ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 630 - (show annotations)
Fri Jul 22 10:00:10 2011 UTC (9 years, 7 months ago) by ph10
File size: 207393 byte(s)
Make (*MARK) work in positive assertions.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.13 30-Apr-2011
5 ------------------------
7 1. The Unicode data tables have been updated to Unicode 6.0.0.
9 2. Two minor typos in pcre_internal.h have been fixed.
11 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
12 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
13 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
15 4. There were a number of related bugs in the code for matching backrefences
16 caselessly in UTF-8 mode when codes for the characters concerned were
17 different numbers of bytes. For example, U+023A and U+2C65 are an upper
18 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
19 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
20 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
21 2-byte code at the end of the subject (it thought there wasn't enough data
22 left).
24 5. Comprehensive information about what went wrong is now returned by
25 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
26 as the output vector has at least 2 elements. The offset of the start of
27 the failing character and a reason code are placed in the vector.
29 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
30 now returned is for the first byte of the failing character, instead of the
31 last byte inspected. This is an incompatible change, but I hope it is small
32 enough not to be a problem. It makes the returned offset consistent with
33 pcre_exec() and pcre_dfa_exec().
35 7. pcretest now gives a text phrase as well as the error number when
36 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
37 failure, the offset and reason code are output.
39 8. When \R was used with a maximizing quantifier it failed to skip backwards
40 over a \r\n pair if the subsequent match failed. Instead, it just skipped
41 back over a single character (\n). This seems wrong (because it treated the
42 two characters as a single entity when going forwards), conflicts with the
43 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
44 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
45 has been changed.
47 9. Some internal refactoring has changed the processing so that the handling
48 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
49 time (the PCRE_DOTALL option was changed this way some time ago: version
50 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
51 which was always a bit of a fudge. It also means that there is one less
52 argument for the match() function, which reduces its stack requirements
53 slightly. This change also fixes an incompatibility with Perl: the pattern
54 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
56 10. More internal refactoring has drastically reduced the number of recursive
57 calls to match() for possessively repeated groups such as (abc)++ when
58 using pcre_exec().
60 11. While implementing 10, a number of bugs in the handling of groups were
61 discovered and fixed:
63 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
64 (a|)*(?1) gave a compile-time internal error.
65 ((a|)+)+ did not notice that the outer group could match an empty string.
66 (^a|^)+ was not marked as anchored.
67 (.*a|.*)+ was not marked as matching at start or after a newline.
69 12. Yet more internal refactoring has removed another argument from the match()
70 function. Special calls to this function are now indicated by setting a
71 value in a variable in the "match data" data block.
73 13. Be more explicit in pcre_study() instead of relying on "default" for
74 opcodes that mean there is no starting character; this means that when new
75 ones are added and accidentally left out of pcre_study(), testing should
76 pick them up.
78 14. The -s option of pcretest has been documented for ages as being an old
79 synonym of -m (show memory usage). I have changed it to mean "force study
80 for every regex", that is, assume /S for every regex. This is similar to -i
81 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
82 using it. It makes it easier to run collections of tests with and without
83 study enabled, and thereby test pcre_study() more easily. All the standard
84 tests are now run with and without -s (but some patterns can be marked as
85 "never study" - see 20 below).
87 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
88 restoration of the capturing data to the outer values was not happening
89 correctly.
91 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
92 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
93 pattern had matched an empty string, and so incorrectly returned a no
94 match.
96 17. There was optimizing code for the last branch of non-capturing parentheses,
97 and also for the obeyed branch of a conditional subexpression, which used
98 tail recursion to cut down on stack usage. Unfortunately, not that there is
99 the possibility of (*THEN) occurring in these branches, tail recursion is
100 no longer possible because the return has to be checked for (*THEN). These
101 two optimizations have therefore been removed.
103 18. If a pattern containing \R was studied, it was assumed that \R always
104 matched two bytes, thus causing the minimum subject length to be
105 incorrectly computed because \R can also match just one byte.
107 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
108 was incorrectly computed.
110 20. If /S is present twice on a test pattern in pcretest input, it *disables*
111 studying, thereby overriding the use of -s on the command line. This is
112 necessary for one or two tests to keep the output identical in both cases.
114 21. When (*ACCEPT) was used in an assertion that matched an empty string and
115 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
117 22. When an atomic group that contained a capturing parenthesis was
118 successfully matched, but the branch in which it appeared failed, the
119 capturing was not being forgotten if a higher numbered group was later
120 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
121 group 1 to "a", when in fact it should be unset. This applied to multi-
122 branched capturing and non-capturing groups, repeated or not, and also to
123 positive assertions (capturing in negative assertions is not well defined
124 in PCRE) and also to nested atomic groups.
126 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
127 subject after a captured substring (to make it easier to tell which of a
128 number of identical substrings has been captured).
130 24. The way atomic groups are processed by pcre_exec() has been changed so that
131 if they are repeated, backtracking one repetition now resets captured
132 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
133 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
134 "aaa". Previously, it would have been "a". As part of this code
135 refactoring, the way recursive calls are handled has also been changed.
137 24. If an assertion condition captured any substrings, they were not passed
138 back unless some other capturing happened later. For example, if
139 (?(?=(a))a) was matched against "a", no capturing was returned.
141 25. When studying a pattern that contained subroutine calls or assertions,
142 the code for finding the minimum length of a possible match was handling
143 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
144 group 1 called group 2 while simultaneously a separate group 2 called group
145 1). A stack overflow occurred in this case. I have fixed this by limiting
146 the recursion depth to 10.
148 26. Updated RunTest.bat in the distribution to the version supplied by Tom
149 Fortmann. This supports explicit test numbers on the command line, and has
150 argument validation and error reporting.
152 27. An instance of \X with an unlimited repeat could fail if at any point the
153 first character it looked at was a mark character.
155 28. Some minor code refactoring concerning Unicode properties and scripts
156 should reduce the stack requirement of match() slightly.
158 29. Added the '=' option to pcretest to check the setting of unused capturing
159 slots at the end of the pattern, which are documented as being -1, but are
160 not included in the return count.
162 30. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
163 compiled something random. Now it gives a compile-time error (as does
164 Perl).
166 31. A *MARK encountered during the processing of a positive assertion is now
167 recorded and passed back (compatible with Perl).
170 Version 8.12 15-Jan-2011
171 ------------------------
173 1. Fixed some typos in the markup of the man pages, and wrote a script that
174 checks for such things as part of the documentation building process.
176 2. On a big-endian 64-bit system, pcregrep did not correctly process the
177 --match-limit and --recursion-limit options (added for 8.11). In
178 particular, this made one of the standard tests fail. (The integer value
179 went into the wrong half of a long int.)
181 3. If the --colour option was given to pcregrep with -v (invert match), it
182 did strange things, either producing crazy output, or crashing. It should,
183 of course, ignore a request for colour when reporting lines that do not
184 match.
186 4. Another pcregrep bug caused similar problems if --colour was specified with
187 -M (multiline) and the pattern match finished with a line ending.
189 5. In pcregrep, when a pattern that ended with a literal newline sequence was
190 matched in multiline mode, the following line was shown as part of the
191 match. This seems wrong, so I have changed it.
193 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
194 the check for further matches in the same line (so they could be coloured)
195 to overrun the end of the current line. If another match was found, it was
196 incorrectly shown (and then shown again when found in the next line).
198 7. If pcregrep was compiled under Windows, there was a reference to the
199 function pcregrep_exit() before it was defined. I am assuming this was
200 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
201 reported by a user. I've moved the definition above the reference.
204 Version 8.11 10-Dec-2010
205 ------------------------
207 1. (*THEN) was not working properly if there were untried alternatives prior
208 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
209 backtracked to try for "b" instead of moving to the next alternative branch
210 at the same level (in this case, to look for "c"). The Perl documentation
211 is clear that when (*THEN) is backtracked onto, it goes to the "next
212 alternative in the innermost enclosing group".
214 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
215 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
216 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
217 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
218 (*THEN).
220 3. If \s appeared in a character class, it removed the VT character from
221 the class, even if it had been included by some previous item, for example
222 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
223 of \s, but is part of the POSIX "space" class.)
225 4. A partial match never returns an empty string (because you can always
226 match an empty string at the end of the subject); however the checking for
227 an empty string was starting at the "start of match" point. This has been
228 changed to the "earliest inspected character" point, because the returned
229 data for a partial match starts at this character. This means that, for
230 example, /(?<=abc)def/ gives a partial match for the subject "abc"
231 (previously it gave "no match").
233 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
234 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
235 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
236 has an implication that the given string is incomplete (because a partial
237 match is preferred over a full match). For this reason, these items now
238 give a partial match in this situation. [Aside: previously, the one case
239 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
240 match rather than a full match, which was wrong by the old rules, but is
241 now correct.]
243 6. There was a bug in the handling of #-introduced comments, recognized when
244 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
245 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
246 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
247 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
248 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
249 places in pcre_compile().
251 7. Related to (6) above, when pcre_compile() was skipping #-introduced
252 comments when looking ahead for named forward references to subpatterns,
253 the only newline sequence it recognized was NL. It now handles newlines
254 according to the set newline convention.
256 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
257 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
258 cater for a lack of strerror(). These oversights have been fixed.
260 9. Added --match-limit and --recursion-limit to pcregrep.
262 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
264 11. When the -o option was used, pcregrep was setting a return code of 1, even
265 when matches were found, and --line-buffered was not being honoured.
267 12. Added an optional parentheses number to the -o and --only-matching options
268 of pcregrep.
270 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
271 can match an empty string. The code to do it in pcretest and pcredemo
272 needed fixing:
274 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
275 only one byte after an empty string match just before CRLF (this case
276 just got forgotten; "any" and "anycrlf" were OK).
278 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
279 mode when an empty string match preceded an ASCII character followed by
280 a non-ASCII character. (The code for advancing by one character rather
281 than one byte was nonsense.)
283 (c) The pcredemo.c sample program did not have any code at all to handle
284 the cases when CRLF is a valid newline sequence.
286 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
287 as a starting offset was within the subject string. There is now a new
288 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
289 negative or greater than the length of the string. In order to test this,
290 pcretest is extended to allow the setting of negative starting offsets.
292 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
293 starting offset points to the beginning of a UTF-8 character was
294 unnecessarily clumsy. I tidied it up.
296 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
297 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
299 17. Nobody had reported that the --include_dir option, which was added in
300 release 7.7 should have been called --include-dir (hyphen, not underscore)
301 for compatibility with GNU grep. I have changed it to --include-dir, but
302 left --include_dir as an undocumented synonym, and the same for
303 --exclude-dir, though that is not available in GNU grep, at least as of
304 release 2.5.4.
306 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
307 characters from a string of bytes) have been redefined so as not to use
308 loops, in order to improve performance in some environments. At the same
309 time, I abstracted some of the common code into auxiliary macros to save
310 repetition (this should not affect the compiled code).
312 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
313 compile-time error is now given if \c is not followed by an ASCII
314 character, that is, a byte less than 128. (In EBCDIC mode, the code is
315 different, and any byte value is allowed.)
317 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
318 START_OPTIMIZE option, which is now allowed at compile time - but just
319 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
320 to pcregrep and other applications that have no direct access to PCRE
321 options. The new /Y option in pcretest sets this option when calling
322 pcre_compile().
324 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
325 back references. Groups containing recursive back references were forced to
326 be atomic by that change, but in the case of named groups, the amount of
327 memory required was incorrectly computed, leading to "Failed: internal
328 error: code overflow". This has been fixed.
330 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
331 pcretest.c, to avoid build problems in some Borland environments.
334 Version 8.10 25-Jun-2010
335 ------------------------
337 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
338 THEN.
340 2. (*ACCEPT) was not working when inside an atomic group.
342 3. Inside a character class, \B is treated as a literal by default, but
343 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
344 causes the error). The code is unchanged, but I tidied the documentation.
346 4. Inside a character class, PCRE always treated \R and \X as literals,
347 whereas Perl faults them if its -w option is set. I have changed PCRE so
348 that it faults them when PCRE_EXTRA is set.
350 5. Added support for \N, which always matches any character other than
351 newline. (It is the same as "." when PCRE_DOTALL is not set.)
353 6. When compiling pcregrep with newer versions of gcc which may have
354 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
355 declared with attribute warn_unused_result" were given. Just casting the
356 result to (void) does not stop the warnings; a more elaborate fudge is
357 needed. I've used a macro to implement this.
359 7. Minor change to pcretest.c to avoid a compiler warning.
361 8. Added four artifical Unicode properties to help with an option to make
362 \s etc use properties (see next item). The new properties are: Xan
363 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
365 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
366 use Unicode properties. (*UCP) at the start of a pattern can be used to set
367 this option. Modified pcretest to add /W to test this facility. Added
368 REG_UCP to make it available via the POSIX interface.
370 10. Added --line-buffered to pcregrep.
372 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
373 studied, and the match started with a letter with a code point greater than
374 127 whose first byte was different to the first byte of the other case of
375 the letter, the other case of this starting letter was not recognized
376 (#976).
378 12. If a pattern that was studied started with a repeated Unicode property
379 test, for example, \p{Nd}+, there was the theoretical possibility of
380 setting up an incorrect bitmap of starting bytes, but fortunately it could
381 not have actually happened in practice until change 8 above was made (it
382 added property types that matched character-matching opcodes).
384 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
385 possible starting bytes for non-anchored patterns.
387 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
388 \R, and also a number of cases that involve Unicode properties, both
389 explicit and implicit when PCRE_UCP is set.
391 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
392 input, it could crash or give wrong results if characters with values
393 greater than 0xc0 were present in the subject string. (Detail: it assumed
394 UTF-8 input when processing these items.)
396 16. Added a lot of (int) casts to avoid compiler warnings in systems where
397 size_t is 64-bit (#991).
399 17. Added a check for running out of memory when PCRE is compiled with
400 --disable-stack-for-recursion (#990).
402 18. If the last data line in a file for pcretest does not have a newline on
403 the end, a newline was missing in the output.
405 19. The default pcre_chartables.c file recognizes only ASCII characters (values
406 less than 128) in its various bitmaps. However, there is a facility for
407 generating tables according to the current locale when PCRE is compiled. It
408 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
409 characters, are recognized by isspace() and therefore were getting set in
410 these tables, and indeed these tables seem to approximate to ISO 8859. This
411 caused a problem in UTF-8 mode when pcre_study() was used to create a list
412 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
413 which of course cannot start UTF-8 characters. I have changed the code so
414 that only real ASCII characters (less than 128) and the correct starting
415 bytes for UTF-8 encodings are set for characters greater than 127 when in
416 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
417 altogether.)
419 20. Added the /T option to pcretest so as to be able to run tests with non-
420 standard character tables, thus making it possible to include the tests
421 used for 19 above in the standard set of tests.
423 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
424 reference to a subpattern the other side of a comment that contains an
425 opening parenthesis caused either an internal compiling error, or a
426 reference to the wrong subpattern.
429 Version 8.02 19-Mar-2010
430 ------------------------
432 1. The Unicode data tables have been updated to Unicode 5.2.0.
434 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
435 configured.
437 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
438 original author of that file, following a query about its status.
440 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
441 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
443 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
444 quantifier applied to a forward-referencing subroutine call, could compile
445 incorrect code or give the error "internal error: previously-checked
446 referenced subpattern not found".
448 6. Both MS Visual Studio and Symbian OS have problems with initializing
449 variables to point to external functions. For these systems, therefore,
450 pcre_malloc etc. are now initialized to local functions that call the
451 relevant global functions.
453 7. There were two entries missing in the vectors called coptable and poptable
454 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
455 I've fixed the data, and added a kludgy way of testing at compile time that
456 the lengths are correct (equal to the number of opcodes).
458 8. Following on from 7, I added a similar kludge to check the length of the
459 eint vector in pcreposix.c.
461 9. Error texts for pcre_compile() are held as one long string to avoid too
462 much relocation at load time. To find a text, the string is searched,
463 counting zeros. There was no check for running off the end of the string,
464 which could happen if a new error number was added without updating the
465 string.
467 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
469 11. \K was not working if it appeared in an atomic group or in a group that
470 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
471 \K is "not well defined" if used in an assertion. PCRE now accepts it if
472 the assertion is positive, but not if it is negative.
474 12. Change 11 fortuitously reduced the size of the stack frame used in the
475 "match()" function of pcre_exec.c by one pointer. Forthcoming
476 implementation of support for (*MARK) will need an extra pointer on the
477 stack; I have reserved it now, so that the stack frame size does not
478 decrease.
480 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
481 item in branch that calls a recursion is a subroutine call - as in the
482 second branch in the above example - was incorrectly given the compile-
483 time error "recursive call could loop indefinitely" because pcre_compile()
484 was not correctly checking the subroutine for matching a non-empty string.
486 14. The checks for overrunning compiling workspace could trigger after an
487 overrun had occurred. This is a "should never occur" error, but it can be
488 triggered by pathological patterns such as hundreds of nested parentheses.
489 The checks now trigger 100 bytes before the end of the workspace.
491 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
494 Version 8.01 19-Jan-2010
495 ------------------------
497 1. If a pattern contained a conditional subpattern with only one branch (in
498 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
499 computed the wrong minimum data length (which is of course zero for such
500 subpatterns). This could cause incorrect "no match" results.
502 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
503 the pattern is reset in the first branch, pcre_compile() failed with
504 "internal error: code overflow at offset...". This happened only when
505 the reset was to the original external option setting. (An optimization
506 abstracts leading options settings into an external setting, which was the
507 cause of this.)
509 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
510 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
511 assertion pattern did not match (meaning that the assertion was true), it
512 was incorrectly treated as false if the SKIP had been reached during the
513 matching. This also applied to assertions used as conditions.
515 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
516 assertion subpattern, including such a pattern used as a condition,
517 unpredictable results occurred, instead of the error return
520 5. The C++ GlobalReplace function was not working like Perl for the special
521 situation when an empty string is matched. It now does the fancy magic
522 stuff that is necessary.
524 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
525 removed. (These were left over from very, very early versions of PCRE.)
527 7. Some cosmetic changes to the code to make life easier when compiling it
528 as part of something else:
530 (a) Change DEBUG to PCRE_DEBUG.
532 (b) In pcre_compile(), rename the member of the "branch_chain" structure
533 called "current" as "current_branch", to prevent a collision with the
534 Linux macro when compiled as a kernel module.
536 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
537 prevent a collision with the Linux macro when compiled as a kernel
538 module.
540 8. In pcre_compile() there are some checks for integer overflows that used to
541 cast potentially large values to (double). This has been changed to that
542 when building, a check for int64_t is made, and if it is found, it is used
543 instead, thus avoiding the use of floating point arithmetic. (There is no
544 other use of FP in PCRE.) If int64_t is not found, the fallback is to
545 double.
547 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
548 2005 (difference between two addresses compared to an unsigned value).
550 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
551 custom one, because of the following reported problem in Windows:
553 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
554 under Win32.
555 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
556 therefore missing the function definition.
557 - The compiler thus generates a "C" signature for the test function.
558 - The linker fails to find the "C" function.
559 - PCRE fails to configure if asked to do so against libbz2.
561 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
562 messages were output:
564 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
565 rerunning libtoolize, to keep the correct libtool macros in-tree.
566 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
568 I have done both of these things.
570 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
571 most of the time, it *can* run out if it is given a pattern that contains a
572 runaway infinite recursion. I updated the discussion in the pcrestack man
573 page.
575 13. Now that we have gone to the x.xx style of version numbers, the minor
576 version may start with zero. Using 08 or 09 is a bad idea because users
577 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
578 interpreted as invalid octal numbers. I've updated the previous comment in
579 configure.ac, and also added a check that gives an error if 08 or 09 are
580 used.
582 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
583 causing partial matching to fail when the end of the subject matched \W
584 in a UTF-8 pattern where \W was quantified with a minimum of 3.
586 15. There were some discrepancies between the declarations in pcre_internal.h
587 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
588 their definitions. The declarations used "const uschar *" and the
589 definitions used USPTR. Even though USPTR is normally defined as "const
590 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
591 reported that: "This difference in casting confuses some C++ compilers, for
592 example, SunCC recognizes above declarations as different functions and
593 generates broken code for hbpcre." I have changed the declarations to use
594 USPTR.
596 16. GNU libtool is named differently on some systems. The autogen.sh script now
597 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
598 (FreeBSD).
600 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
601 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
602 comment: "Figure out how to create a longlong from a string: strtoll and
603 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
604 instance, but it only takes 2 args instead of 3!"
606 18. A subtle bug concerned with back references has been fixed by a change of
607 specification, with a corresponding code fix. A pattern such as
608 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
609 refers, was giving matches when it shouldn't. For example, xa=xaaa would
610 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
611 same bug. Such groups have to be quantified to be useful, or contained
612 inside another quantified group. (If there's no repetition, the reference
613 can never match.) The problem arises because, having left the group and
614 moved on to the rest of the pattern, a later failure that backtracks into
615 the group uses the captured value from the final iteration of the group
616 rather than the correct earlier one. I have fixed this in PCRE by forcing
617 any group that contains a reference to itself to be an atomic group; that
618 is, there cannot be any backtracking into it once it has completed. This is
619 similar to recursive and subroutine calls.
622 Version 8.00 19-Oct-09
623 ----------------------
625 1. The table for translating pcre_compile() error codes into POSIX error codes
626 was out-of-date, and there was no check on the pcre_compile() error code
627 being within the table. This could lead to an OK return being given in
628 error.
630 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
631 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
632 in a Windows environment.
634 3. The pcregrep --count option prints the count for each file even when it is
635 zero, as does GNU grep. However, pcregrep was also printing all files when
636 --files-with-matches was added. Now, when both options are given, it prints
637 counts only for those files that have at least one match. (GNU grep just
638 prints the file name in this circumstance, but including the count seems
639 more useful - otherwise, why use --count?) Also ensured that the
640 combination -clh just lists non-zero counts, with no names.
642 4. The long form of the pcregrep -F option was incorrectly implemented as
643 --fixed_strings instead of --fixed-strings. This is an incompatible change,
644 but it seems right to fix it, and I didn't think it was worth preserving
645 the old behaviour.
647 5. The command line items --regex=pattern and --regexp=pattern were not
648 recognized by pcregrep, which required --regex pattern or --regexp pattern
649 (with a space rather than an '='). The man page documented the '=' forms,
650 which are compatible with GNU grep; these now work.
652 6. No libpcreposix.pc file was created for pkg-config; there was just
653 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
655 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
656 when UCP support is not needed, by modifying the Python script that
657 generates it from Unicode data files. This should not matter if the module
658 is correctly used as a library, but I received one complaint about 50K of
659 unwanted data. My guess is that the person linked everything into his
660 program rather than using a library. Anyway, it does no harm.
662 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
663 was a minimum greater than 1 for a wide character in a possessive
664 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
665 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
666 character. Chaos in the form of incorrect output or a compiling loop could
667 result.
669 9. The restrictions on what a pattern can contain when partial matching is
670 requested for pcre_exec() have been removed. All patterns can now be
671 partially matched by this function. In addition, if there are at least two
672 slots in the offset vector, the offset of the earliest inspected character
673 for the match and the offset of the end of the subject are set in them when
674 PCRE_ERROR_PARTIAL is returned.
676 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
677 synonymous with PCRE_PARTIAL, for backwards compatibility, and
678 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
679 and may be more useful for multi-segment matching.
681 11. Partial matching with pcre_exec() is now more intuitive. A partial match
682 used to be given if ever the end of the subject was reached; now it is
683 given only if matching could not proceed because another character was
684 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
685 string "Z", which now yields "no match" instead of "partial match". In the
686 case of pcre_dfa_exec(), "no match" is given if every matching path for the
687 final character ended with (*FAIL).
689 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
690 if the pattern had a "must contain" character that was already found in the
691 earlier partial match, unless partial matching was again requested. For
692 example, with the pattern /dog.(body)?/, the "must contain" character is
693 "g". If the first part-match was for the string "dog", restarting with
694 "sbody" failed. This bug has been fixed.
696 13. The string returned by pcre_dfa_exec() after a partial match has been
697 changed so that it starts at the first inspected character rather than the
698 first character of the match. This makes a difference only if the pattern
699 starts with a lookbehind assertion or \b or \B (\K is not supported by
700 pcre_dfa_exec()). It's an incompatible change, but it makes the two
701 matching functions compatible, and I think it's the right thing to do.
703 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
704 so that the demonstration program is easily available in environments where
705 PCRE has not been installed from source.
707 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
708 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
709 library.
711 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
712 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
713 is not the first non-POSIX option to be added. Clearly some people find
714 these options useful.
716 17. If a caller to the POSIX matching function regexec() passes a non-zero
717 value for nmatch with a NULL value for pmatch, the value of
718 nmatch is forced to zero.
720 18. RunGrepTest did not have a test for the availability of the -u option of
721 the diff command, as RunTest does. It now checks in the same way as
722 RunTest, and also checks for the -b option.
724 19. If an odd number of negated classes containing just a single character
725 interposed, within parentheses, between a forward reference to a named
726 subpattern and the definition of the subpattern, compilation crashed with
727 an internal error, complaining that it could not find the referenced
728 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
729 [The bug was that it was starting one character too far in when skipping
730 over the character class, thus treating the ] as data rather than
731 terminating the class. This meant it could skip too much.]
733 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
734 /g option in pcretest when the pattern contains \K, which makes it possible
735 to have an empty string match not at the start, even when the pattern is
736 anchored. Updated pcretest and pcredemo to use this option.
738 21. If the maximum number of capturing subpatterns in a recursion was greater
739 than the maximum at the outer level, the higher number was returned, but
740 with unset values at the outer level. The correct (outer level) value is
741 now given.
743 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
744 PCRE did not set those parentheses (unlike Perl). I have now found a way to
745 make it do so. The string so far is captured, making this feature
746 compatible with Perl.
748 23. The tests have been re-organized, adding tests 11 and 12, to make it
749 possible to check the Perl 5.10 features against Perl 5.10.
751 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
752 pattern matches a fixed length string. PCRE did not allow this; now it
753 does. Neither allows recursion.
755 25. I finally figured out how to implement a request to provide the minimum
756 length of subject string that was needed in order to match a given pattern.
757 (It was back references and recursion that I had previously got hung up
758 on.) This code has now been added to pcre_study(); it finds a lower bound
759 to the length of subject needed. It is not necessarily the greatest lower
760 bound, but using it to avoid searching strings that are too short does give
761 some useful speed-ups. The value is available to calling programs via
762 pcre_fullinfo().
764 26. While implementing 25, I discovered to my embarrassment that pcretest had
765 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
766 study optimizations had never been tested with that matching function.
767 Oops. What is worse, even when it was passed study data, there was a bug in
768 pcre_dfa_exec() that meant it never actually used it. Double oops. There
769 were also very few tests of studied patterns with pcre_dfa_exec().
771 27. If (?| is used to create subpatterns with duplicate numbers, they are now
772 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
773 on the other side of the coin, they are no longer allowed to have different
774 names, because these cannot be distinguished in PCRE, and this has caused
775 confusion. (This is a difference from Perl.)
777 28. When duplicate subpattern names are present (necessarily with different
778 numbers, as required by 27 above), and a test is made by name in a
779 conditional pattern, either for a subpattern having been matched, or for
780 recursion in such a pattern, all the associated numbered subpatterns are
781 tested, and the overall condition is true if the condition is true for any
782 one of them. This is the way Perl works, and is also more like the way
783 testing by number works.
786 Version 7.9 11-Apr-09
787 ---------------------
789 1. When building with support for bzlib/zlib (pcregrep) and/or readline
790 (pcretest), all targets were linked against these libraries. This included
791 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
792 libraries. This caused unwanted dependencies to be created. This problem
793 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
794 pcretest is linked with readline.
796 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
797 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
798 moved outside it again, because FALSE and TRUE are already defined in AIX,
799 but BOOL is not.
801 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
802 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
804 4. The pcregrep documentation said spaces were inserted as well as colons (or
805 hyphens) following file names and line numbers when outputting matching
806 lines. This is not true; no spaces are inserted. I have also clarified the
807 wording for the --colour (or --color) option.
809 5. In pcregrep, when --colour was used with -o, the list of matching strings
810 was not coloured; this is different to GNU grep, so I have changed it to be
811 the same.
813 6. When --colo(u)r was used in pcregrep, only the first matching substring in
814 each matching line was coloured. Now it goes on to look for further matches
815 of any of the test patterns, which is the same behaviour as GNU grep.
817 7. A pattern that could match an empty string could cause pcregrep to loop; it
818 doesn't make sense to accept an empty string match in pcregrep, so I have
819 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
820 seems to be how GNU grep behaves.
822 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
823 start or after a newline", because the conditional assertion was not being
824 correctly handled. The rule now is that both the assertion and what follows
825 in the first alternative must satisfy the test.
827 9. If auto-callout was enabled in a pattern with a conditional group whose
828 condition was an assertion, PCRE could crash during matching, both with
829 pcre_exec() and pcre_dfa_exec().
831 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
832 used for matching.
834 11. Unicode property support in character classes was not working for
835 characters (bytes) greater than 127 when not in UTF-8 mode.
837 12. Added the -M command line option to pcretest.
839 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
841 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
843 16. Added comments and documentation about mis-use of no_arg in the C++
844 wrapper.
846 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
847 from Martin Jerabek that uses macro names for all relevant character and
848 string constants.
850 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
851 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
852 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
853 these, but not everybody uses configure.
855 19. A conditional group that had only one branch was not being correctly
856 recognized as an item that could match an empty string. This meant that an
857 enclosing group might also not be so recognized, causing infinite looping
858 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
859 with the subject "ab", where knowledge that the repeated group can match
860 nothing is needed in order to break the loop.
862 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
863 exec(), but without supplying a callout function, matching went wrong.
865 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
866 leak if the size of the offset vector was greater than 30. When the vector
867 is smaller, the saved offsets during recursion go onto a local stack
868 vector, but for larger vectors malloc() is used. It was failing to free
869 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
870 error, in fact).
872 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
873 heapframe that is used only when UTF-8 support is enabled. This caused no
874 problem, but was untidy.
876 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
877 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
878 included within another project.
880 24. Steven Van Ingelgem's patches to add more options to the CMake support,
881 slightly modified by me:
883 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
884 not building pcregrep.
886 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
887 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
889 25. Forward references, both numeric and by name, in patterns that made use of
890 duplicate group numbers, could behave incorrectly or give incorrect errors,
891 because when scanning forward to find the reference group, PCRE was not
892 taking into account the duplicate group numbers. A pattern such as
893 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
895 26. Changed a few more instances of "const unsigned char *" to USPTR, making
896 the feature of a custom pointer more persuasive (as requested by a user).
898 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
899 pcretest.c, inside #ifndefs, because it seems they are sometimes already
900 pre-defined.
902 28. Added support for (*UTF8) at the start of a pattern.
904 29. Arrange for flags added by the "release type" setting in CMake to be shown
905 in the configuration summary.
908 Version 7.8 05-Sep-08
909 ---------------------
911 1. Replaced UCP searching code with optimized version as implemented for Ad
912 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
913 stage table and inline lookup instead of a function, giving speed ups of 2
914 to 5 times on some simple patterns that I tested. Permission was given to
915 distribute the MultiStage2.py script that generates the tables (it's not in
916 the tarball, but is in the Subversion repository).
918 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
919 scripts.
921 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
922 a group with a zero qualifier. The result of the study could be incorrect,
923 or the function might crash, depending on the pattern.
925 4. Caseless matching was not working for non-ASCII characters in back
926 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
927 It now works when Unicode Property Support is available.
929 5. In pcretest, an escape such as \x{de} in the data was always generating
930 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
931 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
932 truncation.
934 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
936 7. Added two (int) casts to pcregrep when printing the difference of two
937 pointers, in case they are 64-bit values.
939 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
940 test 2 if it fails.
942 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
943 and a #define of that name to empty if it is not externally set. This is to
944 allow users of MSVC to set it if necessary.
946 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
947 the convenience functions in the pcre_get.c source file.
949 11. An option change at the start of a pattern that had top-level alternatives
950 could cause overwriting and/or a crash. This command provoked a crash in
951 some environments:
953 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
955 This potential security problem was recorded as CVE-2008-2371.
957 12. For a pattern where the match had to start at the beginning or immediately
958 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
959 pcre_dfa_exec() could read past the end of the passed subject if there was
960 no match. To help with detecting such bugs (e.g. with valgrind), I modified
961 pcretest so that it places the subject at the end of its malloc-ed buffer.
963 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
964 exec() might read past the end of the data buffer in UTF-8 mode.
966 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
967 the data contained the byte 0x85 as part of a UTF-8 character within its
968 first line. This applied both to normal and DFA matching.
970 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
971 /^[^d]*?$/8 failed to match "abc".
973 16. Added a missing copyright notice to pcrecpp_internal.h.
975 17. Make it more clear in the documentation that values returned from
976 pcre_exec() in ovector are byte offsets, not character counts.
978 18. Tidied a few places to stop certain compilers from issuing warnings.
980 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
981 supplied by Stefan Weber. I made a further small update for 7.8 because
982 there is a change of source arrangements: the pcre_searchfuncs.c module is
983 replaced by pcre_ucd.c.
986 Version 7.7 07-May-08
987 ---------------------
989 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
990 a string to a long long, pretend we don't even have a long long." This is
991 done by checking for the strtoq, strtoll, and _strtoi64 functions.
993 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
994 pre-7.6 versions, which defined a global no_arg variable instead of putting
995 it in the RE class. (See also #8 below.)
997 3. Remove a line of dead code, identified by coverity and reported by Nuno
998 Lopes.
1000 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1002 (1) The include/exclude patterns were being applied to the whole pathnames
1003 of files, instead of just to the final components.
1005 (2) If there was more than one level of directory, the subdirectories were
1006 skipped unless they satisfied the include/exclude conditions. This is
1007 inconsistent with GNU grep (and could even be seen as contrary to the
1008 pcregrep specification - which I improved to make it absolutely clear).
1009 The action now is always to scan all levels of directory, and just
1010 apply the include/exclude patterns to regular files.
1012 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1013 --exclude_dir in the tests to avoid scanning .svn directories.
1015 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1016 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1017 doesn't support NULs in patterns.
1019 7. Added some missing "const"s to declarations of static tables in
1020 pcre_compile.c and pcre_dfa_exec.c.
1022 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1023 caused by fix #2 above. (Subsequently also a second patch to fix the
1024 first patch. And a third patch - this was a messy problem.)
1026 9. Applied Craig's patch to remove the use of push_back().
1028 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1029 matching function regexec().
1031 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1032 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1033 references. PCRE supports relative numbers with this syntax (I don't think
1034 Oniguruma does).
1036 12. Previously, a group with a zero repeat such as (...){0} was completely
1037 omitted from the compiled regex. However, this means that if the group
1038 was called as a subroutine from elsewhere in the pattern, things went wrong
1039 (an internal error was given). Such groups are now left in the compiled
1040 pattern, with a new opcode that causes them to be skipped at execution
1041 time.
1043 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1044 to the way PCRE behaves:
1046 (a) A lone ] character is dis-allowed (Perl treats it as data).
1048 (b) A back reference to an unmatched subpattern matches an empty string
1049 (Perl fails the current match path).
1051 (c) A data ] in a character class must be notated as \] because if the
1052 first data character in a class is ], it defines an empty class. (In
1053 Perl it is not possible to have an empty class.) The empty class []
1054 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1055 The negative empty class [^] matches any one character, independently
1056 of the DOTALL setting.
1058 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1059 non-existent subpattern following a character class starting with ']' and
1060 containing () gave an internal compiling error instead of "reference to
1061 non-existent subpattern". Fortunately, when the pattern did exist, the
1062 compiled code was correct. (When scanning forwards to check for the
1063 existencd of the subpattern, it was treating the data ']' as terminating
1064 the class, so got the count wrong. When actually compiling, the reference
1065 was subsequently set up correctly.)
1067 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1068 it was being rejected as not supported by pcre_dfa_exec(), even though
1069 other assertions are supported. I have made pcre_dfa_exec() support
1070 (*FAIL).
1072 16. The implementation of 13c above involved the invention of a new opcode,
1073 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1074 cannot be changed at match time, I realized I could make a small
1075 improvement to matching performance by compiling OP_ALLANY instead of
1076 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1077 on the OP_ANY path.
1079 17. Compiling pcretest on Windows with readline support failed without the
1080 following two fixes: (1) Make the unistd.h include conditional on
1081 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1083 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1084 ncurses library to be included for pcretest when ReadLine support is
1085 requested, but also to allow for it to be overridden. This patch came from
1086 Daniel Bergström.
1088 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1089 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1090 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1091 spotting this.
1094 Version 7.6 28-Jan-08
1095 ---------------------
1097 1. A character class containing a very large number of characters with
1098 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1099 overflow.
1101 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1102 HAVE_LONG_LONG is not defined.
1104 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1105 bring them up to date and include new features. This patch includes:
1107 - Fixed PH's badly added libz and libbz2 support.
1108 - Fixed a problem with static linking.
1109 - Added pcredemo. [But later removed - see 7 below.]
1110 - Fixed dftables problem and added an option.
1111 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1113 - Added readline support for pcretest.
1114 - Added an listing of the option settings after cmake has run.
1116 4. A user submitted a patch to Makefile that makes it easy to create
1117 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1118 Makefile.am that cause it to include this special target, without
1119 affecting anything else. Note that the same mingw target plus all
1120 the other distribution libraries and programs are now supported
1121 when configuring with CMake (see 6 below) instead of with
1122 Configure/Make.
1124 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1125 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1126 exported in the Windows port". It has not yet been confirmed that the patch
1127 solves the problem, but it does no harm.
1129 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1130 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1131 with CMake, and also correct the comment about stack recursion.
1133 7. Remove the automatic building of pcredemo from the ./configure system and
1134 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1135 of a program that users should build themselves after PCRE is installed, so
1136 building it automatically is not really right. What is more, it gave
1137 trouble in some build environments.
1139 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1142 Version 7.5 10-Jan-08
1143 ---------------------
1145 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1146 values in parens when parsing an RE using the C++ wrapper."
1148 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1149 Characters greater than 255 were excluded from the class instead of being
1150 included.
1152 3. The same bug as (2) above applied to negated POSIX classes such as
1153 [:^space:].
1155 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1156 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1157 I have changed it.
1159 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1160 first named subpattern) and a construct such as (?&a) would reference the
1161 first named subpattern whose name started with "a" (in other words, the
1162 length check was missing). Both these problems are fixed. "Subpattern name
1163 expected" is now given for (?&) (a zero-length name), and this patch also
1164 makes it give the same error for \k'' (previously it complained that that
1165 was a reference to a non-existent subpattern).
1167 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1168 this is right because (?- can be followed by option settings as well as by
1169 digits. I have, however, made the messages clearer.
1171 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1172 than the number used in the conditional) now cause a compile-time error.
1173 This is actually not compatible with Perl, which accepts such patterns, but
1174 treats the conditional as always being FALSE (as PCRE used to), but it
1175 seems to me that giving a diagnostic is better.
1177 8. Change "alphameric" to the more common word "alphanumeric" in comments
1178 and messages.
1180 9. Fix two occurrences of "backslash" in comments that should have been
1181 "backspace".
1183 10. Remove two redundant lines of code that can never be obeyed (their function
1184 was moved elsewhere).
1186 11. The program that makes PCRE's Unicode character property table had a bug
1187 which caused it to generate incorrect table entries for sequences of
1188 characters that have the same character type, but are in different scripts.
1189 It amalgamated them into a single range, with the script of the first of
1190 them. In other words, some characters were in the wrong script. There were
1191 thirteen such cases, affecting characters in the following ranges:
1193 U+002b0 - U+002c1
1194 U+0060c - U+0060d
1195 U+0061e - U+00612
1196 U+0064b - U+0065e
1197 U+0074d - U+0076d
1198 U+01800 - U+01805
1199 U+01d00 - U+01d77
1200 U+01d9b - U+01dbf
1201 U+0200b - U+0200f
1202 U+030fc - U+030fe
1203 U+03260 - U+0327f
1204 U+0fb46 - U+0fbb1
1205 U+10450 - U+1049d
1207 12. The -o option (show only the matching part of a line) for pcregrep was not
1208 compatible with GNU grep in that, if there was more than one match in a
1209 line, it showed only the first of them. It now behaves in the same way as
1210 GNU grep.
1212 13. If the -o and -v options were combined for pcregrep, it printed a blank
1213 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1214 does the same. The return code can be used to tell if there were any
1215 non-matching lines.
1217 14. Added --file-offsets and --line-offsets to pcregrep.
1219 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1220 infinitely looping recursion. The bug was that positive lookaheads were not
1221 being skipped when checking for a possible empty match (negative lookaheads
1222 and both kinds of lookbehind were skipped).
1224 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1225 inclusion of <windows.h> to before rather than after the definition of
1226 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1228 17. Specifying a possessive quantifier with a specific limit for a Unicode
1229 character property caused pcre_compile() to compile bad code, which led at
1230 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1231 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1232 caused the error; without that there was no problem.
1234 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1236 19. Added --enable-pcretest-libreadline.
1238 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1239 RE::GlobalReplace(). As a result, the number of replacements returned was
1240 double what it should be. I removed one of the increments, but Craig sent a
1241 later patch that removed the other one (the right fix) and added unit tests
1242 that check the return values (which was not done before).
1244 21. Several CMake things:
1246 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1247 the names libpcre and libpcreposix, not just pcre and pcreposix.
1249 (2) The above change means that pcretest and pcregrep are now correctly
1250 linked with the newly-built libraries, not previously installed ones.
1254 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1255 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1256 UTF-8 newline character). The key issue is that the pattern starts .*;
1257 this means that the match must be either at the beginning, or after a
1258 newline. The bug was in the code for advancing after a failed match and
1259 checking that the new position followed a newline. It was not taking
1260 account of UTF-8 characters correctly.
1262 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1263 character classes. PCRE was not treating the sequence [:...:] as a
1264 character class unless the ... were all letters. Perl, however, seems to
1265 allow any characters between [: and :], though of course it rejects as
1266 unknown any "names" that contain non-letters, because all the known class
1267 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1268 for example, whereas PCRE did not - it did not recognize a POSIX character
1269 class. This seemed a bit dangerous, so the code has been changed to be
1270 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1271 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1272 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1273 Perl does, and where it didn't before.
1275 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1276 Windows environments %n is disabled by default.
1279 Version 7.4 21-Sep-07
1280 ---------------------
1282 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1283 means that a class such as [\s] counted as "explicit reference to CR or
1284 LF". That isn't really right - the whole point of the change was to try to
1285 help when there was an actual mention of one of the two characters. So now
1286 the change happens only if \r or \n (or a literal CR or LF) character is
1287 encountered.
1289 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1290 of both had grown to the point where there were only 3 bits left.
1291 Fortunately, there was spare space in the data structure, and so I have
1292 moved the internal flags into a new 16-bit field to free up more option
1293 bits.
1295 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1296 but did not set the internal JCHANGED flag - either of these is enough to
1297 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1298 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1299 start sets both bits.
1301 4. Added options (at build time, compile time, exec time) to change \R from
1302 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1304 5. doc/pcresyntax.html was missing from the distribution.
1306 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1307 compatibility, even though it is no longer used.
1309 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1310 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1311 windows.h file is present (where different names are used). [This was
1312 reversed later after testing - see 16 below.]
1314 8. Changed all #include <config.h> to #include "config.h". There were also
1315 some further <pcre.h> cases that I changed to "pcre.h".
1317 9. When pcregrep was used with the --colour option, it missed the line ending
1318 sequence off the lines that it output.
1320 10. It was pointed out to me that arrays of string pointers cause lots of
1321 relocations when a shared library is dynamically loaded. A technique of
1322 using a single long string with a table of offsets can drastically reduce
1323 these. I have refactored PCRE in four places to do this. The result is
1324 dramatic:
1326 Originally: 290
1327 After changing UCP table: 187
1328 After changing error message table: 43
1329 After changing table of "verbs" 36
1330 After changing table of Posix names 22
1332 Thanks to the folks working on Gregex for glib for this insight.
1334 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1335 unicode-properties was also set.
1337 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1339 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1340 checked only for CRLF.
1342 14. Added casts to pcretest.c to avoid compiler warnings.
1344 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1346 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1347 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1348 entirely. This removes changes made in 7 above.
1350 17. The CMake files have been updated, and there is now more information about
1351 building with CMake in the NON-UNIX-USE document.
1354 Version 7.3 28-Aug-07
1355 ---------------------
1357 1. In the rejigging of the build system that eventually resulted in 7.1, the
1358 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1359 brackets there is not right, since it causes compilers to look for an
1360 installed pcre.h, not the version that is in the source that is being
1361 compiled (which of course may be different). I have changed it back to:
1363 #include "pcre.h"
1365 I have a vague recollection that the change was concerned with compiling in
1366 different directories, but in the new build system, that is taken care of
1367 by the VPATH setting the Makefile.
1369 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1370 when the subject happened to end in the byte 0x85 (e.g. if the last
1371 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1372 characters but of course it shouldn't be taken as a newline when it is part
1373 of another character. The bug was that, for an unlimited repeat of . in
1374 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1375 characters when looking for a newline.
1377 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1379 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1380 in debug output.
1382 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1383 long printing in the pcrecpp unittest when running under MinGW.
1385 6. ESC_K was left out of the EBCDIC table.
1387 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1388 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1389 limit also applies to "virtual nesting" when a pattern is recursive, and in
1390 this case 1000 isn't so big. I have been able to remove this limit at the
1391 expense of backing off one optimization in certain circumstances. Normally,
1392 when pcre_exec() would call its internal match() function recursively and
1393 immediately return the result unconditionally, it uses a "tail recursion"
1394 feature to save stack. However, when a subpattern that can match an empty
1395 string has an unlimited repetition quantifier, it no longer makes this
1396 optimization. That gives it a stack frame in which to save the data for
1397 checking that an empty string has been matched. Previously this was taken
1398 from the 1000-entry workspace that had been reserved. So now there is no
1399 explicit limit, but more stack is used.
1401 8. Applied Daniel's patches to solve problems with the import/export magic
1402 syntax that is required for Windows, and which was going wrong for the
1403 pcreposix and pcrecpp parts of the library. These were overlooked when this
1404 problem was solved for the main library.
1406 9. There were some crude static tests to avoid integer overflow when computing
1407 the size of patterns that contain repeated groups with explicit upper
1408 limits. As the maximum quantifier is 65535, the maximum group length was
1409 set at 30,000 so that the product of these two numbers did not overflow a
1410 32-bit integer. However, it turns out that people want to use groups that
1411 are longer than 30,000 bytes (though not repeat them that many times).
1412 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1413 made it possible to implement the integer overflow checks in a much more
1414 dynamic way, which I have now done. The artificial limitation on group
1415 length has been removed - we now have only the limit on the total length of
1416 the compiled pattern, which depends on the LINK_SIZE setting.
1418 10. Fixed a bug in the documentation for get/copy named substring when
1419 duplicate names are permitted. If none of the named substrings are set, the
1420 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1421 empty string.
1423 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1424 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1425 because the ] is interpreted as the first data character and the
1426 terminating ] is not found. PCRE has been made compatible with Perl in this
1427 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1428 cause memory overwriting.
1430 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1431 string has been matched (to stop an infinite loop). It was not recognizing
1432 a conditional subpattern that could match an empty string if that
1433 subpattern was within another subpattern. For example, it looped when
1434 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1435 condition was not nested. This bug has been fixed.
1437 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1438 past the start of the subject in the presence of bytes with the top bit
1439 set, for example "\x8aBCD".
1441 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1442 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1444 14. Optimized (?!) to (*FAIL).
1446 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1447 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1448 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1449 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1450 does: it's just the validity check that is more restrictive.
1452 16. Inserted checks for integer overflows during escape sequence (backslash)
1453 processing, and also fixed erroneous offset values for syntax errors during
1454 backslash processing.
1456 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1457 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1459 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1460 caused an overrun.
1462 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1463 something other than just ASCII characters) inside a group that had an
1464 unlimited repeat caused a loop at compile time (while checking to see
1465 whether the group could match an empty string).
1467 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1468 [\P{Any}] did so. (Error in the code for printing property names.)
1470 21. An orphan \E inside a character class could cause a crash.
1472 22. A repeated capturing bracket such as (A)? could cause a wild memory
1473 reference during compilation.
1475 23. There are several functions in pcre_compile() that scan along a compiled
1476 expression for various reasons (e.g. to see if it's fixed length for look
1477 behind). There were bugs in these functions when a repeated \p or \P was
1478 present in the pattern. These operators have additional parameters compared
1479 with \d, etc, and these were not being taken into account when moving along
1480 the compiled data. Specifically:
1482 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1483 length.
1485 (b) An item such as \pL+ within a repeated group could cause crashes or
1486 loops.
1488 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1489 "reference to non-existent subpattern" error.
1491 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1493 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1494 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1496 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1498 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1499 character were causing crashes (broken optimization).
1501 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1502 \p or \P) caused a compile-time loop.
1504 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1505 break. For example, the unstudied pattern [\r\n]A does not match the string
1506 "\r\nA" because change 7.0/46 below moves the current point on by two
1507 characters after failing to match at the start. However, the pattern \nA
1508 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1509 the same is true. There doesn't seem any very clean way out of this, but
1510 what I have chosen to do makes the common cases work: PCRE now takes note
1511 of whether there can be an explicit match for \r or \n anywhere in the
1512 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1513 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1514 pattern has explicit CR or LF references.
1516 29. Added (*CR) etc for changing newline setting at start of pattern.
1519 Version 7.2 19-Jun-07
1520 ---------------------
1522 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1523 which is apparently normally available under Windows.
1525 2. Re-jig the pcregrep tests with different newline settings in an attempt
1526 to make them independent of the local environment's newline setting.
1528 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1530 4. Some of the "internals" tests were previously cut out when the link size
1531 was not 2, because the output contained actual offsets. The recent new
1532 "Z" feature of pcretest means that these can be cut out, making the tests
1533 usable with all link sizes.
1535 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1536 stack recursion. This gives a massive performance boost under BSD, but just
1537 a small improvement under Linux. However, it saves one field in the frame
1538 in all cases.
1540 6. Added more features from the forthcoming Perl 5.10:
1542 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1543 recursion call. It refers to the nth most recently opened parentheses.
1545 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1546 to be opened parentheses.
1548 (c) Conditions that refer to capturing parentheses can be specified
1549 relatively, for example, (?(-2)... or (?(+3)...
1551 (d) \K resets the start of the current match so that everything before
1552 is not part of it.
1554 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1556 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1557 reference syntax.
1559 (g) (?| introduces a group in which the numbering of parentheses in each
1560 alternative starts with the same number.
1562 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1564 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1567 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1568 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1569 for detecting groups that can match an empty string.
1571 9. A pattern with a very large number of alternatives (more than several
1572 hundred) was running out of internal workspace during the pre-compile
1573 phase, where pcre_compile() figures out how much memory will be needed. A
1574 bit of new cunning has reduced the workspace needed for groups with
1575 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1576 workspace instead of running out of the 4096 that are available.
1578 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1580 11. Applied patch from Google to remove an optimization that didn't quite work.
1581 The report of the bug said:
1583 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1584 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1585 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1587 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1588 it matched the wrong number of bytes.
1591 Version 7.1 24-Apr-07
1592 ---------------------
1594 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1595 that is more "standard", making use of automake and other Autotools. There
1596 is some re-arrangement of the files and adjustment of comments consequent
1597 on this.
1599 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1600 for recursive directory scanning broke on some systems because the files
1601 are not scanned in any specific order and on different systems the order
1602 was different. A call to "sort" has been inserted into RunGrepTest for the
1603 approprate test as a short-term fix. In the longer term there may be an
1604 alternative.
1606 3. I had an email from Eric Raymond about problems translating some of PCRE's
1607 man pages to HTML (despite the fact that I distribute HTML pages, some
1608 people do their own conversions for various reasons). The problems
1609 concerned the use of low-level troff macros .br and .in. I have therefore
1610 removed all such uses from the man pages (some were redundant, some could
1611 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1612 HTML has been updated to handle .nf/.fi and to complain if it encounters
1613 .br or .in.
1615 4. Updated comments in configure.ac that get placed in config.h.in and also
1616 arranged for config.h to be included in the distribution, with the name
1617 config.h.generic, for the benefit of those who have to compile without
1618 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1620 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1621 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1622 makevp.bat for the current PCRE, using the additional files
1623 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1625 6. A Windows user reported a minor discrepancy with test 2, which turned out
1626 to be caused by a trailing space on an input line that had got lost in his
1627 copy. The trailing space was an accident, so I've just removed it.
1629 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1630 that is needed.
1632 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1633 as "const" (a) because they are and (b) because it helps the PHP
1634 maintainers who have recently made a script to detect big data structures
1635 in the php code that should be moved to the .rodata section. I remembered
1636 to update Builducptable as well, so it won't revert if ucptable.h is ever
1637 re-created.
1639 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1640 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1641 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1642 support is not required. This saves 1.5-2K of code, which is important in
1643 some applications.
1645 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1646 so as not to refer to the tables, even though these functions will never be
1647 called when UTF-8 support is disabled. Otherwise there are problems with a
1648 shared library.
1650 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1652 (a) It was defining its arguments as char * instead of void *.
1654 (b) It was assuming that all moves were upwards in memory; this was true
1655 a long time ago when I wrote it, but is no longer the case.
1657 The emulated memove() is provided for those environments that have neither
1658 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1659 is clearly not the case, as these two bugs were recently reported.
1661 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1662 and Detrail to create the HTML documentation, the .txt form of the man
1663 pages, and it removes trailing spaces from listed files. It also creates
1664 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1665 case, it wraps all the #defines with #ifndefs. This script should be run
1666 before "make dist".
1668 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1669 with Unicode property support.
1671 (a) For a maximizing quantifier, if the two different cases of the
1672 character were of different lengths in their UTF-8 codings (there are
1673 some cases like this - I found 11), and the matching function had to
1674 back up over a mixture of the two cases, it incorrectly assumed they
1675 were both the same length.
1677 (b) When PCRE was configured to use the heap rather than the stack for
1678 recursion during matching, it was not correctly preserving the data for
1679 the other case of a UTF-8 character when checking ahead for a match
1680 while processing a minimizing repeat. If the check also involved
1681 matching a wide character, but failed, corruption could cause an
1682 erroneous result when trying to check for a repeat of the original
1683 character.
1685 13. Some tidying changes to the testing mechanism:
1687 (a) The RunTest script now detects the internal link size and whether there
1688 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1689 values substituted by "configure". (The RunGrepTest script already did
1690 this for UTF-8.) The configure.ac script no longer substitutes the
1691 relevant variables.
1693 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1694 with length and offset values. This means that the output is different
1695 for different internal link sizes. Test 2 is skipped for link sizes
1696 other than 2 because of this, bypassing the problem. Unfortunately,
1697 there was also a test in test 3 (the locale tests) that used /B and
1698 failed for link sizes other than 2. Rather than cut the whole test out,
1699 I have added a new /Z option to pcretest that replaces the length and
1700 offset values with spaces. This is now used to make test 3 independent
1701 of link size. (Test 2 will be tidied up later.)
1703 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1704 segmentation fault instead of returning the appropriate error message.
1706 15. In multiline mode when the newline sequence was set to "any", the pattern
1707 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1708 This doesn't seem right; it now treats the CRLF combination as the line
1709 ending, and so does not match in that case. It's only a pattern such as ^$
1710 that would hit this one: something like ^ABC$ would have failed after \r
1711 and then tried again after \r\n.
1713 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1714 in an attempt to make files that differ only in their line terminators
1715 compare equal. This works on Linux.
1717 17. Under certain error circumstances pcregrep might try to free random memory
1718 as it exited. This is now fixed, thanks to valgrind.
1720 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1721 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1722 was because its rules for how to advance for /g after matching an empty
1723 string at the end of a line did not allow for this case. They now check for
1724 it specially.
1726 20. pcretest is supposed to handle patterns and data of any length, by
1727 extending its buffers when necessary. It was getting this wrong when the
1728 buffer for a data line had to be extended.
1730 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1731 CRLF as a newline sequence.
1733 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1734 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1735 I have nevertheless tidied it up.
1737 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1739 24. Added a man page for pcre-config.
1742 Version 7.0 19-Dec-06
1743 ---------------------
1745 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1746 moving to gcc 4.1.1.
1748 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1749 sys/time.h, which is documented as needed for this function. It doesn't
1750 seem to matter on Linux, but it showed up on some releases of OS X.
1752 3. It seems that there are systems where bytes whose values are greater than
1753 127 match isprint() in the "C" locale. The "C" locale should be the
1754 default when a C program starts up. In most systems, only ASCII printing
1755 characters match isprint(). This difference caused the output from pcretest
1756 to vary, making some of the tests fail. I have changed pcretest so that:
1758 (a) When it is outputting text in the compiled version of a pattern, bytes
1759 other than 32-126 are always shown as hex escapes.
1761 (b) When it is outputting text that is a matched part of a subject string,
1762 it does the same, unless a different locale has been set for the match
1763 (using the /L modifier). In this case, it uses isprint() to decide.
1765 4. Fixed a major bug that caused incorrect computation of the amount of memory
1766 required for a compiled pattern when options that changed within the
1767 pattern affected the logic of the preliminary scan that determines the
1768 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1769 that the computed length was too small. The symptoms of this bug were
1770 either the PCRE error "internal error: code overflow" from pcre_compile(),
1771 or a glibc crash with a message such as "pcretest: free(): invalid next
1772 size (fast)". Examples of patterns that provoked this bug (shown in
1773 pcretest format) are:
1775 /(?-x: )/x
1776 /(?x)(?-x: \s*#\s*)/
1777 /((?i)[\x{c0}])/8
1778 /(?i:[\x{c0}])/8
1780 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1781 is now done differently.
1783 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1784 wrapper classes; (b) implement a new function in the C++ scanner that is
1785 more efficient than the old way of doing things because it avoids levels of
1786 recursion in the regex matching; (c) add a paragraph to the documentation
1787 for the FullMatch() function.
1789 6. The escape sequence \n was being treated as whatever was defined as
1790 "newline". Not only was this contrary to the documentation, which states
1791 that \n is character 10 (hex 0A), but it also went horribly wrong when
1792 "newline" was defined as CRLF. This has been fixed.
1794 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1795 was being set to -1 for the "end of line" case (supposedly a value that no
1796 character can have). Though this value is never used (the check for end of
1797 line is "zero bytes in current character"), it caused compiler complaints.
1798 I've changed it to 0xffffffff.
1800 8. In pcre_version.c, the version string was being built by a sequence of
1801 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1802 string (as it is for production releases) called a macro with an empty
1803 argument. The C standard says the result of this is undefined. The gcc
1804 compiler treats it as an empty string (which was what was wanted) but it is
1805 reported that Visual C gives an error. The source has been hacked around to
1806 avoid this problem.
1808 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1809 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1810 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1811 of them did).
1813 10. Originally, pcretest opened its input and output without "b"; then I was
1814 told that "b" was needed in some environments, so it was added for release
1815 5.0 to both the input and output. (It makes no difference on Unix-like
1816 systems.) Later I was told that it is wrong for the input on Windows. I've
1817 now abstracted the modes into two macros, to make it easier to fiddle with
1818 them, and removed "b" from the input mode under Windows.
1820 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1822 12. Added -help and --help to pcretest as an official way of being reminded
1823 of the options.
1825 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1826 and pcrecpp.cc because they annoy compilers at high warning levels.
1828 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1830 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1831 scripts are not C programs :-) and which was not noticed because it works
1832 on Linux.
1834 16. pcretest is supposed to handle any length of pattern and data line (as one
1835 line or as a continued sequence of lines) by extending its input buffer if
1836 necessary. This feature was broken for very long pattern lines, leading to
1837 a string of junk being passed to pcre_compile() if the pattern was longer
1838 than about 50K.
1840 17. I have done a major re-factoring of the way pcre_compile() computes the
1841 amount of memory needed for a compiled pattern. Previously, there was code
1842 that made a preliminary scan of the pattern in order to do this. That was
1843 OK when PCRE was new, but as the facilities have expanded, it has become
1844 harder and harder to keep it in step with the real compile phase, and there
1845 have been a number of bugs (see for example, 4 above). I have now found a
1846 cunning way of running the real compile function in a "fake" mode that
1847 enables it to compute how much memory it would need, while actually only
1848 ever using a few hundred bytes of working memory and without too many
1849 tests of the mode. This should make future maintenance and development
1850 easier. A side effect of this work is that the limit of 200 on the nesting
1851 depth of parentheses has been removed (though this was never a serious
1852 limitation, I suspect). However, there is a downside: pcre_compile() now
1853 runs more slowly than before (30% or more, depending on the pattern). I
1854 hope this isn't a big issue. There is no effect on runtime performance.
1856 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1857 newline (only possible for the last line of a file) and it was a
1858 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1860 19. Added additional timing features to pcretest. (1) The -tm option now times
1861 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1862 separate command line item, by a number that specifies the number of
1863 repeats to use when timing. The default is 50000; this gives better
1864 precision, but takes uncomfortably long for very large patterns.
1866 20. Extended pcre_study() to be more clever in cases where a branch of a
1867 subpattern has no definite first character. For example, (a*|b*)[cd] would
1868 previously give no result from pcre_study(). Now it recognizes that the
1869 first character must be a, b, c, or d.
1871 21. There was an incorrect error "recursive call could loop indefinitely" if
1872 a subpattern (or the entire pattern) that was being tested for matching an
1873 empty string contained only one non-empty item after a nested subpattern.
1874 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1875 incorrectly, because the \d was being skipped in the check.
1877 22. The pcretest program now has a new pattern option /B and a command line
1878 option -b, which is equivalent to adding /B to every pattern. This causes
1879 it to show the compiled bytecode, without the additional information that
1880 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1881 is the same as /B/I).
1883 23. A new optimization is now able automatically to treat some sequences such
1884 as a*b as a*+b. More specifically, if something simple (such as a character
1885 or a simple class like \d) has an unlimited quantifier, and is followed by
1886 something that cannot possibly match the quantified thing, the quantifier
1887 is automatically "possessified".
1889 24. A recursive reference to a subpattern whose number was greater than 39
1890 went wrong under certain circumstances in UTF-8 mode. This bug could also
1891 have affected the operation of pcre_study().
1893 25. Realized that a little bit of performance could be had by replacing
1894 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1896 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1898 27. Possessive quantifiers such as a++ were previously implemented by turning
1899 them into atomic groups such as ($>a+). Now they have their own opcodes,
1900 which improves performance. This includes the automatically created ones
1901 from 23 above.
1903 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1904 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1905 the first matched character to be a colon. This applied both to named and
1906 numbered groups.
1908 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1910 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1911 building PCRE on Windows easier, so I have included it in the distribution.
1913 31. There is now a check in pcretest against a ridiculously large number being
1914 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1915 loop, the loop is abandoned.
1917 32. Forward references to subpatterns in conditions such as (?(2)...) where
1918 subpattern 2 is defined later cause pcre_compile() to search forwards in
1919 the pattern for the relevant set of parentheses. This search went wrong
1920 when there were unescaped parentheses in a character class, parentheses
1921 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1923 33. "Subroutine" calls and backreferences were previously restricted to
1924 referencing subpatterns earlier in the regex. This restriction has now
1925 been removed.
1927 34. Added a number of extra features that are going to be in Perl 5.10. On the
1928 whole, these are just syntactic alternatives for features that PCRE had
1929 previously implemented using the Python syntax or my own invention. The
1930 other formats are all retained for compatibility.
1932 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1933 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1934 also .NET compatible.
1936 (b) A recursion or subroutine call to a named group can now be defined as
1937 (?&name) as well as (?P>name).
1939 (c) A backreference to a named group can now be defined as \k<name> or
1940 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1941 5.10, are also .NET compatible.
1943 (d) A conditional reference to a named group can now use the syntax
1944 (?(<name>) or (?('name') as well as (?(name).
1946 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1947 groups (named and numbered) that are never evaluated inline, but can be
1948 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1949 is always false. There may be only one alternative in such a group.
1951 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1952 as the simple (?(R). The condition is true only if the most recent
1953 recursion is that of the given number or name. It does not search out
1954 through the entire recursion stack.
1956 (g) The escape \gN or \g{N} has been added, where N is a positive or
1957 negative number, specifying an absolute or relative reference.
1959 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1960 some "unreachable code" warnings.
1962 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1963 things, this adds five new scripts.
1965 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1966 There were also incompatibilities regarding the handling of \Q..\E inside
1967 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1968 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1970 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1971 matches an empty string, and forcibly breaks the loop. There were bugs in
1972 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1973 against aaaa the result was just "a" rather than "aaaa", for example. Two
1974 separate and independent bugs (that affected different cases) have been
1975 fixed.
1977 39. Refactored the code to abolish the use of different opcodes for small
1978 capturing bracket numbers. This is a tidy that I avoided doing when I
1979 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1980 The new approach is not only tidier, it makes it possible to reduce the
1981 memory needed to fix the previous bug (38).
1983 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1984 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1985 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1986 mode.
1988 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1989 report.
1991 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1992 copy construction and assignment in the C++ wrapper.
1994 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1995 couple of bugs that could have given wrong results in the "--newline=crlf"
1996 case.
1998 44. Added a number of casts and did some reorganization of signed/unsigned int
1999 variables following suggestions from Dair Grant. Also renamed the variable
2000 "this" as "item" because it is a C++ keyword.
2002 45. Arranged for dftables to add
2004 #include "pcre_internal.h"
2006 to pcre_chartables.c because without it, gcc 4.x may remove the array
2007 definition from the final binary if PCRE is built into a static library and
2008 dead code stripping is activated.
2010 46. For an unanchored pattern, if a match attempt fails at the start of a
2011 newline sequence, and the newline setting is CRLF or ANY, and the next two
2012 characters are CRLF, advance by two characters instead of one.
2015 Version 6.7 04-Jul-06
2016 ---------------------
2018 1. In order to handle tests when input lines are enormously long, pcretest has
2019 been re-factored so that it automatically extends its buffers when
2020 necessary. The code is crude, but this _is_ just a test program. The
2021 default size has been increased from 32K to 50K.
2023 2. The code in pcre_study() was using the value of the re argument before
2024 testing it for NULL. (Of course, in any sensible call of the function, it
2025 won't be NULL.)
2027 3. The memmove() emulation function in pcre_internal.h, which is used on
2028 systems that lack both memmove() and bcopy() - that is, hardly ever -
2029 was missing a "static" storage class specifier.
2031 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2032 containing an extended class (one that cannot be represented by a bitmap
2033 because it contains high-valued characters or Unicode property items, e.g.
2034 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2035 pattern, but PCRE should not loop if you do not (it no longer does).
2036 [Detail: two cases were found: (a) a repeated subpattern containing an
2037 extended class; (b) a recursive reference to a subpattern that followed a
2038 previous extended class. It wasn't skipping over the extended class
2039 correctly when UTF-8 mode was not set.]
2041 5. A negated single-character class was not being recognized as fixed-length
2042 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2043 compile error "lookbehind assertion is not fixed length".
2045 6. The RunPerlTest auxiliary script was showing an unexpected difference
2046 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2047 write a Perl script that can interpret lines of an input file either as
2048 byte characters or as UTF-8, which is what "perltest" was being required to
2049 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2050 can't do is switch easily at run time between having the "use utf8;" pragma
2051 or not. In the end, I fudged it by using the RunPerlTest script to insert
2052 "use utf8;" explicitly for the UTF-8 tests.
2054 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2055 the end of the subject string, contrary to the documentation and to what
2056 Perl does. This was true of both matching functions. Now it matches only at
2057 the start of the subject and immediately after *internal* newlines.
2059 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2060 a pointer to an int instead of a pointer to an unsigned long int. This
2061 caused problems on 64-bit systems.
2063 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2064 instance of the 'standard' template library not being so standard".
2066 10. There was no check on the number of named subpatterns nor the maximum
2067 length of a subpattern name. The product of these values is used to compute
2068 the size of the memory block for a compiled pattern. By supplying a very
2069 long subpattern name and a large number of named subpatterns, the size
2070 computation could be caused to overflow. This is now prevented by limiting
2071 the length of names to 32 characters, and the number of named subpatterns
2072 to 10,000.
2074 11. Subpatterns that are repeated with specific counts have to be replicated in
2075 the compiled pattern. The size of memory for this was computed from the
2076 length of the subpattern and the repeat count. The latter is limited to
2077 65535, but there was no limit on the former, meaning that integer overflow
2078 could in principle occur. The compiled length of a repeated subpattern is
2079 now limited to 30,000 bytes in order to prevent this.
2081 12. Added the optional facility to have named substrings with the same name.
2083 13. Added the ability to use a named substring as a condition, using the
2084 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2085 are numbers (not recommended). Forward references are permitted.
2087 14. Added forward references in named backreferences (if you see what I mean).
2089 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2090 pattern could run off the end of the subject. For example, the pattern
2091 "(?s)(.{1,5})"8 did this with the subject "ab".
2093 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2094 PCRE_CASELESS was set when matching characters that were quantified with ?
2095 or *.
2097 17. A character class other than a single negated character that had a minimum
2098 but no maximum quantifier - for example [ab]{6,} - was not handled
2099 correctly by pce_dfa_exec(). It would match only one character.
2101 18. A valid (though odd) pattern that looked like a POSIX character
2102 class but used an invalid character after [ (for example [[,abc,]]) caused
2103 pcre_compile() to give the error "Failed: internal error: code overflow" or
2104 in some cases to crash with a glibc free() error. This could even happen if
2105 the pattern terminated after [[ but there just happened to be a sequence of
2106 letters, a binary zero, and a closing ] in the memory that followed.
2108 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2109 over the years. Originally (before any Unicode support), just the bottom 8
2110 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2111 output from "man perlunicode" includes this:
2113 The regular expression compiler produces polymorphic opcodes. That
2114 is, the pattern adapts to the data and automatically switches to
2115 the Unicode character scheme when presented with Unicode data--or
2116 instead uses a traditional byte scheme when presented with byte
2117 data.
2119 Sadly, a wide octal escape does not cause a switch, and in a string with
2120 no other multibyte characters, these octal escapes are treated as before.
2121 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2122 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2123 Unicode string.
2125 I have not perpetrated such confusion in PCRE. Up till now, it took just
2126 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2127 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2128 translate to the appropriate multibyte character.
2130 29. Applied some refactoring to reduce the number of warnings from Microsoft
2131 and Borland compilers. This has included removing the fudge introduced
2132 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2133 a warning about an unused variable.
2135 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2136 characters since release 4.0, because Perl (from release 5.004) does not.
2137 [Or at least, is documented not to: some releases seem to be in conflict
2138 with the documentation.] However, when a pattern was studied with
2139 pcre_study() and all its branches started with \s, PCRE still included VT
2140 as a possible starting character. Of course, this did no harm; it just
2141 caused an unnecessary match attempt.
2143 22. Removed a now-redundant internal flag bit that recorded the fact that case
2144 dependency changed within the pattern. This was once needed for "required
2145 byte" processing, but is no longer used. This recovers a now-scarce options
2146 bit. Also moved the least significant internal flag bit to the most-
2147 significant bit of the word, which was not previously used (hangover from
2148 the days when it was an int rather than a uint) to free up another bit for
2149 the future.
2151 23. Added support for CRLF line endings as well as CR and LF. As well as the
2152 default being selectable at build time, it can now be changed at runtime
2153 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2154 specify that it is scanning data with non-default line endings.
2156 24. Changed the definition of CXXLINK to make it agree with the definition of
2157 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2159 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2160 recursions. This makes a big different to stack usage for some patterns.
2162 26. If a subpattern containing a named recursion or subroutine reference such
2163 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2164 the space required for the compiled pattern went wrong and gave too small a
2165 value. Depending on the environment, this could lead to "Failed: internal
2166 error: code overflow at offset 49" or "glibc detected double free or
2167 corruption" errors.
2169 27. Applied patches from Google (a) to support the new newline modes and (b) to
2170 advance over multibyte UTF-8 characters in GlobalReplace.
2172 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2173 difference for some implementation of PCRE in some Windows version.
2175 29. Added some extra testing facilities to pcretest:
2177 \q<number> in a data line sets the "match limit" value
2178 \Q<number> in a data line sets the "match recursion limt" value
2179 -S <number> sets the stack size, where <number> is in megabytes
2181 The -S option isn't available for Windows.
2184 Version 6.6 06-Feb-06
2185 ---------------------
2187 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2188 in pcreposix.h. I have copied the definition from pcre.h.
2190 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2191 because pcre.h is no longer a built file.
2193 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2194 not normally included in the compiled code.
2197 Version 6.5 01-Feb-06
2198 ---------------------
2200 1. When using the partial match feature with pcre_dfa_exec(), it was not
2201 anchoring the second and subsequent partial matches at the new starting
2202 point. This could lead to incorrect results. For example, with the pattern
2203 /1234/, partially matching against "123" and then "a4" gave a match.
2205 2. Changes to pcregrep:
2207 (a) All non-match returns from pcre_exec() were being treated as failures
2208 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2209 error message is output. Some extra information is given for the
2211 probably the only errors that are likely to be caused by users (by
2212 specifying a regex that has nested indefinite repeats, for instance).
2213 If there are more than 20 of these errors, pcregrep is abandoned.
2215 (b) A binary zero was treated as data while matching, but terminated the
2216 output line if it was written out. This has been fixed: binary zeroes
2217 are now no different to any other data bytes.
2219 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2220 used to set a locale for matching. The --locale=xxxx long option has
2221 been added (no short equivalent) to specify a locale explicitly on the
2222 pcregrep command, overriding the environment variables.
2224 (d) When -B was used with -n, some line numbers in the output were one less
2225 than they should have been.
2227 (e) Added the -o (--only-matching) option.
2229 (f) If -A or -C was used with -c (count only), some lines of context were
2230 accidentally printed for the final match.
2232 (g) Added the -H (--with-filename) option.
2234 (h) The combination of options -rh failed to suppress file names for files
2235 that were found from directory arguments.
2237 (i) Added the -D (--devices) and -d (--directories) options.
2239 (j) Added the -F (--fixed-strings) option.
2241 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2243 (l) Added the --colo(u)r option.
2245 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2246 is not present by default.
2248 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2249 items such as (?R) or (?1), when the recursion could match a number of
2250 alternatives. If it matched one of the alternatives, but subsequently,
2251 outside the recursion, there was a failure, the code tried to back up into
2252 the recursion. However, because of the way PCRE is implemented, this is not
2253 possible, and the result was an incorrect result from the match.
2255 In order to prevent this happening, the specification of recursion has
2256 been changed so that all such subpatterns are automatically treated as
2257 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2259 4. I had overlooked the fact that, in some locales, there are characters for
2260 which isalpha() is true but neither isupper() nor islower() are true. In
2261 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2262 and ordfeminine) are like this. This affected the treatment of \w and \W
2263 when they appeared in character classes, but not when they appeared outside
2264 a character class. The bit map for "word" characters is now created
2265 separately from the results of isalnum() instead of just taking it from the
2266 upper, lower, and digit maps. (Plus the underscore character, of course.)
2268 5. The above bug also affected the handling of POSIX character classes such as
2269 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2270 permanent tables. Instead, the bit maps for such a class were previously
2271 created as the appropriate unions of the upper, lower, and digit bitmaps.
2272 Now they are created by subtraction from the [[:word:]] class, which has
2273 its own bitmap.
2275 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2276 It is created by subtracting the vertical space characters (\x09, \x0a,
2277 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2278 subtraction was done in the overall bitmap for a character class, meaning
2279 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2280 be recognized. This bug has been fixed.
2282 7. Patches from the folks at Google:
2284 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2285 real life, but is still worth protecting against".
2287 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2288 regular expressions".
2290 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2291 have it.
2293 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2294 "configure" and the latter not, in order to fix a problem somebody had
2295 with compiling the Arg class on HP-UX.
2297 (e) Improve the error-handling of the C++ wrapper a little bit.
2299 (f) New tests for checking recursion limiting.
2301 8. The pcre_memmove() function, which is used only if the environment does not
2302 have a standard memmove() function (and is therefore rarely compiled),
2303 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2304 returning a result (though PCRE never actually uses the result).
2306 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2307 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2308 returned instead of calling malloc() with an overflowing number that would
2309 most likely cause subsequent chaos.
2311 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2313 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2314 with this option is matched, the nmatch and pmatch options of regexec() are
2315 ignored.
2317 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2318 provided in case anyone wants to the the POSIX interface with UTF-8
2319 strings.
2321 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2322 C++ linking (needed for some HP-UX environments).
2324 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2325 (unused parameter) and in the pcre_printint() function (omitted "default"
2326 switch label when the default is to do nothing).
2328 15. Added some code to make it possible, when PCRE is compiled as a C++
2329 library, to replace subject pointers for pcre_exec() with a smart pointer
2330 class, thus making it possible to process discontinuous strings.
2332 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2333 much the same function. They were added by different people who were trying
2334 to make PCRE easy to compile on non-Unix systems. It has been suggested
2335 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2336 for compiling on Windows systems. I have therefore replaced it with
2337 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2338 defaults to "extern" for C or "extern C" for C++, which works fine on
2339 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2340 SCOPE with something explicit in config.h. In addition:
2342 (a) pcreposix.h still had just "extern" instead of either of these macros;
2343 I have replaced it with PCRE_DATA_SCOPE.
2345 (b) Functions such as _pcre_xclass(), which are internal to the library,
2346 but external in the C sense, all had PCRE_EXPORT in their definitions.
2347 This is apparently wrong for the Windows case, so I have removed it.
2348 (It makes no difference on Unix-like systems.)
2350 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2351 of recursive calls to match(). This is different to MATCH_LIMIT because
2352 that limits the total number of calls to match(), not all of which increase
2353 the depth of recursion. Limiting the recursion depth limits the amount of
2354 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2355 when PCRE is compiled, and changed at run time. A patch from Google adds
2356 this functionality to the C++ interface.
2358 18. Changes to the handling of Unicode character properties:
2360 (a) Updated the table to Unicode 4.1.0.
2362 (b) Recognize characters that are not in the table as "Cn" (undefined).
2364 (c) I revised the way the table is implemented to a much improved format
2365 which includes recognition of ranges. It now supports the ranges that
2366 are defined in UnicodeData.txt, and it also amalgamates other
2367 characters into ranges. This has reduced the number of entries in the
2368 table from around 16,000 to around 3,000, thus reducing its size
2369 considerably. I realized I did not need to use a tree structure after
2370 all - a binary chop search is just as efficient. Having reduced the
2371 number of entries, I extended their size from 6 bytes to 8 bytes to
2372 allow for more data.
2374 (d) Added support for Unicode script names via properties such as \p{Han}.
2376 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2377 matching that character.
2379 20. When matching a repeated Unicode property with a minimum greater than zero,
2380 (for example \pL{2,}), PCRE could look past the end of the subject if it
2381 reached it while seeking the minimum number of characters. This could
2382 happen only if some of the characters were more than one byte long, because
2383 there is a check for at least the minimum number of bytes.
2385 21. Refactored the implementation of \p and \P so as to be more general, to
2386 allow for more different types of property in future. This has changed the
2387 compiled form incompatibly. Anybody with saved compiled patterns that use
2388 \p or \P will have to recompile them.
2390 22. Added "Any" and "L&" to the supported property types.
2392 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2393 but give a compile time error if the value is greater than 0xff.
2395 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2396 accidentally not being installed or uninstalled.
2398 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2399 made were to insert the current release number. This seemed silly, because
2400 it made things harder for people building PCRE on systems that don't run
2401 "configure". I have turned pcre.h into a distributed file, no longer built
2402 by "configure", with the version identification directly included. There is
2403 no longer a pcre.h.in file.
2405 However, this change necessitated a change to the pcre-config script as
2406 well. It is built from pcre-config.in, and one of the substitutions was the
2407 release number. I have updated configure.ac so that ./configure now finds
2408 the release number by grepping pcre.h.
2410 26. Added the ability to run the tests under valgrind.
2413 Version 6.4 05-Sep-05
2414 ---------------------
2416 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2417 "--" to be printed when multiple files were scanned, even when none of the
2418 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2419 consider it to be a bug, and have restored the previous behaviour.
2421 2. A couple of code tidies to get rid of compiler warnings.
2423 3. The pcretest program used to cheat by referring to symbols in the library
2424 whose names begin with _pcre_. These are internal symbols that are not
2425 really supposed to be visible externally, and in some environments it is
2426 possible to suppress them. The cheating is now confined to including
2427 certain files from the library's source, which is a bit cleaner.
2429 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2430 file's purpose clearer.
2432 5. Reorganized pcre_ucp_findchar().
2435 Version 6.3 15-Aug-05
2436 ---------------------
2438 1. The file libpcre.pc.in did not have general read permission in the tarball.
2440 2. There were some problems when building without C++ support:
2442 (a) If C++ support was not built, "make install" and "make test" still
2443 tried to test it.
2445 (b) There were problems when the value of CXX was explicitly set. Some
2446 changes have been made to try to fix these, and ...
2448 (c) --disable-cpp can now be used to explicitly disable C++ support.
2450 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2451 backslash in a target when C++ was disabled. This confuses some
2452 versions of "make", apparently. Using an intermediate variable solves
2453 this. (Same for CPP_LOBJ.)
2455 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2456 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2457 necessary on certain architectures.
2459 4. Added a setting of -export-symbols-regex to the link command to remove
2460 those symbols that are exported in the C sense, but actually are local
2461 within the library, and not documented. Their names all begin with
2462 "_pcre_". This is not a perfect job, because (a) we have to except some
2463 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2464 available (and never for static libraries). I have made a note to try to
2465 find a way round (a) in the future.
2468 Version 6.2 01-Aug-05
2469 ---------------------
2471 1. There was no test for integer overflow of quantifier values. A construction
2472 such as {1111111111111111} would give undefined results. What is worse, if
2473 a minimum quantifier for a parenthesized subpattern overflowed and became
2474 negative, the calculation of the memory size went wrong. This could have
2475 led to memory overwriting.
2477 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2479 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2480 operating environments where this matters.
2482 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2483 PCRE options from within the C++ wrapper.
2485 5. Named capturing subpatterns were not being correctly counted when a pattern
2486 was compiled. This caused two problems: (a) If there were more than 100
2487 such subpatterns, the calculation of the memory needed for the whole
2488 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2489 back references of the form \12, where the number was greater than 9, were
2490 not recognized as back references, even though there were sufficient
2491 previous subpatterns.
2493 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2494 versions of gcc, e.g. 2.95.4.
2497 Version 6.1 21-Jun-05
2498 ---------------------
2500 1. There was one reference to the variable "posix" in pcretest.c that was not
2501 surrounded by "#if !defined NOPOSIX".
2503 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2504 the cross-check on the old pcre_info() function, for the benefit of the
2505 cut-down version of PCRE that is currently imported into Exim.
2507 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2508 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2509 patterns that start (?i) but I don't think that matters. The use of (?i) is
2510 just an example; this all applies to the other options as well.
2512 4. Since libtool seems to echo the compile commands it is issuing, the output
2513 from "make" can be reduced a bit by putting "@" in front of each libtool
2514 compile command.
2516 5. Patch from the folks at Google for configure.in to be a bit more thorough
2517 in checking for a suitable C++ installation before trying to compile the
2518 C++ stuff. This should fix a reported problem when a compiler was present,
2519 but no suitable headers.
2521 6. The man pages all had just "PCRE" as their title. I have changed them to
2522 be the relevant file name. I have also arranged that these names are
2523 retained in the file doc/pcre.txt, which is a concatenation in text format
2524 of all the man pages except the little individual ones for each function.
2526 7. The NON-UNIX-USE file had not been updated for the different set of source
2527 files that come with release 6. I also added a few comments about the C++
2528 wrapper.
2531 Version 6.0 07-Jun-05
2532 ---------------------
2534 1. Some minor internal re-organization to help with my DFA experiments.
2536 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2537 didn't matter for the library itself when fully configured, but did matter
2538 when compiling without UCP support, or within Exim, where the ucp files are
2539 not imported.
2541 3. Refactoring of the library code to split up the various functions into
2542 different source modules. The addition of the new DFA matching code (see
2543 below) to a single monolithic source would have made it really too
2544 unwieldy, quite apart from causing all the code to be include in a
2545 statically linked application, when only some functions are used. This is
2546 relevant even without the DFA addition now that patterns can be compiled in
2547 one application and matched in another.
2549 The downside of splitting up is that there have to be some external
2550 functions and data tables that are used internally in different modules of
2551 the library but which are not part of the API. These have all had their
2552 names changed to start with "_pcre_" so that they are unlikely to clash
2553 with other external names.
2555 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2556 a different (DFA) algorithm. Although it is slower than the original
2557 function, it does have some advantages for certain types of matching
2558 problem.
2560 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2561 including restarting after a partial match.
2563 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2564 defined when compiling for Windows was sent to me. I have put it into the
2565 code, though I have no means of testing or verifying it.
2567 7. Added the pcre_refcount() auxiliary function.
2569 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2570 match before or at the first newline in the subject string. In pcretest,
2571 the /f option on a pattern can be used to set this.
2573 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2574 would behave wrongly. This has been present in PCRE since release 4.0.
2576 10. A number of changes to the pcregrep command:
2578 (a) Refactored how -x works; insert ^(...)$ instead of setting
2579 PCRE_ANCHORED and checking the length, in preparation for adding
2580 something similar for -w.
2582 (b) Added the -w (match as a word) option.
2584 (c) Refactored the way lines are read and buffered so as to have more
2585 than one at a time available.
2587 (d) Implemented a pcregrep test script.
2589 (e) Added the -M (multiline match) option. This allows patterns to match
2590 over several lines of the subject. The buffering ensures that at least
2591 8K, or the rest of the document (whichever is the shorter) is available
2592 for matching (and similarly the previous 8K for lookbehind assertions).
2594 (f) Changed the --help output so that it now says
2596 -w, --word-regex(p)
2598 instead of two lines, one with "regex" and the other with "regexp"
2599 because that confused at least one person since the short forms are the
2600 same. (This required a bit of code, as the output is generated
2601 automatically from a table. It wasn't just a text change.)
2603 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2604 option but starts with a hyphen. Could be a pattern or a path name
2605 starting with a hyphen, for instance.
2607 (h) "-" can be given as a file name to represent stdin.
2609 (i) When file names are being printed, "(standard input)" is used for
2610 the standard input, for compatibility with GNU grep. Previously
2611 "<stdin>" was used.
2613 (j) The option --label=xxx can be used to supply a name to be used for
2614 stdin when file names are being printed. There is no short form.
2616 (k) Re-factored the options decoding logic because we are going to add
2617 two more options that take data. Such options can now be given in four
2618 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2620 (l) Added the -A, -B, and -C options for requesting that lines of context
2621 around matches be printed.
2623 (m) Added the -L option to print the names of files that do not contain
2624 any matching lines, that is, the complement of -l.
2626 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2627 continue to scan other files.
2629 (o) The -s option was incorrectly implemented. For compatibility with other
2630 greps, it now suppresses the error message for a non-existent or non-
2631 accessible file (but not the return code). There is a new option called
2632 -q that suppresses the output of matching lines, which was what -s was
2633 previously doing.
2635 (p) Added --include and --exclude options to specify files for inclusion
2636 and exclusion when recursing.
2638 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2639 Hopefully, it now does.
2641 12. Missing cast in pcre_study().
2643 13. Added an "uninstall" target to the makefile.
2645 14. Replaced "extern" in the function prototypes in Makefile.in with
2646 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2647 world, but is set differently for Windows.
2649 15. Added a second compiling function called pcre_compile2(). The only
2650 difference is that it has an extra argument, which is a pointer to an
2651 integer error code. When there is a compile-time failure, this is set
2652 non-zero, in addition to the error test pointer being set to point to an
2653 error message. The new argument may be NULL if no error number is required
2654 (but then you may as well call pcre_compile(), which is now just a
2655 wrapper). This facility is provided because some applications need a
2656 numeric error indication, but it has also enabled me to tidy up the way
2657 compile-time errors are handled in the POSIX wrapper.
2659 16. Added VPATH=.libs to the makefile; this should help when building with one
2660 prefix path and installing with another. (Or so I'm told by someone who
2661 knows more about this stuff than I do.)
2663 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2664 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2665 match everything, including newlines. This is not POSIX-compatible, but
2666 somebody wanted the feature. From pcretest it can be activated by using
2667 both the P and the s flags.
2669 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2671 19. libpcre.pc was being incorrectly installed as executable.
2673 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2674 it now also looks for '\r' so that it will work unmodified on Windows.
2676 21. Added Google's contributed C++ wrapper to the distribution.
2678 22. Added some untidy missing memory free() calls in pcretest, to keep
2679 Electric Fence happy when testing.
2683 Version 5.0 13-Sep-04
2684 ---------------------
2686 1. Internal change: literal characters are no longer packed up into items
2687 containing multiple characters in a single byte-string. Each character
2688 is now matched using a separate opcode. However, there may be more than one
2689 byte in the character in UTF-8 mode.
2691 2. The pcre_callout_block structure has two new fields: pattern_position and
2692 next_item_length. These contain the offset in the pattern to the next match
2693 item, and its length, respectively.
2695 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2696 insertion of callouts before each pattern item. Added the /C option to
2697 pcretest to make use of this.
2699 4. On the advice of a Windows user, the lines
2701 #if defined(_WIN32) || defined(WIN32)
2702 _setmode( _fileno( stdout ), 0x8000 );
2703 #endif /* defined(_WIN32) || defined(WIN32) */
2705 have been added to the source of pcretest. This apparently does useful
2706 magic in relation to line terminators.
2708 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2709 for the benefit of those environments where the "b" makes a difference.
2711 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2712 to know about it. I have put a hack into configure.in that adds in code
2713 to set GCC=yes if CC=icc. This seems to end up at a point in the
2714 generated configure script that is early enough to affect the setting of
2715 compiler options, which is what is needed, but I have no means of testing
2716 whether it really works. (The user who reported this had patched the
2717 generated configure script, which of course I cannot do.)
2719 LATER: After change 22 below (new libtool files), the configure script
2720 seems to know about icc (and also ecc). Therefore, I have commented out
2721 this hack in configure.in.
2723 7. Added support for pkg-config (2 patches were sent in).
2725 8. Negated POSIX character classes that used a combination of internal tables
2726 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2727 [[:^ascii]]. Typically, they would match almost any characters. The other
2728 POSIX classes were not broken in this way.
2730 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2731 to find the match, as PCRE was deluded into thinking that the match had to
2732 start at the start point or following a newline. The same bug applied to
2733 patterns with negative forward assertions or any backward assertions
2734 preceding ".*" at the start, unless the pattern required a fixed first
2735 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2737 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2738 starting at the last subject character, bytes beyond the end of the subject
2739 string were read.
2741 11. Renamed the variable "class" as "classbits" to make life easier for C++
2742 users. (Previously there was a macro definition, but it apparently wasn't
2743 enough.)
2745 12. Added the new field "tables" to the extra data so that tables can be passed
2746 in at exec time, or the internal tables can be re-selected. This allows
2747 a compiled regex to be saved and re-used at a later time by a different
2748 program that might have everything at different addresses.
2750 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2751 -R library as well as a -L library.
2753 14. The debugging options of pcretest (-d on the command line or D on a
2754 pattern) showed incorrect output for anything following an extended class
2755 that contained multibyte characters and which was followed by a quantifier.
2757 15. Added optional support for general category Unicode character properties
2758 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2759 support. It adds about 90K to the size of the library. The meanings of the
2760 inbuilt class escapes such as \d and \s have NOT been changed.
2762 16. Updated pcredemo.c to include calls to free() to release the memory for the
2763 compiled pattern.
2765 17. The generated file chartables.c was being created in the source directory
2766 instead of in the building directory. This caused the build to fail if the
2767 source directory was different from the building directory, and was
2768 read-only.
2770 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2771 file. No doubt somebody will tell me if they don't make sense... Also added
2772 Dan Mooney's comments about building on OpenVMS.
2774 19. Added support for partial matching via the PCRE_PARTIAL option for
2775 pcre_exec() and the \P data escape in pcretest.
2777 20. Extended pcretest with 3 new pattern features:
2779 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2780 write the compiled pattern to the file whose name is "rest-of-line".
2781 This is a straight binary dump of the data, with the saved pointer to
2782 the character tables forced to be NULL. The study data, if any, is
2783 written too. After writing, pcretest reads a new pattern.
2785 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2786 compiled pattern from the given file. There must not be any
2787 occurrences of "<" in the file name (pretty unlikely); if there are,
2788 pcretest will instead treat the initial "<" as a pattern delimiter.
2789 After reading in the pattern, pcretest goes on to read data lines as
2790 usual.
2792 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2793 and 16-bit fields in a compiled pattern, to simulate a pattern that
2794 was compiled on a host of opposite endianness.
2796 21. The pcre-exec() function can now cope with patterns that were compiled on
2797 hosts of opposite endianness, with this restriction:
2799 As for any compiled expression that is saved and used later, the tables
2800 pointer field cannot be preserved; the extra_data field in the arguments
2801 to pcre_exec() should be used to pass in a tables address if a value
2802 other than the default internal tables were used at compile time.
2804 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2805 now diagnosed as an error. Previously, most of the time, a negative number
2806 would have been treated as zero, but if in addition "ovector" was passed as
2807 NULL, a crash could occur.
2809 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2810 new versions from the libtool 1.5 distribution (the last one is a copy of
2811 a file called libtool.m4). This seems to have fixed the need to patch
2812 "configure" to support Darwin 1.3 (which I used to do). However, I still
2813 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2814 workstation).
2816 24. Changed the PCRE licence to be the more standard "BSD" licence.
2819 Version 4.5 01-Dec-03
2820 ---------------------
2822 1. There has been some re-arrangement of the code for the match() function so
2823 that it can be compiled in a version that does not call itself recursively.
2824 Instead, it keeps those local variables that need separate instances for
2825 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2826 needs to "recurse". Keeping track of where control must go is done by means
2827 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2828 hide most of the details from the main code, and operates only if
2829 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2830 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2831 operating.
2833 To make it easier for callers to provide specially tailored get/free
2834 functions for this usage, two new functions, pcre_stack_malloc, and
2835 pcre_stack_free, are used. They are always called in strict stacking order,
2836 and the size of block requested is always the same.
2838 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2839 PCRE has been compiled to use the stack or the heap for recursion. The
2840 -C option of pcretest uses this to show which version is compiled.
2842 A new data escape \S, is added to pcretest; it causes the amounts of store
2843 obtained and freed by both kinds of malloc/free at match time to be added
2844 to the output.
2846 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2847 what's available on my current Linux desktop machine.
2849 3. When matching a UTF-8 string, the test for a valid string at the start has
2850 been extended. If start_offset is not zero, PCRE now checks that it points
2851 to a byte that is the start of a UTF-8 character. If not, it returns
2852 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2853 this is necessary because there may be backward assertions in the pattern.
2854 When matching the same subject several times, it may save resources to use
2855 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2857 4. The code for checking the validity of UTF-8 strings has been tightened so
2858 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2859 containing "overlong sequences".
2861 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2862 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2863 should have been "&", but it just so happened that all the cases this let
2864 through by mistake were picked up later in the function.
2866 6. I had used a variable called "isblank" - this is a C99 function, causing
2867 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2869 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2870 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2871 so the version is shown; (c) stop "make test" echoing "./RunTest".
2873 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2875 9. The prototype for memmove() for systems that don't have it was using
2876 size_t, but the inclusion of the header that defines size_t was later. I've
2877 moved the #includes for the C headers earlier to avoid this.
2879 10. Added some adjustments to the code to make it easier to compiler on certain
2880 special systems:
2882 (a) Some "const" qualifiers were missing.
2883 (b) Added the macro EXPORT before all exported functions; by default this
2884 is defined to be empty.
2885 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2886 that it reads its output file name as an argument instead of writing
2887 to the standard output and assuming this can be redirected.
2889 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2890 class containing characters with values greater than 255, PCRE compilation
2891 went into a loop.
2893 12. A recursive reference to a subpattern that was within another subpattern
2894 that had a minimum quantifier of zero caused PCRE to crash. For example,
2895 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2896 recursion. If the recursively-called subpattern itself had a zero repeat,
2897 that was OK.
2899 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2900 buffer into which it was copied (for escape processing) was still set at
2901 1024, so long lines caused crashes.
2903 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2904 "internal error: code overflow...". This applied to any character class
2905 that was followed by a possessive quantifier.
2907 15. Modified the Makefile to add libpcre.la as a prerequisite for
2908 libpcreposix.la because I was told this is needed for a parallel build to
2909 work.
2911 16. If a pattern that contained .* following optional items at the start was
2912 studied, the wrong optimizing data was generated, leading to matching
2913 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2914 matching string must start with a or b or c. The correct conclusion for
2915 this pattern is that a match can start with any character.
2918 Version 4.4 13-Aug-03
2919 ---------------------
2921 1. In UTF-8 mode, a character class containing characters with values between
2922 127 and 255 was not handled correctly if the compiled pattern was studied.
2923 In fixing this, I have also improved the studying algorithm for such
2924 classes (slightly).
2926 2. Three internal functions had redundant arguments passed to them. Removal
2927 might give a very teeny performance improvement.
2929 3. Documentation bug: the value of the capture_top field in a callout is *one
2930 more than* the number of the hightest numbered captured substring.
2932 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2933 in incorrectly linking with a previously installed version. They now link
2934 explicitly with libpcre.la.
2936 5. configure.in no longer needs to recognize Cygwin specially.
2938 6. A problem in pcre.in for Windows platforms is fixed.
2940 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2941 pcretest, it used to include the size of the study block as part of its
2942 output. Unfortunately, the structure contains a field that has a different
2943 size on different hardware architectures. This meant that the tests that
2944 showed this size failed. As the block is currently always of a fixed size,
2945 this information isn't actually particularly useful in pcretest output, so
2946 I have just removed it.
2948 8. Three pre-processor statements accidentally did not start in column 1.
2949 Sadly, there are *still* compilers around that complain, even though
2950 standard C has not required this for well over a decade. Sigh.
2952 9. In pcretest, the code for checking callouts passed small integers in the
2953 callout_data field, which is a void * field. However, some picky compilers
2954 complained about the casts involved for this on 64-bit systems. Now
2955 pcretest passes the address of the small integer instead, which should get
2956 rid of the warnings.
2958 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2959 both compile and run time, and gives an error if an invalid UTF-8 sequence
2960 is found. There is a option for disabling this check in cases where the
2961 string is known to be correct and/or the maximum performance is wanted.
2963 11. In response to a bug report, I changed one line in Makefile.in from
2965 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2966 to
2967 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2969 to look similar to other lines, but I have no way of telling whether this
2970 is the right thing to do, as I do not use Windows. No doubt I'll get told
2971 if it's wrong...
2974 Version 4.3 21-May-03
2975 ---------------------
2977 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2978 Makefile.
2980 2. Some refactoring to improve the quality of the code:
2982 (i) The utf8_table... variables are now declared "const".
2984 (ii) The code for \cx, which used the "case flipping" table to upper case
2985 lower case letters, now just substracts 32. This is ASCII-specific,
2986 but the whole concept of \cx is ASCII-specific, so it seems
2987 reasonable.
2989 (iii) PCRE was using its character types table to recognize decimal and
2990 hexadecimal digits in the pattern. This is silly, because it handles
2991 only 0-9, a-f, and A-F, but the character types table is locale-
2992 specific, which means strange things might happen. A private
2993 table is now used for this - though it costs 256 bytes, a table is
2994 much faster than multiple explicit tests. Of course, the standard
2995 character types table is still used for matching digits in subject
2996 strings against \d.
2998 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2999 ending in _t are). So I've renamed it as ESC_tee.
3001 3. The first argument for regexec() in the POSIX wrapper should have been
3002 defined as "const".
3004 4. Changed pcretest to use malloc() for its buffers so that they can be
3005 Electric Fenced for debugging.
3007 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3008 to read one or more bytes before the start of the subject string. Often this
3009 had no effect on PCRE's behaviour, but in some circumstances it could
3010 provoke a segmentation fault.
3012 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3013 to try to read one or more bytes before the start of the subject string.
3015 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3016 UTF-8 support could misbehave in various ways if the subject string
3017 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3018 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3019 back over UTF-8 characters.)
3022 Version 4.2 14-Apr-03
3023 ---------------------
3025 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3027 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3028 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3029 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3030 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3031 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3033 Note: automatic setting of the BUILD variables is not yet working
3034 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3035 compile-time but not at link-time
3036 [LINK]: use for linking executables only
3037 make different versions for Windows and non-Windows
3038 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3039 libraries
3040 [LINK_FOR_BUILD]: new variable
3041 [OBJEXT]: use throughout
3042 [EXEEXT]: use throughout
3043 <winshared>: new target
3044 <wininstall>: new target
3045 <dftables.o>: use native compiler
3046 <dftables>: use native linker
3047 <install>: handle Windows platform correctly
3048 <clean>: ditto
3049 <check>: ditto
3050 copy DLL to top builddir before testing
3052 As part of these changes, -no-undefined was removed again. This was reported
3053 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3054 in any case.
3056 3. Some tidies to get rid of compiler warnings:
3058 . In the match_data structure, match_limit was an unsigned long int, whereas
3059 match_call_count was an int. I've made them both unsigned long ints.
3061 . In pcretest the fact that a const uschar * doesn't automatically cast to
3062 a void * provoked a warning.
3064 . Turning on some more compiler warnings threw up some "shadow" variables
3065 and a few more missing casts.
3067 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3068 option, a class that contained a single character with a value between 128
3069 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3071 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3072 option, a class that contained several characters, but with at least one
3073 whose value was between 128 and 255 caused PCRE to crash.
3076 Version 4.1 12-Mar-03
3077 ---------------------
3079 1. Compiling with gcc -pedantic found a couple of places where casts were
3080 needed, and a string in dftables.c that was longer than standard compilers are
3081 required to support.
3083 2. Compiling with Sun's compiler found a few more places where the code could
3084 be tidied up in order to avoid warnings.
3086 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3087 first of these names is deprecated in the latest Autoconf in favour of the name
3088 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3089 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3090 analogy I have changed it to CFLAGS_FOR_BUILD.
3092 4. Added -no-undefined to the linking command in the Makefile, because this is
3093 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3094 linking step for the pcreposix library.
3096 5. PCRE was failing to diagnose the case of two named groups with the same
3097 name.
3099 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3100 literal character that is needed in the subject for a match, and scans along to
3101 ensure that it is present before embarking on the full matching process. This
3102 saves time in cases of nested unlimited repeats that are never going to match.
3103 Problem: the scan can take a lot of time if the subject is very long (e.g.
3104 megabytes), thus penalizing straightforward matches. It is now done only if the
3105 amount of subject to be scanned is less than 1000 bytes.
3107 7. A lesser problem with the same optimization is that it was recording the
3108 first character of an anchored pattern as "needed", thus provoking a search
3109 right along the subject, even when the first match of the pattern was going to
3110 fail. The "needed" character is now not set for anchored patterns, unless it
3111 follows something in the pattern that is of non-fixed length. Thus, it still
3112 fulfils its original purpose of finding quick non-matches in cases of nested
3113 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3116 Version 4.0 17-Feb-03
3117 ---------------------
3119 1. If a comment in an extended regex that started immediately after a meta-item
3120 extended to the end of string, PCRE compiled incorrect data. This could lead to
3121 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3123 2. Moved to autoconf 2.53 and libtool 1.4.2.
3125 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3126 the special perltest8 script is no longer needed - all the tests can be run
3127 from a single perltest script.
3129 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3130 by \s. It has now been removed in PCRE. This means it isn't recognized as
3131 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3132 class [:space:] *does* include VT, thereby creating a mess.
3134 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3135 space and tab.
3137 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3138 its new features into the main test script, reducing the number of scripts.
3140 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3141 were backward compatible, and made the (?i) apply to the whole pattern, as if
3142 /i were given. Now it behaves more logically, and applies the option setting
3143 only to what follows. PCRE has been changed to follow suit. However, if it
3144 finds options settings right at the start of the pattern, it extracts them into
3145 the global options, as before. Thus, they show up in the info data.
3147 8. Added support for the \Q...\E escape sequence. Characters in between are
3148 treated as literals. This is slightly different from Perl in that $ and @ are
3149 also handled as literals inside the quotes. In Perl, they will cause variable
3150 interpolation. Note the following examples:
3152 Pattern PCRE matches Perl matches
3154 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3155 \Qabc\$xyz\E abc\$xyz abc\$xyz
3156 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3158 For compatibility with Perl, \Q...\E sequences are recognized inside character
3159 classes as well as outside them.
3161 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3162 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3163 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3164 signed/unsigned warnings.
3166 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3167 option for pcretest, so I've replaced it by a simple function that does just
3168 that job.
3170 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3171 "pcregrep -".
3173 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3174 Java package. This provides some syntactic sugar for simple cases of what my
3175 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3176 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3177 item, you can use this simplified notation. Note that only makes sense with
3178 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3179 greediness, whatever the setting of the PCRE_UNGREEDY option.
3181 13. A change of greediness default within a pattern was not taking effect at
3182 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3183 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3184 was abstracted outside.
3186 14. PCRE now supports the \G assertion. It is true when the current matching
3187 position is at the start point of the match. This differs from \A when the
3188 starting offset is non-zero. Used with the /g option of pcretest (or similar
3189 code), it works in the same way as it does for Perl's /g option. If all
3190 alternatives of a regex begin with \G, the expression is anchored to the start
3191 match position, and the "anchored" flag is set in the compiled expression.
3193 15. Some bugs concerning the handling of certain option changes within patterns
3194 have been fixed. These applied to options other than (?ims). For example,
3195 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3196 been the other way round. Some of this was related to change 7 above.
3198 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3199 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3200 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3201 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3203 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3204 mode. Unlike ".", it always matches newline, whatever the setting of
3205 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3206 assertions. Perl allows it, but it doesn't (in general) work because it can't
3207 calculate the length of the lookbehind. At least, that's the case for Perl
3208 5.8.0 - I've been told they are going to document that it doesn't work in
3209 future.
3211 18. Added an error diagnosis for escapes that PCRE does not support: these are
3212 \L, \l, \N, \P, \p, \U, \u, and \X.
3214 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3215 reading past the end of the pattern in cases such as /[abcd/.
3217 20. PCRE was getting more memory than necessary for patterns with classes that
3218 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3220 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3221 compiling PCRE for use with Virtual Pascal.
3223 22. Small fix to the Makefile to make it work properly if the build is done
3224 outside the source tree.
3226 23. Added a new extension: a condition to go with recursion. If a conditional
3227 subpattern starts with (?(R) the "true" branch is used if recursion has
3228 happened, whereas the "false" branch is used only at the top level.
3230 24. When there was a very long string of literal characters (over 255 bytes
3231 without UTF support, over 250 bytes with UTF support), the computation of how
3232 much memory was required could be incorrect, leading to segfaults or other
3233 strange effects.
3235 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3236 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3237 there was a subsequent back reference to those brackets. This meant that, for
3238 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3239 possible to check for precisely this case. All we can do is abandon the
3240 optimization if .* occurs inside capturing brackets when there are any back
3241 references whatsoever. (See below for a better fix that came later.)
3243 26. The handling of the optimization for finding the first character of a
3244 non-anchored pattern, and for finding a character that is required later in the
3245 match were failing in some cases. This didn't break the matching; it just
3246 failed to optimize when it could. The way this is done has been re-implemented.
3248 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3250 28. Added a new feature that provides some of the functionality that Perl
3251 provides with (?{...}). The facility is termed a "callout". The way it is done
3252 in PCRE is for the caller to provide an optional function, by setting
3253 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3254 global variable. By default it is unset, which disables all calling out. To get
3255 the function called, the regex must include (?C) at appropriate points. This
3256 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3257 This provides a means of identifying different callout points. When PCRE
3258 reaches such a point in the regex, if pcre_callout has been set, the external
3259 function is called. It is provided with data in a structure called
3260 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3261 matching continues; if it returns a non-zero value, the match at the current
3262 point fails. However, backtracking will occur if possible. [This was changed
3263 later and other features added - see item 49 below.]
3265 29. pcretest is upgraded to test the callout functionality. It provides a
3266 callout function that displays information. By default, it shows the start of
3267 the match and the current position in the text. There are some new data escapes
3268 to vary what happens:
3270 \C+ in addition, show current contents of captured substrings
3271 \C- do not supply a callout function
3272 \C!n return 1 when callout number n is reached
3273 \C!n!m return 1 when callout number n is reached for the mth time
3275 30. If pcregrep was called with the -l option and just a single file name, it
3276 output "<stdin>" if a match was found, instead of the file name.
3278 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3279 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3280 pcre_exec(). This saves a malloc/free per call. The default value of
3281 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3282 when configuring.
3284 32. The default maximum size of a compiled pattern is 64K. There have been a
3285 few cases of people hitting this limit. The code now uses macros to handle the
3286 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3287 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3288 configuring. Tests 2 and 5 work only with 2-byte links because they output
3289 debugging information about compiled patterns.
3291 33. Internal code re-arrangements:
3293 (a) Moved the debugging function for printing out a compiled regex into
3294 its own source file (printint.c) and used #include to pull it into
3295 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3296 separate copies.
3298 (b) Defined the list of op-code names for debugging as a macro in
3299 internal.h so that it is next to the definition of the opcodes.
3301 (c) Defined a table of op-code lengths for simpler skipping along compiled
3302 code. This is again a macro in internal.h so that it is next to the
3303 definition of the opcodes.
3305 34. Added support for recursive calls to individual subpatterns, along the
3306 lines of Robin Houston's patch (but implemented somewhat differently).
3308 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3309 allow it to read and process whole directories in Win32. This code was
3310 contributed by Lionel Fourquaux; it has not been tested by me.
3312 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3313 used to name a group. Names consist of alphanumerics and underscores, and must
3314 be unique. Back references use the syntax (?P=name) and recursive calls use
3315 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3316 numbers. The function pcre_fullinfo() can be used after compilation to extract
3317 a name/number map. There are three relevant calls:
3319 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3320 PCRE_INFO_NAMECOUNT yields the number of entries
3321 PCRE_INFO_NAMETABLE yields a pointer to the map.
3323 The map is a vector of fixed-size entries. The size of each entry depends on
3324 the length of the longest name used. The first two bytes of each entry are the
3325 group number, most significant byte first. There follows the corresponding
3326 name, zero terminated. The names are in alphabetical order.
3328 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3329 case instead of 255. Making it the same both with and without UTF-8 support
3330 means that the same test output works with both.
3332 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3333 calling malloc() with a zero argument.
3335 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3336 optimization. I've improved things by keeping a bitmap of backreferences with
3337 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3338 fact referenced, the optimization can be applied. It is unlikely that a
3339 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3340 the match to follow \n) will appear inside brackets with a number greater than
3341 31, but if it does, any back reference > 31 suppresses the optimization.
3343 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3344 of disabling numbered capturing parentheses. Any opening parenthesis that is
3345 not followed by ? behaves as if it were followed by ?: but named parentheses
3346 can still be used for capturing (and they will acquire numbers in the usual
3347 way).
3349 41. Redesigned the return codes from the match() function into yes/no/error so
3350 that errors can be passed back from deep inside the nested calls. A malloc
3351 failure while inside a recursive subpattern call now causes the
3352 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3354 42. It is now possible to set a limit on the number of times the match()
3355 function is called in a call to pcre_exec(). This facility makes it possible to
3356 limit the amount of recursion and backtracking, though not in a directly
3357 obvious way, because the match() function is used in a number of different
3358 circumstances. The count starts from zero for each position in the subject
3359 string (for non-anchored patterns). The default limit is, for compatibility, a
3360 large number, namely 10 000 000. You can change this in two ways:
3362 (a) When configuring PCRE before making, you can use --with-match-limit=n
3363 to set a default value for the compiled library.
3365 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3366 a different value is set. See 45 below.
3368 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3370 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3371 of things that can be changed at compile time. The first argument specifies
3372 what is wanted and the second points to where the information is to be placed.
3373 The current list of available information is:
3377 The output is an integer that is set to one if UTF-8 support is available;
3378 otherwise it is set to zero.
3382 The output is an integer that it set to the value of the code that is used for
3383 newline. It is either LF (10) or CR (13).
3387 The output is an integer that contains the number of bytes used for internal
3388 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3392 The output is an integer that contains the threshold above which the POSIX
3393 interface uses malloc() for output vectors. See item 31 above.
3397 The output is an unsigned integer that contains the default limit of the number
3398 of match() calls in a pcre_exec() execution. See 42 above.
3400 44. pcretest has been upgraded by the addition of the -C option. This causes it
3401 to extract all the available output from the new pcre_config() function, and to
3402 output it. The program then exits immediately.
3404 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3405 order to support additional features. One way would have been to define
3406 pcre_exec2() (for example) with extra arguments, but this would not have been
3407 extensible, and would also have required all calls to the original function to
3408 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3409 is used for passing in "extra" data from pcre_study().
3411 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3412 contains the following fields:
3414 flags a bitmap indicating which of the following fields are set
3415 study_data opaque data from pcre_study()
3416 match_limit a way of specifying a limit on match() calls for a specific
3417 call to pcre_exec()
3418 callout_data data for callouts (see 49 below)
3420 The flag bits are also defined in pcre.h, and are
3426 The pcre_study() function now returns one of these new pcre_extra blocks, with
3427 the actual study data pointed to by the study_data field, and the
3428 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3429 before. That is, this change is entirely upwards-compatible and requires no
3430 change to existing code.
3432 If you want to pass in additional data to pcre_exec(), you can either place it
3433 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3434 block.
3436 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3437 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3438 times with different match limits, until it finds the minimum value needed for
3439 pcre_exec() to complete. The value is then output. This can be instructive; for
3440 most simple matches the number is quite small, but for pathological cases it
3441 gets very large very quickly.
3443 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3444 returns the size of the data block pointed to by the study_data field in a
3445 pcre_extra block, that is, the value that was passed as the argument to
3446 pcre_malloc() when PCRE was getting memory in which to place the information
3447 created by pcre_study(). The fourth argument should point to a size_t variable.
3448 pcretest has been extended so that this information is shown after a successful
3449 pcre_study() call when information about the compiled regex is being displayed.
3451 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3452 because what follows is always an absolute path. (Later: it turns out that this
3453 is more than cosmetic for MinGW, because it doesn't like empty path
3454 components.)
3456 49. Some changes have been made to the callout feature (see 28 above):
3458 (i) A callout function now has three choices for what it returns:
3460 0 => success, carry on matching
3461 > 0 => failure at this point, but backtrack if possible
3462 < 0 => serious error, return this value from pcre_exec()
3464 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3465 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3466 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3467 use by callout functions. It will never be used by PCRE itself.
3469 (ii) The pcre_extra structure (see 45 above) has a void * field called
3470 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3471 pcre_callout_block structure has a field of the same name. The contents of
3472 the field passed in the pcre_extra structure are passed to the callout
3473 function in the corresponding field in the callout block. This makes it
3474 easier to use the same callout-containing regex from multiple threads. For
3475 testing, the pcretest program has a new data escape
3477 \C*n pass the number n (may be negative) as callout_data
3479 If the callout function in pcretest receives a non-zero value as
3480 callout_data, it returns that value.
3482 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3483 there were some redundant $(CFLAGS) in commands that are now specified as
3484 $(LINK), which already includes $(CFLAGS).
3486 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3487 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3488 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3489 one-byte characters throughout. Note that case-insensitive matching applies
3490 only to characters whose values are less than 256. PCRE doesn't support the
3491 notion of cases for higher-valued characters.
3493 (i) A character class whose characters are all within 0-255 is handled as
3494 a bit map, and the map is inverted for negative classes. Previously, a
3495 character > 255 always failed to match such a class; however it should
3496 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3498 (ii) A negated character class with a single character < 255 is coded as
3499 "not this character" (OP_NOT). This wasn't working properly when the test
3500 character was multibyte, either singly or repeated.
3502 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3503 mode, for example: \x{100}{2,3}.
3505 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3506 singly or repeated) now correctly test multibyte characters. However,
3507 PCRE doesn't recognize any characters with values greater than 255 as
3508 digits, spaces, or word characters. Such characters always match \D, \S,
3509 and \W, and never match \d, \s, or \w.
3511 (v) Classes may now contain characters and character ranges with values
3512 greater than 255. For example: [ab\x{100}-\x{400}].
3514 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3515 PCRE in UTF-8 mode.
3517 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3518 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3519 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3520 value.)
3522 53. The single man page has become too large. I have therefore split it up into
3523 a number of separate man pages. These also give rise to individual HTML pages;
3524 these are now put in a separate directory, and there is an index.html page that
3525 lists them all. Some hyperlinking between the pages has been installed.
3527 54. Added convenience functions for handling named capturing parentheses.
3529 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3530 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3531 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3532 are faulted.
3534 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3535 calling configure. These values are used when compiling the dftables.c program
3536 which is run to generate the source of the default character tables. They
3537 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3538 you will need to set these values.
3540 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3543 Version 3.9 02-Jan-02
3544 ---------------------
3546 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3548 2. If --disable-static was given, the building process failed when trying to
3549 build pcretest and pcregrep. (For some reason it was using libtool to compile
3550 them, which is not right, as they aren't part of the library.)
3553 Version 3.8 18-Dec-01
3554 ---------------------
3556 1. The experimental UTF-8 code was completely screwed up. It was packing the
3557 bytes in the wrong order. How dumb can you get?
3560 Version 3.7 29-Oct-01
3561 ---------------------
3563 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3564 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3565 this didn't happen under Solaris 8, where I normally test things.
3567 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3568 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3569 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3570 this fix an hour or so after the initial 3.7 release.)
3573 Version 3.6 23-Oct-01
3574 ---------------------
3576 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3577 offsets passed as NULL with zero offset count.
3579 2. The config.guess and config.sub files had not been updated when I moved to
3580 the latest autoconf.
3583 Version 3.5 15-Aug-01
3584 ---------------------
3586 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3587 had been forgotten.
3589 2. By using declared but undefined structures, we can avoid using "void"
3590 definitions in pcre.h while keeping the internal definitions of the structures
3591 private.
3593 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3594 user point of view, this means that both static and shared libraries are built
3595 by default, but this can be individually controlled. More of the work of
3596 handling this static/shared cases is now inside libtool instead of PCRE's make
3597 file.
3599 4. The pcretest utility is now installed along with pcregrep because it is
3600 useful for users (to test regexs) and by doing this, it automatically gets
3601 relinked by libtool. The documentation has been turned into a man page, so
3602 there are now .1, .txt, and .html versions in /doc.
3604 5. Upgrades to pcregrep:
3605 (i) Added long-form option names like gnu grep.
3606 (ii) Added --help to list all options with an explanatory phrase.
3607 (iii) Added -r, --recursive to recurse into sub-directories.
3608 (iv) Added -f, --file to read patterns from a file.
3610 6. pcre_exec() was referring to its "code" argument before testing that
3611 argument for NULL (and giving an error if it was NULL).
3613 7. Upgraded Makefile.in to allow for compiling in a different directory from
3614 the source directory.
3616 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3617 options bits, the pointer it was passed was to an int instead of to an unsigned
3618 long int. This mattered only on 64-bit systems.
3620 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3621 generated) instead of pcre.in, which it its source. Also made the same change
3622 in several of the .c files.
3624 10. A new release of gcc defines printf() as a macro, which broke pcretest
3625 because it had an ifdef in the middle of a string argument for printf(). Fixed
3626 by using separate calls to printf().
3628 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3629 script, to force use of CR or LF instead of \n in the source. On non-Unix
3630 systems, the value can be set in config.h.
3632 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3633 absolute limit. Changed the text of the error message to make this clear, and
3634 likewise updated the man page.
3636 13. The limit of 99 on the number of capturing subpatterns has been removed.
3637 The new limit is 65535, which I hope will not be a "real" limit.
3640 Version 3.4 22-Aug-00
3641 ---------------------
3643 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3645 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3648 Version 3.3 01-Aug-00
3649 ---------------------
3651 1. If an octal character was given, but the value was greater than \377, it
3652 was not getting masked to the least significant bits, as documented. This could
3653 lead to crashes in some systems.
3655 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3656 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3658 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3659 These just pass their arguments on to (pcre_free)(), but they are provided
3660 because some uses of PCRE bind it to non-C systems that can call its functions,
3661 but cannot call free() or pcre_free() directly.
3663 4. Add "make test" as a synonym for "make check". Corrected some comments in
3664 the Makefile.
3666 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3667 Makefile.
3669 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3670 command called pgrep for grepping around the active processes.
3672 7. Added the beginnings of support for UTF-8 character strings.
3674 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3675 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3676 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3677 out for the ar command.)
3680 Version 3.2 12-May-00
3681 ---------------------
3683 This is purely a bug fixing release.
3685 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3686 of ZA. This was just one example of several cases that could provoke this bug,
3687 which was introduced by change 9 of version 2.00. The code for breaking
3688 infinite loops after an iteration that matches an empty string was't working
3689 correctly.
3691 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3692 when matched against abbab (for example). After matching an empty string, it
3693 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3694 caused it to match further down the string than it should.
3696 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3697 was there because it doesn't seem to be needed, and it causes trouble on some
3698 systems, as it is not a Standard C header. It has been removed.
3700 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3701 were reported on the Macintosh. The changes were from
3703 while ((c = *(++ptr)) != 0 && c != '\n');
3704 to
3705 while ((c = *(++ptr)) != 0 && c != '\n') ;
3707 Totally extraordinary, but if that's what it takes...
3709 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3710 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3711 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3712 assumes the way PCRE uses memmove() (always moving upwards).
3714 6. PCRE is being used in one environment where strchr() is not available. There
3715 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3716 faster code anyway.
3719 Version 3.1 09-Feb-00
3720 ---------------------
3722 The only change in this release is the fixing of some bugs in Makefile.in for
3723 the "install" target:
3725 (1) It was failing to install pcreposix.h.
3727 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3730 Version 3.0 01-Feb-00
3731 ---------------------
3733 1. Add support for the /+ modifier to perltest (to output $` like it does in
3734 pcretest).
3736 2. Add support for the /g modifier to perltest.
3738 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3739 matches null strings.
3741 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3742 pattern. Perl treats empty patterns specially - it reuses the most recent
3743 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3744 effect.
3746 5. The POSIX interface was broken in that it was just handing over the POSIX
3747 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3748 required a bigger vector, with some working space on the end. This means that
3749 the POSIX wrapper now has to get and free some memory, and copy the results.
3751 6. Added some simple autoconf support, placing the test data and the
3752 documentation in separate directories, re-organizing some of the
3753 information files, and making it build pcre-config (a GNU standard). Also added
3754 libtool support for building PCRE as a shared library, which is now the
3755 default.
3757 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3758 09 are not valid octal constants. Single digits will be used for minor values
3759 less than 10.
3761 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3762 existing programs that set these in the POSIX interface can use PCRE without
3763 modification.
3765 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3766 return all that pcre_info() returns, plus additional data. The pcre_info()
3767 function is retained for compatibility, but is considered to be obsolete.
3769 10. Added experimental recursion feature (?R) to handle one common case that
3770 Perl 5.6 will be able to do with (?p{...}).
3772 11. Added support for POSIX character classes like [:alpha:], which Perl is
3773 adopting.
3776 Version 2.08 31-Aug-99
3777 ----------------------
3779 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3780 trying to match at the startoffset position, but instead was moving forward to
3781 the next newline as if a previous match had failed.
3783 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3784 and could get into a loop if a null string was matched other than at the start
3785 of the subject.
3787 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3788 be distinguished at compile time, and for completeness also added PCRE_DATE.
3790 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3791 in GnuWin32 environments.
3794 Version 2.07 29-Jul-99
3795 ----------------------
3797 1. The documentation is now supplied in plain text form and HTML as well as in
3798 the form of man page sources.
3800 2. C++ compilers don't like assigning (void *) values to other pointer types.
3801 In particular this affects malloc(). Although there is no problem in Standard
3802 C, I've put in casts to keep C++ compilers happy.
3804 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3805 should be (const char *).
3807 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3808 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3809 However, I haven't made this a standard facility. The documentation doesn't
3810 mention it, and the Makefile doesn't support it.
3812 5. The Makefile now contains an "install" target, with editable destinations at
3813 the top of the file. The pcretest program is not installed.
3815 6. pgrep -V now gives the PCRE version number and date.
3817 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3818 causing the entire string to be ignored, instead of just the last character.
3820 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3821 non-matching string, it can take a very, very long time, even for strings of
3822 quite modest length, because of the nested recursion. PCRE now does better in
3823 some of these cases. It does this by remembering the last required literal
3824 character in the pattern, and pre-searching the subject to ensure it is present
3825 before running the real match. In other words, it applies a heuristic to detect
3826 some types of certain failure quickly, and in the above example, if presented
3827 with a string that has no trailing " it gives "no match" very quickly.
3829 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3830 other alternatives are tried instead.
3833 Version 2.06 09-Jun-99
3834 ----------------------
3836 1. Change pcretest's output for amount of store used to show just the code
3837 space, because the remainder (the data block) varies in size between 32-bit and
3838 64-bit systems.
3840 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3841 start matching at. This allows lookbehinds to work when searching for multiple
3842 occurrences in a string.
3844 3. Added additional options to pcretest for testing multiple occurrences:
3846 /+ outputs the rest of the string that follows a match
3847 /g loops for multiple occurrences, using the new startoffset argument
3848 /G loops for multiple occurrences by passing an incremented pointer
3850 4. PCRE wasn't doing the "first character" optimization for patterns starting
3851 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3852 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3853 the letter 'x'. On long subject strings, this gives a significant speed-up.
3856 Version 2.05 21-Apr-99
3857 ----------------------
3859 1. Changed the type of magic_number from int to long int so that it works
3860 properly on 16-bit systems.
3862 2. Fixed a bug which caused patterns starting with .* not to work correctly
3863 when the subject string contained newline characters. PCRE was assuming
3864 anchoring for such patterns in all cases, which is not correct because .* will
3865 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3866 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3867 must be retried after every newline in the subject.
3870 Version 2.04 18-Feb-99
3871 ----------------------
3873 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3874 computation of the store needed to hold the pattern was incorrect (too large).
3875 If such patterns were nested a few deep, this could multiply and become a real
3876 problem.
3878 2. Added /M option to pcretest to show the memory requirement of a specific
3879 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3881 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3882 compiled in such a way that the backtracking after subsequent failure was
3883 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3884 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3887 Version 2.03 02-Feb-99
3888 ----------------------
3890 1. Fixed typo and small mistake in man page.
3892 2. Added 4th condition (GPL supersedes if conflict) and created separate
3893 LICENCE file containing the conditions.
3895 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3896 Perl, that is the internal \ allows the delimiter to be included in the
3897 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3898 the final delimiter, add \ to the end of the pattern (to test the error).
3900 4. Added the convenience functions for extracting substrings after a successful
3901 match. Updated pcretest to make it able to test these functions.
3904 Version 2.02 14-Jan-99
3905 ----------------------
3907 1. Initialized the working variables associated with each extraction so that
3908 their saving and restoring doesn't refer to uninitialized store.
3910 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3911 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3912 fix the problem.
3914 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3915 calls, and wasn't printing the correct value for compiling calls. Increased the
3916 default value of LOOPREPEAT, and the number of significant figures in the
3917 times.
3919 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3921 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3922 a building problem on Windows NT with a FAT file system.
3925 Version 2.01 21-Oct-98
3926 ----------------------
3928 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3929 to character tables built by pcre_maketables() in the current locale. If NULL
3930 is passed, the default tables are used.
3933 Version 2.00 24-Sep-98
3934 ----------------------
3936 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3937 it any more.
3939 2. Allow quantification of (?>) groups, and make it work correctly.
3941 3. The first character computation wasn't working for (?>) groups.
3943 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3944 end of the subject) and add 5.005's \z, which really does match only at the
3945 very end of the subject.
3947 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3949 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3950 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3951 localized options. All options to pcre_study() were also removed.
3953 7. Add other new features from 5.005:
3955 $(?<= positive lookbehind
3956 $(?<! negative lookbehind
3957 (?imsx-imsx) added the unsetting capability
3958 such a setting is global if at outer level; local otherwise
3959 (?imsx-imsx:) non-capturing groups with option setting
3960 (?(cond)re|re) conditional pattern matching
3962 A backreference to itself in a repeated group matches the previous
3963 captured string.
3965 8. General tidying up of studying (both automatic and via "study")
3966 consequential on the addition of new assertions.
3968 9. As in 5.005, unlimited repeated groups that could match an empty substring
3969 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3970 runtime if any iteration does actually match an empty substring.
3972 10. Include the RunTest script in the distribution.
3974 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3975 discrepancies, some of which were old and were also with respect to 5.004. They
3976 have now been fixed.
3979 Version 1.09 28-Apr-98
3980 ----------------------
3982 1. A negated single character class followed by a quantifier with a minimum
3983 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3984 program crashes, or just wrong answers. This did not apply to negated classes
3985 containing more than one character, or to minima other than one.
3988 Version 1.08 27-Mar-98
3989 ----------------------
3991 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3993 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3994 latter must appear before anything that relies on it in the pattern.
3997 Version 1.07 16-Feb-98
3998 ----------------------
4000 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4001 repeat of a potentially empty string).
4004 Version 1.06 23-Jan-98
4005 ----------------------
4007 1. Added Markus Oberhumer's little patches for C++.
4009 2. Literal strings longer than 255 characters were broken.
4012 Version 1.05 23-Dec-97
4013 ----------------------
4015 1. Negated character classes containing more than one character were failing if
4016 PCRE_CASELESS was set at run time.
4019 Version 1.04 19-Dec-97
4020 ----------------------
4022 1. Corrected the man page, where some "const" qualifiers had been omitted.
4024 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4025 input syntax.
4027 3. Fixed memory leak which occurred when a regex with back references was
4028 matched with an offsets vector that wasn't big enough. The temporary memory
4029 that is used in this case wasn't being freed if the match failed.
4031 4. Tidied pcretest to ensure it frees memory that it gets.
4033 5. Temporary memory was being obtained in the case where the passed offsets
4034 vector was exactly big enough.
4036 6. Corrected definition of offsetof() from change 5 below.
4038 7. I had screwed up change 6 below and broken the rules for the use of
4039 setjmp(). Now fixed.
4042 Version 1.03 18-Dec-97
4043 ----------------------
4045 1. A erroneous regex with a missing opening parenthesis was correctly
4046 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4047 on some systems.
4049 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4050 it was reported that one broken compiler failed on the former because "code" is
4051 also an independent variable.
4053 3. The erroneous regex a[]b caused an array overrun reference.
4055 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4056 fail on data ending with that character. (It was going on too far, and checking
4057 the next character, typically a binary zero.) This was specific to the
4058 optimized code for single-character negative classes.
4060 5. Added a contributed patch from the TIN world which does the following:
4062 + Add an undef for memmove, in case the the system defines a macro for it.
4064 + Add a definition of offsetof(), in case there isn't one. (I don't know
4065 the reason behind this - offsetof() is part of the ANSI standard - but
4066 it does no harm).
4068 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4069 most of the places where whitespace preceded '#'. I have given up and
4070 allowed the remaining 2 cases to be at the margin.
4072 + Rename some variables in pcre to eliminate shadowing. This seems very
4073 pedantic, but does no harm, of course.
4075 6. Moved the call to setjmp() into its own function, to get rid of warnings
4076 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4078 7. Constructs such as \d{8,} were compiling into the equivalent of
4079 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4080 outcome, but in this particular case used more store than had been allocated,
4081 which caused the bug to be discovered because it threw up an internal error.
4083 8. The debugging code in both pcre and pcretest for outputting the compiled
4084 form of a regex was going wrong in the case of back references followed by
4085 curly-bracketed repeats.
4088 Version 1.02 12-Dec-97
4089 ----------------------
4091 1. Typos in pcre.3 and comments in the source fixed.
4093 2. Applied a contributed patch to get rid of places where it used to remove
4094 'const' from variables, and fixed some signed/unsigned and uninitialized
4095 variable warnings.
4097 3. Added the "runtest" target to Makefile.
4099 4. Set default compiler flag to -O2 rather than just -O.
4102 Version 1.01 19-Nov-97
4103 ----------------------
4105 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4106 like /([ab]*)*/, that is, for classes with more than one character in them.
4108 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4109 as /((?>a*))*/ (a PCRE_EXTRA facility).
4112 Version 1.00 18-Nov-97
4113 ----------------------
4115 1. Added compile-time macros to support systems such as SunOS4 which don't have
4116 memmove() or strerror() but have other things that can be used instead.
4118 2. Arranged that "make clean" removes the executables.
4121 Version 0.99 27-Oct-97
4122 ----------------------
4124 1. Fixed bug in code for optimizing classes with only one character. It was
4125 initializing a 32-byte map regardless, which could cause it to run off the end
4126 of the memory it had got.
4128 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4131 Version 0.98 22-Oct-97
4132 ----------------------
4134 1. Fixed bug in code for handling temporary memory usage when there are more
4135 back references than supplied space in the ovector. This could cause segfaults.
4138 Version 0.97 21-Oct-97
4139 ----------------------
4141 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4143 2. Optimized negated single characters not to use a bit map.
4145 3. Brought error texts together as macro definitions; clarified some of them;
4146 fixed one that was wrong - it said "range out of order" when it meant "invalid
4147 escape sequence".
4149 4. Changed some char * arguments to const char *.
4151 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4153 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4154 pcretest.
4157 Version 0.96 16-Oct-97
4158 ----------------------
4160 1. Added a simple "pgrep" utility to the distribution.
4162 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4163 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4164 where "ddd" means "one or more decimal digits".
4166 3. Fixed serious bug. If a pattern had a back reference, but the call to
4167 pcre_exec() didn't supply a large enough ovector to record the related
4168 identifying subpattern, the match always failed. PCRE now remembers the number
4169 of the largest back reference, and gets some temporary memory in which to save
4170 the offsets during matching if necessary, in order to ensure that
4171 backreferences always work.
4173 4. Increased the compatibility with Perl in a number of ways:
4175 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4176 to request this handling. The option can be set at compile or exec time.
4178 (b) $ matches before a terminating newline by default; an option
4179 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4180 mode). The option can be set at compile or exec time.
4182 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4183 the same as Perl's. If the decimal number it represents is less than 10
4184 or there aren't that many previous left capturing parentheses, an octal
4185 escape is read. Inside a character class, it's always an octal escape,
4186 even if it is a single digit.
4188 (d) An escaped but undefined alphabetic character is taken as a literal,
4189 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4190 escapes.
4192 (e) {0} is now permitted. (The previous item is removed from the compiled
4193 pattern).
4195 5. Changed all the names of code files so that the basic parts are no longer
4196 than 10 characters, and abolished the teeny "globals.c" file.
4198 6. Changed the handling of character classes; they are now done with a 32-byte
4199 bit map always.
4201 7. Added the -d and /D options to pcretest to make it possible to look at the
4202 internals of compilation without having to recompile pcre.
4205 Version 0.95 23-Sep-97
4206 ----------------------
4208 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4209 \x20 at the start of a run of normal characters. These were being treated as
4210 real characters, instead of the source characters being re-checked.
4213 Version 0.94 18-Sep-97
4214 ----------------------
4216 1. The functions are now thread-safe, with the caveat that the global variables
4217 containing pointers to malloc() and free() or alternative functions are the
4218 same for all threads.
4220 2. Get pcre_study() to generate a bitmap of initial characters for non-
4221 anchored patterns when this is possible, and use it if passed to pcre_exec().
4224 Version 0.93 15-Sep-97
4225 ----------------------
4227 1. /(b)|(:+)/ was computing an incorrect first character.
4229 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4230 but not actually doing anything yet.
4232 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4233 as Perl does (e.g. [-az] or [az-]).
4235 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4236 all possible positions.
4238 5. Split up into different modules to avoid including unneeded functions in a
4239 compiled binary. However, compile and exec are still in one module. The "study"
4240 function is split off.
4242 6. The character tables are now in a separate module whose source is generated
4243 by an auxiliary program - but can then be edited by hand if required. There are
4244 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4245 toupper() in the code.
4247 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4248 make them global. Abolish the function for setting them, as the caller can now
4249 set them directly.
4252 Version 0.92 11-Sep-97
4253 ----------------------
4255 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4256 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4258 2. Caseless matching was not working in character classes if the characters in
4259 the pattern were in upper case.
4261 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4263 4. Make PCRE_ANCHORED public and accept as a compile option.
4265 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4266 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4267 pass them.
4269 6. Give an error if bad option bits passed at compile or run time.
4271 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4272 pcretest to cause it to pass that flag.
4274 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4275 options, and the first character, if set.
4277 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4280 Version 0.91 10-Sep-97
4281 ----------------------
4283 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4284 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4286 2. PCRE was looping on encountering an indefinitely repeated back reference to
4287 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4288 Perl does - treats the match as successful.
4290 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5