ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 643 - (show annotations)
Fri Jul 29 15:56:39 2011 UTC (10 years, 1 month ago) by ph10
File size: 209984 byte(s)
Error occurred while calculating annotation data.
Allow all characters except closing parens in MARK:NAME etc.
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).
169 32. If --only-matching or --colour was set on a pcregrep call whose pattern
170 had alternative anchored branches, the search for a second match in a line
171 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
172 matched the line "0102" twice. The same bug affected patterns that started
173 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
174 twice.
176 33. Previously, PCRE did not allow quantification of assertions. However, Perl
177 does, and because of capturing effects, quantifying parenthesized
178 assertions may at times be useful. Quantifiers are now allowed for
179 parenthesized assertions.
181 34. A minor code tidy in pcre_compile() when checking options for \R usage.
183 35. \g was being checked for fancy things in a character class, when it should
184 just be a literal "g".
186 36. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
187 appearance of a nested POSIX class supersedes an apparent external class.
188 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
189 unescaped square brackets may also appear as part of class names. For
190 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
191 more like Perl.
193 37. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
194 was because it thought it was \N{name}, which is not supported).
196 38. Add minix to OS list not supporting the -S option in pcretest.
198 39. PCRE tries to detect cases of infinite recursion at compile time, but it
199 cannot analyze patterns in sufficient detail to catch mutual recursions
200 such as ((?1))((?2)). There is now a runtime test that gives an error if a
201 subgroup is called recursively as a subpattern for a second time at the
202 same position in the subject string. In previous releases this might have
203 been caught by the recursion limit, or it might have run out of stack.
205 40. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
206 happen only once. PCRE was, however incorrectly giving a compile time error
207 "recursive call could loop indefinitely" because it cannot analyze the
208 pattern in sufficient detail. The compile time test no longer happens when
209 PCRE is compiling a conditional subpattern, but actual runaway loops are
210 now caught at runtime (see 39 above).
212 41. It seems that Perl allows any characters other than a closing parenthesis
213 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
214 has been changed to be the same.
217 Version 8.12 15-Jan-2011
218 ------------------------
220 1. Fixed some typos in the markup of the man pages, and wrote a script that
221 checks for such things as part of the documentation building process.
223 2. On a big-endian 64-bit system, pcregrep did not correctly process the
224 --match-limit and --recursion-limit options (added for 8.11). In
225 particular, this made one of the standard tests fail. (The integer value
226 went into the wrong half of a long int.)
228 3. If the --colour option was given to pcregrep with -v (invert match), it
229 did strange things, either producing crazy output, or crashing. It should,
230 of course, ignore a request for colour when reporting lines that do not
231 match.
233 4. Another pcregrep bug caused similar problems if --colour was specified with
234 -M (multiline) and the pattern match finished with a line ending.
236 5. In pcregrep, when a pattern that ended with a literal newline sequence was
237 matched in multiline mode, the following line was shown as part of the
238 match. This seems wrong, so I have changed it.
240 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
241 the check for further matches in the same line (so they could be coloured)
242 to overrun the end of the current line. If another match was found, it was
243 incorrectly shown (and then shown again when found in the next line).
245 7. If pcregrep was compiled under Windows, there was a reference to the
246 function pcregrep_exit() before it was defined. I am assuming this was
247 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
248 reported by a user. I've moved the definition above the reference.
251 Version 8.11 10-Dec-2010
252 ------------------------
254 1. (*THEN) was not working properly if there were untried alternatives prior
255 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
256 backtracked to try for "b" instead of moving to the next alternative branch
257 at the same level (in this case, to look for "c"). The Perl documentation
258 is clear that when (*THEN) is backtracked onto, it goes to the "next
259 alternative in the innermost enclosing group".
261 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
262 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
263 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
264 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
265 (*THEN).
267 3. If \s appeared in a character class, it removed the VT character from
268 the class, even if it had been included by some previous item, for example
269 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
270 of \s, but is part of the POSIX "space" class.)
272 4. A partial match never returns an empty string (because you can always
273 match an empty string at the end of the subject); however the checking for
274 an empty string was starting at the "start of match" point. This has been
275 changed to the "earliest inspected character" point, because the returned
276 data for a partial match starts at this character. This means that, for
277 example, /(?<=abc)def/ gives a partial match for the subject "abc"
278 (previously it gave "no match").
280 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
281 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
282 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
283 has an implication that the given string is incomplete (because a partial
284 match is preferred over a full match). For this reason, these items now
285 give a partial match in this situation. [Aside: previously, the one case
286 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
287 match rather than a full match, which was wrong by the old rules, but is
288 now correct.]
290 6. There was a bug in the handling of #-introduced comments, recognized when
291 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
292 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
293 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
294 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
295 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
296 places in pcre_compile().
298 7. Related to (6) above, when pcre_compile() was skipping #-introduced
299 comments when looking ahead for named forward references to subpatterns,
300 the only newline sequence it recognized was NL. It now handles newlines
301 according to the set newline convention.
303 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
304 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
305 cater for a lack of strerror(). These oversights have been fixed.
307 9. Added --match-limit and --recursion-limit to pcregrep.
309 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
311 11. When the -o option was used, pcregrep was setting a return code of 1, even
312 when matches were found, and --line-buffered was not being honoured.
314 12. Added an optional parentheses number to the -o and --only-matching options
315 of pcregrep.
317 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
318 can match an empty string. The code to do it in pcretest and pcredemo
319 needed fixing:
321 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
322 only one byte after an empty string match just before CRLF (this case
323 just got forgotten; "any" and "anycrlf" were OK).
325 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
326 mode when an empty string match preceded an ASCII character followed by
327 a non-ASCII character. (The code for advancing by one character rather
328 than one byte was nonsense.)
330 (c) The pcredemo.c sample program did not have any code at all to handle
331 the cases when CRLF is a valid newline sequence.
333 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
334 as a starting offset was within the subject string. There is now a new
335 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
336 negative or greater than the length of the string. In order to test this,
337 pcretest is extended to allow the setting of negative starting offsets.
339 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
340 starting offset points to the beginning of a UTF-8 character was
341 unnecessarily clumsy. I tidied it up.
343 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
344 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
346 17. Nobody had reported that the --include_dir option, which was added in
347 release 7.7 should have been called --include-dir (hyphen, not underscore)
348 for compatibility with GNU grep. I have changed it to --include-dir, but
349 left --include_dir as an undocumented synonym, and the same for
350 --exclude-dir, though that is not available in GNU grep, at least as of
351 release 2.5.4.
353 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
354 characters from a string of bytes) have been redefined so as not to use
355 loops, in order to improve performance in some environments. At the same
356 time, I abstracted some of the common code into auxiliary macros to save
357 repetition (this should not affect the compiled code).
359 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
360 compile-time error is now given if \c is not followed by an ASCII
361 character, that is, a byte less than 128. (In EBCDIC mode, the code is
362 different, and any byte value is allowed.)
364 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
365 START_OPTIMIZE option, which is now allowed at compile time - but just
366 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
367 to pcregrep and other applications that have no direct access to PCRE
368 options. The new /Y option in pcretest sets this option when calling
369 pcre_compile().
371 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
372 back references. Groups containing recursive back references were forced to
373 be atomic by that change, but in the case of named groups, the amount of
374 memory required was incorrectly computed, leading to "Failed: internal
375 error: code overflow". This has been fixed.
377 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
378 pcretest.c, to avoid build problems in some Borland environments.
381 Version 8.10 25-Jun-2010
382 ------------------------
384 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
385 THEN.
387 2. (*ACCEPT) was not working when inside an atomic group.
389 3. Inside a character class, \B is treated as a literal by default, but
390 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
391 causes the error). The code is unchanged, but I tidied the documentation.
393 4. Inside a character class, PCRE always treated \R and \X as literals,
394 whereas Perl faults them if its -w option is set. I have changed PCRE so
395 that it faults them when PCRE_EXTRA is set.
397 5. Added support for \N, which always matches any character other than
398 newline. (It is the same as "." when PCRE_DOTALL is not set.)
400 6. When compiling pcregrep with newer versions of gcc which may have
401 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
402 declared with attribute warn_unused_result" were given. Just casting the
403 result to (void) does not stop the warnings; a more elaborate fudge is
404 needed. I've used a macro to implement this.
406 7. Minor change to pcretest.c to avoid a compiler warning.
408 8. Added four artifical Unicode properties to help with an option to make
409 \s etc use properties (see next item). The new properties are: Xan
410 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
412 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
413 use Unicode properties. (*UCP) at the start of a pattern can be used to set
414 this option. Modified pcretest to add /W to test this facility. Added
415 REG_UCP to make it available via the POSIX interface.
417 10. Added --line-buffered to pcregrep.
419 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
420 studied, and the match started with a letter with a code point greater than
421 127 whose first byte was different to the first byte of the other case of
422 the letter, the other case of this starting letter was not recognized
423 (#976).
425 12. If a pattern that was studied started with a repeated Unicode property
426 test, for example, \p{Nd}+, there was the theoretical possibility of
427 setting up an incorrect bitmap of starting bytes, but fortunately it could
428 not have actually happened in practice until change 8 above was made (it
429 added property types that matched character-matching opcodes).
431 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
432 possible starting bytes for non-anchored patterns.
434 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
435 \R, and also a number of cases that involve Unicode properties, both
436 explicit and implicit when PCRE_UCP is set.
438 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
439 input, it could crash or give wrong results if characters with values
440 greater than 0xc0 were present in the subject string. (Detail: it assumed
441 UTF-8 input when processing these items.)
443 16. Added a lot of (int) casts to avoid compiler warnings in systems where
444 size_t is 64-bit (#991).
446 17. Added a check for running out of memory when PCRE is compiled with
447 --disable-stack-for-recursion (#990).
449 18. If the last data line in a file for pcretest does not have a newline on
450 the end, a newline was missing in the output.
452 19. The default pcre_chartables.c file recognizes only ASCII characters (values
453 less than 128) in its various bitmaps. However, there is a facility for
454 generating tables according to the current locale when PCRE is compiled. It
455 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
456 characters, are recognized by isspace() and therefore were getting set in
457 these tables, and indeed these tables seem to approximate to ISO 8859. This
458 caused a problem in UTF-8 mode when pcre_study() was used to create a list
459 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
460 which of course cannot start UTF-8 characters. I have changed the code so
461 that only real ASCII characters (less than 128) and the correct starting
462 bytes for UTF-8 encodings are set for characters greater than 127 when in
463 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
464 altogether.)
466 20. Added the /T option to pcretest so as to be able to run tests with non-
467 standard character tables, thus making it possible to include the tests
468 used for 19 above in the standard set of tests.
470 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
471 reference to a subpattern the other side of a comment that contains an
472 opening parenthesis caused either an internal compiling error, or a
473 reference to the wrong subpattern.
476 Version 8.02 19-Mar-2010
477 ------------------------
479 1. The Unicode data tables have been updated to Unicode 5.2.0.
481 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
482 configured.
484 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
485 original author of that file, following a query about its status.
487 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
488 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
490 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
491 quantifier applied to a forward-referencing subroutine call, could compile
492 incorrect code or give the error "internal error: previously-checked
493 referenced subpattern not found".
495 6. Both MS Visual Studio and Symbian OS have problems with initializing
496 variables to point to external functions. For these systems, therefore,
497 pcre_malloc etc. are now initialized to local functions that call the
498 relevant global functions.
500 7. There were two entries missing in the vectors called coptable and poptable
501 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
502 I've fixed the data, and added a kludgy way of testing at compile time that
503 the lengths are correct (equal to the number of opcodes).
505 8. Following on from 7, I added a similar kludge to check the length of the
506 eint vector in pcreposix.c.
508 9. Error texts for pcre_compile() are held as one long string to avoid too
509 much relocation at load time. To find a text, the string is searched,
510 counting zeros. There was no check for running off the end of the string,
511 which could happen if a new error number was added without updating the
512 string.
514 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
516 11. \K was not working if it appeared in an atomic group or in a group that
517 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
518 \K is "not well defined" if used in an assertion. PCRE now accepts it if
519 the assertion is positive, but not if it is negative.
521 12. Change 11 fortuitously reduced the size of the stack frame used in the
522 "match()" function of pcre_exec.c by one pointer. Forthcoming
523 implementation of support for (*MARK) will need an extra pointer on the
524 stack; I have reserved it now, so that the stack frame size does not
525 decrease.
527 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
528 item in branch that calls a recursion is a subroutine call - as in the
529 second branch in the above example - was incorrectly given the compile-
530 time error "recursive call could loop indefinitely" because pcre_compile()
531 was not correctly checking the subroutine for matching a non-empty string.
533 14. The checks for overrunning compiling workspace could trigger after an
534 overrun had occurred. This is a "should never occur" error, but it can be
535 triggered by pathological patterns such as hundreds of nested parentheses.
536 The checks now trigger 100 bytes before the end of the workspace.
538 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
541 Version 8.01 19-Jan-2010
542 ------------------------
544 1. If a pattern contained a conditional subpattern with only one branch (in
545 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
546 computed the wrong minimum data length (which is of course zero for such
547 subpatterns). This could cause incorrect "no match" results.
549 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
550 the pattern is reset in the first branch, pcre_compile() failed with
551 "internal error: code overflow at offset...". This happened only when
552 the reset was to the original external option setting. (An optimization
553 abstracts leading options settings into an external setting, which was the
554 cause of this.)
556 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
557 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
558 assertion pattern did not match (meaning that the assertion was true), it
559 was incorrectly treated as false if the SKIP had been reached during the
560 matching. This also applied to assertions used as conditions.
562 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
563 assertion subpattern, including such a pattern used as a condition,
564 unpredictable results occurred, instead of the error return
567 5. The C++ GlobalReplace function was not working like Perl for the special
568 situation when an empty string is matched. It now does the fancy magic
569 stuff that is necessary.
571 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
572 removed. (These were left over from very, very early versions of PCRE.)
574 7. Some cosmetic changes to the code to make life easier when compiling it
575 as part of something else:
577 (a) Change DEBUG to PCRE_DEBUG.
579 (b) In pcre_compile(), rename the member of the "branch_chain" structure
580 called "current" as "current_branch", to prevent a collision with the
581 Linux macro when compiled as a kernel module.
583 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
584 prevent a collision with the Linux macro when compiled as a kernel
585 module.
587 8. In pcre_compile() there are some checks for integer overflows that used to
588 cast potentially large values to (double). This has been changed to that
589 when building, a check for int64_t is made, and if it is found, it is used
590 instead, thus avoiding the use of floating point arithmetic. (There is no
591 other use of FP in PCRE.) If int64_t is not found, the fallback is to
592 double.
594 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
595 2005 (difference between two addresses compared to an unsigned value).
597 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
598 custom one, because of the following reported problem in Windows:
600 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
601 under Win32.
602 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
603 therefore missing the function definition.
604 - The compiler thus generates a "C" signature for the test function.
605 - The linker fails to find the "C" function.
606 - PCRE fails to configure if asked to do so against libbz2.
608 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
609 messages were output:
611 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
612 rerunning libtoolize, to keep the correct libtool macros in-tree.
613 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
615 I have done both of these things.
617 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
618 most of the time, it *can* run out if it is given a pattern that contains a
619 runaway infinite recursion. I updated the discussion in the pcrestack man
620 page.
622 13. Now that we have gone to the x.xx style of version numbers, the minor
623 version may start with zero. Using 08 or 09 is a bad idea because users
624 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
625 interpreted as invalid octal numbers. I've updated the previous comment in
626 configure.ac, and also added a check that gives an error if 08 or 09 are
627 used.
629 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
630 causing partial matching to fail when the end of the subject matched \W
631 in a UTF-8 pattern where \W was quantified with a minimum of 3.
633 15. There were some discrepancies between the declarations in pcre_internal.h
634 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
635 their definitions. The declarations used "const uschar *" and the
636 definitions used USPTR. Even though USPTR is normally defined as "const
637 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
638 reported that: "This difference in casting confuses some C++ compilers, for
639 example, SunCC recognizes above declarations as different functions and
640 generates broken code for hbpcre." I have changed the declarations to use
641 USPTR.
643 16. GNU libtool is named differently on some systems. The autogen.sh script now
644 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
645 (FreeBSD).
647 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
648 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
649 comment: "Figure out how to create a longlong from a string: strtoll and
650 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
651 instance, but it only takes 2 args instead of 3!"
653 18. A subtle bug concerned with back references has been fixed by a change of
654 specification, with a corresponding code fix. A pattern such as
655 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
656 refers, was giving matches when it shouldn't. For example, xa=xaaa would
657 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
658 same bug. Such groups have to be quantified to be useful, or contained
659 inside another quantified group. (If there's no repetition, the reference
660 can never match.) The problem arises because, having left the group and
661 moved on to the rest of the pattern, a later failure that backtracks into
662 the group uses the captured value from the final iteration of the group
663 rather than the correct earlier one. I have fixed this in PCRE by forcing
664 any group that contains a reference to itself to be an atomic group; that
665 is, there cannot be any backtracking into it once it has completed. This is
666 similar to recursive and subroutine calls.
669 Version 8.00 19-Oct-09
670 ----------------------
672 1. The table for translating pcre_compile() error codes into POSIX error codes
673 was out-of-date, and there was no check on the pcre_compile() error code
674 being within the table. This could lead to an OK return being given in
675 error.
677 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
678 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
679 in a Windows environment.
681 3. The pcregrep --count option prints the count for each file even when it is
682 zero, as does GNU grep. However, pcregrep was also printing all files when
683 --files-with-matches was added. Now, when both options are given, it prints
684 counts only for those files that have at least one match. (GNU grep just
685 prints the file name in this circumstance, but including the count seems
686 more useful - otherwise, why use --count?) Also ensured that the
687 combination -clh just lists non-zero counts, with no names.
689 4. The long form of the pcregrep -F option was incorrectly implemented as
690 --fixed_strings instead of --fixed-strings. This is an incompatible change,
691 but it seems right to fix it, and I didn't think it was worth preserving
692 the old behaviour.
694 5. The command line items --regex=pattern and --regexp=pattern were not
695 recognized by pcregrep, which required --regex pattern or --regexp pattern
696 (with a space rather than an '='). The man page documented the '=' forms,
697 which are compatible with GNU grep; these now work.
699 6. No libpcreposix.pc file was created for pkg-config; there was just
700 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
702 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
703 when UCP support is not needed, by modifying the Python script that
704 generates it from Unicode data files. This should not matter if the module
705 is correctly used as a library, but I received one complaint about 50K of
706 unwanted data. My guess is that the person linked everything into his
707 program rather than using a library. Anyway, it does no harm.
709 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
710 was a minimum greater than 1 for a wide character in a possessive
711 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
712 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
713 character. Chaos in the form of incorrect output or a compiling loop could
714 result.
716 9. The restrictions on what a pattern can contain when partial matching is
717 requested for pcre_exec() have been removed. All patterns can now be
718 partially matched by this function. In addition, if there are at least two
719 slots in the offset vector, the offset of the earliest inspected character
720 for the match and the offset of the end of the subject are set in them when
721 PCRE_ERROR_PARTIAL is returned.
723 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
724 synonymous with PCRE_PARTIAL, for backwards compatibility, and
725 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
726 and may be more useful for multi-segment matching.
728 11. Partial matching with pcre_exec() is now more intuitive. A partial match
729 used to be given if ever the end of the subject was reached; now it is
730 given only if matching could not proceed because another character was
731 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
732 string "Z", which now yields "no match" instead of "partial match". In the
733 case of pcre_dfa_exec(), "no match" is given if every matching path for the
734 final character ended with (*FAIL).
736 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
737 if the pattern had a "must contain" character that was already found in the
738 earlier partial match, unless partial matching was again requested. For
739 example, with the pattern /dog.(body)?/, the "must contain" character is
740 "g". If the first part-match was for the string "dog", restarting with
741 "sbody" failed. This bug has been fixed.
743 13. The string returned by pcre_dfa_exec() after a partial match has been
744 changed so that it starts at the first inspected character rather than the
745 first character of the match. This makes a difference only if the pattern
746 starts with a lookbehind assertion or \b or \B (\K is not supported by
747 pcre_dfa_exec()). It's an incompatible change, but it makes the two
748 matching functions compatible, and I think it's the right thing to do.
750 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
751 so that the demonstration program is easily available in environments where
752 PCRE has not been installed from source.
754 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
755 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
756 library.
758 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
759 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
760 is not the first non-POSIX option to be added. Clearly some people find
761 these options useful.
763 17. If a caller to the POSIX matching function regexec() passes a non-zero
764 value for nmatch with a NULL value for pmatch, the value of
765 nmatch is forced to zero.
767 18. RunGrepTest did not have a test for the availability of the -u option of
768 the diff command, as RunTest does. It now checks in the same way as
769 RunTest, and also checks for the -b option.
771 19. If an odd number of negated classes containing just a single character
772 interposed, within parentheses, between a forward reference to a named
773 subpattern and the definition of the subpattern, compilation crashed with
774 an internal error, complaining that it could not find the referenced
775 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
776 [The bug was that it was starting one character too far in when skipping
777 over the character class, thus treating the ] as data rather than
778 terminating the class. This meant it could skip too much.]
780 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
781 /g option in pcretest when the pattern contains \K, which makes it possible
782 to have an empty string match not at the start, even when the pattern is
783 anchored. Updated pcretest and pcredemo to use this option.
785 21. If the maximum number of capturing subpatterns in a recursion was greater
786 than the maximum at the outer level, the higher number was returned, but
787 with unset values at the outer level. The correct (outer level) value is
788 now given.
790 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
791 PCRE did not set those parentheses (unlike Perl). I have now found a way to
792 make it do so. The string so far is captured, making this feature
793 compatible with Perl.
795 23. The tests have been re-organized, adding tests 11 and 12, to make it
796 possible to check the Perl 5.10 features against Perl 5.10.
798 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
799 pattern matches a fixed length string. PCRE did not allow this; now it
800 does. Neither allows recursion.
802 25. I finally figured out how to implement a request to provide the minimum
803 length of subject string that was needed in order to match a given pattern.
804 (It was back references and recursion that I had previously got hung up
805 on.) This code has now been added to pcre_study(); it finds a lower bound
806 to the length of subject needed. It is not necessarily the greatest lower
807 bound, but using it to avoid searching strings that are too short does give
808 some useful speed-ups. The value is available to calling programs via
809 pcre_fullinfo().
811 26. While implementing 25, I discovered to my embarrassment that pcretest had
812 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
813 study optimizations had never been tested with that matching function.
814 Oops. What is worse, even when it was passed study data, there was a bug in
815 pcre_dfa_exec() that meant it never actually used it. Double oops. There
816 were also very few tests of studied patterns with pcre_dfa_exec().
818 27. If (?| is used to create subpatterns with duplicate numbers, they are now
819 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
820 on the other side of the coin, they are no longer allowed to have different
821 names, because these cannot be distinguished in PCRE, and this has caused
822 confusion. (This is a difference from Perl.)
824 28. When duplicate subpattern names are present (necessarily with different
825 numbers, as required by 27 above), and a test is made by name in a
826 conditional pattern, either for a subpattern having been matched, or for
827 recursion in such a pattern, all the associated numbered subpatterns are
828 tested, and the overall condition is true if the condition is true for any
829 one of them. This is the way Perl works, and is also more like the way
830 testing by number works.
833 Version 7.9 11-Apr-09
834 ---------------------
836 1. When building with support for bzlib/zlib (pcregrep) and/or readline
837 (pcretest), all targets were linked against these libraries. This included
838 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
839 libraries. This caused unwanted dependencies to be created. This problem
840 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
841 pcretest is linked with readline.
843 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
844 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
845 moved outside it again, because FALSE and TRUE are already defined in AIX,
846 but BOOL is not.
848 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
849 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
851 4. The pcregrep documentation said spaces were inserted as well as colons (or
852 hyphens) following file names and line numbers when outputting matching
853 lines. This is not true; no spaces are inserted. I have also clarified the
854 wording for the --colour (or --color) option.
856 5. In pcregrep, when --colour was used with -o, the list of matching strings
857 was not coloured; this is different to GNU grep, so I have changed it to be
858 the same.
860 6. When --colo(u)r was used in pcregrep, only the first matching substring in
861 each matching line was coloured. Now it goes on to look for further matches
862 of any of the test patterns, which is the same behaviour as GNU grep.
864 7. A pattern that could match an empty string could cause pcregrep to loop; it
865 doesn't make sense to accept an empty string match in pcregrep, so I have
866 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
867 seems to be how GNU grep behaves.
869 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
870 start or after a newline", because the conditional assertion was not being
871 correctly handled. The rule now is that both the assertion and what follows
872 in the first alternative must satisfy the test.
874 9. If auto-callout was enabled in a pattern with a conditional group whose
875 condition was an assertion, PCRE could crash during matching, both with
876 pcre_exec() and pcre_dfa_exec().
878 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
879 used for matching.
881 11. Unicode property support in character classes was not working for
882 characters (bytes) greater than 127 when not in UTF-8 mode.
884 12. Added the -M command line option to pcretest.
886 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
888 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
890 16. Added comments and documentation about mis-use of no_arg in the C++
891 wrapper.
893 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
894 from Martin Jerabek that uses macro names for all relevant character and
895 string constants.
897 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
898 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
899 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
900 these, but not everybody uses configure.
902 19. A conditional group that had only one branch was not being correctly
903 recognized as an item that could match an empty string. This meant that an
904 enclosing group might also not be so recognized, causing infinite looping
905 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
906 with the subject "ab", where knowledge that the repeated group can match
907 nothing is needed in order to break the loop.
909 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
910 exec(), but without supplying a callout function, matching went wrong.
912 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
913 leak if the size of the offset vector was greater than 30. When the vector
914 is smaller, the saved offsets during recursion go onto a local stack
915 vector, but for larger vectors malloc() is used. It was failing to free
916 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
917 error, in fact).
919 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
920 heapframe that is used only when UTF-8 support is enabled. This caused no
921 problem, but was untidy.
923 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
924 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
925 included within another project.
927 24. Steven Van Ingelgem's patches to add more options to the CMake support,
928 slightly modified by me:
930 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
931 not building pcregrep.
933 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
934 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
936 25. Forward references, both numeric and by name, in patterns that made use of
937 duplicate group numbers, could behave incorrectly or give incorrect errors,
938 because when scanning forward to find the reference group, PCRE was not
939 taking into account the duplicate group numbers. A pattern such as
940 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
942 26. Changed a few more instances of "const unsigned char *" to USPTR, making
943 the feature of a custom pointer more persuasive (as requested by a user).
945 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
946 pcretest.c, inside #ifndefs, because it seems they are sometimes already
947 pre-defined.
949 28. Added support for (*UTF8) at the start of a pattern.
951 29. Arrange for flags added by the "release type" setting in CMake to be shown
952 in the configuration summary.
955 Version 7.8 05-Sep-08
956 ---------------------
958 1. Replaced UCP searching code with optimized version as implemented for Ad
959 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
960 stage table and inline lookup instead of a function, giving speed ups of 2
961 to 5 times on some simple patterns that I tested. Permission was given to
962 distribute the MultiStage2.py script that generates the tables (it's not in
963 the tarball, but is in the Subversion repository).
965 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
966 scripts.
968 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
969 a group with a zero qualifier. The result of the study could be incorrect,
970 or the function might crash, depending on the pattern.
972 4. Caseless matching was not working for non-ASCII characters in back
973 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
974 It now works when Unicode Property Support is available.
976 5. In pcretest, an escape such as \x{de} in the data was always generating
977 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
978 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
979 truncation.
981 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
983 7. Added two (int) casts to pcregrep when printing the difference of two
984 pointers, in case they are 64-bit values.
986 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
987 test 2 if it fails.
989 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
990 and a #define of that name to empty if it is not externally set. This is to
991 allow users of MSVC to set it if necessary.
993 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
994 the convenience functions in the pcre_get.c source file.
996 11. An option change at the start of a pattern that had top-level alternatives
997 could cause overwriting and/or a crash. This command provoked a crash in
998 some environments:
1000 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1002 This potential security problem was recorded as CVE-2008-2371.
1004 12. For a pattern where the match had to start at the beginning or immediately
1005 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1006 pcre_dfa_exec() could read past the end of the passed subject if there was
1007 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1008 pcretest so that it places the subject at the end of its malloc-ed buffer.
1010 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1011 exec() might read past the end of the data buffer in UTF-8 mode.
1013 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1014 the data contained the byte 0x85 as part of a UTF-8 character within its
1015 first line. This applied both to normal and DFA matching.
1017 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1018 /^[^d]*?$/8 failed to match "abc".
1020 16. Added a missing copyright notice to pcrecpp_internal.h.
1022 17. Make it more clear in the documentation that values returned from
1023 pcre_exec() in ovector are byte offsets, not character counts.
1025 18. Tidied a few places to stop certain compilers from issuing warnings.
1027 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1028 supplied by Stefan Weber. I made a further small update for 7.8 because
1029 there is a change of source arrangements: the pcre_searchfuncs.c module is
1030 replaced by pcre_ucd.c.
1033 Version 7.7 07-May-08
1034 ---------------------
1036 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1037 a string to a long long, pretend we don't even have a long long." This is
1038 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1040 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1041 pre-7.6 versions, which defined a global no_arg variable instead of putting
1042 it in the RE class. (See also #8 below.)
1044 3. Remove a line of dead code, identified by coverity and reported by Nuno
1045 Lopes.
1047 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1049 (1) The include/exclude patterns were being applied to the whole pathnames
1050 of files, instead of just to the final components.
1052 (2) If there was more than one level of directory, the subdirectories were
1053 skipped unless they satisfied the include/exclude conditions. This is
1054 inconsistent with GNU grep (and could even be seen as contrary to the
1055 pcregrep specification - which I improved to make it absolutely clear).
1056 The action now is always to scan all levels of directory, and just
1057 apply the include/exclude patterns to regular files.
1059 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1060 --exclude_dir in the tests to avoid scanning .svn directories.
1062 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1063 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1064 doesn't support NULs in patterns.
1066 7. Added some missing "const"s to declarations of static tables in
1067 pcre_compile.c and pcre_dfa_exec.c.
1069 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1070 caused by fix #2 above. (Subsequently also a second patch to fix the
1071 first patch. And a third patch - this was a messy problem.)
1073 9. Applied Craig's patch to remove the use of push_back().
1075 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1076 matching function regexec().
1078 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1079 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1080 references. PCRE supports relative numbers with this syntax (I don't think
1081 Oniguruma does).
1083 12. Previously, a group with a zero repeat such as (...){0} was completely
1084 omitted from the compiled regex. However, this means that if the group
1085 was called as a subroutine from elsewhere in the pattern, things went wrong
1086 (an internal error was given). Such groups are now left in the compiled
1087 pattern, with a new opcode that causes them to be skipped at execution
1088 time.
1090 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1091 to the way PCRE behaves:
1093 (a) A lone ] character is dis-allowed (Perl treats it as data).
1095 (b) A back reference to an unmatched subpattern matches an empty string
1096 (Perl fails the current match path).
1098 (c) A data ] in a character class must be notated as \] because if the
1099 first data character in a class is ], it defines an empty class. (In
1100 Perl it is not possible to have an empty class.) The empty class []
1101 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1102 The negative empty class [^] matches any one character, independently
1103 of the DOTALL setting.
1105 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1106 non-existent subpattern following a character class starting with ']' and
1107 containing () gave an internal compiling error instead of "reference to
1108 non-existent subpattern". Fortunately, when the pattern did exist, the
1109 compiled code was correct. (When scanning forwards to check for the
1110 existencd of the subpattern, it was treating the data ']' as terminating
1111 the class, so got the count wrong. When actually compiling, the reference
1112 was subsequently set up correctly.)
1114 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1115 it was being rejected as not supported by pcre_dfa_exec(), even though
1116 other assertions are supported. I have made pcre_dfa_exec() support
1117 (*FAIL).
1119 16. The implementation of 13c above involved the invention of a new opcode,
1120 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1121 cannot be changed at match time, I realized I could make a small
1122 improvement to matching performance by compiling OP_ALLANY instead of
1123 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1124 on the OP_ANY path.
1126 17. Compiling pcretest on Windows with readline support failed without the
1127 following two fixes: (1) Make the unistd.h include conditional on
1128 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1130 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1131 ncurses library to be included for pcretest when ReadLine support is
1132 requested, but also to allow for it to be overridden. This patch came from
1133 Daniel Bergström.
1135 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1136 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1137 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1138 spotting this.
1141 Version 7.6 28-Jan-08
1142 ---------------------
1144 1. A character class containing a very large number of characters with
1145 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1146 overflow.
1148 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1149 HAVE_LONG_LONG is not defined.
1151 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1152 bring them up to date and include new features. This patch includes:
1154 - Fixed PH's badly added libz and libbz2 support.
1155 - Fixed a problem with static linking.
1156 - Added pcredemo. [But later removed - see 7 below.]
1157 - Fixed dftables problem and added an option.
1158 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1160 - Added readline support for pcretest.
1161 - Added an listing of the option settings after cmake has run.
1163 4. A user submitted a patch to Makefile that makes it easy to create
1164 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1165 Makefile.am that cause it to include this special target, without
1166 affecting anything else. Note that the same mingw target plus all
1167 the other distribution libraries and programs are now supported
1168 when configuring with CMake (see 6 below) instead of with
1169 Configure/Make.
1171 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1172 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1173 exported in the Windows port". It has not yet been confirmed that the patch
1174 solves the problem, but it does no harm.
1176 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1177 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1178 with CMake, and also correct the comment about stack recursion.
1180 7. Remove the automatic building of pcredemo from the ./configure system and
1181 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1182 of a program that users should build themselves after PCRE is installed, so
1183 building it automatically is not really right. What is more, it gave
1184 trouble in some build environments.
1186 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1189 Version 7.5 10-Jan-08
1190 ---------------------
1192 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1193 values in parens when parsing an RE using the C++ wrapper."
1195 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1196 Characters greater than 255 were excluded from the class instead of being
1197 included.
1199 3. The same bug as (2) above applied to negated POSIX classes such as
1200 [:^space:].
1202 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1203 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1204 I have changed it.
1206 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1207 first named subpattern) and a construct such as (?&a) would reference the
1208 first named subpattern whose name started with "a" (in other words, the
1209 length check was missing). Both these problems are fixed. "Subpattern name
1210 expected" is now given for (?&) (a zero-length name), and this patch also
1211 makes it give the same error for \k'' (previously it complained that that
1212 was a reference to a non-existent subpattern).
1214 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1215 this is right because (?- can be followed by option settings as well as by
1216 digits. I have, however, made the messages clearer.
1218 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1219 than the number used in the conditional) now cause a compile-time error.
1220 This is actually not compatible with Perl, which accepts such patterns, but
1221 treats the conditional as always being FALSE (as PCRE used to), but it
1222 seems to me that giving a diagnostic is better.
1224 8. Change "alphameric" to the more common word "alphanumeric" in comments
1225 and messages.
1227 9. Fix two occurrences of "backslash" in comments that should have been
1228 "backspace".
1230 10. Remove two redundant lines of code that can never be obeyed (their function
1231 was moved elsewhere).
1233 11. The program that makes PCRE's Unicode character property table had a bug
1234 which caused it to generate incorrect table entries for sequences of
1235 characters that have the same character type, but are in different scripts.
1236 It amalgamated them into a single range, with the script of the first of
1237 them. In other words, some characters were in the wrong script. There were
1238 thirteen such cases, affecting characters in the following ranges:
1240 U+002b0 - U+002c1
1241 U+0060c - U+0060d
1242 U+0061e - U+00612
1243 U+0064b - U+0065e
1244 U+0074d - U+0076d
1245 U+01800 - U+01805
1246 U+01d00 - U+01d77
1247 U+01d9b - U+01dbf
1248 U+0200b - U+0200f
1249 U+030fc - U+030fe
1250 U+03260 - U+0327f
1251 U+0fb46 - U+0fbb1
1252 U+10450 - U+1049d
1254 12. The -o option (show only the matching part of a line) for pcregrep was not
1255 compatible with GNU grep in that, if there was more than one match in a
1256 line, it showed only the first of them. It now behaves in the same way as
1257 GNU grep.
1259 13. If the -o and -v options were combined for pcregrep, it printed a blank
1260 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1261 does the same. The return code can be used to tell if there were any
1262 non-matching lines.
1264 14. Added --file-offsets and --line-offsets to pcregrep.
1266 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1267 infinitely looping recursion. The bug was that positive lookaheads were not
1268 being skipped when checking for a possible empty match (negative lookaheads
1269 and both kinds of lookbehind were skipped).
1271 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1272 inclusion of <windows.h> to before rather than after the definition of
1273 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1275 17. Specifying a possessive quantifier with a specific limit for a Unicode
1276 character property caused pcre_compile() to compile bad code, which led at
1277 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1278 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1279 caused the error; without that there was no problem.
1281 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1283 19. Added --enable-pcretest-libreadline.
1285 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1286 RE::GlobalReplace(). As a result, the number of replacements returned was
1287 double what it should be. I removed one of the increments, but Craig sent a
1288 later patch that removed the other one (the right fix) and added unit tests
1289 that check the return values (which was not done before).
1291 21. Several CMake things:
1293 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1294 the names libpcre and libpcreposix, not just pcre and pcreposix.
1296 (2) The above change means that pcretest and pcregrep are now correctly
1297 linked with the newly-built libraries, not previously installed ones.
1301 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1302 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1303 UTF-8 newline character). The key issue is that the pattern starts .*;
1304 this means that the match must be either at the beginning, or after a
1305 newline. The bug was in the code for advancing after a failed match and
1306 checking that the new position followed a newline. It was not taking
1307 account of UTF-8 characters correctly.
1309 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1310 character classes. PCRE was not treating the sequence [:...:] as a
1311 character class unless the ... were all letters. Perl, however, seems to
1312 allow any characters between [: and :], though of course it rejects as
1313 unknown any "names" that contain non-letters, because all the known class
1314 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1315 for example, whereas PCRE did not - it did not recognize a POSIX character
1316 class. This seemed a bit dangerous, so the code has been changed to be
1317 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1318 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1319 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1320 Perl does, and where it didn't before.
1322 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1323 Windows environments %n is disabled by default.
1326 Version 7.4 21-Sep-07
1327 ---------------------
1329 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1330 means that a class such as [\s] counted as "explicit reference to CR or
1331 LF". That isn't really right - the whole point of the change was to try to
1332 help when there was an actual mention of one of the two characters. So now
1333 the change happens only if \r or \n (or a literal CR or LF) character is
1334 encountered.
1336 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1337 of both had grown to the point where there were only 3 bits left.
1338 Fortunately, there was spare space in the data structure, and so I have
1339 moved the internal flags into a new 16-bit field to free up more option
1340 bits.
1342 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1343 but did not set the internal JCHANGED flag - either of these is enough to
1344 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1345 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1346 start sets both bits.
1348 4. Added options (at build time, compile time, exec time) to change \R from
1349 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1351 5. doc/pcresyntax.html was missing from the distribution.
1353 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1354 compatibility, even though it is no longer used.
1356 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1357 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1358 windows.h file is present (where different names are used). [This was
1359 reversed later after testing - see 16 below.]
1361 8. Changed all #include <config.h> to #include "config.h". There were also
1362 some further <pcre.h> cases that I changed to "pcre.h".
1364 9. When pcregrep was used with the --colour option, it missed the line ending
1365 sequence off the lines that it output.
1367 10. It was pointed out to me that arrays of string pointers cause lots of
1368 relocations when a shared library is dynamically loaded. A technique of
1369 using a single long string with a table of offsets can drastically reduce
1370 these. I have refactored PCRE in four places to do this. The result is
1371 dramatic:
1373 Originally: 290
1374 After changing UCP table: 187
1375 After changing error message table: 43
1376 After changing table of "verbs" 36
1377 After changing table of Posix names 22
1379 Thanks to the folks working on Gregex for glib for this insight.
1381 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1382 unicode-properties was also set.
1384 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1386 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1387 checked only for CRLF.
1389 14. Added casts to pcretest.c to avoid compiler warnings.
1391 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1393 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1394 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1395 entirely. This removes changes made in 7 above.
1397 17. The CMake files have been updated, and there is now more information about
1398 building with CMake in the NON-UNIX-USE document.
1401 Version 7.3 28-Aug-07
1402 ---------------------
1404 1. In the rejigging of the build system that eventually resulted in 7.1, the
1405 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1406 brackets there is not right, since it causes compilers to look for an
1407 installed pcre.h, not the version that is in the source that is being
1408 compiled (which of course may be different). I have changed it back to:
1410 #include "pcre.h"
1412 I have a vague recollection that the change was concerned with compiling in
1413 different directories, but in the new build system, that is taken care of
1414 by the VPATH setting the Makefile.
1416 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1417 when the subject happened to end in the byte 0x85 (e.g. if the last
1418 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1419 characters but of course it shouldn't be taken as a newline when it is part
1420 of another character. The bug was that, for an unlimited repeat of . in
1421 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1422 characters when looking for a newline.
1424 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1426 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1427 in debug output.
1429 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1430 long printing in the pcrecpp unittest when running under MinGW.
1432 6. ESC_K was left out of the EBCDIC table.
1434 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1435 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1436 limit also applies to "virtual nesting" when a pattern is recursive, and in
1437 this case 1000 isn't so big. I have been able to remove this limit at the
1438 expense of backing off one optimization in certain circumstances. Normally,
1439 when pcre_exec() would call its internal match() function recursively and
1440 immediately return the result unconditionally, it uses a "tail recursion"
1441 feature to save stack. However, when a subpattern that can match an empty
1442 string has an unlimited repetition quantifier, it no longer makes this
1443 optimization. That gives it a stack frame in which to save the data for
1444 checking that an empty string has been matched. Previously this was taken
1445 from the 1000-entry workspace that had been reserved. So now there is no
1446 explicit limit, but more stack is used.
1448 8. Applied Daniel's patches to solve problems with the import/export magic
1449 syntax that is required for Windows, and which was going wrong for the
1450 pcreposix and pcrecpp parts of the library. These were overlooked when this
1451 problem was solved for the main library.
1453 9. There were some crude static tests to avoid integer overflow when computing
1454 the size of patterns that contain repeated groups with explicit upper
1455 limits. As the maximum quantifier is 65535, the maximum group length was
1456 set at 30,000 so that the product of these two numbers did not overflow a
1457 32-bit integer. However, it turns out that people want to use groups that
1458 are longer than 30,000 bytes (though not repeat them that many times).
1459 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1460 made it possible to implement the integer overflow checks in a much more
1461 dynamic way, which I have now done. The artificial limitation on group
1462 length has been removed - we now have only the limit on the total length of
1463 the compiled pattern, which depends on the LINK_SIZE setting.
1465 10. Fixed a bug in the documentation for get/copy named substring when
1466 duplicate names are permitted. If none of the named substrings are set, the
1467 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1468 empty string.
1470 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1471 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1472 because the ] is interpreted as the first data character and the
1473 terminating ] is not found. PCRE has been made compatible with Perl in this
1474 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1475 cause memory overwriting.
1477 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1478 string has been matched (to stop an infinite loop). It was not recognizing
1479 a conditional subpattern that could match an empty string if that
1480 subpattern was within another subpattern. For example, it looped when
1481 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1482 condition was not nested. This bug has been fixed.
1484 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1485 past the start of the subject in the presence of bytes with the top bit
1486 set, for example "\x8aBCD".
1488 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1489 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1491 14. Optimized (?!) to (*FAIL).
1493 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1494 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1495 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1496 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1497 does: it's just the validity check that is more restrictive.
1499 16. Inserted checks for integer overflows during escape sequence (backslash)
1500 processing, and also fixed erroneous offset values for syntax errors during
1501 backslash processing.
1503 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1504 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1506 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1507 caused an overrun.
1509 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1510 something other than just ASCII characters) inside a group that had an
1511 unlimited repeat caused a loop at compile time (while checking to see
1512 whether the group could match an empty string).
1514 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1515 [\P{Any}] did so. (Error in the code for printing property names.)
1517 21. An orphan \E inside a character class could cause a crash.
1519 22. A repeated capturing bracket such as (A)? could cause a wild memory
1520 reference during compilation.
1522 23. There are several functions in pcre_compile() that scan along a compiled
1523 expression for various reasons (e.g. to see if it's fixed length for look
1524 behind). There were bugs in these functions when a repeated \p or \P was
1525 present in the pattern. These operators have additional parameters compared
1526 with \d, etc, and these were not being taken into account when moving along
1527 the compiled data. Specifically:
1529 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1530 length.
1532 (b) An item such as \pL+ within a repeated group could cause crashes or
1533 loops.
1535 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1536 "reference to non-existent subpattern" error.
1538 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1540 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1541 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1543 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1545 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1546 character were causing crashes (broken optimization).
1548 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1549 \p or \P) caused a compile-time loop.
1551 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1552 break. For example, the unstudied pattern [\r\n]A does not match the string
1553 "\r\nA" because change 7.0/46 below moves the current point on by two
1554 characters after failing to match at the start. However, the pattern \nA
1555 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1556 the same is true. There doesn't seem any very clean way out of this, but
1557 what I have chosen to do makes the common cases work: PCRE now takes note
1558 of whether there can be an explicit match for \r or \n anywhere in the
1559 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1560 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1561 pattern has explicit CR or LF references.
1563 29. Added (*CR) etc for changing newline setting at start of pattern.
1566 Version 7.2 19-Jun-07
1567 ---------------------
1569 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1570 which is apparently normally available under Windows.
1572 2. Re-jig the pcregrep tests with different newline settings in an attempt
1573 to make them independent of the local environment's newline setting.
1575 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1577 4. Some of the "internals" tests were previously cut out when the link size
1578 was not 2, because the output contained actual offsets. The recent new
1579 "Z" feature of pcretest means that these can be cut out, making the tests
1580 usable with all link sizes.
1582 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1583 stack recursion. This gives a massive performance boost under BSD, but just
1584 a small improvement under Linux. However, it saves one field in the frame
1585 in all cases.
1587 6. Added more features from the forthcoming Perl 5.10:
1589 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1590 recursion call. It refers to the nth most recently opened parentheses.
1592 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1593 to be opened parentheses.
1595 (c) Conditions that refer to capturing parentheses can be specified
1596 relatively, for example, (?(-2)... or (?(+3)...
1598 (d) \K resets the start of the current match so that everything before
1599 is not part of it.
1601 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1603 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1604 reference syntax.
1606 (g) (?| introduces a group in which the numbering of parentheses in each
1607 alternative starts with the same number.
1609 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1611 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1614 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1615 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1616 for detecting groups that can match an empty string.
1618 9. A pattern with a very large number of alternatives (more than several
1619 hundred) was running out of internal workspace during the pre-compile
1620 phase, where pcre_compile() figures out how much memory will be needed. A
1621 bit of new cunning has reduced the workspace needed for groups with
1622 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1623 workspace instead of running out of the 4096 that are available.
1625 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1627 11. Applied patch from Google to remove an optimization that didn't quite work.
1628 The report of the bug said:
1630 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1631 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1632 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1634 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1635 it matched the wrong number of bytes.
1638 Version 7.1 24-Apr-07
1639 ---------------------
1641 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1642 that is more "standard", making use of automake and other Autotools. There
1643 is some re-arrangement of the files and adjustment of comments consequent
1644 on this.
1646 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1647 for recursive directory scanning broke on some systems because the files
1648 are not scanned in any specific order and on different systems the order
1649 was different. A call to "sort" has been inserted into RunGrepTest for the
1650 approprate test as a short-term fix. In the longer term there may be an
1651 alternative.
1653 3. I had an email from Eric Raymond about problems translating some of PCRE's
1654 man pages to HTML (despite the fact that I distribute HTML pages, some
1655 people do their own conversions for various reasons). The problems
1656 concerned the use of low-level troff macros .br and .in. I have therefore
1657 removed all such uses from the man pages (some were redundant, some could
1658 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1659 HTML has been updated to handle .nf/.fi and to complain if it encounters
1660 .br or .in.
1662 4. Updated comments in configure.ac that get placed in config.h.in and also
1663 arranged for config.h to be included in the distribution, with the name
1664 config.h.generic, for the benefit of those who have to compile without
1665 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1667 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1668 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1669 makevp.bat for the current PCRE, using the additional files
1670 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1672 6. A Windows user reported a minor discrepancy with test 2, which turned out
1673 to be caused by a trailing space on an input line that had got lost in his
1674 copy. The trailing space was an accident, so I've just removed it.
1676 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1677 that is needed.
1679 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1680 as "const" (a) because they are and (b) because it helps the PHP
1681 maintainers who have recently made a script to detect big data structures
1682 in the php code that should be moved to the .rodata section. I remembered
1683 to update Builducptable as well, so it won't revert if ucptable.h is ever
1684 re-created.
1686 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1687 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1688 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1689 support is not required. This saves 1.5-2K of code, which is important in
1690 some applications.
1692 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1693 so as not to refer to the tables, even though these functions will never be
1694 called when UTF-8 support is disabled. Otherwise there are problems with a
1695 shared library.
1697 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1699 (a) It was defining its arguments as char * instead of void *.
1701 (b) It was assuming that all moves were upwards in memory; this was true
1702 a long time ago when I wrote it, but is no longer the case.
1704 The emulated memove() is provided for those environments that have neither
1705 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1706 is clearly not the case, as these two bugs were recently reported.
1708 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1709 and Detrail to create the HTML documentation, the .txt form of the man
1710 pages, and it removes trailing spaces from listed files. It also creates
1711 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1712 case, it wraps all the #defines with #ifndefs. This script should be run
1713 before "make dist".
1715 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1716 with Unicode property support.
1718 (a) For a maximizing quantifier, if the two different cases of the
1719 character were of different lengths in their UTF-8 codings (there are
1720 some cases like this - I found 11), and the matching function had to
1721 back up over a mixture of the two cases, it incorrectly assumed they
1722 were both the same length.
1724 (b) When PCRE was configured to use the heap rather than the stack for
1725 recursion during matching, it was not correctly preserving the data for
1726 the other case of a UTF-8 character when checking ahead for a match
1727 while processing a minimizing repeat. If the check also involved
1728 matching a wide character, but failed, corruption could cause an
1729 erroneous result when trying to check for a repeat of the original
1730 character.
1732 13. Some tidying changes to the testing mechanism:
1734 (a) The RunTest script now detects the internal link size and whether there
1735 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1736 values substituted by "configure". (The RunGrepTest script already did
1737 this for UTF-8.) The configure.ac script no longer substitutes the
1738 relevant variables.
1740 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1741 with length and offset values. This means that the output is different
1742 for different internal link sizes. Test 2 is skipped for link sizes
1743 other than 2 because of this, bypassing the problem. Unfortunately,
1744 there was also a test in test 3 (the locale tests) that used /B and
1745 failed for link sizes other than 2. Rather than cut the whole test out,
1746 I have added a new /Z option to pcretest that replaces the length and
1747 offset values with spaces. This is now used to make test 3 independent
1748 of link size. (Test 2 will be tidied up later.)
1750 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1751 segmentation fault instead of returning the appropriate error message.
1753 15. In multiline mode when the newline sequence was set to "any", the pattern
1754 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1755 This doesn't seem right; it now treats the CRLF combination as the line
1756 ending, and so does not match in that case. It's only a pattern such as ^$
1757 that would hit this one: something like ^ABC$ would have failed after \r
1758 and then tried again after \r\n.
1760 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1761 in an attempt to make files that differ only in their line terminators
1762 compare equal. This works on Linux.
1764 17. Under certain error circumstances pcregrep might try to free random memory
1765 as it exited. This is now fixed, thanks to valgrind.
1767 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1768 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1769 was because its rules for how to advance for /g after matching an empty
1770 string at the end of a line did not allow for this case. They now check for
1771 it specially.
1773 20. pcretest is supposed to handle patterns and data of any length, by
1774 extending its buffers when necessary. It was getting this wrong when the
1775 buffer for a data line had to be extended.
1777 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1778 CRLF as a newline sequence.
1780 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1781 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1782 I have nevertheless tidied it up.
1784 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1786 24. Added a man page for pcre-config.
1789 Version 7.0 19-Dec-06
1790 ---------------------
1792 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1793 moving to gcc 4.1.1.
1795 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1796 sys/time.h, which is documented as needed for this function. It doesn't
1797 seem to matter on Linux, but it showed up on some releases of OS X.
1799 3. It seems that there are systems where bytes whose values are greater than
1800 127 match isprint() in the "C" locale. The "C" locale should be the
1801 default when a C program starts up. In most systems, only ASCII printing
1802 characters match isprint(). This difference caused the output from pcretest
1803 to vary, making some of the tests fail. I have changed pcretest so that:
1805 (a) When it is outputting text in the compiled version of a pattern, bytes
1806 other than 32-126 are always shown as hex escapes.
1808 (b) When it is outputting text that is a matched part of a subject string,
1809 it does the same, unless a different locale has been set for the match
1810 (using the /L modifier). In this case, it uses isprint() to decide.
1812 4. Fixed a major bug that caused incorrect computation of the amount of memory
1813 required for a compiled pattern when options that changed within the
1814 pattern affected the logic of the preliminary scan that determines the
1815 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1816 that the computed length was too small. The symptoms of this bug were
1817 either the PCRE error "internal error: code overflow" from pcre_compile(),
1818 or a glibc crash with a message such as "pcretest: free(): invalid next
1819 size (fast)". Examples of patterns that provoked this bug (shown in
1820 pcretest format) are:
1822 /(?-x: )/x
1823 /(?x)(?-x: \s*#\s*)/
1824 /((?i)[\x{c0}])/8
1825 /(?i:[\x{c0}])/8
1827 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1828 is now done differently.
1830 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1831 wrapper classes; (b) implement a new function in the C++ scanner that is
1832 more efficient than the old way of doing things because it avoids levels of
1833 recursion in the regex matching; (c) add a paragraph to the documentation
1834 for the FullMatch() function.
1836 6. The escape sequence \n was being treated as whatever was defined as
1837 "newline". Not only was this contrary to the documentation, which states
1838 that \n is character 10 (hex 0A), but it also went horribly wrong when
1839 "newline" was defined as CRLF. This has been fixed.
1841 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1842 was being set to -1 for the "end of line" case (supposedly a value that no
1843 character can have). Though this value is never used (the check for end of
1844 line is "zero bytes in current character"), it caused compiler complaints.
1845 I've changed it to 0xffffffff.
1847 8. In pcre_version.c, the version string was being built by a sequence of
1848 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1849 string (as it is for production releases) called a macro with an empty
1850 argument. The C standard says the result of this is undefined. The gcc
1851 compiler treats it as an empty string (which was what was wanted) but it is
1852 reported that Visual C gives an error. The source has been hacked around to
1853 avoid this problem.
1855 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1856 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1857 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1858 of them did).
1860 10. Originally, pcretest opened its input and output without "b"; then I was
1861 told that "b" was needed in some environments, so it was added for release
1862 5.0 to both the input and output. (It makes no difference on Unix-like
1863 systems.) Later I was told that it is wrong for the input on Windows. I've
1864 now abstracted the modes into two macros, to make it easier to fiddle with
1865 them, and removed "b" from the input mode under Windows.
1867 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1869 12. Added -help and --help to pcretest as an official way of being reminded
1870 of the options.
1872 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1873 and pcrecpp.cc because they annoy compilers at high warning levels.
1875 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1877 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1878 scripts are not C programs :-) and which was not noticed because it works
1879 on Linux.
1881 16. pcretest is supposed to handle any length of pattern and data line (as one
1882 line or as a continued sequence of lines) by extending its input buffer if
1883 necessary. This feature was broken for very long pattern lines, leading to
1884 a string of junk being passed to pcre_compile() if the pattern was longer
1885 than about 50K.
1887 17. I have done a major re-factoring of the way pcre_compile() computes the
1888 amount of memory needed for a compiled pattern. Previously, there was code
1889 that made a preliminary scan of the pattern in order to do this. That was
1890 OK when PCRE was new, but as the facilities have expanded, it has become
1891 harder and harder to keep it in step with the real compile phase, and there
1892 have been a number of bugs (see for example, 4 above). I have now found a
1893 cunning way of running the real compile function in a "fake" mode that
1894 enables it to compute how much memory it would need, while actually only
1895 ever using a few hundred bytes of working memory and without too many
1896 tests of the mode. This should make future maintenance and development
1897 easier. A side effect of this work is that the limit of 200 on the nesting
1898 depth of parentheses has been removed (though this was never a serious
1899 limitation, I suspect). However, there is a downside: pcre_compile() now
1900 runs more slowly than before (30% or more, depending on the pattern). I
1901 hope this isn't a big issue. There is no effect on runtime performance.
1903 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1904 newline (only possible for the last line of a file) and it was a
1905 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1907 19. Added additional timing features to pcretest. (1) The -tm option now times
1908 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1909 separate command line item, by a number that specifies the number of
1910 repeats to use when timing. The default is 50000; this gives better
1911 precision, but takes uncomfortably long for very large patterns.
1913 20. Extended pcre_study() to be more clever in cases where a branch of a
1914 subpattern has no definite first character. For example, (a*|b*)[cd] would
1915 previously give no result from pcre_study(). Now it recognizes that the
1916 first character must be a, b, c, or d.
1918 21. There was an incorrect error "recursive call could loop indefinitely" if
1919 a subpattern (or the entire pattern) that was being tested for matching an
1920 empty string contained only one non-empty item after a nested subpattern.
1921 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1922 incorrectly, because the \d was being skipped in the check.
1924 22. The pcretest program now has a new pattern option /B and a command line
1925 option -b, which is equivalent to adding /B to every pattern. This causes
1926 it to show the compiled bytecode, without the additional information that
1927 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1928 is the same as /B/I).
1930 23. A new optimization is now able automatically to treat some sequences such
1931 as a*b as a*+b. More specifically, if something simple (such as a character
1932 or a simple class like \d) has an unlimited quantifier, and is followed by
1933 something that cannot possibly match the quantified thing, the quantifier
1934 is automatically "possessified".
1936 24. A recursive reference to a subpattern whose number was greater than 39
1937 went wrong under certain circumstances in UTF-8 mode. This bug could also
1938 have affected the operation of pcre_study().
1940 25. Realized that a little bit of performance could be had by replacing
1941 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1943 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1945 27. Possessive quantifiers such as a++ were previously implemented by turning
1946 them into atomic groups such as ($>a+). Now they have their own opcodes,
1947 which improves performance. This includes the automatically created ones
1948 from 23 above.
1950 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1951 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1952 the first matched character to be a colon. This applied both to named and
1953 numbered groups.
1955 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1957 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1958 building PCRE on Windows easier, so I have included it in the distribution.
1960 31. There is now a check in pcretest against a ridiculously large number being
1961 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1962 loop, the loop is abandoned.
1964 32. Forward references to subpatterns in conditions such as (?(2)...) where
1965 subpattern 2 is defined later cause pcre_compile() to search forwards in
1966 the pattern for the relevant set of parentheses. This search went wrong
1967 when there were unescaped parentheses in a character class, parentheses
1968 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1970 33. "Subroutine" calls and backreferences were previously restricted to
1971 referencing subpatterns earlier in the regex. This restriction has now
1972 been removed.
1974 34. Added a number of extra features that are going to be in Perl 5.10. On the
1975 whole, these are just syntactic alternatives for features that PCRE had
1976 previously implemented using the Python syntax or my own invention. The
1977 other formats are all retained for compatibility.
1979 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1980 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1981 also .NET compatible.
1983 (b) A recursion or subroutine call to a named group can now be defined as
1984 (?&name) as well as (?P>name).
1986 (c) A backreference to a named group can now be defined as \k<name> or
1987 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1988 5.10, are also .NET compatible.
1990 (d) A conditional reference to a named group can now use the syntax
1991 (?(<name>) or (?('name') as well as (?(name).
1993 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1994 groups (named and numbered) that are never evaluated inline, but can be
1995 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1996 is always false. There may be only one alternative in such a group.
1998 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1999 as the simple (?(R). The condition is true only if the most recent
2000 recursion is that of the given number or name. It does not search out
2001 through the entire recursion stack.
2003 (g) The escape \gN or \g{N} has been added, where N is a positive or
2004 negative number, specifying an absolute or relative reference.
2006 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2007 some "unreachable code" warnings.
2009 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2010 things, this adds five new scripts.
2012 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2013 There were also incompatibilities regarding the handling of \Q..\E inside
2014 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2015 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2017 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2018 matches an empty string, and forcibly breaks the loop. There were bugs in
2019 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2020 against aaaa the result was just "a" rather than "aaaa", for example. Two
2021 separate and independent bugs (that affected different cases) have been
2022 fixed.
2024 39. Refactored the code to abolish the use of different opcodes for small
2025 capturing bracket numbers. This is a tidy that I avoided doing when I
2026 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2027 The new approach is not only tidier, it makes it possible to reduce the
2028 memory needed to fix the previous bug (38).
2030 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2031 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2032 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2033 mode.
2035 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2036 report.
2038 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2039 copy construction and assignment in the C++ wrapper.
2041 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2042 couple of bugs that could have given wrong results in the "--newline=crlf"
2043 case.
2045 44. Added a number of casts and did some reorganization of signed/unsigned int
2046 variables following suggestions from Dair Grant. Also renamed the variable
2047 "this" as "item" because it is a C++ keyword.
2049 45. Arranged for dftables to add
2051 #include "pcre_internal.h"
2053 to pcre_chartables.c because without it, gcc 4.x may remove the array
2054 definition from the final binary if PCRE is built into a static library and
2055 dead code stripping is activated.
2057 46. For an unanchored pattern, if a match attempt fails at the start of a
2058 newline sequence, and the newline setting is CRLF or ANY, and the next two
2059 characters are CRLF, advance by two characters instead of one.
2062 Version 6.7 04-Jul-06
2063 ---------------------
2065 1. In order to handle tests when input lines are enormously long, pcretest has
2066 been re-factored so that it automatically extends its buffers when
2067 necessary. The code is crude, but this _is_ just a test program. The
2068 default size has been increased from 32K to 50K.
2070 2. The code in pcre_study() was using the value of the re argument before
2071 testing it for NULL. (Of course, in any sensible call of the function, it
2072 won't be NULL.)
2074 3. The memmove() emulation function in pcre_internal.h, which is used on
2075 systems that lack both memmove() and bcopy() - that is, hardly ever -
2076 was missing a "static" storage class specifier.
2078 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2079 containing an extended class (one that cannot be represented by a bitmap
2080 because it contains high-valued characters or Unicode property items, e.g.
2081 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2082 pattern, but PCRE should not loop if you do not (it no longer does).
2083 [Detail: two cases were found: (a) a repeated subpattern containing an
2084 extended class; (b) a recursive reference to a subpattern that followed a
2085 previous extended class. It wasn't skipping over the extended class
2086 correctly when UTF-8 mode was not set.]
2088 5. A negated single-character class was not being recognized as fixed-length
2089 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2090 compile error "lookbehind assertion is not fixed length".
2092 6. The RunPerlTest auxiliary script was showing an unexpected difference
2093 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2094 write a Perl script that can interpret lines of an input file either as
2095 byte characters or as UTF-8, which is what "perltest" was being required to
2096 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2097 can't do is switch easily at run time between having the "use utf8;" pragma
2098 or not. In the end, I fudged it by using the RunPerlTest script to insert
2099 "use utf8;" explicitly for the UTF-8 tests.
2101 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2102 the end of the subject string, contrary to the documentation and to what
2103 Perl does. This was true of both matching functions. Now it matches only at
2104 the start of the subject and immediately after *internal* newlines.
2106 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2107 a pointer to an int instead of a pointer to an unsigned long int. This
2108 caused problems on 64-bit systems.
2110 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2111 instance of the 'standard' template library not being so standard".
2113 10. There was no check on the number of named subpatterns nor the maximum
2114 length of a subpattern name. The product of these values is used to compute
2115 the size of the memory block for a compiled pattern. By supplying a very
2116 long subpattern name and a large number of named subpatterns, the size
2117 computation could be caused to overflow. This is now prevented by limiting
2118 the length of names to 32 characters, and the number of named subpatterns
2119 to 10,000.
2121 11. Subpatterns that are repeated with specific counts have to be replicated in
2122 the compiled pattern. The size of memory for this was computed from the
2123 length of the subpattern and the repeat count. The latter is limited to
2124 65535, but there was no limit on the former, meaning that integer overflow
2125 could in principle occur. The compiled length of a repeated subpattern is
2126 now limited to 30,000 bytes in order to prevent this.
2128 12. Added the optional facility to have named substrings with the same name.
2130 13. Added the ability to use a named substring as a condition, using the
2131 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2132 are numbers (not recommended). Forward references are permitted.
2134 14. Added forward references in named backreferences (if you see what I mean).
2136 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2137 pattern could run off the end of the subject. For example, the pattern
2138 "(?s)(.{1,5})"8 did this with the subject "ab".
2140 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2141 PCRE_CASELESS was set when matching characters that were quantified with ?
2142 or *.
2144 17. A character class other than a single negated character that had a minimum
2145 but no maximum quantifier - for example [ab]{6,} - was not handled
2146 correctly by pce_dfa_exec(). It would match only one character.
2148 18. A valid (though odd) pattern that looked like a POSIX character
2149 class but used an invalid character after [ (for example [[,abc,]]) caused
2150 pcre_compile() to give the error "Failed: internal error: code overflow" or
2151 in some cases to crash with a glibc free() error. This could even happen if
2152 the pattern terminated after [[ but there just happened to be a sequence of
2153 letters, a binary zero, and a closing ] in the memory that followed.
2155 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2156 over the years. Originally (before any Unicode support), just the bottom 8
2157 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2158 output from "man perlunicode" includes this:
2160 The regular expression compiler produces polymorphic opcodes. That
2161 is, the pattern adapts to the data and automatically switches to
2162 the Unicode character scheme when presented with Unicode data--or
2163 instead uses a traditional byte scheme when presented with byte
2164 data.
2166 Sadly, a wide octal escape does not cause a switch, and in a string with
2167 no other multibyte characters, these octal escapes are treated as before.
2168 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2169 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2170 Unicode string.
2172 I have not perpetrated such confusion in PCRE. Up till now, it took just
2173 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2174 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2175 translate to the appropriate multibyte character.
2177 29. Applied some refactoring to reduce the number of warnings from Microsoft
2178 and Borland compilers. This has included removing the fudge introduced
2179 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2180 a warning about an unused variable.
2182 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2183 characters since release 4.0, because Perl (from release 5.004) does not.
2184 [Or at least, is documented not to: some releases seem to be in conflict
2185 with the documentation.] However, when a pattern was studied with
2186 pcre_study() and all its branches started with \s, PCRE still included VT
2187 as a possible starting character. Of course, this did no harm; it just
2188 caused an unnecessary match attempt.
2190 22. Removed a now-redundant internal flag bit that recorded the fact that case
2191 dependency changed within the pattern. This was once needed for "required
2192 byte" processing, but is no longer used. This recovers a now-scarce options
2193 bit. Also moved the least significant internal flag bit to the most-
2194 significant bit of the word, which was not previously used (hangover from
2195 the days when it was an int rather than a uint) to free up another bit for
2196 the future.
2198 23. Added support for CRLF line endings as well as CR and LF. As well as the
2199 default being selectable at build time, it can now be changed at runtime
2200 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2201 specify that it is scanning data with non-default line endings.
2203 24. Changed the definition of CXXLINK to make it agree with the definition of
2204 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2206 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2207 recursions. This makes a big different to stack usage for some patterns.
2209 26. If a subpattern containing a named recursion or subroutine reference such
2210 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2211 the space required for the compiled pattern went wrong and gave too small a
2212 value. Depending on the environment, this could lead to "Failed: internal
2213 error: code overflow at offset 49" or "glibc detected double free or
2214 corruption" errors.
2216 27. Applied patches from Google (a) to support the new newline modes and (b) to
2217 advance over multibyte UTF-8 characters in GlobalReplace.
2219 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2220 difference for some implementation of PCRE in some Windows version.
2222 29. Added some extra testing facilities to pcretest:
2224 \q<number> in a data line sets the "match limit" value
2225 \Q<number> in a data line sets the "match recursion limt" value
2226 -S <number> sets the stack size, where <number> is in megabytes
2228 The -S option isn't available for Windows.
2231 Version 6.6 06-Feb-06
2232 ---------------------
2234 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2235 in pcreposix.h. I have copied the definition from pcre.h.
2237 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2238 because pcre.h is no longer a built file.
2240 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2241 not normally included in the compiled code.
2244 Version 6.5 01-Feb-06
2245 ---------------------
2247 1. When using the partial match feature with pcre_dfa_exec(), it was not
2248 anchoring the second and subsequent partial matches at the new starting
2249 point. This could lead to incorrect results. For example, with the pattern
2250 /1234/, partially matching against "123" and then "a4" gave a match.
2252 2. Changes to pcregrep:
2254 (a) All non-match returns from pcre_exec() were being treated as failures
2255 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2256 error message is output. Some extra information is given for the
2258 probably the only errors that are likely to be caused by users (by
2259 specifying a regex that has nested indefinite repeats, for instance).
2260 If there are more than 20 of these errors, pcregrep is abandoned.
2262 (b) A binary zero was treated as data while matching, but terminated the
2263 output line if it was written out. This has been fixed: binary zeroes
2264 are now no different to any other data bytes.
2266 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2267 used to set a locale for matching. The --locale=xxxx long option has
2268 been added (no short equivalent) to specify a locale explicitly on the
2269 pcregrep command, overriding the environment variables.
2271 (d) When -B was used with -n, some line numbers in the output were one less
2272 than they should have been.
2274 (e) Added the -o (--only-matching) option.
2276 (f) If -A or -C was used with -c (count only), some lines of context were
2277 accidentally printed for the final match.
2279 (g) Added the -H (--with-filename) option.
2281 (h) The combination of options -rh failed to suppress file names for files
2282 that were found from directory arguments.
2284 (i) Added the -D (--devices) and -d (--directories) options.
2286 (j) Added the -F (--fixed-strings) option.
2288 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2290 (l) Added the --colo(u)r option.
2292 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2293 is not present by default.
2295 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2296 items such as (?R) or (?1), when the recursion could match a number of
2297 alternatives. If it matched one of the alternatives, but subsequently,
2298 outside the recursion, there was a failure, the code tried to back up into
2299 the recursion. However, because of the way PCRE is implemented, this is not
2300 possible, and the result was an incorrect result from the match.
2302 In order to prevent this happening, the specification of recursion has
2303 been changed so that all such subpatterns are automatically treated as
2304 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2306 4. I had overlooked the fact that, in some locales, there are characters for
2307 which isalpha() is true but neither isupper() nor islower() are true. In
2308 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2309 and ordfeminine) are like this. This affected the treatment of \w and \W
2310 when they appeared in character classes, but not when they appeared outside
2311 a character class. The bit map for "word" characters is now created
2312 separately from the results of isalnum() instead of just taking it from the
2313 upper, lower, and digit maps. (Plus the underscore character, of course.)
2315 5. The above bug also affected the handling of POSIX character classes such as
2316 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2317 permanent tables. Instead, the bit maps for such a class were previously
2318 created as the appropriate unions of the upper, lower, and digit bitmaps.
2319 Now they are created by subtraction from the [[:word:]] class, which has
2320 its own bitmap.
2322 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2323 It is created by subtracting the vertical space characters (\x09, \x0a,
2324 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2325 subtraction was done in the overall bitmap for a character class, meaning
2326 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2327 be recognized. This bug has been fixed.
2329 7. Patches from the folks at Google:
2331 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2332 real life, but is still worth protecting against".
2334 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2335 regular expressions".
2337 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2338 have it.
2340 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2341 "configure" and the latter not, in order to fix a problem somebody had
2342 with compiling the Arg class on HP-UX.
2344 (e) Improve the error-handling of the C++ wrapper a little bit.
2346 (f) New tests for checking recursion limiting.
2348 8. The pcre_memmove() function, which is used only if the environment does not
2349 have a standard memmove() function (and is therefore rarely compiled),
2350 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2351 returning a result (though PCRE never actually uses the result).
2353 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2354 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2355 returned instead of calling malloc() with an overflowing number that would
2356 most likely cause subsequent chaos.
2358 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2360 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2361 with this option is matched, the nmatch and pmatch options of regexec() are
2362 ignored.
2364 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2365 provided in case anyone wants to the the POSIX interface with UTF-8
2366 strings.
2368 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2369 C++ linking (needed for some HP-UX environments).
2371 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2372 (unused parameter) and in the pcre_printint() function (omitted "default"
2373 switch label when the default is to do nothing).
2375 15. Added some code to make it possible, when PCRE is compiled as a C++
2376 library, to replace subject pointers for pcre_exec() with a smart pointer
2377 class, thus making it possible to process discontinuous strings.
2379 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2380 much the same function. They were added by different people who were trying
2381 to make PCRE easy to compile on non-Unix systems. It has been suggested
2382 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2383 for compiling on Windows systems. I have therefore replaced it with
2384 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2385 defaults to "extern" for C or "extern C" for C++, which works fine on
2386 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2387 SCOPE with something explicit in config.h. In addition:
2389 (a) pcreposix.h still had just "extern" instead of either of these macros;
2390 I have replaced it with PCRE_DATA_SCOPE.
2392 (b) Functions such as _pcre_xclass(), which are internal to the library,
2393 but external in the C sense, all had PCRE_EXPORT in their definitions.
2394 This is apparently wrong for the Windows case, so I have removed it.
2395 (It makes no difference on Unix-like systems.)
2397 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2398 of recursive calls to match(). This is different to MATCH_LIMIT because
2399 that limits the total number of calls to match(), not all of which increase
2400 the depth of recursion. Limiting the recursion depth limits the amount of
2401 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2402 when PCRE is compiled, and changed at run time. A patch from Google adds
2403 this functionality to the C++ interface.
2405 18. Changes to the handling of Unicode character properties:
2407 (a) Updated the table to Unicode 4.1.0.
2409 (b) Recognize characters that are not in the table as "Cn" (undefined).
2411 (c) I revised the way the table is implemented to a much improved format
2412 which includes recognition of ranges. It now supports the ranges that
2413 are defined in UnicodeData.txt, and it also amalgamates other
2414 characters into ranges. This has reduced the number of entries in the
2415 table from around 16,000 to around 3,000, thus reducing its size
2416 considerably. I realized I did not need to use a tree structure after
2417 all - a binary chop search is just as efficient. Having reduced the
2418 number of entries, I extended their size from 6 bytes to 8 bytes to
2419 allow for more data.
2421 (d) Added support for Unicode script names via properties such as \p{Han}.
2423 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2424 matching that character.
2426 20. When matching a repeated Unicode property with a minimum greater than zero,
2427 (for example \pL{2,}), PCRE could look past the end of the subject if it
2428 reached it while seeking the minimum number of characters. This could
2429 happen only if some of the characters were more than one byte long, because
2430 there is a check for at least the minimum number of bytes.
2432 21. Refactored the implementation of \p and \P so as to be more general, to
2433 allow for more different types of property in future. This has changed the
2434 compiled form incompatibly. Anybody with saved compiled patterns that use
2435 \p or \P will have to recompile them.
2437 22. Added "Any" and "L&" to the supported property types.
2439 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2440 but give a compile time error if the value is greater than 0xff.
2442 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2443 accidentally not being installed or uninstalled.
2445 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2446 made were to insert the current release number. This seemed silly, because
2447 it made things harder for people building PCRE on systems that don't run
2448 "configure". I have turned pcre.h into a distributed file, no longer built
2449 by "configure", with the version identification directly included. There is
2450 no longer a pcre.h.in file.
2452 However, this change necessitated a change to the pcre-config script as
2453 well. It is built from pcre-config.in, and one of the substitutions was the
2454 release number. I have updated configure.ac so that ./configure now finds
2455 the release number by grepping pcre.h.
2457 26. Added the ability to run the tests under valgrind.
2460 Version 6.4 05-Sep-05
2461 ---------------------
2463 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2464 "--" to be printed when multiple files were scanned, even when none of the
2465 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2466 consider it to be a bug, and have restored the previous behaviour.
2468 2. A couple of code tidies to get rid of compiler warnings.
2470 3. The pcretest program used to cheat by referring to symbols in the library
2471 whose names begin with _pcre_. These are internal symbols that are not
2472 really supposed to be visible externally, and in some environments it is
2473 possible to suppress them. The cheating is now confined to including
2474 certain files from the library's source, which is a bit cleaner.
2476 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2477 file's purpose clearer.
2479 5. Reorganized pcre_ucp_findchar().
2482 Version 6.3 15-Aug-05
2483 ---------------------
2485 1. The file libpcre.pc.in did not have general read permission in the tarball.
2487 2. There were some problems when building without C++ support:
2489 (a) If C++ support was not built, "make install" and "make test" still
2490 tried to test it.
2492 (b) There were problems when the value of CXX was explicitly set. Some
2493 changes have been made to try to fix these, and ...
2495 (c) --disable-cpp can now be used to explicitly disable C++ support.
2497 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2498 backslash in a target when C++ was disabled. This confuses some
2499 versions of "make", apparently. Using an intermediate variable solves
2500 this. (Same for CPP_LOBJ.)
2502 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2503 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2504 necessary on certain architectures.
2506 4. Added a setting of -export-symbols-regex to the link command to remove
2507 those symbols that are exported in the C sense, but actually are local
2508 within the library, and not documented. Their names all begin with
2509 "_pcre_". This is not a perfect job, because (a) we have to except some
2510 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2511 available (and never for static libraries). I have made a note to try to
2512 find a way round (a) in the future.
2515 Version 6.2 01-Aug-05
2516 ---------------------
2518 1. There was no test for integer overflow of quantifier values. A construction
2519 such as {1111111111111111} would give undefined results. What is worse, if
2520 a minimum quantifier for a parenthesized subpattern overflowed and became
2521 negative, the calculation of the memory size went wrong. This could have
2522 led to memory overwriting.
2524 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2526 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2527 operating environments where this matters.
2529 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2530 PCRE options from within the C++ wrapper.
2532 5. Named capturing subpatterns were not being correctly counted when a pattern
2533 was compiled. This caused two problems: (a) If there were more than 100
2534 such subpatterns, the calculation of the memory needed for the whole
2535 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2536 back references of the form \12, where the number was greater than 9, were
2537 not recognized as back references, even though there were sufficient
2538 previous subpatterns.
2540 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2541 versions of gcc, e.g. 2.95.4.
2544 Version 6.1 21-Jun-05
2545 ---------------------
2547 1. There was one reference to the variable "posix" in pcretest.c that was not
2548 surrounded by "#if !defined NOPOSIX".
2550 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2551 the cross-check on the old pcre_info() function, for the benefit of the
2552 cut-down version of PCRE that is currently imported into Exim.
2554 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2555 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2556 patterns that start (?i) but I don't think that matters. The use of (?i) is
2557 just an example; this all applies to the other options as well.
2559 4. Since libtool seems to echo the compile commands it is issuing, the output
2560 from "make" can be reduced a bit by putting "@" in front of each libtool
2561 compile command.
2563 5. Patch from the folks at Google for configure.in to be a bit more thorough
2564 in checking for a suitable C++ installation before trying to compile the
2565 C++ stuff. This should fix a reported problem when a compiler was present,
2566 but no suitable headers.
2568 6. The man pages all had just "PCRE" as their title. I have changed them to
2569 be the relevant file name. I have also arranged that these names are
2570 retained in the file doc/pcre.txt, which is a concatenation in text format
2571 of all the man pages except the little individual ones for each function.
2573 7. The NON-UNIX-USE file had not been updated for the different set of source
2574 files that come with release 6. I also added a few comments about the C++
2575 wrapper.
2578 Version 6.0 07-Jun-05
2579 ---------------------
2581 1. Some minor internal re-organization to help with my DFA experiments.
2583 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2584 didn't matter for the library itself when fully configured, but did matter
2585 when compiling without UCP support, or within Exim, where the ucp files are
2586 not imported.
2588 3. Refactoring of the library code to split up the various functions into
2589 different source modules. The addition of the new DFA matching code (see
2590 below) to a single monolithic source would have made it really too
2591 unwieldy, quite apart from causing all the code to be include in a
2592 statically linked application, when only some functions are used. This is
2593 relevant even without the DFA addition now that patterns can be compiled in
2594 one application and matched in another.
2596 The downside of splitting up is that there have to be some external
2597 functions and data tables that are used internally in different modules of
2598 the library but which are not part of the API. These have all had their
2599 names changed to start with "_pcre_" so that they are unlikely to clash
2600 with other external names.
2602 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2603 a different (DFA) algorithm. Although it is slower than the original
2604 function, it does have some advantages for certain types of matching
2605 problem.
2607 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2608 including restarting after a partial match.
2610 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2611 defined when compiling for Windows was sent to me. I have put it into the
2612 code, though I have no means of testing or verifying it.
2614 7. Added the pcre_refcount() auxiliary function.
2616 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2617 match before or at the first newline in the subject string. In pcretest,
2618 the /f option on a pattern can be used to set this.
2620 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2621 would behave wrongly. This has been present in PCRE since release 4.0.
2623 10. A number of changes to the pcregrep command:
2625 (a) Refactored how -x works; insert ^(...)$ instead of setting
2626 PCRE_ANCHORED and checking the length, in preparation for adding
2627 something similar for -w.
2629 (b) Added the -w (match as a word) option.
2631 (c) Refactored the way lines are read and buffered so as to have more
2632 than one at a time available.
2634 (d) Implemented a pcregrep test script.
2636 (e) Added the -M (multiline match) option. This allows patterns to match
2637 over several lines of the subject. The buffering ensures that at least
2638 8K, or the rest of the document (whichever is the shorter) is available
2639 for matching (and similarly the previous 8K for lookbehind assertions).
2641 (f) Changed the --help output so that it now says
2643 -w, --word-regex(p)
2645 instead of two lines, one with "regex" and the other with "regexp"
2646 because that confused at least one person since the short forms are the
2647 same. (This required a bit of code, as the output is generated
2648 automatically from a table. It wasn't just a text change.)
2650 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2651 option but starts with a hyphen. Could be a pattern or a path name
2652 starting with a hyphen, for instance.
2654 (h) "-" can be given as a file name to represent stdin.
2656 (i) When file names are being printed, "(standard input)" is used for
2657 the standard input, for compatibility with GNU grep. Previously
2658 "<stdin>" was used.
2660 (j) The option --label=xxx can be used to supply a name to be used for
2661 stdin when file names are being printed. There is no short form.
2663 (k) Re-factored the options decoding logic because we are going to add
2664 two more options that take data. Such options can now be given in four
2665 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2667 (l) Added the -A, -B, and -C options for requesting that lines of context
2668 around matches be printed.
2670 (m) Added the -L option to print the names of files that do not contain
2671 any matching lines, that is, the complement of -l.
2673 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2674 continue to scan other files.
2676 (o) The -s option was incorrectly implemented. For compatibility with other
2677 greps, it now suppresses the error message for a non-existent or non-
2678 accessible file (but not the return code). There is a new option called
2679 -q that suppresses the output of matching lines, which was what -s was
2680 previously doing.
2682 (p) Added --include and --exclude options to specify files for inclusion
2683 and exclusion when recursing.
2685 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2686 Hopefully, it now does.
2688 12. Missing cast in pcre_study().
2690 13. Added an "uninstall" target to the makefile.
2692 14. Replaced "extern" in the function prototypes in Makefile.in with
2693 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2694 world, but is set differently for Windows.
2696 15. Added a second compiling function called pcre_compile2(). The only
2697 difference is that it has an extra argument, which is a pointer to an
2698 integer error code. When there is a compile-time failure, this is set
2699 non-zero, in addition to the error test pointer being set to point to an
2700 error message. The new argument may be NULL if no error number is required
2701 (but then you may as well call pcre_compile(), which is now just a
2702 wrapper). This facility is provided because some applications need a
2703 numeric error indication, but it has also enabled me to tidy up the way
2704 compile-time errors are handled in the POSIX wrapper.
2706 16. Added VPATH=.libs to the makefile; this should help when building with one
2707 prefix path and installing with another. (Or so I'm told by someone who
2708 knows more about this stuff than I do.)
2710 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2711 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2712 match everything, including newlines. This is not POSIX-compatible, but
2713 somebody wanted the feature. From pcretest it can be activated by using
2714 both the P and the s flags.
2716 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2718 19. libpcre.pc was being incorrectly installed as executable.
2720 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2721 it now also looks for '\r' so that it will work unmodified on Windows.
2723 21. Added Google's contributed C++ wrapper to the distribution.
2725 22. Added some untidy missing memory free() calls in pcretest, to keep
2726 Electric Fence happy when testing.
2730 Version 5.0 13-Sep-04
2731 ---------------------
2733 1. Internal change: literal characters are no longer packed up into items
2734 containing multiple characters in a single byte-string. Each character
2735 is now matched using a separate opcode. However, there may be more than one
2736 byte in the character in UTF-8 mode.
2738 2. The pcre_callout_block structure has two new fields: pattern_position and
2739 next_item_length. These contain the offset in the pattern to the next match
2740 item, and its length, respectively.
2742 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2743 insertion of callouts before each pattern item. Added the /C option to
2744 pcretest to make use of this.
2746 4. On the advice of a Windows user, the lines
2748 #if defined(_WIN32) || defined(WIN32)
2749 _setmode( _fileno( stdout ), 0x8000 );
2750 #endif /* defined(_WIN32) || defined(WIN32) */
2752 have been added to the source of pcretest. This apparently does useful
2753 magic in relation to line terminators.
2755 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2756 for the benefit of those environments where the "b" makes a difference.
2758 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2759 to know about it. I have put a hack into configure.in that adds in code
2760 to set GCC=yes if CC=icc. This seems to end up at a point in the
2761 generated configure script that is early enough to affect the setting of
2762 compiler options, which is what is needed, but I have no means of testing
2763 whether it really works. (The user who reported this had patched the
2764 generated configure script, which of course I cannot do.)
2766 LATER: After change 22 below (new libtool files), the configure script
2767 seems to know about icc (and also ecc). Therefore, I have commented out
2768 this hack in configure.in.
2770 7. Added support for pkg-config (2 patches were sent in).
2772 8. Negated POSIX character classes that used a combination of internal tables
2773 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2774 [[:^ascii]]. Typically, they would match almost any characters. The other
2775 POSIX classes were not broken in this way.
2777 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2778 to find the match, as PCRE was deluded into thinking that the match had to
2779 start at the start point or following a newline. The same bug applied to
2780 patterns with negative forward assertions or any backward assertions
2781 preceding ".*" at the start, unless the pattern required a fixed first
2782 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2784 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2785 starting at the last subject character, bytes beyond the end of the subject
2786 string were read.
2788 11. Renamed the variable "class" as "classbits" to make life easier for C++
2789 users. (Previously there was a macro definition, but it apparently wasn't
2790 enough.)
2792 12. Added the new field "tables" to the extra data so that tables can be passed
2793 in at exec time, or the internal tables can be re-selected. This allows
2794 a compiled regex to be saved and re-used at a later time by a different
2795 program that might have everything at different addresses.
2797 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2798 -R library as well as a -L library.
2800 14. The debugging options of pcretest (-d on the command line or D on a
2801 pattern) showed incorrect output for anything following an extended class
2802 that contained multibyte characters and which was followed by a quantifier.
2804 15. Added optional support for general category Unicode character properties
2805 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2806 support. It adds about 90K to the size of the library. The meanings of the
2807 inbuilt class escapes such as \d and \s have NOT been changed.
2809 16. Updated pcredemo.c to include calls to free() to release the memory for the
2810 compiled pattern.
2812 17. The generated file chartables.c was being created in the source directory
2813 instead of in the building directory. This caused the build to fail if the
2814 source directory was different from the building directory, and was
2815 read-only.
2817 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2818 file. No doubt somebody will tell me if they don't make sense... Also added
2819 Dan Mooney's comments about building on OpenVMS.
2821 19. Added support for partial matching via the PCRE_PARTIAL option for
2822 pcre_exec() and the \P data escape in pcretest.
2824 20. Extended pcretest with 3 new pattern features:
2826 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2827 write the compiled pattern to the file whose name is "rest-of-line".
2828 This is a straight binary dump of the data, with the saved pointer to
2829 the character tables forced to be NULL. The study data, if any, is
2830 written too. After writing, pcretest reads a new pattern.
2832 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2833 compiled pattern from the given file. There must not be any
2834 occurrences of "<" in the file name (pretty unlikely); if there are,
2835 pcretest will instead treat the initial "<" as a pattern delimiter.
2836 After reading in the pattern, pcretest goes on to read data lines as
2837 usual.
2839 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2840 and 16-bit fields in a compiled pattern, to simulate a pattern that
2841 was compiled on a host of opposite endianness.
2843 21. The pcre-exec() function can now cope with patterns that were compiled on
2844 hosts of opposite endianness, with this restriction:
2846 As for any compiled expression that is saved and used later, the tables
2847 pointer field cannot be preserved; the extra_data field in the arguments
2848 to pcre_exec() should be used to pass in a tables address if a value
2849 other than the default internal tables were used at compile time.
2851 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2852 now diagnosed as an error. Previously, most of the time, a negative number
2853 would have been treated as zero, but if in addition "ovector" was passed as
2854 NULL, a crash could occur.
2856 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2857 new versions from the libtool 1.5 distribution (the last one is a copy of
2858 a file called libtool.m4). This seems to have fixed the need to patch
2859 "configure" to support Darwin 1.3 (which I used to do). However, I still
2860 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2861 workstation).
2863 24. Changed the PCRE licence to be the more standard "BSD" licence.
2866 Version 4.5 01-Dec-03
2867 ---------------------
2869 1. There has been some re-arrangement of the code for the match() function so
2870 that it can be compiled in a version that does not call itself recursively.
2871 Instead, it keeps those local variables that need separate instances for
2872 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2873 needs to "recurse". Keeping track of where control must go is done by means
2874 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2875 hide most of the details from the main code, and operates only if
2876 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2877 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2878 operating.
2880 To make it easier for callers to provide specially tailored get/free
2881 functions for this usage, two new functions, pcre_stack_malloc, and
2882 pcre_stack_free, are used. They are always called in strict stacking order,
2883 and the size of block requested is always the same.
2885 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2886 PCRE has been compiled to use the stack or the heap for recursion. The
2887 -C option of pcretest uses this to show which version is compiled.
2889 A new data escape \S, is added to pcretest; it causes the amounts of store
2890 obtained and freed by both kinds of malloc/free at match time to be added
2891 to the output.
2893 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2894 what's available on my current Linux desktop machine.
2896 3. When matching a UTF-8 string, the test for a valid string at the start has
2897 been extended. If start_offset is not zero, PCRE now checks that it points
2898 to a byte that is the start of a UTF-8 character. If not, it returns
2899 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2900 this is necessary because there may be backward assertions in the pattern.
2901 When matching the same subject several times, it may save resources to use
2902 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2904 4. The code for checking the validity of UTF-8 strings has been tightened so
2905 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2906 containing "overlong sequences".
2908 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2909 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2910 should have been "&", but it just so happened that all the cases this let
2911 through by mistake were picked up later in the function.
2913 6. I had used a variable called "isblank" - this is a C99 function, causing
2914 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2916 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2917 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2918 so the version is shown; (c) stop "make test" echoing "./RunTest".
2920 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2922 9. The prototype for memmove() for systems that don't have it was using
2923 size_t, but the inclusion of the header that defines size_t was later. I've
2924 moved the #includes for the C headers earlier to avoid this.
2926 10. Added some adjustments to the code to make it easier to compiler on certain
2927 special systems:
2929 (a) Some "const" qualifiers were missing.
2930 (b) Added the macro EXPORT before all exported functions; by default this
2931 is defined to be empty.
2932 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2933 that it reads its output file name as an argument instead of writing
2934 to the standard output and assuming this can be redirected.
2936 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2937 class containing characters with values greater than 255, PCRE compilation
2938 went into a loop.
2940 12. A recursive reference to a subpattern that was within another subpattern
2941 that had a minimum quantifier of zero caused PCRE to crash. For example,
2942 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2943 recursion. If the recursively-called subpattern itself had a zero repeat,
2944 that was OK.
2946 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2947 buffer into which it was copied (for escape processing) was still set at
2948 1024, so long lines caused crashes.
2950 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2951 "internal error: code overflow...". This applied to any character class
2952 that was followed by a possessive quantifier.
2954 15. Modified the Makefile to add libpcre.la as a prerequisite for
2955 libpcreposix.la because I was told this is needed for a parallel build to
2956 work.
2958 16. If a pattern that contained .* following optional items at the start was
2959 studied, the wrong optimizing data was generated, leading to matching
2960 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2961 matching string must start with a or b or c. The correct conclusion for
2962 this pattern is that a match can start with any character.
2965 Version 4.4 13-Aug-03
2966 ---------------------
2968 1. In UTF-8 mode, a character class containing characters with values between
2969 127 and 255 was not handled correctly if the compiled pattern was studied.
2970 In fixing this, I have also improved the studying algorithm for such
2971 classes (slightly).
2973 2. Three internal functions had redundant arguments passed to them. Removal
2974 might give a very teeny performance improvement.
2976 3. Documentation bug: the value of the capture_top field in a callout is *one
2977 more than* the number of the hightest numbered captured substring.
2979 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2980 in incorrectly linking with a previously installed version. They now link
2981 explicitly with libpcre.la.
2983 5. configure.in no longer needs to recognize Cygwin specially.
2985 6. A problem in pcre.in for Windows platforms is fixed.
2987 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2988 pcretest, it used to include the size of the study block as part of its
2989 output. Unfortunately, the structure contains a field that has a different
2990 size on different hardware architectures. This meant that the tests that
2991 showed this size failed. As the block is currently always of a fixed size,
2992 this information isn't actually particularly useful in pcretest output, so
2993 I have just removed it.
2995 8. Three pre-processor statements accidentally did not start in column 1.
2996 Sadly, there are *still* compilers around that complain, even though
2997 standard C has not required this for well over a decade. Sigh.
2999 9. In pcretest, the code for checking callouts passed small integers in the
3000 callout_data field, which is a void * field. However, some picky compilers
3001 complained about the casts involved for this on 64-bit systems. Now
3002 pcretest passes the address of the small integer instead, which should get
3003 rid of the warnings.
3005 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3006 both compile and run time, and gives an error if an invalid UTF-8 sequence
3007 is found. There is a option for disabling this check in cases where the
3008 string is known to be correct and/or the maximum performance is wanted.
3010 11. In response to a bug report, I changed one line in Makefile.in from
3012 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3013 to
3014 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3016 to look similar to other lines, but I have no way of telling whether this
3017 is the right thing to do, as I do not use Windows. No doubt I'll get told
3018 if it's wrong...
3021 Version 4.3 21-May-03
3022 ---------------------
3024 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3025 Makefile.
3027 2. Some refactoring to improve the quality of the code:
3029 (i) The utf8_table... variables are now declared "const".
3031 (ii) The code for \cx, which used the "case flipping" table to upper case
3032 lower case letters, now just substracts 32. This is ASCII-specific,
3033 but the whole concept of \cx is ASCII-specific, so it seems
3034 reasonable.
3036 (iii) PCRE was using its character types table to recognize decimal and
3037 hexadecimal digits in the pattern. This is silly, because it handles
3038 only 0-9, a-f, and A-F, but the character types table is locale-
3039 specific, which means strange things might happen. A private
3040 table is now used for this - though it costs 256 bytes, a table is
3041 much faster than multiple explicit tests. Of course, the standard
3042 character types table is still used for matching digits in subject
3043 strings against \d.
3045 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3046 ending in _t are). So I've renamed it as ESC_tee.
3048 3. The first argument for regexec() in the POSIX wrapper should have been
3049 defined as "const".
3051 4. Changed pcretest to use malloc() for its buffers so that they can be
3052 Electric Fenced for debugging.
3054 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3055 to read one or more bytes before the start of the subject string. Often this
3056 had no effect on PCRE's behaviour, but in some circumstances it could
3057 provoke a segmentation fault.
3059 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3060 to try to read one or more bytes before the start of the subject string.
3062 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3063 UTF-8 support could misbehave in various ways if the subject string
3064 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3065 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3066 back over UTF-8 characters.)
3069 Version 4.2 14-Apr-03
3070 ---------------------
3072 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3074 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3075 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3076 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3077 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3078 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3080 Note: automatic setting of the BUILD variables is not yet working
3081 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3082 compile-time but not at link-time
3083 [LINK]: use for linking executables only
3084 make different versions for Windows and non-Windows
3085 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3086 libraries
3087 [LINK_FOR_BUILD]: new variable
3088 [OBJEXT]: use throughout
3089 [EXEEXT]: use throughout
3090 <winshared>: new target
3091 <wininstall>: new target
3092 <dftables.o>: use native compiler
3093 <dftables>: use native linker
3094 <install>: handle Windows platform correctly
3095 <clean>: ditto
3096 <check>: ditto
3097 copy DLL to top builddir before testing
3099 As part of these changes, -no-undefined was removed again. This was reported
3100 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3101 in any case.
3103 3. Some tidies to get rid of compiler warnings:
3105 . In the match_data structure, match_limit was an unsigned long int, whereas
3106 match_call_count was an int. I've made them both unsigned long ints.
3108 . In pcretest the fact that a const uschar * doesn't automatically cast to
3109 a void * provoked a warning.
3111 . Turning on some more compiler warnings threw up some "shadow" variables
3112 and a few more missing casts.
3114 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3115 option, a class that contained a single character with a value between 128
3116 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3118 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3119 option, a class that contained several characters, but with at least one
3120 whose value was between 128 and 255 caused PCRE to crash.
3123 Version 4.1 12-Mar-03
3124 ---------------------
3126 1. Compiling with gcc -pedantic found a couple of places where casts were
3127 needed, and a string in dftables.c that was longer than standard compilers are
3128 required to support.
3130 2. Compiling with Sun's compiler found a few more places where the code could
3131 be tidied up in order to avoid warnings.
3133 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3134 first of these names is deprecated in the latest Autoconf in favour of the name
3135 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3136 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3137 analogy I have changed it to CFLAGS_FOR_BUILD.
3139 4. Added -no-undefined to the linking command in the Makefile, because this is
3140 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3141 linking step for the pcreposix library.
3143 5. PCRE was failing to diagnose the case of two named groups with the same
3144 name.
3146 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3147 literal character that is needed in the subject for a match, and scans along to
3148 ensure that it is present before embarking on the full matching process. This
3149 saves time in cases of nested unlimited repeats that are never going to match.
3150 Problem: the scan can take a lot of time if the subject is very long (e.g.
3151 megabytes), thus penalizing straightforward matches. It is now done only if the
3152 amount of subject to be scanned is less than 1000 bytes.
3154 7. A lesser problem with the same optimization is that it was recording the
3155 first character of an anchored pattern as "needed", thus provoking a search
3156 right along the subject, even when the first match of the pattern was going to
3157 fail. The "needed" character is now not set for anchored patterns, unless it
3158 follows something in the pattern that is of non-fixed length. Thus, it still
3159 fulfils its original purpose of finding quick non-matches in cases of nested
3160 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3163 Version 4.0 17-Feb-03
3164 ---------------------
3166 1. If a comment in an extended regex that started immediately after a meta-item
3167 extended to the end of string, PCRE compiled incorrect data. This could lead to
3168 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3170 2. Moved to autoconf 2.53 and libtool 1.4.2.
3172 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3173 the special perltest8 script is no longer needed - all the tests can be run
3174 from a single perltest script.
3176 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3177 by \s. It has now been removed in PCRE. This means it isn't recognized as
3178 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3179 class [:space:] *does* include VT, thereby creating a mess.
3181 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3182 space and tab.
3184 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3185 its new features into the main test script, reducing the number of scripts.
3187 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3188 were backward compatible, and made the (?i) apply to the whole pattern, as if
3189 /i were given. Now it behaves more logically, and applies the option setting
3190 only to what follows. PCRE has been changed to follow suit. However, if it
3191 finds options settings right at the start of the pattern, it extracts them into
3192 the global options, as before. Thus, they show up in the info data.
3194 8. Added support for the \Q...\E escape sequence. Characters in between are
3195 treated as literals. This is slightly different from Perl in that $ and @ are
3196 also handled as literals inside the quotes. In Perl, they will cause variable
3197 interpolation. Note the following examples:
3199 Pattern PCRE matches Perl matches
3201 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3202 \Qabc\$xyz\E abc\$xyz abc\$xyz
3203 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3205 For compatibility with Perl, \Q...\E sequences are recognized inside character
3206 classes as well as outside them.
3208 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3209 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3210 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3211 signed/unsigned warnings.
3213 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3214 option for pcretest, so I've replaced it by a simple function that does just
3215 that job.
3217 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3218 "pcregrep -".
3220 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3221 Java package. This provides some syntactic sugar for simple cases of what my
3222 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3223 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3224 item, you can use this simplified notation. Note that only makes sense with
3225 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3226 greediness, whatever the setting of the PCRE_UNGREEDY option.
3228 13. A change of greediness default within a pattern was not taking effect at
3229 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3230 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3231 was abstracted outside.
3233 14. PCRE now supports the \G assertion. It is true when the current matching
3234 position is at the start point of the match. This differs from \A when the
3235 starting offset is non-zero. Used with the /g option of pcretest (or similar
3236 code), it works in the same way as it does for Perl's /g option. If all
3237 alternatives of a regex begin with \G, the expression is anchored to the start
3238 match position, and the "anchored" flag is set in the compiled expression.
3240 15. Some bugs concerning the handling of certain option changes within patterns
3241 have been fixed. These applied to options other than (?ims). For example,
3242 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3243 been the other way round. Some of this was related to change 7 above.
3245 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3246 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3247 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3248 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3250 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3251 mode. Unlike ".", it always matches newline, whatever the setting of
3252 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3253 assertions. Perl allows it, but it doesn't (in general) work because it can't
3254 calculate the length of the lookbehind. At least, that's the case for Perl
3255 5.8.0 - I've been told they are going to document that it doesn't work in
3256 future.
3258 18. Added an error diagnosis for escapes that PCRE does not support: these are
3259 \L, \l, \N, \P, \p, \U, \u, and \X.
3261 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3262 reading past the end of the pattern in cases such as /[abcd/.
3264 20. PCRE was getting more memory than necessary for patterns with classes that
3265 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3267 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3268 compiling PCRE for use with Virtual Pascal.
3270 22. Small fix to the Makefile to make it work properly if the build is done
3271 outside the source tree.
3273 23. Added a new extension: a condition to go with recursion. If a conditional
3274 subpattern starts with (?(R) the "true" branch is used if recursion has
3275 happened, whereas the "false" branch is used only at the top level.
3277 24. When there was a very long string of literal characters (over 255 bytes
3278 without UTF support, over 250 bytes with UTF support), the computation of how
3279 much memory was required could be incorrect, leading to segfaults or other
3280 strange effects.
3282 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3283 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3284 there was a subsequent back reference to those brackets. This meant that, for
3285 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3286 possible to check for precisely this case. All we can do is abandon the
3287 optimization if .* occurs inside capturing brackets when there are any back
3288 references whatsoever. (See below for a better fix that came later.)
3290 26. The handling of the optimization for finding the first character of a
3291 non-anchored pattern, and for finding a character that is required later in the
3292 match were failing in some cases. This didn't break the matching; it just
3293 failed to optimize when it could. The way this is done has been re-implemented.
3295 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3297 28. Added a new feature that provides some of the functionality that Perl
3298 provides with (?{...}). The facility is termed a "callout". The way it is done
3299 in PCRE is for the caller to provide an optional function, by setting
3300 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3301 global variable. By default it is unset, which disables all calling out. To get
3302 the function called, the regex must include (?C) at appropriate points. This
3303 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3304 This provides a means of identifying different callout points. When PCRE
3305 reaches such a point in the regex, if pcre_callout has been set, the external
3306 function is called. It is provided with data in a structure called
3307 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3308 matching continues; if it returns a non-zero value, the match at the current
3309 point fails. However, backtracking will occur if possible. [This was changed
3310 later and other features added - see item 49 below.]
3312 29. pcretest is upgraded to test the callout functionality. It provides a
3313 callout function that displays information. By default, it shows the start of
3314 the match and the current position in the text. There are some new data escapes
3315 to vary what happens:
3317 \C+ in addition, show current contents of captured substrings
3318 \C- do not supply a callout function
3319 \C!n return 1 when callout number n is reached
3320 \C!n!m return 1 when callout number n is reached for the mth time
3322 30. If pcregrep was called with the -l option and just a single file name, it
3323 output "<stdin>" if a match was found, instead of the file name.
3325 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3326 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3327 pcre_exec(). This saves a malloc/free per call. The default value of
3328 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3329 when configuring.
3331 32. The default maximum size of a compiled pattern is 64K. There have been a
3332 few cases of people hitting this limit. The code now uses macros to handle the
3333 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3334 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3335 configuring. Tests 2 and 5 work only with 2-byte links because they output
3336 debugging information about compiled patterns.
3338 33. Internal code re-arrangements:
3340 (a) Moved the debugging function for printing out a compiled regex into
3341 its own source file (printint.c) and used #include to pull it into
3342 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3343 separate copies.
3345 (b) Defined the list of op-code names for debugging as a macro in
3346 internal.h so that it is next to the definition of the opcodes.
3348 (c) Defined a table of op-code lengths for simpler skipping along compiled
3349 code. This is again a macro in internal.h so that it is next to the
3350 definition of the opcodes.
3352 34. Added support for recursive calls to individual subpatterns, along the
3353 lines of Robin Houston's patch (but implemented somewhat differently).
3355 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3356 allow it to read and process whole directories in Win32. This code was
3357 contributed by Lionel Fourquaux; it has not been tested by me.
3359 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3360 used to name a group. Names consist of alphanumerics and underscores, and must
3361 be unique. Back references use the syntax (?P=name) and recursive calls use
3362 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3363 numbers. The function pcre_fullinfo() can be used after compilation to extract
3364 a name/number map. There are three relevant calls:
3366 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3367 PCRE_INFO_NAMECOUNT yields the number of entries
3368 PCRE_INFO_NAMETABLE yields a pointer to the map.
3370 The map is a vector of fixed-size entries. The size of each entry depends on
3371 the length of the longest name used. The first two bytes of each entry are the
3372 group number, most significant byte first. There follows the corresponding
3373 name, zero terminated. The names are in alphabetical order.
3375 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3376 case instead of 255. Making it the same both with and without UTF-8 support
3377 means that the same test output works with both.
3379 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3380 calling malloc() with a zero argument.
3382 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3383 optimization. I've improved things by keeping a bitmap of backreferences with
3384 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3385 fact referenced, the optimization can be applied. It is unlikely that a
3386 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3387 the match to follow \n) will appear inside brackets with a number greater than
3388 31, but if it does, any back reference > 31 suppresses the optimization.
3390 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3391 of disabling numbered capturing parentheses. Any opening parenthesis that is
3392 not followed by ? behaves as if it were followed by ?: but named parentheses
3393 can still be used for capturing (and they will acquire numbers in the usual
3394 way).
3396 41. Redesigned the return codes from the match() function into yes/no/error so
3397 that errors can be passed back from deep inside the nested calls. A malloc
3398 failure while inside a recursive subpattern call now causes the
3399 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3401 42. It is now possible to set a limit on the number of times the match()
3402 function is called in a call to pcre_exec(). This facility makes it possible to
3403 limit the amount of recursion and backtracking, though not in a directly
3404 obvious way, because the match() function is used in a number of different
3405 circumstances. The count starts from zero for each position in the subject
3406 string (for non-anchored patterns). The default limit is, for compatibility, a
3407 large number, namely 10 000 000. You can change this in two ways:
3409 (a) When configuring PCRE before making, you can use --with-match-limit=n
3410 to set a default value for the compiled library.
3412 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3413 a different value is set. See 45 below.
3415 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3417 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3418 of things that can be changed at compile time. The first argument specifies
3419 what is wanted and the second points to where the information is to be placed.
3420 The current list of available information is:
3424 The output is an integer that is set to one if UTF-8 support is available;
3425 otherwise it is set to zero.
3429 The output is an integer that it set to the value of the code that is used for
3430 newline. It is either LF (10) or CR (13).
3434 The output is an integer that contains the number of bytes used for internal
3435 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3439 The output is an integer that contains the threshold above which the POSIX
3440 interface uses malloc() for output vectors. See item 31 above.
3444 The output is an unsigned integer that contains the default limit of the number
3445 of match() calls in a pcre_exec() execution. See 42 above.
3447 44. pcretest has been upgraded by the addition of the -C option. This causes it
3448 to extract all the available output from the new pcre_config() function, and to
3449 output it. The program then exits immediately.
3451 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3452 order to support additional features. One way would have been to define
3453 pcre_exec2() (for example) with extra arguments, but this would not have been
3454 extensible, and would also have required all calls to the original function to
3455 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3456 is used for passing in "extra" data from pcre_study().
3458 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3459 contains the following fields:
3461 flags a bitmap indicating which of the following fields are set
3462 study_data opaque data from pcre_study()
3463 match_limit a way of specifying a limit on match() calls for a specific
3464 call to pcre_exec()
3465 callout_data data for callouts (see 49 below)
3467 The flag bits are also defined in pcre.h, and are
3473 The pcre_study() function now returns one of these new pcre_extra blocks, with
3474 the actual study data pointed to by the study_data field, and the
3475 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3476 before. That is, this change is entirely upwards-compatible and requires no
3477 change to existing code.
3479 If you want to pass in additional data to pcre_exec(), you can either place it
3480 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3481 block.
3483 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3484 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3485 times with different match limits, until it finds the minimum value needed for
3486 pcre_exec() to complete. The value is then output. This can be instructive; for
3487 most simple matches the number is quite small, but for pathological cases it
3488 gets very large very quickly.
3490 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3491 returns the size of the data block pointed to by the study_data field in a
3492 pcre_extra block, that is, the value that was passed as the argument to
3493 pcre_malloc() when PCRE was getting memory in which to place the information
3494 created by pcre_study(). The fourth argument should point to a size_t variable.
3495 pcretest has been extended so that this information is shown after a successful
3496 pcre_study() call when information about the compiled regex is being displayed.
3498 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3499 because what follows is always an absolute path. (Later: it turns out that this
3500 is more than cosmetic for MinGW, because it doesn't like empty path
3501 components.)
3503 49. Some changes have been made to the callout feature (see 28 above):
3505 (i) A callout function now has three choices for what it returns:
3507 0 => success, carry on matching
3508 > 0 => failure at this point, but backtrack if possible
3509 < 0 => serious error, return this value from pcre_exec()
3511 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3512 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3513 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3514 use by callout functions. It will never be used by PCRE itself.
3516 (ii) The pcre_extra structure (see 45 above) has a void * field called
3517 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3518 pcre_callout_block structure has a field of the same name. The contents of
3519 the field passed in the pcre_extra structure are passed to the callout
3520 function in the corresponding field in the callout block. This makes it
3521 easier to use the same callout-containing regex from multiple threads. For
3522 testing, the pcretest program has a new data escape
3524 \C*n pass the number n (may be negative) as callout_data
3526 If the callout function in pcretest receives a non-zero value as
3527 callout_data, it returns that value.
3529 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3530 there were some redundant $(CFLAGS) in commands that are now specified as
3531 $(LINK), which already includes $(CFLAGS).
3533 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3534 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3535 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3536 one-byte characters throughout. Note that case-insensitive matching applies
3537 only to characters whose values are less than 256. PCRE doesn't support the
3538 notion of cases for higher-valued characters.
3540 (i) A character class whose characters are all within 0-255 is handled as
3541 a bit map, and the map is inverted for negative classes. Previously, a
3542 character > 255 always failed to match such a class; however it should
3543 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3545 (ii) A negated character class with a single character < 255 is coded as
3546 "not this character" (OP_NOT). This wasn't working properly when the test
3547 character was multibyte, either singly or repeated.
3549 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3550 mode, for example: \x{100}{2,3}.
3552 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3553 singly or repeated) now correctly test multibyte characters. However,
3554 PCRE doesn't recognize any characters with values greater than 255 as
3555 digits, spaces, or word characters. Such characters always match \D, \S,
3556 and \W, and never match \d, \s, or \w.
3558 (v) Classes may now contain characters and character ranges with values
3559 greater than 255. For example: [ab\x{100}-\x{400}].
3561 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3562 PCRE in UTF-8 mode.
3564 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3565 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3566 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3567 value.)
3569 53. The single man page has become too large. I have therefore split it up into
3570 a number of separate man pages. These also give rise to individual HTML pages;
3571 these are now put in a separate directory, and there is an index.html page that
3572 lists them all. Some hyperlinking between the pages has been installed.
3574 54. Added convenience functions for handling named capturing parentheses.
3576 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3577 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3578 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3579 are faulted.
3581 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3582 calling configure. These values are used when compiling the dftables.c program
3583 which is run to generate the source of the default character tables. They
3584 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3585 you will need to set these values.
3587 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3590 Version 3.9 02-Jan-02
3591 ---------------------
3593 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3595 2. If --disable-static was given, the building process failed when trying to
3596 build pcretest and pcregrep. (For some reason it was using libtool to compile
3597 them, which is not right, as they aren't part of the library.)
3600 Version 3.8 18-Dec-01
3601 ---------------------
3603 1. The experimental UTF-8 code was completely screwed up. It was packing the
3604 bytes in the wrong order. How dumb can you get?
3607 Version 3.7 29-Oct-01
3608 ---------------------
3610 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3611 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3612 this didn't happen under Solaris 8, where I normally test things.
3614 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3615 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3616 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3617 this fix an hour or so after the initial 3.7 release.)
3620 Version 3.6 23-Oct-01
3621 ---------------------
3623 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3624 offsets passed as NULL with zero offset count.
3626 2. The config.guess and config.sub files had not been updated when I moved to
3627 the latest autoconf.
3630 Version 3.5 15-Aug-01
3631 ---------------------
3633 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3634 had been forgotten.
3636 2. By using declared but undefined structures, we can avoid using "void"
3637 definitions in pcre.h while keeping the internal definitions of the structures
3638 private.
3640 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3641 user point of view, this means that both static and shared libraries are built
3642 by default, but this can be individually controlled. More of the work of
3643 handling this static/shared cases is now inside libtool instead of PCRE's make
3644 file.
3646 4. The pcretest utility is now installed along with pcregrep because it is
3647 useful for users (to test regexs) and by doing this, it automatically gets
3648 relinked by libtool. The documentation has been turned into a man page, so
3649 there are now .1, .txt, and .html versions in /doc.
3651 5. Upgrades to pcregrep:
3652 (i) Added long-form option names like gnu grep.
3653 (ii) Added --help to list all options with an explanatory phrase.
3654 (iii) Added -r, --recursive to recurse into sub-directories.
3655 (iv) Added -f, --file to read patterns from a file.
3657 6. pcre_exec() was referring to its "code" argument before testing that
3658 argument for NULL (and giving an error if it was NULL).
3660 7. Upgraded Makefile.in to allow for compiling in a different directory from
3661 the source directory.
3663 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3664 options bits, the pointer it was passed was to an int instead of to an unsigned
3665 long int. This mattered only on 64-bit systems.
3667 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3668 generated) instead of pcre.in, which it its source. Also made the same change
3669 in several of the .c files.
3671 10. A new release of gcc defines printf() as a macro, which broke pcretest
3672 because it had an ifdef in the middle of a string argument for printf(). Fixed
3673 by using separate calls to printf().
3675 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3676 script, to force use of CR or LF instead of \n in the source. On non-Unix
3677 systems, the value can be set in config.h.
3679 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3680 absolute limit. Changed the text of the error message to make this clear, and
3681 likewise updated the man page.
3683 13. The limit of 99 on the number of capturing subpatterns has been removed.
3684 The new limit is 65535, which I hope will not be a "real" limit.
3687 Version 3.4 22-Aug-00
3688 ---------------------
3690 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3692 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3695 Version 3.3 01-Aug-00
3696 ---------------------
3698 1. If an octal character was given, but the value was greater than \377, it
3699 was not getting masked to the least significant bits, as documented. This could
3700 lead to crashes in some systems.
3702 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3703 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3705 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3706 These just pass their arguments on to (pcre_free)(), but they are provided
3707 because some uses of PCRE bind it to non-C systems that can call its functions,
3708 but cannot call free() or pcre_free() directly.
3710 4. Add "make test" as a synonym for "make check". Corrected some comments in
3711 the Makefile.
3713 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3714 Makefile.
3716 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3717 command called pgrep for grepping around the active processes.
3719 7. Added the beginnings of support for UTF-8 character strings.
3721 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3722 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3723 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3724 out for the ar command.)
3727 Version 3.2 12-May-00
3728 ---------------------
3730 This is purely a bug fixing release.
3732 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3733 of ZA. This was just one example of several cases that could provoke this bug,
3734 which was introduced by change 9 of version 2.00. The code for breaking
3735 infinite loops after an iteration that matches an empty string was't working
3736 correctly.
3738 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3739 when matched against abbab (for example). After matching an empty string, it
3740 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3741 caused it to match further down the string than it should.
3743 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3744 was there because it doesn't seem to be needed, and it causes trouble on some
3745 systems, as it is not a Standard C header. It has been removed.
3747 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3748 were reported on the Macintosh. The changes were from
3750 while ((c = *(++ptr)) != 0 && c != '\n');
3751 to
3752 while ((c = *(++ptr)) != 0 && c != '\n') ;
3754 Totally extraordinary, but if that's what it takes...
3756 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3757 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3758 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3759 assumes the way PCRE uses memmove() (always moving upwards).
3761 6. PCRE is being used in one environment where strchr() is not available. There
3762 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3763 faster code anyway.
3766 Version 3.1 09-Feb-00
3767 ---------------------
3769 The only change in this release is the fixing of some bugs in Makefile.in for
3770 the "install" target:
3772 (1) It was failing to install pcreposix.h.
3774 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3777 Version 3.0 01-Feb-00
3778 ---------------------
3780 1. Add support for the /+ modifier to perltest (to output $` like it does in
3781 pcretest).
3783 2. Add support for the /g modifier to perltest.
3785 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3786 matches null strings.
3788 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3789 pattern. Perl treats empty patterns specially - it reuses the most recent
3790 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3791 effect.
3793 5. The POSIX interface was broken in that it was just handing over the POSIX
3794 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3795 required a bigger vector, with some working space on the end. This means that
3796 the POSIX wrapper now has to get and free some memory, and copy the results.
3798 6. Added some simple autoconf support, placing the test data and the
3799 documentation in separate directories, re-organizing some of the
3800 information files, and making it build pcre-config (a GNU standard). Also added
3801 libtool support for building PCRE as a shared library, which is now the
3802 default.
3804 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3805 09 are not valid octal constants. Single digits will be used for minor values
3806 less than 10.
3808 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3809 existing programs that set these in the POSIX interface can use PCRE without
3810 modification.
3812 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3813 return all that pcre_info() returns, plus additional data. The pcre_info()
3814 function is retained for compatibility, but is considered to be obsolete.
3816 10. Added experimental recursion feature (?R) to handle one common case that
3817 Perl 5.6 will be able to do with (?p{...}).
3819 11. Added support for POSIX character classes like [:alpha:], which Perl is
3820 adopting.
3823 Version 2.08 31-Aug-99
3824 ----------------------
3826 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3827 trying to match at the startoffset position, but instead was moving forward to
3828 the next newline as if a previous match had failed.
3830 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3831 and could get into a loop if a null string was matched other than at the start
3832 of the subject.
3834 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3835 be distinguished at compile time, and for completeness also added PCRE_DATE.
3837 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3838 in GnuWin32 environments.
3841 Version 2.07 29-Jul-99
3842 ----------------------
3844 1. The documentation is now supplied in plain text form and HTML as well as in
3845 the form of man page sources.
3847 2. C++ compilers don't like assigning (void *) values to other pointer types.
3848 In particular this affects malloc(). Although there is no problem in Standard
3849 C, I've put in casts to keep C++ compilers happy.
3851 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3852 should be (const char *).
3854 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3855 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3856 However, I haven't made this a standard facility. The documentation doesn't
3857 mention it, and the Makefile doesn't support it.
3859 5. The Makefile now contains an "install" target, with editable destinations at
3860 the top of the file. The pcretest program is not installed.
3862 6. pgrep -V now gives the PCRE version number and date.
3864 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3865 causing the entire string to be ignored, instead of just the last character.
3867 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3868 non-matching string, it can take a very, very long time, even for strings of
3869 quite modest length, because of the nested recursion. PCRE now does better in
3870 some of these cases. It does this by remembering the last required literal
3871 character in the pattern, and pre-searching the subject to ensure it is present
3872 before running the real match. In other words, it applies a heuristic to detect
3873 some types of certain failure quickly, and in the above example, if presented
3874 with a string that has no trailing " it gives "no match" very quickly.
3876 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3877 other alternatives are tried instead.
3880 Version 2.06 09-Jun-99
3881 ----------------------
3883 1. Change pcretest's output for amount of store used to show just the code
3884 space, because the remainder (the data block) varies in size between 32-bit and
3885 64-bit systems.
3887 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3888 start matching at. This allows lookbehinds to work when searching for multiple
3889 occurrences in a string.
3891 3. Added additional options to pcretest for testing multiple occurrences:
3893 /+ outputs the rest of the string that follows a match
3894 /g loops for multiple occurrences, using the new startoffset argument
3895 /G loops for multiple occurrences by passing an incremented pointer
3897 4. PCRE wasn't doing the "first character" optimization for patterns starting
3898 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3899 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3900 the letter 'x'. On long subject strings, this gives a significant speed-up.
3903 Version 2.05 21-Apr-99
3904 ----------------------
3906 1. Changed the type of magic_number from int to long int so that it works
3907 properly on 16-bit systems.
3909 2. Fixed a bug which caused patterns starting with .* not to work correctly
3910 when the subject string contained newline characters. PCRE was assuming
3911 anchoring for such patterns in all cases, which is not correct because .* will
3912 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3913 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3914 must be retried after every newline in the subject.
3917 Version 2.04 18-Feb-99
3918 ----------------------
3920 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3921 computation of the store needed to hold the pattern was incorrect (too large).
3922 If such patterns were nested a few deep, this could multiply and become a real
3923 problem.
3925 2. Added /M option to pcretest to show the memory requirement of a specific
3926 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3928 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3929 compiled in such a way that the backtracking after subsequent failure was
3930 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3931 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3934 Version 2.03 02-Feb-99
3935 ----------------------
3937 1. Fixed typo and small mistake in man page.
3939 2. Added 4th condition (GPL supersedes if conflict) and created separate
3940 LICENCE file containing the conditions.
3942 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3943 Perl, that is the internal \ allows the delimiter to be included in the
3944 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3945 the final delimiter, add \ to the end of the pattern (to test the error).
3947 4. Added the convenience functions for extracting substrings after a successful
3948 match. Updated pcretest to make it able to test these functions.
3951 Version 2.02 14-Jan-99
3952 ----------------------
3954 1. Initialized the working variables associated with each extraction so that
3955 their saving and restoring doesn't refer to uninitialized store.
3957 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3958 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3959 fix the problem.
3961 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3962 calls, and wasn't printing the correct value for compiling calls. Increased the
3963 default value of LOOPREPEAT, and the number of significant figures in the
3964 times.
3966 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3968 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3969 a building problem on Windows NT with a FAT file system.
3972 Version 2.01 21-Oct-98
3973 ----------------------
3975 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3976 to character tables built by pcre_maketables() in the current locale. If NULL
3977 is passed, the default tables are used.
3980 Version 2.00 24-Sep-98
3981 ----------------------
3983 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3984 it any more.
3986 2. Allow quantification of (?>) groups, and make it work correctly.
3988 3. The first character computation wasn't working for (?>) groups.
3990 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3991 end of the subject) and add 5.005's \z, which really does match only at the
3992 very end of the subject.
3994 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3996 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3997 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3998 localized options. All options to pcre_study() were also removed.
4000 7. Add other new features from 5.005:
4002 $(?<= positive lookbehind
4003 $(?<! negative lookbehind
4004 (?imsx-imsx) added the unsetting capability
4005 such a setting is global if at outer level; local otherwise
4006 (?imsx-imsx:) non-capturing groups with option setting
4007 (?(cond)re|re) conditional pattern matching
4009 A backreference to itself in a repeated group matches the previous
4010 captured string.
4012 8. General tidying up of studying (both automatic and via "study")
4013 consequential on the addition of new assertions.
4015 9. As in 5.005, unlimited repeated groups that could match an empty substring
4016 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4017 runtime if any iteration does actually match an empty substring.
4019 10. Include the RunTest script in the distribution.
4021 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4022 discrepancies, some of which were old and were also with respect to 5.004. They
4023 have now been fixed.
4026 Version 1.09 28-Apr-98
4027 ----------------------
4029 1. A negated single character class followed by a quantifier with a minimum
4030 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4031 program crashes, or just wrong answers. This did not apply to negated classes
4032 containing more than one character, or to minima other than one.
4035 Version 1.08 27-Mar-98
4036 ----------------------
4038 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4040 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4041 latter must appear before anything that relies on it in the pattern.
4044 Version 1.07 16-Feb-98
4045 ----------------------
4047 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4048 repeat of a potentially empty string).
4051 Version 1.06 23-Jan-98
4052 ----------------------
4054 1. Added Markus Oberhumer's little patches for C++.
4056 2. Literal strings longer than 255 characters were broken.
4059 Version 1.05 23-Dec-97
4060 ----------------------
4062 1. Negated character classes containing more than one character were failing if
4063 PCRE_CASELESS was set at run time.
4066 Version 1.04 19-Dec-97
4067 ----------------------
4069 1. Corrected the man page, where some "const" qualifiers had been omitted.
4071 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4072 input syntax.
4074 3. Fixed memory leak which occurred when a regex with back references was
4075 matched with an offsets vector that wasn't big enough. The temporary memory
4076 that is used in this case wasn't being freed if the match failed.
4078 4. Tidied pcretest to ensure it frees memory that it gets.
4080 5. Temporary memory was being obtained in the case where the passed offsets
4081 vector was exactly big enough.
4083 6. Corrected definition of offsetof() from change 5 below.
4085 7. I had screwed up change 6 below and broken the rules for the use of
4086 setjmp(). Now fixed.
4089 Version 1.03 18-Dec-97
4090 ----------------------
4092 1. A erroneous regex with a missing opening parenthesis was correctly
4093 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4094 on some systems.
4096 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4097 it was reported that one broken compiler failed on the former because "code" is
4098 also an independent variable.
4100 3. The erroneous regex a[]b caused an array overrun reference.
4102 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4103 fail on data ending with that character. (It was going on too far, and checking
4104 the next character, typically a binary zero.) This was specific to the
4105 optimized code for single-character negative classes.
4107 5. Added a contributed patch from the TIN world which does the following:
4109 + Add an undef for memmove, in case the the system defines a macro for it.
4111 + Add a definition of offsetof(), in case there isn't one. (I don't know
4112 the reason behind this - offsetof() is part of the ANSI standard - but
4113 it does no harm).
4115 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4116 most of the places where whitespace preceded '#'. I have given up and
4117 allowed the remaining 2 cases to be at the margin.
4119 + Rename some variables in pcre to eliminate shadowing. This seems very
4120 pedantic, but does no harm, of course.
4122 6. Moved the call to setjmp() into its own function, to get rid of warnings
4123 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4125 7. Constructs such as \d{8,} were compiling into the equivalent of
4126 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4127 outcome, but in this particular case used more store than had been allocated,
4128 which caused the bug to be discovered because it threw up an internal error.
4130 8. The debugging code in both pcre and pcretest for outputting the compiled
4131 form of a regex was going wrong in the case of back references followed by
4132 curly-bracketed repeats.
4135 Version 1.02 12-Dec-97
4136 ----------------------
4138 1. Typos in pcre.3 and comments in the source fixed.
4140 2. Applied a contributed patch to get rid of places where it used to remove
4141 'const' from variables, and fixed some signed/unsigned and uninitialized
4142 variable warnings.
4144 3. Added the "runtest" target to Makefile.
4146 4. Set default compiler flag to -O2 rather than just -O.
4149 Version 1.01 19-Nov-97
4150 ----------------------
4152 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4153 like /([ab]*)*/, that is, for classes with more than one character in them.
4155 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4156 as /((?>a*))*/ (a PCRE_EXTRA facility).
4159 Version 1.00 18-Nov-97
4160 ----------------------
4162 1. Added compile-time macros to support systems such as SunOS4 which don't have
4163 memmove() or strerror() but have other things that can be used instead.
4165 2. Arranged that "make clean" removes the executables.
4168 Version 0.99 27-Oct-97
4169 ----------------------
4171 1. Fixed bug in code for optimizing classes with only one character. It was
4172 initializing a 32-byte map regardless, which could cause it to run off the end
4173 of the memory it had got.
4175 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4178 Version 0.98 22-Oct-97
4179 ----------------------
4181 1. Fixed bug in code for handling temporary memory usage when there are more
4182 back references than supplied space in the ovector. This could cause segfaults.
4185 Version 0.97 21-Oct-97
4186 ----------------------
4188 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4190 2. Optimized negated single characters not to use a bit map.
4192 3. Brought error texts together as macro definitions; clarified some of them;
4193 fixed one that was wrong - it said "range out of order" when it meant "invalid
4194 escape sequence".
4196 4. Changed some char * arguments to const char *.
4198 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4200 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4201 pcretest.
4204 Version 0.96 16-Oct-97
4205 ----------------------
4207 1. Added a simple "pgrep" utility to the distribution.
4209 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4210 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4211 where "ddd" means "one or more decimal digits".
4213 3. Fixed serious bug. If a pattern had a back reference, but the call to
4214 pcre_exec() didn't supply a large enough ovector to record the related
4215 identifying subpattern, the match always failed. PCRE now remembers the number
4216 of the largest back reference, and gets some temporary memory in which to save
4217 the offsets during matching if necessary, in order to ensure that
4218 backreferences always work.
4220 4. Increased the compatibility with Perl in a number of ways:
4222 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4223 to request this handling. The option can be set at compile or exec time.
4225 (b) $ matches before a terminating newline by default; an option
4226 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4227 mode). The option can be set at compile or exec time.
4229 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4230 the same as Perl's. If the decimal number it represents is less than 10
4231 or there aren't that many previous left capturing parentheses, an octal
4232 escape is read. Inside a character class, it's always an octal escape,
4233 even if it is a single digit.
4235 (d) An escaped but undefined alphabetic character is taken as a literal,
4236 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4237 escapes.
4239 (e) {0} is now permitted. (The previous item is removed from the compiled
4240 pattern).
4242 5. Changed all the names of code files so that the basic parts are no longer
4243 than 10 characters, and abolished the teeny "globals.c" file.
4245 6. Changed the handling of character classes; they are now done with a 32-byte
4246 bit map always.
4248 7. Added the -d and /D options to pcretest to make it possible to look at the
4249 internals of compilation without having to recompile pcre.
4252 Version 0.95 23-Sep-97
4253 ----------------------
4255 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4256 \x20 at the start of a run of normal characters. These were being treated as
4257 real characters, instead of the source characters being re-checked.
4260 Version 0.94 18-Sep-97
4261 ----------------------
4263 1. The functions are now thread-safe, with the caveat that the global variables
4264 containing pointers to malloc() and free() or alternative functions are the
4265 same for all threads.
4267 2. Get pcre_study() to generate a bitmap of initial characters for non-
4268 anchored patterns when this is possible, and use it if passed to pcre_exec().
4271 Version 0.93 15-Sep-97
4272 ----------------------
4274 1. /(b)|(:+)/ was computing an incorrect first character.
4276 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4277 but not actually doing anything yet.
4279 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4280 as Perl does (e.g. [-az] or [az-]).
4282 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4283 all possible positions.
4285 5. Split up into different modules to avoid including unneeded functions in a
4286 compiled binary. However, compile and exec are still in one module. The "study"
4287 function is split off.
4289 6. The character tables are now in a separate module whose source is generated
4290 by an auxiliary program - but can then be edited by hand if required. There are
4291 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4292 toupper() in the code.
4294 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4295 make them global. Abolish the function for setting them, as the caller can now
4296 set them directly.
4299 Version 0.92 11-Sep-97
4300 ----------------------
4302 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4303 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4305 2. Caseless matching was not working in character classes if the characters in
4306 the pattern were in upper case.
4308 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4310 4. Make PCRE_ANCHORED public and accept as a compile option.
4312 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4313 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4314 pass them.
4316 6. Give an error if bad option bits passed at compile or run time.
4318 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4319 pcretest to cause it to pass that flag.
4321 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4322 options, and the first character, if set.
4324 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4327 Version 0.91 10-Sep-97
4328 ----------------------
4330 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4331 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4333 2. PCRE was looping on encountering an indefinitely repeated back reference to
4334 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4335 Perl does - treats the match as successful.
4337 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5