
3268 pcre_malloc() when PCRE was getting memory in which to place the information
3269 created by pcre_study(). The fourth argument should point to a size_t variable.
3270 pcretest has been extended so that this information is shown after a successful
3271 pcre_study() call when information about the compiled regex is being displayed.
3273 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3274 because what follows is always an absolute path. (Later: it turns out that this
3275 is more than cosmetic for MinGW, because it doesn't like empty path
3276 components.)
3278 49. Some changes have been made to the callout feature (see 28 above):
3280 (i) A callout function now has three choices for what it returns:
3282 0 => success, carry on matching
3283 > 0 => failure at this point, but backtrack if possible
3284 < 0 => serious error, return this value from pcre_exec()
3286 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3287 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3288 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3289 use by callout functions. It will never be used by PCRE itself.
3291 (ii) The pcre_extra structure (see 45 above) has a void * field called
3292 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3293 pcre_callout_block structure has a field of the same name. The contents of
3294 the field passed in the pcre_extra structure are passed to the callout
3295 function in the corresponding field in the callout block. This makes it
3296 easier to use the same callout-containing regex from multiple threads. For
3297 testing, the pcretest program has a new data escape
3299 \C*n pass the number n (may be negative) as callout_data
3301 If the callout function in pcretest receives a non-zero value as
3302 callout_data, it returns that value.
3304 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3305 there were some redundant $(CFLAGS) in commands that are now specified as
3306 $(LINK), which already includes $(CFLAGS).
3308 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3309 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3310 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3311 one-byte characters throughout. Note that case-insensitive matching applies
3312 only to characters whose values are less than 256. PCRE doesn't support the
3313 notion of cases for higher-valued characters.
3315 (i) A character class whose characters are all within 0-255 is handled as
3316 a bit map, and the map is inverted for negative classes. Previously, a
3317 character > 255 always failed to match such a class; however it should
3318 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3320 (ii) A negated character class with a single character < 255 is coded as
3321 "not this character" (OP_NOT). This wasn't working properly when the test
3322 character was multibyte, either singly or repeated.
3324 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3325 mode, for example: \x{100}{2,3}.
3327 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3328 singly or repeated) now correctly test multibyte characters. However,
3329 PCRE doesn't recognize any characters with values greater than 255 as
3330 digits, spaces, or word characters. Such characters always match \D, \S,
3331 and \W, and never match \d, \s, or \w.
3333 (v) Classes may now contain characters and character ranges with values
3334 greater than 255. For example: [ab\x{100}-\x{400}].
3336 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3337 PCRE in UTF-8 mode.
3339 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3340 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3341 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3342 value.)
3344 53. The single man page has become too large. I have therefore split it up into
3345 a number of separate man pages. These also give rise to individual HTML pages;
3346 these are now put in a separate directory, and there is an index.html page that
3347 lists them all. Some hyperlinking between the pages has been installed.
3349 54. Added convenience functions for handling named capturing parentheses.
3351 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3352 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3353 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3354 are faulted.
3356 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3357 calling configure. These values are used when compiling the dftables.c program
3358 which is run to generate the source of the default character tables. They
3359 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3360 you will need to set these values.
3362 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3365 Version 3.9 02-Jan-02
3366 ---------------------
3368 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3370 2. If --disable-static was given, the building process failed when trying to
3371 build pcretest and pcregrep. (For some reason it was using libtool to compile
3372 them, which is not right, as they aren't part of the library.)
3375 Version 3.8 18-Dec-01
3376 ---------------------
3378 1. The experimental UTF-8 code was completely screwed up. It was packing the
3379 bytes in the wrong order. How dumb can you get?
3382 Version 3.7 29-Oct-01
3383 ---------------------
3385 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3386 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3387 this didn't happen under Solaris 8, where I normally test things.
3389 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3390 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3391 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3392 this fix an hour or so after the initial 3.7 release.)
3395 Version 3.6 23-Oct-01
3396 ---------------------
3398 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3399 offsets passed as NULL with zero offset count.
3401 2. The config.guess and config.sub files had not been updated when I moved to
3402 the latest autoconf.
3405 Version 3.5 15-Aug-01
3406 ---------------------
3408 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3409 had been forgotten.
3411 2. By using declared but undefined structures, we can avoid using "void"
3412 definitions in pcre.h while keeping the internal definitions of the structures
3413 private.
3415 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3416 user point of view, this means that both static and shared libraries are built
3417 by default, but this can be individually controlled. More of the work of
3418 handling this static/shared cases is now inside libtool instead of PCRE's make
3419 file.
3421 4. The pcretest utility is now installed along with pcregrep because it is
3422 useful for users (to test regexs) and by doing this, it automatically gets
3423 relinked by libtool. The documentation has been turned into a man page, so
3424 there are now .1, .txt, and .html versions in /doc.
3426 5. Upgrades to pcregrep:
3427 (i) Added long-form option names like gnu grep.
3428 (ii) Added --help to list all options with an explanatory phrase.
3429 (iii) Added -r, --recursive to recurse into sub-directories.
3430 (iv) Added -f, --file to read patterns from a file.
3432 6. pcre_exec() was referring to its "code" argument before testing that
3433 argument for NULL (and giving an error if it was NULL).
3435 7. Upgraded Makefile.in to allow for compiling in a different directory from
3436 the source directory.
3438 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3439 options bits, the pointer it was passed was to an int instead of to an unsigned
3440 long int. This mattered only on 64-bit systems.
3442 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3443 generated) instead of pcre.in, which it its source. Also made the same change
3444 in several of the .c files.
3446 10. A new release of gcc defines printf() as a macro, which broke pcretest
3447 because it had an ifdef in the middle of a string argument for printf(). Fixed
3448 by using separate calls to printf().
3450 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3451 script, to force use of CR or LF instead of \n in the source. On non-Unix
3452 systems, the value can be set in config.h.
3454 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3455 absolute limit. Changed the text of the error message to make this clear, and
3456 likewise updated the man page.
3458 13. The limit of 99 on the number of capturing subpatterns has been removed.
3459 The new limit is 65535, which I hope will not be a "real" limit.
3462 Version 3.4 22-Aug-00
3463 ---------------------
3465 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3467 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3470 Version 3.3 01-Aug-00
3471 ---------------------
3473 1. If an octal character was given, but the value was greater than \377, it
3474 was not getting masked to the least significant bits, as documented. This could
3475 lead to crashes in some systems.
3477 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3478 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3480 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3481 These just pass their arguments on to (pcre_free)(), but they are provided
3482 because some uses of PCRE bind it to non-C systems that can call its functions,
3483 but cannot call free() or pcre_free() directly.
3485 4. Add "make test" as a synonym for "make check". Corrected some comments in
3486 the Makefile.
3488 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3489 Makefile.
3491 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3492 command called pgrep for grepping around the active processes.
3494 7. Added the beginnings of support for UTF-8 character strings.
3496 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3497 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3498 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3499 out for the ar command.)
3502 Version 3.2 12-May-00
3503 ---------------------
3505 This is purely a bug fixing release.
3507 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3508 of ZA. This was just one example of several cases that could provoke this bug,
3509 which was introduced by change 9 of version 2.00. The code for breaking
3510 infinite loops after an iteration that matches an empty string was't working
3511 correctly.
3513 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3514 when matched against abbab (for example). After matching an empty string, it
3515 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3516 caused it to match further down the string than it should.
3518 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3519 was there because it doesn't seem to be needed, and it causes trouble on some
3520 systems, as it is not a Standard C header. It has been removed.
3522 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3523 were reported on the Macintosh. The changes were from
3525 while ((c = *(++ptr)) != 0 && c != '\n');
3526 to
3527 while ((c = *(++ptr)) != 0 && c != '\n') ;
3529 Totally extraordinary, but if that's what it takes...
3531 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3532 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3533 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3534 assumes the way PCRE uses memmove() (always moving upwards).
3536 6. PCRE is being used in one environment where strchr() is not available. There
3537 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3538 faster code anyway.
3541 Version 3.1 09-Feb-00
3542 ---------------------
3544 The only change in this release is the fixing of some bugs in Makefile.in for
3545 the "install" target:
3547 (1) It was failing to install pcreposix.h.
3549 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3552 Version 3.0 01-Feb-00
3553 ---------------------
3555 1. Add support for the /+ modifier to perltest (to output $` like it does in
3556 pcretest).
3558 2. Add support for the /g modifier to perltest.
3560 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3561 matches null strings.
3563 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3564 pattern. Perl treats empty patterns specially - it reuses the most recent
3565 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3566 effect.
3568 5. The POSIX interface was broken in that it was just handing over the POSIX
3569 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3570 required a bigger vector, with some working space on the end. This means that
3571 the POSIX wrapper now has to get and free some memory, and copy the results.
3573 6. Added some simple autoconf support, placing the test data and the
3574 documentation in separate directories, re-organizing some of the
3575 information files, and making it build pcre-config (a GNU standard). Also added
3576 libtool support for building PCRE as a shared library, which is now the
3577 default.
3579 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3580 09 are not valid octal constants. Single digits will be used for minor values
3581 less than 10.
3583 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3584 existing programs that set these in the POSIX interface can use PCRE without
3585 modification.
3587 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3588 return all that pcre_info() returns, plus additional data. The pcre_info()
3589 function is retained for compatibility, but is considered to be obsolete.
3591 10. Added experimental recursion feature (?R) to handle one common case that
3592 Perl 5.6 will be able to do with (?p{...}).
3594 11. Added support for POSIX character classes like [:alpha:], which Perl is
3595 adopting.
3598 Version 2.08 31-Aug-99
3599 ----------------------
3601 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3602 trying to match at the startoffset position, but instead was moving forward to
3603 the next newline as if a previous match had failed.
3605 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3606 and could get into a loop if a null string was matched other than at the start
3607 of the subject.
3609 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3610 be distinguished at compile time, and for completeness also added PCRE_DATE.
3612 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3613 in GnuWin32 environments.
3616 Version 2.07 29-Jul-99
3617 ----------------------
3619 1. The documentation is now supplied in plain text form and HTML as well as in
3620 the form of man page sources.
3622 2. C++ compilers don't like assigning (void *) values to other pointer types.
3623 In particular this affects malloc(). Although there is no problem in Standard
3624 C, I've put in casts to keep C++ compilers happy.
3626 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3627 should be (const char *).
3629 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3630 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3631 However, I haven't made this a standard facility. The documentation doesn't
3632 mention it, and the Makefile doesn't support it.
3634 5. The Makefile now contains an "install" target, with editable destinations at
3635 the top of the file. The pcretest program is not installed.
3637 6. pgrep -V now gives the PCRE version number and date.
3639 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3640 causing the entire string to be ignored, instead of just the last character.
3642 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3643 non-matching string, it can take a very, very long time, even for strings of
3644 quite modest length, because of the nested recursion. PCRE now does better in
3645 some of these cases. It does this by remembering the last required literal
3646 character in the pattern, and pre-searching the subject to ensure it is present
3647 before running the real match. In other words, it applies a heuristic to detect
3648 some types of certain failure quickly, and in the above example, if presented
3649 with a string that has no trailing " it gives "no match" very quickly.
3651 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3652 other alternatives are tried instead.
3655 Version 2.06 09-Jun-99
3656 ----------------------
3658 1. Change pcretest's output for amount of store used to show just the code
3659 space, because the remainder (the data block) varies in size between 32-bit and
3660 64-bit systems.
3662 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3663 start matching at. This allows lookbehinds to work when searching for multiple
3664 occurrences in a string.
3666 3. Added additional options to pcretest for testing multiple occurrences:
3668 /+ outputs the rest of the string that follows a match
3669 /g loops for multiple occurrences, using the new startoffset argument
3670 /G loops for multiple occurrences by passing an incremented pointer
3672 4. PCRE wasn't doing the "first character" optimization for patterns starting
3673 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3674 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3675 the letter 'x'. On long subject strings, this gives a significant speed-up.
3678 Version 2.05 21-Apr-99
3679 ----------------------
3681 1. Changed the type of magic_number from int to long int so that it works
3682 properly on 16-bit systems.
3684 2. Fixed a bug which caused patterns starting with .* not to work correctly
3685 when the subject string contained newline characters. PCRE was assuming
3686 anchoring for such patterns in all cases, which is not correct because .* will
3687 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3688 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3689 must be retried after every newline in the subject.
3692 Version 2.04 18-Feb-99
3693 ----------------------
3695 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3696 computation of the store needed to hold the pattern was incorrect (too large).
3697 If such patterns were nested a few deep, this could multiply and become a real
3698 problem.
3700 2. Added /M option to pcretest to show the memory requirement of a specific
3701 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3703 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3704 compiled in such a way that the backtracking after subsequent failure was
3705 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3706 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3709 Version 2.03 02-Feb-99
3710 ----------------------
3712 1. Fixed typo and small mistake in man page.
3714 2. Added 4th condition (GPL supersedes if conflict) and created separate
3715 LICENCE file containing the conditions.
3717 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3718 Perl, that is the internal \ allows the delimiter to be included in the
3719 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3720 the final delimiter, add \ to the end of the pattern (to test the error).
3722 4. Added the convenience functions for extracting substrings after a successful
3723 match. Updated pcretest to make it able to test these functions.
3726 Version 2.02 14-Jan-99
3727 ----------------------
3729 1. Initialized the working variables associated with each extraction so that
3730 their saving and restoring doesn't refer to uninitialized store.
3732 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3733 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3734 fix the problem.
3736 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3737 calls, and wasn't printing the correct value for compiling calls. Increased the
3738 default value of LOOPREPEAT, and the number of significant figures in the
3739 times.
3741 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3743 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3744 a building problem on Windows NT with a FAT file system.
3747 Version 2.01 21-Oct-98
3748 ----------------------
3750 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3751 to character tables built by pcre_maketables() in the current locale. If NULL
3752 is passed, the default tables are used.
3755 Version 2.00 24-Sep-98
3756 ----------------------
3758 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3759 it any more.
3761 2. Allow quantification of (?>) groups, and make it work correctly.
3763 3. The first character computation wasn't working for (?>) groups.
3765 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3766 end of the subject) and add 5.005's \z, which really does match only at the
3767 very end of the subject.
3769 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3771 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3772 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3773 localized options. All options to pcre_study() were also removed.
3775 7. Add other new features from 5.005:
3777 $(?<= positive lookbehind
3778 $(?<! negative lookbehind
3779 (?imsx-imsx) added the unsetting capability
3780 such a setting is global if at outer level; local otherwise
3781 (?imsx-imsx:) non-capturing groups with option setting
3782 (?(cond)re|re) conditional pattern matching
3784 A backreference to itself in a repeated group matches the previous
3785 captured string.
3787 8. General tidying up of studying (both automatic and via "study")
3788 consequential on the addition of new assertions.
3790 9. As in 5.005, unlimited repeated groups that could match an empty substring
3791 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3792 runtime if any iteration does actually match an empty substring.
3794 10. Include the RunTest script in the distribution.
3796 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3797 discrepancies, some of which were old and were also with respect to 5.004. They
3798 have now been fixed.
3801 Version 1.09 28-Apr-98
3802 ----------------------
3804 1. A negated single character class followed by a quantifier with a minimum
3805 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3806 program crashes, or just wrong answers. This did not apply to negated classes
3807 containing more than one character, or to minima other than one.
3810 Version 1.08 27-Mar-98
3811 ----------------------
3813 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3815 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3816 latter must appear before anything that relies on it in the pattern.
3819 Version 1.07 16-Feb-98
3820 ----------------------
3822 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3823 repeat of a potentially empty string).
3826 Version 1.06 23-Jan-98
3827 ----------------------
3829 1. Added Markus Oberhumer's little patches for C++.
3831 2. Literal strings longer than 255 characters were broken.
3834 Version 1.05 23-Dec-97
3835 ----------------------
3837 1. Negated character classes containing more than one character were failing if
3838 PCRE_CASELESS was set at run time.
3841 Version 1.04 19-Dec-97
3842 ----------------------
3844 1. Corrected the man page, where some "const" qualifiers had been omitted.
3846 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3847 input syntax.
3849 3. Fixed memory leak which occurred when a regex with back references was
3850 matched with an offsets vector that wasn't big enough. The temporary memory
3851 that is used in this case wasn't being freed if the match failed.
3853 4. Tidied pcretest to ensure it frees memory that it gets.
3855 5. Temporary memory was being obtained in the case where the passed offsets
3856 vector was exactly big enough.
3858 6. Corrected definition of offsetof() from change 5 below.
3860 7. I had screwed up change 6 below and broken the rules for the use of
3861 setjmp(). Now fixed.
3864 Version 1.03 18-Dec-97
3865 ----------------------
3867 1. A erroneous regex with a missing opening parenthesis was correctly
3868 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3869 on some systems.
3871 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3872 it was reported that one broken compiler failed on the former because "code" is
3873 also an independent variable.
3875 3. The erroneous regex a[]b caused an array overrun reference.
3877 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3878 fail on data ending with that character. (It was going on too far, and checking
3879 the next character, typically a binary zero.) This was specific to the
3880 optimized code for single-character negative classes.
3882 5. Added a contributed patch from the TIN world which does the following:
3884 + Add an undef for memmove, in case the the system defines a macro for it.
3886 + Add a definition of offsetof(), in case there isn't one. (I don't know
3887 the reason behind this - offsetof() is part of the ANSI standard - but
3888 it does no harm).
3890 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3891 most of the places where whitespace preceded '#'. I have given up and
3892 allowed the remaining 2 cases to be at the margin.
3894 + Rename some variables in pcre to eliminate shadowing. This seems very
3895 pedantic, but does no harm, of course.
3897 6. Moved the call to setjmp() into its own function, to get rid of warnings
3898 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3900 7. Constructs such as \d{8,} were compiling into the equivalent of
3901 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3902 outcome, but in this particular case used more store than had been allocated,
3903 which caused the bug to be discovered because it threw up an internal error.
3905 8. The debugging code in both pcre and pcretest for outputting the compiled
3906 form of a regex was going wrong in the case of back references followed by
3907 curly-bracketed repeats.
3910 Version 1.02 12-Dec-97
3911 ----------------------
3913 1. Typos in pcre.3 and comments in the source fixed.
3915 2. Applied a contributed patch to get rid of places where it used to remove
3916 'const' from variables, and fixed some signed/unsigned and uninitialized
3917 variable warnings.
3919 3. Added the "runtest" target to Makefile.
3921 4. Set default compiler flag to -O2 rather than just -O.
3924 Version 1.01 19-Nov-97
3925 ----------------------
3927 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3928 like /([ab]*)*/, that is, for classes with more than one character in them.
3930 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3931 as /((?>a*))*/ (a PCRE_EXTRA facility).
3934 Version 1.00 18-Nov-97
3935 ----------------------
3937 1. Added compile-time macros to support systems such as SunOS4 which don't have
3938 memmove() or strerror() but have other things that can be used instead.
3940 2. Arranged that "make clean" removes the executables.
3943 Version 0.99 27-Oct-97
3944 ----------------------
3946 1. Fixed bug in code for optimizing classes with only one character. It was
3947 initializing a 32-byte map regardless, which could cause it to run off the end
3948 of the memory it had got.
3950 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3953 Version 0.98 22-Oct-97
3954 ----------------------
3956 1. Fixed bug in code for handling temporary memory usage when there are more
3957 back references than supplied space in the ovector. This could cause segfaults.
3960 Version 0.97 21-Oct-97
3961 ----------------------
3963 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3965 2. Optimized negated single characters not to use a bit map.
3967 3. Brought error texts together as macro definitions; clarified some of them;
3968 fixed one that was wrong - it said "range out of order" when it meant "invalid
3969 escape sequence".
3971 4. Changed some char * arguments to const char *.
3973 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3975 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3976 pcretest.
3979 Version 0.96 16-Oct-97
3980 ----------------------
3982 1. Added a simple "pgrep" utility to the distribution.
3984 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3985 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3986 where "ddd" means "one or more decimal digits".
3988 3. Fixed serious bug. If a pattern had a back reference, but the call to
3989 pcre_exec() didn't supply a large enough ovector to record the related
3990 identifying subpattern, the match always failed. PCRE now remembers the number
3991 of the largest back reference, and gets some temporary memory in which to save
3992 the offsets during matching if necessary, in order to ensure that
3993 backreferences always work.
3995 4. Increased the compatibility with Perl in a number of ways:
3997 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3998 to request this handling. The option can be set at compile or exec time.
4000 (b) $ matches before a terminating newline by default; an option
4001 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4002 mode). The option can be set at compile or exec time.
4004 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4005 the same as Perl's. If the decimal number it represents is less than 10
4006 or there aren't that many previous left capturing parentheses, an octal
4007 escape is read. Inside a character class, it's always an octal escape,
4008 even if it is a single digit.
4010 (d) An escaped but undefined alphabetic character is taken as a literal,
4011 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4012 escapes.
4014 (e) {0} is now permitted. (The previous item is removed from the compiled
4015 pattern).
4017 5. Changed all the names of code files so that the basic parts are no longer
4018 than 10 characters, and abolished the teeny "globals.c" file.
4020 6. Changed the handling of character classes; they are now done with a 32-byte
4021 bit map always.
4023 7. Added the -d and /D options to pcretest to make it possible to look at the
4024 internals of compilation without having to recompile pcre.
4027 Version 0.95 23-Sep-97
4028 ----------------------
4030 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4031 \x20 at the start of a run of normal characters. These were being treated as
4032 real characters, instead of the source characters being re-checked.
4035 Version 0.94 18-Sep-97
4036 ----------------------
4038 1. The functions are now thread-safe, with the caveat that the global variables
4039 containing pointers to malloc() and free() or alternative functions are the
4040 same for all threads.
4042 2. Get pcre_study() to generate a bitmap of initial characters for non-
4043 anchored patterns when this is possible, and use it if passed to pcre_exec().
4046 Version 0.93 15-Sep-97
4047 ----------------------
4049 1. /(b)|(:+)/ was computing an incorrect first character.
4051 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4052 but not actually doing anything yet.
4054 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4055 as Perl does (e.g. [-az] or [az-]).
4057 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4058 all possible positions.
4060 5. Split up into different modules to avoid including unneeded functions in a
4061 compiled binary. However, compile and exec are still in one module. The "study"
4062 function is split off.
4064 6. The character tables are now in a separate module whose source is generated
4065 by an auxiliary program - but can then be edited by hand if required. There are
4066 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4067 toupper() in the code.
4069 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4070 make them global. Abolish the function for setting them, as the caller can now
4071 set them directly.
4074 Version 0.92 11-Sep-97
4075 ----------------------
4077 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4078 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4080 2. Caseless matching was not working in character classes if the characters in
4081 the pattern were in upper case.
4083 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4085 4. Make PCRE_ANCHORED public and accept as a compile option.
4087 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4088 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4089 pass them.
4091 6. Give an error if bad option bits passed at compile or run time.
4093 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4094 pcretest to cause it to pass that flag.
4096 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4097 options, and the first character, if set.
4099 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4102 Version 0.91 10-Sep-97
4103 ----------------------
4105 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4106 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4108 2. PCRE was looping on encountering an indefinitely repeated back reference to
4109 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4110 Perl does - treats the match as successful.
4112 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5