ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 649 - (show annotations)
Mon Aug 1 11:23:52 2011 UTC (10 years, 2 months ago) by ph10
File size: 211438 byte(s)
Fix partial match bug with /f.*/8s.
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.
216 42. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
217 as not to get warnings when autogen.sh is called. Also changed
218 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
220 43. To help people who use pcregrep to scan files containing exceedingly long
221 lines, the following changes have been made:
223 (a) The default value of the buffer size parameter has been increased from
224 8K to 20K. (A buffer three times this size is actually used.)
226 (b) The default can be changed by ./configure --with-pcregrep-bufsiz when
227 PCRE is built.
229 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
230 to be set at run time.
232 (d) Numerical values in pcregrep options can be followed by K or M, for
233 example --buffer-size=50K.
235 (e) If a line being scanned overflows pcregrep's buffer, an error is now
236 given and the return code is set to 2.
238 44. Add a pointer to the latest mark to the callout data block.
240 45. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
241 partial match of an empty string instead of no match. This was specific to
242 the use of ".".
244 46. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
245 complete match instead of a partial match. This bug was dependent on both
246 the PCRE_UTF8 and PCRE_DOTALL options being set.
249 Version 8.12 15-Jan-2011
250 ------------------------
252 1. Fixed some typos in the markup of the man pages, and wrote a script that
253 checks for such things as part of the documentation building process.
255 2. On a big-endian 64-bit system, pcregrep did not correctly process the
256 --match-limit and --recursion-limit options (added for 8.11). In
257 particular, this made one of the standard tests fail. (The integer value
258 went into the wrong half of a long int.)
260 3. If the --colour option was given to pcregrep with -v (invert match), it
261 did strange things, either producing crazy output, or crashing. It should,
262 of course, ignore a request for colour when reporting lines that do not
263 match.
265 4. Another pcregrep bug caused similar problems if --colour was specified with
266 -M (multiline) and the pattern match finished with a line ending.
268 5. In pcregrep, when a pattern that ended with a literal newline sequence was
269 matched in multiline mode, the following line was shown as part of the
270 match. This seems wrong, so I have changed it.
272 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
273 the check for further matches in the same line (so they could be coloured)
274 to overrun the end of the current line. If another match was found, it was
275 incorrectly shown (and then shown again when found in the next line).
277 7. If pcregrep was compiled under Windows, there was a reference to the
278 function pcregrep_exit() before it was defined. I am assuming this was
279 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
280 reported by a user. I've moved the definition above the reference.
283 Version 8.11 10-Dec-2010
284 ------------------------
286 1. (*THEN) was not working properly if there were untried alternatives prior
287 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
288 backtracked to try for "b" instead of moving to the next alternative branch
289 at the same level (in this case, to look for "c"). The Perl documentation
290 is clear that when (*THEN) is backtracked onto, it goes to the "next
291 alternative in the innermost enclosing group".
293 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
294 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
295 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
296 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
297 (*THEN).
299 3. If \s appeared in a character class, it removed the VT character from
300 the class, even if it had been included by some previous item, for example
301 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
302 of \s, but is part of the POSIX "space" class.)
304 4. A partial match never returns an empty string (because you can always
305 match an empty string at the end of the subject); however the checking for
306 an empty string was starting at the "start of match" point. This has been
307 changed to the "earliest inspected character" point, because the returned
308 data for a partial match starts at this character. This means that, for
309 example, /(?<=abc)def/ gives a partial match for the subject "abc"
310 (previously it gave "no match").
312 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
313 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
314 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
315 has an implication that the given string is incomplete (because a partial
316 match is preferred over a full match). For this reason, these items now
317 give a partial match in this situation. [Aside: previously, the one case
318 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
319 match rather than a full match, which was wrong by the old rules, but is
320 now correct.]
322 6. There was a bug in the handling of #-introduced comments, recognized when
323 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
324 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
325 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
326 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
327 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
328 places in pcre_compile().
330 7. Related to (6) above, when pcre_compile() was skipping #-introduced
331 comments when looking ahead for named forward references to subpatterns,
332 the only newline sequence it recognized was NL. It now handles newlines
333 according to the set newline convention.
335 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
336 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
337 cater for a lack of strerror(). These oversights have been fixed.
339 9. Added --match-limit and --recursion-limit to pcregrep.
341 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
343 11. When the -o option was used, pcregrep was setting a return code of 1, even
344 when matches were found, and --line-buffered was not being honoured.
346 12. Added an optional parentheses number to the -o and --only-matching options
347 of pcregrep.
349 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
350 can match an empty string. The code to do it in pcretest and pcredemo
351 needed fixing:
353 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
354 only one byte after an empty string match just before CRLF (this case
355 just got forgotten; "any" and "anycrlf" were OK).
357 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
358 mode when an empty string match preceded an ASCII character followed by
359 a non-ASCII character. (The code for advancing by one character rather
360 than one byte was nonsense.)
362 (c) The pcredemo.c sample program did not have any code at all to handle
363 the cases when CRLF is a valid newline sequence.
365 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
366 as a starting offset was within the subject string. There is now a new
367 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
368 negative or greater than the length of the string. In order to test this,
369 pcretest is extended to allow the setting of negative starting offsets.
371 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
372 starting offset points to the beginning of a UTF-8 character was
373 unnecessarily clumsy. I tidied it up.
375 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
376 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
378 17. Nobody had reported that the --include_dir option, which was added in
379 release 7.7 should have been called --include-dir (hyphen, not underscore)
380 for compatibility with GNU grep. I have changed it to --include-dir, but
381 left --include_dir as an undocumented synonym, and the same for
382 --exclude-dir, though that is not available in GNU grep, at least as of
383 release 2.5.4.
385 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
386 characters from a string of bytes) have been redefined so as not to use
387 loops, in order to improve performance in some environments. At the same
388 time, I abstracted some of the common code into auxiliary macros to save
389 repetition (this should not affect the compiled code).
391 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
392 compile-time error is now given if \c is not followed by an ASCII
393 character, that is, a byte less than 128. (In EBCDIC mode, the code is
394 different, and any byte value is allowed.)
396 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
397 START_OPTIMIZE option, which is now allowed at compile time - but just
398 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
399 to pcregrep and other applications that have no direct access to PCRE
400 options. The new /Y option in pcretest sets this option when calling
401 pcre_compile().
403 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
404 back references. Groups containing recursive back references were forced to
405 be atomic by that change, but in the case of named groups, the amount of
406 memory required was incorrectly computed, leading to "Failed: internal
407 error: code overflow". This has been fixed.
409 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
410 pcretest.c, to avoid build problems in some Borland environments.
413 Version 8.10 25-Jun-2010
414 ------------------------
416 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
417 THEN.
419 2. (*ACCEPT) was not working when inside an atomic group.
421 3. Inside a character class, \B is treated as a literal by default, but
422 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
423 causes the error). The code is unchanged, but I tidied the documentation.
425 4. Inside a character class, PCRE always treated \R and \X as literals,
426 whereas Perl faults them if its -w option is set. I have changed PCRE so
427 that it faults them when PCRE_EXTRA is set.
429 5. Added support for \N, which always matches any character other than
430 newline. (It is the same as "." when PCRE_DOTALL is not set.)
432 6. When compiling pcregrep with newer versions of gcc which may have
433 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
434 declared with attribute warn_unused_result" were given. Just casting the
435 result to (void) does not stop the warnings; a more elaborate fudge is
436 needed. I've used a macro to implement this.
438 7. Minor change to pcretest.c to avoid a compiler warning.
440 8. Added four artifical Unicode properties to help with an option to make
441 \s etc use properties (see next item). The new properties are: Xan
442 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
444 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
445 use Unicode properties. (*UCP) at the start of a pattern can be used to set
446 this option. Modified pcretest to add /W to test this facility. Added
447 REG_UCP to make it available via the POSIX interface.
449 10. Added --line-buffered to pcregrep.
451 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
452 studied, and the match started with a letter with a code point greater than
453 127 whose first byte was different to the first byte of the other case of
454 the letter, the other case of this starting letter was not recognized
455 (#976).
457 12. If a pattern that was studied started with a repeated Unicode property
458 test, for example, \p{Nd}+, there was the theoretical possibility of
459 setting up an incorrect bitmap of starting bytes, but fortunately it could
460 not have actually happened in practice until change 8 above was made (it
461 added property types that matched character-matching opcodes).
463 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
464 possible starting bytes for non-anchored patterns.
466 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
467 \R, and also a number of cases that involve Unicode properties, both
468 explicit and implicit when PCRE_UCP is set.
470 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
471 input, it could crash or give wrong results if characters with values
472 greater than 0xc0 were present in the subject string. (Detail: it assumed
473 UTF-8 input when processing these items.)
475 16. Added a lot of (int) casts to avoid compiler warnings in systems where
476 size_t is 64-bit (#991).
478 17. Added a check for running out of memory when PCRE is compiled with
479 --disable-stack-for-recursion (#990).
481 18. If the last data line in a file for pcretest does not have a newline on
482 the end, a newline was missing in the output.
484 19. The default pcre_chartables.c file recognizes only ASCII characters (values
485 less than 128) in its various bitmaps. However, there is a facility for
486 generating tables according to the current locale when PCRE is compiled. It
487 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
488 characters, are recognized by isspace() and therefore were getting set in
489 these tables, and indeed these tables seem to approximate to ISO 8859. This
490 caused a problem in UTF-8 mode when pcre_study() was used to create a list
491 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
492 which of course cannot start UTF-8 characters. I have changed the code so
493 that only real ASCII characters (less than 128) and the correct starting
494 bytes for UTF-8 encodings are set for characters greater than 127 when in
495 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
496 altogether.)
498 20. Added the /T option to pcretest so as to be able to run tests with non-
499 standard character tables, thus making it possible to include the tests
500 used for 19 above in the standard set of tests.
502 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
503 reference to a subpattern the other side of a comment that contains an
504 opening parenthesis caused either an internal compiling error, or a
505 reference to the wrong subpattern.
508 Version 8.02 19-Mar-2010
509 ------------------------
511 1. The Unicode data tables have been updated to Unicode 5.2.0.
513 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
514 configured.
516 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
517 original author of that file, following a query about its status.
519 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
520 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
522 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
523 quantifier applied to a forward-referencing subroutine call, could compile
524 incorrect code or give the error "internal error: previously-checked
525 referenced subpattern not found".
527 6. Both MS Visual Studio and Symbian OS have problems with initializing
528 variables to point to external functions. For these systems, therefore,
529 pcre_malloc etc. are now initialized to local functions that call the
530 relevant global functions.
532 7. There were two entries missing in the vectors called coptable and poptable
533 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
534 I've fixed the data, and added a kludgy way of testing at compile time that
535 the lengths are correct (equal to the number of opcodes).
537 8. Following on from 7, I added a similar kludge to check the length of the
538 eint vector in pcreposix.c.
540 9. Error texts for pcre_compile() are held as one long string to avoid too
541 much relocation at load time. To find a text, the string is searched,
542 counting zeros. There was no check for running off the end of the string,
543 which could happen if a new error number was added without updating the
544 string.
546 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
548 11. \K was not working if it appeared in an atomic group or in a group that
549 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
550 \K is "not well defined" if used in an assertion. PCRE now accepts it if
551 the assertion is positive, but not if it is negative.
553 12. Change 11 fortuitously reduced the size of the stack frame used in the
554 "match()" function of pcre_exec.c by one pointer. Forthcoming
555 implementation of support for (*MARK) will need an extra pointer on the
556 stack; I have reserved it now, so that the stack frame size does not
557 decrease.
559 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
560 item in branch that calls a recursion is a subroutine call - as in the
561 second branch in the above example - was incorrectly given the compile-
562 time error "recursive call could loop indefinitely" because pcre_compile()
563 was not correctly checking the subroutine for matching a non-empty string.
565 14. The checks for overrunning compiling workspace could trigger after an
566 overrun had occurred. This is a "should never occur" error, but it can be
567 triggered by pathological patterns such as hundreds of nested parentheses.
568 The checks now trigger 100 bytes before the end of the workspace.
570 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
573 Version 8.01 19-Jan-2010
574 ------------------------
576 1. If a pattern contained a conditional subpattern with only one branch (in
577 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
578 computed the wrong minimum data length (which is of course zero for such
579 subpatterns). This could cause incorrect "no match" results.
581 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
582 the pattern is reset in the first branch, pcre_compile() failed with
583 "internal error: code overflow at offset...". This happened only when
584 the reset was to the original external option setting. (An optimization
585 abstracts leading options settings into an external setting, which was the
586 cause of this.)
588 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
589 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
590 assertion pattern did not match (meaning that the assertion was true), it
591 was incorrectly treated as false if the SKIP had been reached during the
592 matching. This also applied to assertions used as conditions.
594 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
595 assertion subpattern, including such a pattern used as a condition,
596 unpredictable results occurred, instead of the error return
599 5. The C++ GlobalReplace function was not working like Perl for the special
600 situation when an empty string is matched. It now does the fancy magic
601 stuff that is necessary.
603 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
604 removed. (These were left over from very, very early versions of PCRE.)
606 7. Some cosmetic changes to the code to make life easier when compiling it
607 as part of something else:
609 (a) Change DEBUG to PCRE_DEBUG.
611 (b) In pcre_compile(), rename the member of the "branch_chain" structure
612 called "current" as "current_branch", to prevent a collision with the
613 Linux macro when compiled as a kernel module.
615 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
616 prevent a collision with the Linux macro when compiled as a kernel
617 module.
619 8. In pcre_compile() there are some checks for integer overflows that used to
620 cast potentially large values to (double). This has been changed to that
621 when building, a check for int64_t is made, and if it is found, it is used
622 instead, thus avoiding the use of floating point arithmetic. (There is no
623 other use of FP in PCRE.) If int64_t is not found, the fallback is to
624 double.
626 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
627 2005 (difference between two addresses compared to an unsigned value).
629 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
630 custom one, because of the following reported problem in Windows:
632 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
633 under Win32.
634 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
635 therefore missing the function definition.
636 - The compiler thus generates a "C" signature for the test function.
637 - The linker fails to find the "C" function.
638 - PCRE fails to configure if asked to do so against libbz2.
640 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
641 messages were output:
643 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
644 rerunning libtoolize, to keep the correct libtool macros in-tree.
645 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
647 I have done both of these things.
649 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
650 most of the time, it *can* run out if it is given a pattern that contains a
651 runaway infinite recursion. I updated the discussion in the pcrestack man
652 page.
654 13. Now that we have gone to the x.xx style of version numbers, the minor
655 version may start with zero. Using 08 or 09 is a bad idea because users
656 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
657 interpreted as invalid octal numbers. I've updated the previous comment in
658 configure.ac, and also added a check that gives an error if 08 or 09 are
659 used.
661 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
662 causing partial matching to fail when the end of the subject matched \W
663 in a UTF-8 pattern where \W was quantified with a minimum of 3.
665 15. There were some discrepancies between the declarations in pcre_internal.h
666 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
667 their definitions. The declarations used "const uschar *" and the
668 definitions used USPTR. Even though USPTR is normally defined as "const
669 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
670 reported that: "This difference in casting confuses some C++ compilers, for
671 example, SunCC recognizes above declarations as different functions and
672 generates broken code for hbpcre." I have changed the declarations to use
673 USPTR.
675 16. GNU libtool is named differently on some systems. The autogen.sh script now
676 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
677 (FreeBSD).
679 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
680 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
681 comment: "Figure out how to create a longlong from a string: strtoll and
682 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
683 instance, but it only takes 2 args instead of 3!"
685 18. A subtle bug concerned with back references has been fixed by a change of
686 specification, with a corresponding code fix. A pattern such as
687 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
688 refers, was giving matches when it shouldn't. For example, xa=xaaa would
689 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
690 same bug. Such groups have to be quantified to be useful, or contained
691 inside another quantified group. (If there's no repetition, the reference
692 can never match.) The problem arises because, having left the group and
693 moved on to the rest of the pattern, a later failure that backtracks into
694 the group uses the captured value from the final iteration of the group
695 rather than the correct earlier one. I have fixed this in PCRE by forcing
696 any group that contains a reference to itself to be an atomic group; that
697 is, there cannot be any backtracking into it once it has completed. This is
698 similar to recursive and subroutine calls.
701 Version 8.00 19-Oct-09
702 ----------------------
704 1. The table for translating pcre_compile() error codes into POSIX error codes
705 was out-of-date, and there was no check on the pcre_compile() error code
706 being within the table. This could lead to an OK return being given in
707 error.
709 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
710 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
711 in a Windows environment.
713 3. The pcregrep --count option prints the count for each file even when it is
714 zero, as does GNU grep. However, pcregrep was also printing all files when
715 --files-with-matches was added. Now, when both options are given, it prints
716 counts only for those files that have at least one match. (GNU grep just
717 prints the file name in this circumstance, but including the count seems
718 more useful - otherwise, why use --count?) Also ensured that the
719 combination -clh just lists non-zero counts, with no names.
721 4. The long form of the pcregrep -F option was incorrectly implemented as
722 --fixed_strings instead of --fixed-strings. This is an incompatible change,
723 but it seems right to fix it, and I didn't think it was worth preserving
724 the old behaviour.
726 5. The command line items --regex=pattern and --regexp=pattern were not
727 recognized by pcregrep, which required --regex pattern or --regexp pattern
728 (with a space rather than an '='). The man page documented the '=' forms,
729 which are compatible with GNU grep; these now work.
731 6. No libpcreposix.pc file was created for pkg-config; there was just
732 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
734 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
735 when UCP support is not needed, by modifying the Python script that
736 generates it from Unicode data files. This should not matter if the module
737 is correctly used as a library, but I received one complaint about 50K of
738 unwanted data. My guess is that the person linked everything into his
739 program rather than using a library. Anyway, it does no harm.
741 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
742 was a minimum greater than 1 for a wide character in a possessive
743 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
744 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
745 character. Chaos in the form of incorrect output or a compiling loop could
746 result.
748 9. The restrictions on what a pattern can contain when partial matching is
749 requested for pcre_exec() have been removed. All patterns can now be
750 partially matched by this function. In addition, if there are at least two
751 slots in the offset vector, the offset of the earliest inspected character
752 for the match and the offset of the end of the subject are set in them when
753 PCRE_ERROR_PARTIAL is returned.
755 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
756 synonymous with PCRE_PARTIAL, for backwards compatibility, and
757 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
758 and may be more useful for multi-segment matching.
760 11. Partial matching with pcre_exec() is now more intuitive. A partial match
761 used to be given if ever the end of the subject was reached; now it is
762 given only if matching could not proceed because another character was
763 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
764 string "Z", which now yields "no match" instead of "partial match". In the
765 case of pcre_dfa_exec(), "no match" is given if every matching path for the
766 final character ended with (*FAIL).
768 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
769 if the pattern had a "must contain" character that was already found in the
770 earlier partial match, unless partial matching was again requested. For
771 example, with the pattern /dog.(body)?/, the "must contain" character is
772 "g". If the first part-match was for the string "dog", restarting with
773 "sbody" failed. This bug has been fixed.
775 13. The string returned by pcre_dfa_exec() after a partial match has been
776 changed so that it starts at the first inspected character rather than the
777 first character of the match. This makes a difference only if the pattern
778 starts with a lookbehind assertion or \b or \B (\K is not supported by
779 pcre_dfa_exec()). It's an incompatible change, but it makes the two
780 matching functions compatible, and I think it's the right thing to do.
782 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
783 so that the demonstration program is easily available in environments where
784 PCRE has not been installed from source.
786 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
787 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
788 library.
790 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
791 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
792 is not the first non-POSIX option to be added. Clearly some people find
793 these options useful.
795 17. If a caller to the POSIX matching function regexec() passes a non-zero
796 value for nmatch with a NULL value for pmatch, the value of
797 nmatch is forced to zero.
799 18. RunGrepTest did not have a test for the availability of the -u option of
800 the diff command, as RunTest does. It now checks in the same way as
801 RunTest, and also checks for the -b option.
803 19. If an odd number of negated classes containing just a single character
804 interposed, within parentheses, between a forward reference to a named
805 subpattern and the definition of the subpattern, compilation crashed with
806 an internal error, complaining that it could not find the referenced
807 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
808 [The bug was that it was starting one character too far in when skipping
809 over the character class, thus treating the ] as data rather than
810 terminating the class. This meant it could skip too much.]
812 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
813 /g option in pcretest when the pattern contains \K, which makes it possible
814 to have an empty string match not at the start, even when the pattern is
815 anchored. Updated pcretest and pcredemo to use this option.
817 21. If the maximum number of capturing subpatterns in a recursion was greater
818 than the maximum at the outer level, the higher number was returned, but
819 with unset values at the outer level. The correct (outer level) value is
820 now given.
822 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
823 PCRE did not set those parentheses (unlike Perl). I have now found a way to
824 make it do so. The string so far is captured, making this feature
825 compatible with Perl.
827 23. The tests have been re-organized, adding tests 11 and 12, to make it
828 possible to check the Perl 5.10 features against Perl 5.10.
830 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
831 pattern matches a fixed length string. PCRE did not allow this; now it
832 does. Neither allows recursion.
834 25. I finally figured out how to implement a request to provide the minimum
835 length of subject string that was needed in order to match a given pattern.
836 (It was back references and recursion that I had previously got hung up
837 on.) This code has now been added to pcre_study(); it finds a lower bound
838 to the length of subject needed. It is not necessarily the greatest lower
839 bound, but using it to avoid searching strings that are too short does give
840 some useful speed-ups. The value is available to calling programs via
841 pcre_fullinfo().
843 26. While implementing 25, I discovered to my embarrassment that pcretest had
844 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
845 study optimizations had never been tested with that matching function.
846 Oops. What is worse, even when it was passed study data, there was a bug in
847 pcre_dfa_exec() that meant it never actually used it. Double oops. There
848 were also very few tests of studied patterns with pcre_dfa_exec().
850 27. If (?| is used to create subpatterns with duplicate numbers, they are now
851 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
852 on the other side of the coin, they are no longer allowed to have different
853 names, because these cannot be distinguished in PCRE, and this has caused
854 confusion. (This is a difference from Perl.)
856 28. When duplicate subpattern names are present (necessarily with different
857 numbers, as required by 27 above), and a test is made by name in a
858 conditional pattern, either for a subpattern having been matched, or for
859 recursion in such a pattern, all the associated numbered subpatterns are
860 tested, and the overall condition is true if the condition is true for any
861 one of them. This is the way Perl works, and is also more like the way
862 testing by number works.
865 Version 7.9 11-Apr-09
866 ---------------------
868 1. When building with support for bzlib/zlib (pcregrep) and/or readline
869 (pcretest), all targets were linked against these libraries. This included
870 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
871 libraries. This caused unwanted dependencies to be created. This problem
872 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
873 pcretest is linked with readline.
875 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
876 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
877 moved outside it again, because FALSE and TRUE are already defined in AIX,
878 but BOOL is not.
880 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
881 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
883 4. The pcregrep documentation said spaces were inserted as well as colons (or
884 hyphens) following file names and line numbers when outputting matching
885 lines. This is not true; no spaces are inserted. I have also clarified the
886 wording for the --colour (or --color) option.
888 5. In pcregrep, when --colour was used with -o, the list of matching strings
889 was not coloured; this is different to GNU grep, so I have changed it to be
890 the same.
892 6. When --colo(u)r was used in pcregrep, only the first matching substring in
893 each matching line was coloured. Now it goes on to look for further matches
894 of any of the test patterns, which is the same behaviour as GNU grep.
896 7. A pattern that could match an empty string could cause pcregrep to loop; it
897 doesn't make sense to accept an empty string match in pcregrep, so I have
898 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
899 seems to be how GNU grep behaves.
901 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
902 start or after a newline", because the conditional assertion was not being
903 correctly handled. The rule now is that both the assertion and what follows
904 in the first alternative must satisfy the test.
906 9. If auto-callout was enabled in a pattern with a conditional group whose
907 condition was an assertion, PCRE could crash during matching, both with
908 pcre_exec() and pcre_dfa_exec().
910 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
911 used for matching.
913 11. Unicode property support in character classes was not working for
914 characters (bytes) greater than 127 when not in UTF-8 mode.
916 12. Added the -M command line option to pcretest.
918 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
920 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
922 16. Added comments and documentation about mis-use of no_arg in the C++
923 wrapper.
925 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
926 from Martin Jerabek that uses macro names for all relevant character and
927 string constants.
929 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
930 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
931 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
932 these, but not everybody uses configure.
934 19. A conditional group that had only one branch was not being correctly
935 recognized as an item that could match an empty string. This meant that an
936 enclosing group might also not be so recognized, causing infinite looping
937 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
938 with the subject "ab", where knowledge that the repeated group can match
939 nothing is needed in order to break the loop.
941 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
942 exec(), but without supplying a callout function, matching went wrong.
944 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
945 leak if the size of the offset vector was greater than 30. When the vector
946 is smaller, the saved offsets during recursion go onto a local stack
947 vector, but for larger vectors malloc() is used. It was failing to free
948 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
949 error, in fact).
951 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
952 heapframe that is used only when UTF-8 support is enabled. This caused no
953 problem, but was untidy.
955 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
956 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
957 included within another project.
959 24. Steven Van Ingelgem's patches to add more options to the CMake support,
960 slightly modified by me:
962 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
963 not building pcregrep.
965 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
966 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
968 25. Forward references, both numeric and by name, in patterns that made use of
969 duplicate group numbers, could behave incorrectly or give incorrect errors,
970 because when scanning forward to find the reference group, PCRE was not
971 taking into account the duplicate group numbers. A pattern such as
972 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
974 26. Changed a few more instances of "const unsigned char *" to USPTR, making
975 the feature of a custom pointer more persuasive (as requested by a user).
977 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
978 pcretest.c, inside #ifndefs, because it seems they are sometimes already
979 pre-defined.
981 28. Added support for (*UTF8) at the start of a pattern.
983 29. Arrange for flags added by the "release type" setting in CMake to be shown
984 in the configuration summary.
987 Version 7.8 05-Sep-08
988 ---------------------
990 1. Replaced UCP searching code with optimized version as implemented for Ad
991 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
992 stage table and inline lookup instead of a function, giving speed ups of 2
993 to 5 times on some simple patterns that I tested. Permission was given to
994 distribute the MultiStage2.py script that generates the tables (it's not in
995 the tarball, but is in the Subversion repository).
997 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
998 scripts.
1000 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1001 a group with a zero qualifier. The result of the study could be incorrect,
1002 or the function might crash, depending on the pattern.
1004 4. Caseless matching was not working for non-ASCII characters in back
1005 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1006 It now works when Unicode Property Support is available.
1008 5. In pcretest, an escape such as \x{de} in the data was always generating
1009 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1010 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1011 truncation.
1013 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1015 7. Added two (int) casts to pcregrep when printing the difference of two
1016 pointers, in case they are 64-bit values.
1018 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1019 test 2 if it fails.
1021 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1022 and a #define of that name to empty if it is not externally set. This is to
1023 allow users of MSVC to set it if necessary.
1025 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1026 the convenience functions in the pcre_get.c source file.
1028 11. An option change at the start of a pattern that had top-level alternatives
1029 could cause overwriting and/or a crash. This command provoked a crash in
1030 some environments:
1032 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1034 This potential security problem was recorded as CVE-2008-2371.
1036 12. For a pattern where the match had to start at the beginning or immediately
1037 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1038 pcre_dfa_exec() could read past the end of the passed subject if there was
1039 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1040 pcretest so that it places the subject at the end of its malloc-ed buffer.
1042 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1043 exec() might read past the end of the data buffer in UTF-8 mode.
1045 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1046 the data contained the byte 0x85 as part of a UTF-8 character within its
1047 first line. This applied both to normal and DFA matching.
1049 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1050 /^[^d]*?$/8 failed to match "abc".
1052 16. Added a missing copyright notice to pcrecpp_internal.h.
1054 17. Make it more clear in the documentation that values returned from
1055 pcre_exec() in ovector are byte offsets, not character counts.
1057 18. Tidied a few places to stop certain compilers from issuing warnings.
1059 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1060 supplied by Stefan Weber. I made a further small update for 7.8 because
1061 there is a change of source arrangements: the pcre_searchfuncs.c module is
1062 replaced by pcre_ucd.c.
1065 Version 7.7 07-May-08
1066 ---------------------
1068 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1069 a string to a long long, pretend we don't even have a long long." This is
1070 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1072 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1073 pre-7.6 versions, which defined a global no_arg variable instead of putting
1074 it in the RE class. (See also #8 below.)
1076 3. Remove a line of dead code, identified by coverity and reported by Nuno
1077 Lopes.
1079 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1081 (1) The include/exclude patterns were being applied to the whole pathnames
1082 of files, instead of just to the final components.
1084 (2) If there was more than one level of directory, the subdirectories were
1085 skipped unless they satisfied the include/exclude conditions. This is
1086 inconsistent with GNU grep (and could even be seen as contrary to the
1087 pcregrep specification - which I improved to make it absolutely clear).
1088 The action now is always to scan all levels of directory, and just
1089 apply the include/exclude patterns to regular files.
1091 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1092 --exclude_dir in the tests to avoid scanning .svn directories.
1094 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1095 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1096 doesn't support NULs in patterns.
1098 7. Added some missing "const"s to declarations of static tables in
1099 pcre_compile.c and pcre_dfa_exec.c.
1101 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1102 caused by fix #2 above. (Subsequently also a second patch to fix the
1103 first patch. And a third patch - this was a messy problem.)
1105 9. Applied Craig's patch to remove the use of push_back().
1107 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1108 matching function regexec().
1110 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1111 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1112 references. PCRE supports relative numbers with this syntax (I don't think
1113 Oniguruma does).
1115 12. Previously, a group with a zero repeat such as (...){0} was completely
1116 omitted from the compiled regex. However, this means that if the group
1117 was called as a subroutine from elsewhere in the pattern, things went wrong
1118 (an internal error was given). Such groups are now left in the compiled
1119 pattern, with a new opcode that causes them to be skipped at execution
1120 time.
1122 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1123 to the way PCRE behaves:
1125 (a) A lone ] character is dis-allowed (Perl treats it as data).
1127 (b) A back reference to an unmatched subpattern matches an empty string
1128 (Perl fails the current match path).
1130 (c) A data ] in a character class must be notated as \] because if the
1131 first data character in a class is ], it defines an empty class. (In
1132 Perl it is not possible to have an empty class.) The empty class []
1133 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1134 The negative empty class [^] matches any one character, independently
1135 of the DOTALL setting.
1137 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1138 non-existent subpattern following a character class starting with ']' and
1139 containing () gave an internal compiling error instead of "reference to
1140 non-existent subpattern". Fortunately, when the pattern did exist, the
1141 compiled code was correct. (When scanning forwards to check for the
1142 existencd of the subpattern, it was treating the data ']' as terminating
1143 the class, so got the count wrong. When actually compiling, the reference
1144 was subsequently set up correctly.)
1146 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1147 it was being rejected as not supported by pcre_dfa_exec(), even though
1148 other assertions are supported. I have made pcre_dfa_exec() support
1149 (*FAIL).
1151 16. The implementation of 13c above involved the invention of a new opcode,
1152 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1153 cannot be changed at match time, I realized I could make a small
1154 improvement to matching performance by compiling OP_ALLANY instead of
1155 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1156 on the OP_ANY path.
1158 17. Compiling pcretest on Windows with readline support failed without the
1159 following two fixes: (1) Make the unistd.h include conditional on
1160 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1162 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1163 ncurses library to be included for pcretest when ReadLine support is
1164 requested, but also to allow for it to be overridden. This patch came from
1165 Daniel Bergström.
1167 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1168 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1169 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1170 spotting this.
1173 Version 7.6 28-Jan-08
1174 ---------------------
1176 1. A character class containing a very large number of characters with
1177 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1178 overflow.
1180 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1181 HAVE_LONG_LONG is not defined.
1183 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1184 bring them up to date and include new features. This patch includes:
1186 - Fixed PH's badly added libz and libbz2 support.
1187 - Fixed a problem with static linking.
1188 - Added pcredemo. [But later removed - see 7 below.]
1189 - Fixed dftables problem and added an option.
1190 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1192 - Added readline support for pcretest.
1193 - Added an listing of the option settings after cmake has run.
1195 4. A user submitted a patch to Makefile that makes it easy to create
1196 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1197 Makefile.am that cause it to include this special target, without
1198 affecting anything else. Note that the same mingw target plus all
1199 the other distribution libraries and programs are now supported
1200 when configuring with CMake (see 6 below) instead of with
1201 Configure/Make.
1203 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1204 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1205 exported in the Windows port". It has not yet been confirmed that the patch
1206 solves the problem, but it does no harm.
1208 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1209 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1210 with CMake, and also correct the comment about stack recursion.
1212 7. Remove the automatic building of pcredemo from the ./configure system and
1213 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1214 of a program that users should build themselves after PCRE is installed, so
1215 building it automatically is not really right. What is more, it gave
1216 trouble in some build environments.
1218 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1221 Version 7.5 10-Jan-08
1222 ---------------------
1224 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1225 values in parens when parsing an RE using the C++ wrapper."
1227 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1228 Characters greater than 255 were excluded from the class instead of being
1229 included.
1231 3. The same bug as (2) above applied to negated POSIX classes such as
1232 [:^space:].
1234 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1235 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1236 I have changed it.
1238 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1239 first named subpattern) and a construct such as (?&a) would reference the
1240 first named subpattern whose name started with "a" (in other words, the
1241 length check was missing). Both these problems are fixed. "Subpattern name
1242 expected" is now given for (?&) (a zero-length name), and this patch also
1243 makes it give the same error for \k'' (previously it complained that that
1244 was a reference to a non-existent subpattern).
1246 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1247 this is right because (?- can be followed by option settings as well as by
1248 digits. I have, however, made the messages clearer.
1250 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1251 than the number used in the conditional) now cause a compile-time error.
1252 This is actually not compatible with Perl, which accepts such patterns, but
1253 treats the conditional as always being FALSE (as PCRE used to), but it
1254 seems to me that giving a diagnostic is better.
1256 8. Change "alphameric" to the more common word "alphanumeric" in comments
1257 and messages.
1259 9. Fix two occurrences of "backslash" in comments that should have been
1260 "backspace".
1262 10. Remove two redundant lines of code that can never be obeyed (their function
1263 was moved elsewhere).
1265 11. The program that makes PCRE's Unicode character property table had a bug
1266 which caused it to generate incorrect table entries for sequences of
1267 characters that have the same character type, but are in different scripts.
1268 It amalgamated them into a single range, with the script of the first of
1269 them. In other words, some characters were in the wrong script. There were
1270 thirteen such cases, affecting characters in the following ranges:
1272 U+002b0 - U+002c1
1273 U+0060c - U+0060d
1274 U+0061e - U+00612
1275 U+0064b - U+0065e
1276 U+0074d - U+0076d
1277 U+01800 - U+01805
1278 U+01d00 - U+01d77
1279 U+01d9b - U+01dbf
1280 U+0200b - U+0200f
1281 U+030fc - U+030fe
1282 U+03260 - U+0327f
1283 U+0fb46 - U+0fbb1
1284 U+10450 - U+1049d
1286 12. The -o option (show only the matching part of a line) for pcregrep was not
1287 compatible with GNU grep in that, if there was more than one match in a
1288 line, it showed only the first of them. It now behaves in the same way as
1289 GNU grep.
1291 13. If the -o and -v options were combined for pcregrep, it printed a blank
1292 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1293 does the same. The return code can be used to tell if there were any
1294 non-matching lines.
1296 14. Added --file-offsets and --line-offsets to pcregrep.
1298 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1299 infinitely looping recursion. The bug was that positive lookaheads were not
1300 being skipped when checking for a possible empty match (negative lookaheads
1301 and both kinds of lookbehind were skipped).
1303 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1304 inclusion of <windows.h> to before rather than after the definition of
1305 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1307 17. Specifying a possessive quantifier with a specific limit for a Unicode
1308 character property caused pcre_compile() to compile bad code, which led at
1309 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1310 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1311 caused the error; without that there was no problem.
1313 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1315 19. Added --enable-pcretest-libreadline.
1317 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1318 RE::GlobalReplace(). As a result, the number of replacements returned was
1319 double what it should be. I removed one of the increments, but Craig sent a
1320 later patch that removed the other one (the right fix) and added unit tests
1321 that check the return values (which was not done before).
1323 21. Several CMake things:
1325 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1326 the names libpcre and libpcreposix, not just pcre and pcreposix.
1328 (2) The above change means that pcretest and pcregrep are now correctly
1329 linked with the newly-built libraries, not previously installed ones.
1333 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1334 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1335 UTF-8 newline character). The key issue is that the pattern starts .*;
1336 this means that the match must be either at the beginning, or after a
1337 newline. The bug was in the code for advancing after a failed match and
1338 checking that the new position followed a newline. It was not taking
1339 account of UTF-8 characters correctly.
1341 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1342 character classes. PCRE was not treating the sequence [:...:] as a
1343 character class unless the ... were all letters. Perl, however, seems to
1344 allow any characters between [: and :], though of course it rejects as
1345 unknown any "names" that contain non-letters, because all the known class
1346 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1347 for example, whereas PCRE did not - it did not recognize a POSIX character
1348 class. This seemed a bit dangerous, so the code has been changed to be
1349 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1350 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1351 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1352 Perl does, and where it didn't before.
1354 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1355 Windows environments %n is disabled by default.
1358 Version 7.4 21-Sep-07
1359 ---------------------
1361 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1362 means that a class such as [\s] counted as "explicit reference to CR or
1363 LF". That isn't really right - the whole point of the change was to try to
1364 help when there was an actual mention of one of the two characters. So now
1365 the change happens only if \r or \n (or a literal CR or LF) character is
1366 encountered.
1368 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1369 of both had grown to the point where there were only 3 bits left.
1370 Fortunately, there was spare space in the data structure, and so I have
1371 moved the internal flags into a new 16-bit field to free up more option
1372 bits.
1374 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1375 but did not set the internal JCHANGED flag - either of these is enough to
1376 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1377 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1378 start sets both bits.
1380 4. Added options (at build time, compile time, exec time) to change \R from
1381 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1383 5. doc/pcresyntax.html was missing from the distribution.
1385 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1386 compatibility, even though it is no longer used.
1388 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1389 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1390 windows.h file is present (where different names are used). [This was
1391 reversed later after testing - see 16 below.]
1393 8. Changed all #include <config.h> to #include "config.h". There were also
1394 some further <pcre.h> cases that I changed to "pcre.h".
1396 9. When pcregrep was used with the --colour option, it missed the line ending
1397 sequence off the lines that it output.
1399 10. It was pointed out to me that arrays of string pointers cause lots of
1400 relocations when a shared library is dynamically loaded. A technique of
1401 using a single long string with a table of offsets can drastically reduce
1402 these. I have refactored PCRE in four places to do this. The result is
1403 dramatic:
1405 Originally: 290
1406 After changing UCP table: 187
1407 After changing error message table: 43
1408 After changing table of "verbs" 36
1409 After changing table of Posix names 22
1411 Thanks to the folks working on Gregex for glib for this insight.
1413 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1414 unicode-properties was also set.
1416 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1418 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1419 checked only for CRLF.
1421 14. Added casts to pcretest.c to avoid compiler warnings.
1423 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1425 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1426 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1427 entirely. This removes changes made in 7 above.
1429 17. The CMake files have been updated, and there is now more information about
1430 building with CMake in the NON-UNIX-USE document.
1433 Version 7.3 28-Aug-07
1434 ---------------------
1436 1. In the rejigging of the build system that eventually resulted in 7.1, the
1437 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1438 brackets there is not right, since it causes compilers to look for an
1439 installed pcre.h, not the version that is in the source that is being
1440 compiled (which of course may be different). I have changed it back to:
1442 #include "pcre.h"
1444 I have a vague recollection that the change was concerned with compiling in
1445 different directories, but in the new build system, that is taken care of
1446 by the VPATH setting the Makefile.
1448 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1449 when the subject happened to end in the byte 0x85 (e.g. if the last
1450 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1451 characters but of course it shouldn't be taken as a newline when it is part
1452 of another character. The bug was that, for an unlimited repeat of . in
1453 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1454 characters when looking for a newline.
1456 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1458 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1459 in debug output.
1461 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1462 long printing in the pcrecpp unittest when running under MinGW.
1464 6. ESC_K was left out of the EBCDIC table.
1466 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1467 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1468 limit also applies to "virtual nesting" when a pattern is recursive, and in
1469 this case 1000 isn't so big. I have been able to remove this limit at the
1470 expense of backing off one optimization in certain circumstances. Normally,
1471 when pcre_exec() would call its internal match() function recursively and
1472 immediately return the result unconditionally, it uses a "tail recursion"
1473 feature to save stack. However, when a subpattern that can match an empty
1474 string has an unlimited repetition quantifier, it no longer makes this
1475 optimization. That gives it a stack frame in which to save the data for
1476 checking that an empty string has been matched. Previously this was taken
1477 from the 1000-entry workspace that had been reserved. So now there is no
1478 explicit limit, but more stack is used.
1480 8. Applied Daniel's patches to solve problems with the import/export magic
1481 syntax that is required for Windows, and which was going wrong for the
1482 pcreposix and pcrecpp parts of the library. These were overlooked when this
1483 problem was solved for the main library.
1485 9. There were some crude static tests to avoid integer overflow when computing
1486 the size of patterns that contain repeated groups with explicit upper
1487 limits. As the maximum quantifier is 65535, the maximum group length was
1488 set at 30,000 so that the product of these two numbers did not overflow a
1489 32-bit integer. However, it turns out that people want to use groups that
1490 are longer than 30,000 bytes (though not repeat them that many times).
1491 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1492 made it possible to implement the integer overflow checks in a much more
1493 dynamic way, which I have now done. The artificial limitation on group
1494 length has been removed - we now have only the limit on the total length of
1495 the compiled pattern, which depends on the LINK_SIZE setting.
1497 10. Fixed a bug in the documentation for get/copy named substring when
1498 duplicate names are permitted. If none of the named substrings are set, the
1499 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1500 empty string.
1502 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1503 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1504 because the ] is interpreted as the first data character and the
1505 terminating ] is not found. PCRE has been made compatible with Perl in this
1506 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1507 cause memory overwriting.
1509 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1510 string has been matched (to stop an infinite loop). It was not recognizing
1511 a conditional subpattern that could match an empty string if that
1512 subpattern was within another subpattern. For example, it looped when
1513 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1514 condition was not nested. This bug has been fixed.
1516 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1517 past the start of the subject in the presence of bytes with the top bit
1518 set, for example "\x8aBCD".
1520 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1521 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1523 14. Optimized (?!) to (*FAIL).
1525 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1526 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1527 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1528 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1529 does: it's just the validity check that is more restrictive.
1531 16. Inserted checks for integer overflows during escape sequence (backslash)
1532 processing, and also fixed erroneous offset values for syntax errors during
1533 backslash processing.
1535 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1536 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1538 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1539 caused an overrun.
1541 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1542 something other than just ASCII characters) inside a group that had an
1543 unlimited repeat caused a loop at compile time (while checking to see
1544 whether the group could match an empty string).
1546 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1547 [\P{Any}] did so. (Error in the code for printing property names.)
1549 21. An orphan \E inside a character class could cause a crash.
1551 22. A repeated capturing bracket such as (A)? could cause a wild memory
1552 reference during compilation.
1554 23. There are several functions in pcre_compile() that scan along a compiled
1555 expression for various reasons (e.g. to see if it's fixed length for look
1556 behind). There were bugs in these functions when a repeated \p or \P was
1557 present in the pattern. These operators have additional parameters compared
1558 with \d, etc, and these were not being taken into account when moving along
1559 the compiled data. Specifically:
1561 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1562 length.
1564 (b) An item such as \pL+ within a repeated group could cause crashes or
1565 loops.
1567 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1568 "reference to non-existent subpattern" error.
1570 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1572 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1573 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1575 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1577 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1578 character were causing crashes (broken optimization).
1580 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1581 \p or \P) caused a compile-time loop.
1583 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1584 break. For example, the unstudied pattern [\r\n]A does not match the string
1585 "\r\nA" because change 7.0/46 below moves the current point on by two
1586 characters after failing to match at the start. However, the pattern \nA
1587 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1588 the same is true. There doesn't seem any very clean way out of this, but
1589 what I have chosen to do makes the common cases work: PCRE now takes note
1590 of whether there can be an explicit match for \r or \n anywhere in the
1591 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1592 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1593 pattern has explicit CR or LF references.
1595 29. Added (*CR) etc for changing newline setting at start of pattern.
1598 Version 7.2 19-Jun-07
1599 ---------------------
1601 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1602 which is apparently normally available under Windows.
1604 2. Re-jig the pcregrep tests with different newline settings in an attempt
1605 to make them independent of the local environment's newline setting.
1607 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1609 4. Some of the "internals" tests were previously cut out when the link size
1610 was not 2, because the output contained actual offsets. The recent new
1611 "Z" feature of pcretest means that these can be cut out, making the tests
1612 usable with all link sizes.
1614 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1615 stack recursion. This gives a massive performance boost under BSD, but just
1616 a small improvement under Linux. However, it saves one field in the frame
1617 in all cases.
1619 6. Added more features from the forthcoming Perl 5.10:
1621 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1622 recursion call. It refers to the nth most recently opened parentheses.
1624 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1625 to be opened parentheses.
1627 (c) Conditions that refer to capturing parentheses can be specified
1628 relatively, for example, (?(-2)... or (?(+3)...
1630 (d) \K resets the start of the current match so that everything before
1631 is not part of it.
1633 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1635 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1636 reference syntax.
1638 (g) (?| introduces a group in which the numbering of parentheses in each
1639 alternative starts with the same number.
1641 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1643 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1646 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1647 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1648 for detecting groups that can match an empty string.
1650 9. A pattern with a very large number of alternatives (more than several
1651 hundred) was running out of internal workspace during the pre-compile
1652 phase, where pcre_compile() figures out how much memory will be needed. A
1653 bit of new cunning has reduced the workspace needed for groups with
1654 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1655 workspace instead of running out of the 4096 that are available.
1657 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1659 11. Applied patch from Google to remove an optimization that didn't quite work.
1660 The report of the bug said:
1662 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1663 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1664 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1666 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1667 it matched the wrong number of bytes.
1670 Version 7.1 24-Apr-07
1671 ---------------------
1673 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1674 that is more "standard", making use of automake and other Autotools. There
1675 is some re-arrangement of the files and adjustment of comments consequent
1676 on this.
1678 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1679 for recursive directory scanning broke on some systems because the files
1680 are not scanned in any specific order and on different systems the order
1681 was different. A call to "sort" has been inserted into RunGrepTest for the
1682 approprate test as a short-term fix. In the longer term there may be an
1683 alternative.
1685 3. I had an email from Eric Raymond about problems translating some of PCRE's
1686 man pages to HTML (despite the fact that I distribute HTML pages, some
1687 people do their own conversions for various reasons). The problems
1688 concerned the use of low-level troff macros .br and .in. I have therefore
1689 removed all such uses from the man pages (some were redundant, some could
1690 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1691 HTML has been updated to handle .nf/.fi and to complain if it encounters
1692 .br or .in.
1694 4. Updated comments in configure.ac that get placed in config.h.in and also
1695 arranged for config.h to be included in the distribution, with the name
1696 config.h.generic, for the benefit of those who have to compile without
1697 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1699 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1700 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1701 makevp.bat for the current PCRE, using the additional files
1702 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1704 6. A Windows user reported a minor discrepancy with test 2, which turned out
1705 to be caused by a trailing space on an input line that had got lost in his
1706 copy. The trailing space was an accident, so I've just removed it.
1708 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1709 that is needed.
1711 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1712 as "const" (a) because they are and (b) because it helps the PHP
1713 maintainers who have recently made a script to detect big data structures
1714 in the php code that should be moved to the .rodata section. I remembered
1715 to update Builducptable as well, so it won't revert if ucptable.h is ever
1716 re-created.
1718 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1719 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1720 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1721 support is not required. This saves 1.5-2K of code, which is important in
1722 some applications.
1724 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1725 so as not to refer to the tables, even though these functions will never be
1726 called when UTF-8 support is disabled. Otherwise there are problems with a
1727 shared library.
1729 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1731 (a) It was defining its arguments as char * instead of void *.
1733 (b) It was assuming that all moves were upwards in memory; this was true
1734 a long time ago when I wrote it, but is no longer the case.
1736 The emulated memove() is provided for those environments that have neither
1737 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1738 is clearly not the case, as these two bugs were recently reported.
1740 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1741 and Detrail to create the HTML documentation, the .txt form of the man
1742 pages, and it removes trailing spaces from listed files. It also creates
1743 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1744 case, it wraps all the #defines with #ifndefs. This script should be run
1745 before "make dist".
1747 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1748 with Unicode property support.
1750 (a) For a maximizing quantifier, if the two different cases of the
1751 character were of different lengths in their UTF-8 codings (there are
1752 some cases like this - I found 11), and the matching function had to
1753 back up over a mixture of the two cases, it incorrectly assumed they
1754 were both the same length.
1756 (b) When PCRE was configured to use the heap rather than the stack for
1757 recursion during matching, it was not correctly preserving the data for
1758 the other case of a UTF-8 character when checking ahead for a match
1759 while processing a minimizing repeat. If the check also involved
1760 matching a wide character, but failed, corruption could cause an
1761 erroneous result when trying to check for a repeat of the original
1762 character.
1764 13. Some tidying changes to the testing mechanism:
1766 (a) The RunTest script now detects the internal link size and whether there
1767 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1768 values substituted by "configure". (The RunGrepTest script already did
1769 this for UTF-8.) The configure.ac script no longer substitutes the
1770 relevant variables.
1772 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1773 with length and offset values. This means that the output is different
1774 for different internal link sizes. Test 2 is skipped for link sizes
1775 other than 2 because of this, bypassing the problem. Unfortunately,
1776 there was also a test in test 3 (the locale tests) that used /B and
1777 failed for link sizes other than 2. Rather than cut the whole test out,
1778 I have added a new /Z option to pcretest that replaces the length and
1779 offset values with spaces. This is now used to make test 3 independent
1780 of link size. (Test 2 will be tidied up later.)
1782 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1783 segmentation fault instead of returning the appropriate error message.
1785 15. In multiline mode when the newline sequence was set to "any", the pattern
1786 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1787 This doesn't seem right; it now treats the CRLF combination as the line
1788 ending, and so does not match in that case. It's only a pattern such as ^$
1789 that would hit this one: something like ^ABC$ would have failed after \r
1790 and then tried again after \r\n.
1792 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1793 in an attempt to make files that differ only in their line terminators
1794 compare equal. This works on Linux.
1796 17. Under certain error circumstances pcregrep might try to free random memory
1797 as it exited. This is now fixed, thanks to valgrind.
1799 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1800 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1801 was because its rules for how to advance for /g after matching an empty
1802 string at the end of a line did not allow for this case. They now check for
1803 it specially.
1805 20. pcretest is supposed to handle patterns and data of any length, by
1806 extending its buffers when necessary. It was getting this wrong when the
1807 buffer for a data line had to be extended.
1809 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1810 CRLF as a newline sequence.
1812 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1813 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1814 I have nevertheless tidied it up.
1816 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1818 24. Added a man page for pcre-config.
1821 Version 7.0 19-Dec-06
1822 ---------------------
1824 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1825 moving to gcc 4.1.1.
1827 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1828 sys/time.h, which is documented as needed for this function. It doesn't
1829 seem to matter on Linux, but it showed up on some releases of OS X.
1831 3. It seems that there are systems where bytes whose values are greater than
1832 127 match isprint() in the "C" locale. The "C" locale should be the
1833 default when a C program starts up. In most systems, only ASCII printing
1834 characters match isprint(). This difference caused the output from pcretest
1835 to vary, making some of the tests fail. I have changed pcretest so that:
1837 (a) When it is outputting text in the compiled version of a pattern, bytes
1838 other than 32-126 are always shown as hex escapes.
1840 (b) When it is outputting text that is a matched part of a subject string,
1841 it does the same, unless a different locale has been set for the match
1842 (using the /L modifier). In this case, it uses isprint() to decide.
1844 4. Fixed a major bug that caused incorrect computation of the amount of memory
1845 required for a compiled pattern when options that changed within the
1846 pattern affected the logic of the preliminary scan that determines the
1847 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1848 that the computed length was too small. The symptoms of this bug were
1849 either the PCRE error "internal error: code overflow" from pcre_compile(),
1850 or a glibc crash with a message such as "pcretest: free(): invalid next
1851 size (fast)". Examples of patterns that provoked this bug (shown in
1852 pcretest format) are:
1854 /(?-x: )/x
1855 /(?x)(?-x: \s*#\s*)/
1856 /((?i)[\x{c0}])/8
1857 /(?i:[\x{c0}])/8
1859 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1860 is now done differently.
1862 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1863 wrapper classes; (b) implement a new function in the C++ scanner that is
1864 more efficient than the old way of doing things because it avoids levels of
1865 recursion in the regex matching; (c) add a paragraph to the documentation
1866 for the FullMatch() function.
1868 6. The escape sequence \n was being treated as whatever was defined as
1869 "newline". Not only was this contrary to the documentation, which states
1870 that \n is character 10 (hex 0A), but it also went horribly wrong when
1871 "newline" was defined as CRLF. This has been fixed.
1873 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1874 was being set to -1 for the "end of line" case (supposedly a value that no
1875 character can have). Though this value is never used (the check for end of
1876 line is "zero bytes in current character"), it caused compiler complaints.
1877 I've changed it to 0xffffffff.
1879 8. In pcre_version.c, the version string was being built by a sequence of
1880 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1881 string (as it is for production releases) called a macro with an empty
1882 argument. The C standard says the result of this is undefined. The gcc
1883 compiler treats it as an empty string (which was what was wanted) but it is
1884 reported that Visual C gives an error. The source has been hacked around to
1885 avoid this problem.
1887 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1888 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1889 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1890 of them did).
1892 10. Originally, pcretest opened its input and output without "b"; then I was
1893 told that "b" was needed in some environments, so it was added for release
1894 5.0 to both the input and output. (It makes no difference on Unix-like
1895 systems.) Later I was told that it is wrong for the input on Windows. I've
1896 now abstracted the modes into two macros, to make it easier to fiddle with
1897 them, and removed "b" from the input mode under Windows.
1899 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1901 12. Added -help and --help to pcretest as an official way of being reminded
1902 of the options.
1904 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1905 and pcrecpp.cc because they annoy compilers at high warning levels.
1907 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1909 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1910 scripts are not C programs :-) and which was not noticed because it works
1911 on Linux.
1913 16. pcretest is supposed to handle any length of pattern and data line (as one
1914 line or as a continued sequence of lines) by extending its input buffer if
1915 necessary. This feature was broken for very long pattern lines, leading to
1916 a string of junk being passed to pcre_compile() if the pattern was longer
1917 than about 50K.
1919 17. I have done a major re-factoring of the way pcre_compile() computes the
1920 amount of memory needed for a compiled pattern. Previously, there was code
1921 that made a preliminary scan of the pattern in order to do this. That was
1922 OK when PCRE was new, but as the facilities have expanded, it has become
1923 harder and harder to keep it in step with the real compile phase, and there
1924 have been a number of bugs (see for example, 4 above). I have now found a
1925 cunning way of running the real compile function in a "fake" mode that
1926 enables it to compute how much memory it would need, while actually only
1927 ever using a few hundred bytes of working memory and without too many
1928 tests of the mode. This should make future maintenance and development
1929 easier. A side effect of this work is that the limit of 200 on the nesting
1930 depth of parentheses has been removed (though this was never a serious
1931 limitation, I suspect). However, there is a downside: pcre_compile() now
1932 runs more slowly than before (30% or more, depending on the pattern). I
1933 hope this isn't a big issue. There is no effect on runtime performance.
1935 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1936 newline (only possible for the last line of a file) and it was a
1937 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1939 19. Added additional timing features to pcretest. (1) The -tm option now times
1940 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1941 separate command line item, by a number that specifies the number of
1942 repeats to use when timing. The default is 50000; this gives better
1943 precision, but takes uncomfortably long for very large patterns.
1945 20. Extended pcre_study() to be more clever in cases where a branch of a
1946 subpattern has no definite first character. For example, (a*|b*)[cd] would
1947 previously give no result from pcre_study(). Now it recognizes that the
1948 first character must be a, b, c, or d.
1950 21. There was an incorrect error "recursive call could loop indefinitely" if
1951 a subpattern (or the entire pattern) that was being tested for matching an
1952 empty string contained only one non-empty item after a nested subpattern.
1953 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1954 incorrectly, because the \d was being skipped in the check.
1956 22. The pcretest program now has a new pattern option /B and a command line
1957 option -b, which is equivalent to adding /B to every pattern. This causes
1958 it to show the compiled bytecode, without the additional information that
1959 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1960 is the same as /B/I).
1962 23. A new optimization is now able automatically to treat some sequences such
1963 as a*b as a*+b. More specifically, if something simple (such as a character
1964 or a simple class like \d) has an unlimited quantifier, and is followed by
1965 something that cannot possibly match the quantified thing, the quantifier
1966 is automatically "possessified".
1968 24. A recursive reference to a subpattern whose number was greater than 39
1969 went wrong under certain circumstances in UTF-8 mode. This bug could also
1970 have affected the operation of pcre_study().
1972 25. Realized that a little bit of performance could be had by replacing
1973 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1975 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1977 27. Possessive quantifiers such as a++ were previously implemented by turning
1978 them into atomic groups such as ($>a+). Now they have their own opcodes,
1979 which improves performance. This includes the automatically created ones
1980 from 23 above.
1982 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1983 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1984 the first matched character to be a colon. This applied both to named and
1985 numbered groups.
1987 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1989 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1990 building PCRE on Windows easier, so I have included it in the distribution.
1992 31. There is now a check in pcretest against a ridiculously large number being
1993 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1994 loop, the loop is abandoned.
1996 32. Forward references to subpatterns in conditions such as (?(2)...) where
1997 subpattern 2 is defined later cause pcre_compile() to search forwards in
1998 the pattern for the relevant set of parentheses. This search went wrong
1999 when there were unescaped parentheses in a character class, parentheses
2000 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2002 33. "Subroutine" calls and backreferences were previously restricted to
2003 referencing subpatterns earlier in the regex. This restriction has now
2004 been removed.
2006 34. Added a number of extra features that are going to be in Perl 5.10. On the
2007 whole, these are just syntactic alternatives for features that PCRE had
2008 previously implemented using the Python syntax or my own invention. The
2009 other formats are all retained for compatibility.
2011 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2012 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2013 also .NET compatible.
2015 (b) A recursion or subroutine call to a named group can now be defined as
2016 (?&name) as well as (?P>name).
2018 (c) A backreference to a named group can now be defined as \k<name> or
2019 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2020 5.10, are also .NET compatible.
2022 (d) A conditional reference to a named group can now use the syntax
2023 (?(<name>) or (?('name') as well as (?(name).
2025 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2026 groups (named and numbered) that are never evaluated inline, but can be
2027 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2028 is always false. There may be only one alternative in such a group.
2030 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2031 as the simple (?(R). The condition is true only if the most recent
2032 recursion is that of the given number or name. It does not search out
2033 through the entire recursion stack.
2035 (g) The escape \gN or \g{N} has been added, where N is a positive or
2036 negative number, specifying an absolute or relative reference.
2038 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2039 some "unreachable code" warnings.
2041 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2042 things, this adds five new scripts.
2044 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2045 There were also incompatibilities regarding the handling of \Q..\E inside
2046 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2047 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2049 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2050 matches an empty string, and forcibly breaks the loop. There were bugs in
2051 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2052 against aaaa the result was just "a" rather than "aaaa", for example. Two
2053 separate and independent bugs (that affected different cases) have been
2054 fixed.
2056 39. Refactored the code to abolish the use of different opcodes for small
2057 capturing bracket numbers. This is a tidy that I avoided doing when I
2058 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2059 The new approach is not only tidier, it makes it possible to reduce the
2060 memory needed to fix the previous bug (38).
2062 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2063 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2064 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2065 mode.
2067 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2068 report.
2070 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2071 copy construction and assignment in the C++ wrapper.
2073 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2074 couple of bugs that could have given wrong results in the "--newline=crlf"
2075 case.
2077 44. Added a number of casts and did some reorganization of signed/unsigned int
2078 variables following suggestions from Dair Grant. Also renamed the variable
2079 "this" as "item" because it is a C++ keyword.
2081 45. Arranged for dftables to add
2083 #include "pcre_internal.h"
2085 to pcre_chartables.c because without it, gcc 4.x may remove the array
2086 definition from the final binary if PCRE is built into a static library and
2087 dead code stripping is activated.
2089 46. For an unanchored pattern, if a match attempt fails at the start of a
2090 newline sequence, and the newline setting is CRLF or ANY, and the next two
2091 characters are CRLF, advance by two characters instead of one.
2094 Version 6.7 04-Jul-06
2095 ---------------------
2097 1. In order to handle tests when input lines are enormously long, pcretest has
2098 been re-factored so that it automatically extends its buffers when
2099 necessary. The code is crude, but this _is_ just a test program. The
2100 default size has been increased from 32K to 50K.
2102 2. The code in pcre_study() was using the value of the re argument before
2103 testing it for NULL. (Of course, in any sensible call of the function, it
2104 won't be NULL.)
2106 3. The memmove() emulation function in pcre_internal.h, which is used on
2107 systems that lack both memmove() and bcopy() - that is, hardly ever -
2108 was missing a "static" storage class specifier.
2110 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2111 containing an extended class (one that cannot be represented by a bitmap
2112 because it contains high-valued characters or Unicode property items, e.g.
2113 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2114 pattern, but PCRE should not loop if you do not (it no longer does).
2115 [Detail: two cases were found: (a) a repeated subpattern containing an
2116 extended class; (b) a recursive reference to a subpattern that followed a
2117 previous extended class. It wasn't skipping over the extended class
2118 correctly when UTF-8 mode was not set.]
2120 5. A negated single-character class was not being recognized as fixed-length
2121 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2122 compile error "lookbehind assertion is not fixed length".
2124 6. The RunPerlTest auxiliary script was showing an unexpected difference
2125 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2126 write a Perl script that can interpret lines of an input file either as
2127 byte characters or as UTF-8, which is what "perltest" was being required to
2128 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2129 can't do is switch easily at run time between having the "use utf8;" pragma
2130 or not. In the end, I fudged it by using the RunPerlTest script to insert
2131 "use utf8;" explicitly for the UTF-8 tests.
2133 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2134 the end of the subject string, contrary to the documentation and to what
2135 Perl does. This was true of both matching functions. Now it matches only at
2136 the start of the subject and immediately after *internal* newlines.
2138 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2139 a pointer to an int instead of a pointer to an unsigned long int. This
2140 caused problems on 64-bit systems.
2142 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2143 instance of the 'standard' template library not being so standard".
2145 10. There was no check on the number of named subpatterns nor the maximum
2146 length of a subpattern name. The product of these values is used to compute
2147 the size of the memory block for a compiled pattern. By supplying a very
2148 long subpattern name and a large number of named subpatterns, the size
2149 computation could be caused to overflow. This is now prevented by limiting
2150 the length of names to 32 characters, and the number of named subpatterns
2151 to 10,000.
2153 11. Subpatterns that are repeated with specific counts have to be replicated in
2154 the compiled pattern. The size of memory for this was computed from the
2155 length of the subpattern and the repeat count. The latter is limited to
2156 65535, but there was no limit on the former, meaning that integer overflow
2157 could in principle occur. The compiled length of a repeated subpattern is
2158 now limited to 30,000 bytes in order to prevent this.
2160 12. Added the optional facility to have named substrings with the same name.
2162 13. Added the ability to use a named substring as a condition, using the
2163 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2164 are numbers (not recommended). Forward references are permitted.
2166 14. Added forward references in named backreferences (if you see what I mean).
2168 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2169 pattern could run off the end of the subject. For example, the pattern
2170 "(?s)(.{1,5})"8 did this with the subject "ab".
2172 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2173 PCRE_CASELESS was set when matching characters that were quantified with ?
2174 or *.
2176 17. A character class other than a single negated character that had a minimum
2177 but no maximum quantifier - for example [ab]{6,} - was not handled
2178 correctly by pce_dfa_exec(). It would match only one character.
2180 18. A valid (though odd) pattern that looked like a POSIX character
2181 class but used an invalid character after [ (for example [[,abc,]]) caused
2182 pcre_compile() to give the error "Failed: internal error: code overflow" or
2183 in some cases to crash with a glibc free() error. This could even happen if
2184 the pattern terminated after [[ but there just happened to be a sequence of
2185 letters, a binary zero, and a closing ] in the memory that followed.
2187 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2188 over the years. Originally (before any Unicode support), just the bottom 8
2189 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2190 output from "man perlunicode" includes this:
2192 The regular expression compiler produces polymorphic opcodes. That
2193 is, the pattern adapts to the data and automatically switches to
2194 the Unicode character scheme when presented with Unicode data--or
2195 instead uses a traditional byte scheme when presented with byte
2196 data.
2198 Sadly, a wide octal escape does not cause a switch, and in a string with
2199 no other multibyte characters, these octal escapes are treated as before.
2200 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2201 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2202 Unicode string.
2204 I have not perpetrated such confusion in PCRE. Up till now, it took just
2205 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2206 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2207 translate to the appropriate multibyte character.
2209 29. Applied some refactoring to reduce the number of warnings from Microsoft
2210 and Borland compilers. This has included removing the fudge introduced
2211 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2212 a warning about an unused variable.
2214 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2215 characters since release 4.0, because Perl (from release 5.004) does not.
2216 [Or at least, is documented not to: some releases seem to be in conflict
2217 with the documentation.] However, when a pattern was studied with
2218 pcre_study() and all its branches started with \s, PCRE still included VT
2219 as a possible starting character. Of course, this did no harm; it just
2220 caused an unnecessary match attempt.
2222 22. Removed a now-redundant internal flag bit that recorded the fact that case
2223 dependency changed within the pattern. This was once needed for "required
2224 byte" processing, but is no longer used. This recovers a now-scarce options
2225 bit. Also moved the least significant internal flag bit to the most-
2226 significant bit of the word, which was not previously used (hangover from
2227 the days when it was an int rather than a uint) to free up another bit for
2228 the future.
2230 23. Added support for CRLF line endings as well as CR and LF. As well as the
2231 default being selectable at build time, it can now be changed at runtime
2232 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2233 specify that it is scanning data with non-default line endings.
2235 24. Changed the definition of CXXLINK to make it agree with the definition of
2236 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2238 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2239 recursions. This makes a big different to stack usage for some patterns.
2241 26. If a subpattern containing a named recursion or subroutine reference such
2242 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2243 the space required for the compiled pattern went wrong and gave too small a
2244 value. Depending on the environment, this could lead to "Failed: internal
2245 error: code overflow at offset 49" or "glibc detected double free or
2246 corruption" errors.
2248 27. Applied patches from Google (a) to support the new newline modes and (b) to
2249 advance over multibyte UTF-8 characters in GlobalReplace.
2251 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2252 difference for some implementation of PCRE in some Windows version.
2254 29. Added some extra testing facilities to pcretest:
2256 \q<number> in a data line sets the "match limit" value
2257 \Q<number> in a data line sets the "match recursion limt" value
2258 -S <number> sets the stack size, where <number> is in megabytes
2260 The -S option isn't available for Windows.
2263 Version 6.6 06-Feb-06
2264 ---------------------
2266 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2267 in pcreposix.h. I have copied the definition from pcre.h.
2269 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2270 because pcre.h is no longer a built file.
2272 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2273 not normally included in the compiled code.
2276 Version 6.5 01-Feb-06
2277 ---------------------
2279 1. When using the partial match feature with pcre_dfa_exec(), it was not
2280 anchoring the second and subsequent partial matches at the new starting
2281 point. This could lead to incorrect results. For example, with the pattern
2282 /1234/, partially matching against "123" and then "a4" gave a match.
2284 2. Changes to pcregrep:
2286 (a) All non-match returns from pcre_exec() were being treated as failures
2287 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2288 error message is output. Some extra information is given for the
2290 probably the only errors that are likely to be caused by users (by
2291 specifying a regex that has nested indefinite repeats, for instance).
2292 If there are more than 20 of these errors, pcregrep is abandoned.
2294 (b) A binary zero was treated as data while matching, but terminated the
2295 output line if it was written out. This has been fixed: binary zeroes
2296 are now no different to any other data bytes.
2298 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2299 used to set a locale for matching. The --locale=xxxx long option has
2300 been added (no short equivalent) to specify a locale explicitly on the
2301 pcregrep command, overriding the environment variables.
2303 (d) When -B was used with -n, some line numbers in the output were one less
2304 than they should have been.
2306 (e) Added the -o (--only-matching) option.
2308 (f) If -A or -C was used with -c (count only), some lines of context were
2309 accidentally printed for the final match.
2311 (g) Added the -H (--with-filename) option.
2313 (h) The combination of options -rh failed to suppress file names for files
2314 that were found from directory arguments.
2316 (i) Added the -D (--devices) and -d (--directories) options.
2318 (j) Added the -F (--fixed-strings) option.
2320 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2322 (l) Added the --colo(u)r option.
2324 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2325 is not present by default.
2327 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2328 items such as (?R) or (?1), when the recursion could match a number of
2329 alternatives. If it matched one of the alternatives, but subsequently,
2330 outside the recursion, there was a failure, the code tried to back up into
2331 the recursion. However, because of the way PCRE is implemented, this is not
2332 possible, and the result was an incorrect result from the match.
2334 In order to prevent this happening, the specification of recursion has
2335 been changed so that all such subpatterns are automatically treated as
2336 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2338 4. I had overlooked the fact that, in some locales, there are characters for
2339 which isalpha() is true but neither isupper() nor islower() are true. In
2340 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2341 and ordfeminine) are like this. This affected the treatment of \w and \W
2342 when they appeared in character classes, but not when they appeared outside
2343 a character class. The bit map for "word" characters is now created
2344 separately from the results of isalnum() instead of just taking it from the
2345 upper, lower, and digit maps. (Plus the underscore character, of course.)
2347 5. The above bug also affected the handling of POSIX character classes such as
2348 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2349 permanent tables. Instead, the bit maps for such a class were previously
2350 created as the appropriate unions of the upper, lower, and digit bitmaps.
2351 Now they are created by subtraction from the [[:word:]] class, which has
2352 its own bitmap.
2354 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2355 It is created by subtracting the vertical space characters (\x09, \x0a,
2356 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2357 subtraction was done in the overall bitmap for a character class, meaning
2358 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2359 be recognized. This bug has been fixed.
2361 7. Patches from the folks at Google:
2363 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2364 real life, but is still worth protecting against".
2366 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2367 regular expressions".
2369 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2370 have it.
2372 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2373 "configure" and the latter not, in order to fix a problem somebody had
2374 with compiling the Arg class on HP-UX.
2376 (e) Improve the error-handling of the C++ wrapper a little bit.
2378 (f) New tests for checking recursion limiting.
2380 8. The pcre_memmove() function, which is used only if the environment does not
2381 have a standard memmove() function (and is therefore rarely compiled),
2382 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2383 returning a result (though PCRE never actually uses the result).
2385 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2386 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2387 returned instead of calling malloc() with an overflowing number that would
2388 most likely cause subsequent chaos.
2390 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2392 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2393 with this option is matched, the nmatch and pmatch options of regexec() are
2394 ignored.
2396 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2397 provided in case anyone wants to the the POSIX interface with UTF-8
2398 strings.
2400 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2401 C++ linking (needed for some HP-UX environments).
2403 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2404 (unused parameter) and in the pcre_printint() function (omitted "default"
2405 switch label when the default is to do nothing).
2407 15. Added some code to make it possible, when PCRE is compiled as a C++
2408 library, to replace subject pointers for pcre_exec() with a smart pointer
2409 class, thus making it possible to process discontinuous strings.
2411 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2412 much the same function. They were added by different people who were trying
2413 to make PCRE easy to compile on non-Unix systems. It has been suggested
2414 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2415 for compiling on Windows systems. I have therefore replaced it with
2416 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2417 defaults to "extern" for C or "extern C" for C++, which works fine on
2418 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2419 SCOPE with something explicit in config.h. In addition:
2421 (a) pcreposix.h still had just "extern" instead of either of these macros;
2422 I have replaced it with PCRE_DATA_SCOPE.
2424 (b) Functions such as _pcre_xclass(), which are internal to the library,
2425 but external in the C sense, all had PCRE_EXPORT in their definitions.
2426 This is apparently wrong for the Windows case, so I have removed it.
2427 (It makes no difference on Unix-like systems.)
2429 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2430 of recursive calls to match(). This is different to MATCH_LIMIT because
2431 that limits the total number of calls to match(), not all of which increase
2432 the depth of recursion. Limiting the recursion depth limits the amount of
2433 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2434 when PCRE is compiled, and changed at run time. A patch from Google adds
2435 this functionality to the C++ interface.
2437 18. Changes to the handling of Unicode character properties:
2439 (a) Updated the table to Unicode 4.1.0.
2441 (b) Recognize characters that are not in the table as "Cn" (undefined).
2443 (c) I revised the way the table is implemented to a much improved format
2444 which includes recognition of ranges. It now supports the ranges that
2445 are defined in UnicodeData.txt, and it also amalgamates other
2446 characters into ranges. This has reduced the number of entries in the
2447 table from around 16,000 to around 3,000, thus reducing its size
2448 considerably. I realized I did not need to use a tree structure after
2449 all - a binary chop search is just as efficient. Having reduced the
2450 number of entries, I extended their size from 6 bytes to 8 bytes to
2451 allow for more data.
2453 (d) Added support for Unicode script names via properties such as \p{Han}.
2455 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2456 matching that character.
2458 20. When matching a repeated Unicode property with a minimum greater than zero,
2459 (for example \pL{2,}), PCRE could look past the end of the subject if it
2460 reached it while seeking the minimum number of characters. This could
2461 happen only if some of the characters were more than one byte long, because
2462 there is a check for at least the minimum number of bytes.
2464 21. Refactored the implementation of \p and \P so as to be more general, to
2465 allow for more different types of property in future. This has changed the
2466 compiled form incompatibly. Anybody with saved compiled patterns that use
2467 \p or \P will have to recompile them.
2469 22. Added "Any" and "L&" to the supported property types.
2471 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2472 but give a compile time error if the value is greater than 0xff.
2474 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2475 accidentally not being installed or uninstalled.
2477 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2478 made were to insert the current release number. This seemed silly, because
2479 it made things harder for people building PCRE on systems that don't run
2480 "configure". I have turned pcre.h into a distributed file, no longer built
2481 by "configure", with the version identification directly included. There is
2482 no longer a pcre.h.in file.
2484 However, this change necessitated a change to the pcre-config script as
2485 well. It is built from pcre-config.in, and one of the substitutions was the
2486 release number. I have updated configure.ac so that ./configure now finds
2487 the release number by grepping pcre.h.
2489 26. Added the ability to run the tests under valgrind.
2492 Version 6.4 05-Sep-05
2493 ---------------------
2495 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2496 "--" to be printed when multiple files were scanned, even when none of the
2497 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2498 consider it to be a bug, and have restored the previous behaviour.
2500 2. A couple of code tidies to get rid of compiler warnings.
2502 3. The pcretest program used to cheat by referring to symbols in the library
2503 whose names begin with _pcre_. These are internal symbols that are not
2504 really supposed to be visible externally, and in some environments it is
2505 possible to suppress them. The cheating is now confined to including
2506 certain files from the library's source, which is a bit cleaner.
2508 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2509 file's purpose clearer.
2511 5. Reorganized pcre_ucp_findchar().
2514 Version 6.3 15-Aug-05
2515 ---------------------
2517 1. The file libpcre.pc.in did not have general read permission in the tarball.
2519 2. There were some problems when building without C++ support:
2521 (a) If C++ support was not built, "make install" and "make test" still
2522 tried to test it.
2524 (b) There were problems when the value of CXX was explicitly set. Some
2525 changes have been made to try to fix these, and ...
2527 (c) --disable-cpp can now be used to explicitly disable C++ support.
2529 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2530 backslash in a target when C++ was disabled. This confuses some
2531 versions of "make", apparently. Using an intermediate variable solves
2532 this. (Same for CPP_LOBJ.)
2534 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2535 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2536 necessary on certain architectures.
2538 4. Added a setting of -export-symbols-regex to the link command to remove
2539 those symbols that are exported in the C sense, but actually are local
2540 within the library, and not documented. Their names all begin with
2541 "_pcre_". This is not a perfect job, because (a) we have to except some
2542 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2543 available (and never for static libraries). I have made a note to try to
2544 find a way round (a) in the future.
2547 Version 6.2 01-Aug-05
2548 ---------------------
2550 1. There was no test for integer overflow of quantifier values. A construction
2551 such as {1111111111111111} would give undefined results. What is worse, if
2552 a minimum quantifier for a parenthesized subpattern overflowed and became
2553 negative, the calculation of the memory size went wrong. This could have
2554 led to memory overwriting.
2556 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2558 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2559 operating environments where this matters.
2561 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2562 PCRE options from within the C++ wrapper.
2564 5. Named capturing subpatterns were not being correctly counted when a pattern
2565 was compiled. This caused two problems: (a) If there were more than 100
2566 such subpatterns, the calculation of the memory needed for the whole
2567 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2568 back references of the form \12, where the number was greater than 9, were
2569 not recognized as back references, even though there were sufficient
2570 previous subpatterns.
2572 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2573 versions of gcc, e.g. 2.95.4.
2576 Version 6.1 21-Jun-05
2577 ---------------------
2579 1. There was one reference to the variable "posix" in pcretest.c that was not
2580 surrounded by "#if !defined NOPOSIX".
2582 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2583 the cross-check on the old pcre_info() function, for the benefit of the
2584 cut-down version of PCRE that is currently imported into Exim.
2586 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2587 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2588 patterns that start (?i) but I don't think that matters. The use of (?i) is
2589 just an example; this all applies to the other options as well.
2591 4. Since libtool seems to echo the compile commands it is issuing, the output
2592 from "make" can be reduced a bit by putting "@" in front of each libtool
2593 compile command.
2595 5. Patch from the folks at Google for configure.in to be a bit more thorough
2596 in checking for a suitable C++ installation before trying to compile the
2597 C++ stuff. This should fix a reported problem when a compiler was present,
2598 but no suitable headers.
2600 6. The man pages all had just "PCRE" as their title. I have changed them to
2601 be the relevant file name. I have also arranged that these names are
2602 retained in the file doc/pcre.txt, which is a concatenation in text format
2603 of all the man pages except the little individual ones for each function.
2605 7. The NON-UNIX-USE file had not been updated for the different set of source
2606 files that come with release 6. I also added a few comments about the C++
2607 wrapper.
2610 Version 6.0 07-Jun-05
2611 ---------------------
2613 1. Some minor internal re-organization to help with my DFA experiments.
2615 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2616 didn't matter for the library itself when fully configured, but did matter
2617 when compiling without UCP support, or within Exim, where the ucp files are
2618 not imported.
2620 3. Refactoring of the library code to split up the various functions into
2621 different source modules. The addition of the new DFA matching code (see
2622 below) to a single monolithic source would have made it really too
2623 unwieldy, quite apart from causing all the code to be include in a
2624 statically linked application, when only some functions are used. This is
2625 relevant even without the DFA addition now that patterns can be compiled in
2626 one application and matched in another.
2628 The downside of splitting up is that there have to be some external
2629 functions and data tables that are used internally in different modules of
2630 the library but which are not part of the API. These have all had their
2631 names changed to start with "_pcre_" so that they are unlikely to clash
2632 with other external names.
2634 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2635 a different (DFA) algorithm. Although it is slower than the original
2636 function, it does have some advantages for certain types of matching
2637 problem.
2639 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2640 including restarting after a partial match.
2642 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2643 defined when compiling for Windows was sent to me. I have put it into the
2644 code, though I have no means of testing or verifying it.
2646 7. Added the pcre_refcount() auxiliary function.
2648 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2649 match before or at the first newline in the subject string. In pcretest,
2650 the /f option on a pattern can be used to set this.
2652 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2653 would behave wrongly. This has been present in PCRE since release 4.0.
2655 10. A number of changes to the pcregrep command:
2657 (a) Refactored how -x works; insert ^(...)$ instead of setting
2658 PCRE_ANCHORED and checking the length, in preparation for adding
2659 something similar for -w.
2661 (b) Added the -w (match as a word) option.
2663 (c) Refactored the way lines are read and buffered so as to have more
2664 than one at a time available.
2666 (d) Implemented a pcregrep test script.
2668 (e) Added the -M (multiline match) option. This allows patterns to match
2669 over several lines of the subject. The buffering ensures that at least
2670 8K, or the rest of the document (whichever is the shorter) is available
2671 for matching (and similarly the previous 8K for lookbehind assertions).
2673 (f) Changed the --help output so that it now says
2675 -w, --word-regex(p)
2677 instead of two lines, one with "regex" and the other with "regexp"
2678 because that confused at least one person since the short forms are the
2679 same. (This required a bit of code, as the output is generated
2680 automatically from a table. It wasn't just a text change.)
2682 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2683 option but starts with a hyphen. Could be a pattern or a path name
2684 starting with a hyphen, for instance.
2686 (h) "-" can be given as a file name to represent stdin.
2688 (i) When file names are being printed, "(standard input)" is used for
2689 the standard input, for compatibility with GNU grep. Previously
2690 "<stdin>" was used.
2692 (j) The option --label=xxx can be used to supply a name to be used for
2693 stdin when file names are being printed. There is no short form.
2695 (k) Re-factored the options decoding logic because we are going to add
2696 two more options that take data. Such options can now be given in four
2697 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2699 (l) Added the -A, -B, and -C options for requesting that lines of context
2700 around matches be printed.
2702 (m) Added the -L option to print the names of files that do not contain
2703 any matching lines, that is, the complement of -l.
2705 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2706 continue to scan other files.
2708 (o) The -s option was incorrectly implemented. For compatibility with other
2709 greps, it now suppresses the error message for a non-existent or non-
2710 accessible file (but not the return code). There is a new option called
2711 -q that suppresses the output of matching lines, which was what -s was
2712 previously doing.
2714 (p) Added --include and --exclude options to specify files for inclusion
2715 and exclusion when recursing.
2717 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2718 Hopefully, it now does.
2720 12. Missing cast in pcre_study().
2722 13. Added an "uninstall" target to the makefile.
2724 14. Replaced "extern" in the function prototypes in Makefile.in with
2725 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2726 world, but is set differently for Windows.
2728 15. Added a second compiling function called pcre_compile2(). The only
2729 difference is that it has an extra argument, which is a pointer to an
2730 integer error code. When there is a compile-time failure, this is set
2731 non-zero, in addition to the error test pointer being set to point to an
2732 error message. The new argument may be NULL if no error number is required
2733 (but then you may as well call pcre_compile(), which is now just a
2734 wrapper). This facility is provided because some applications need a
2735 numeric error indication, but it has also enabled me to tidy up the way
2736 compile-time errors are handled in the POSIX wrapper.
2738 16. Added VPATH=.libs to the makefile; this should help when building with one
2739 prefix path and installing with another. (Or so I'm told by someone who
2740 knows more about this stuff than I do.)
2742 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2743 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2744 match everything, including newlines. This is not POSIX-compatible, but
2745 somebody wanted the feature. From pcretest it can be activated by using
2746 both the P and the s flags.
2748 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2750 19. libpcre.pc was being incorrectly installed as executable.
2752 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2753 it now also looks for '\r' so that it will work unmodified on Windows.
2755 21. Added Google's contributed C++ wrapper to the distribution.
2757 22. Added some untidy missing memory free() calls in pcretest, to keep
2758 Electric Fence happy when testing.
2762 Version 5.0 13-Sep-04
2763 ---------------------
2765 1. Internal change: literal characters are no longer packed up into items
2766 containing multiple characters in a single byte-string. Each character
2767 is now matched using a separate opcode. However, there may be more than one
2768 byte in the character in UTF-8 mode.
2770 2. The pcre_callout_block structure has two new fields: pattern_position and
2771 next_item_length. These contain the offset in the pattern to the next match
2772 item, and its length, respectively.
2774 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2775 insertion of callouts before each pattern item. Added the /C option to
2776 pcretest to make use of this.
2778 4. On the advice of a Windows user, the lines
2780 #if defined(_WIN32) || defined(WIN32)
2781 _setmode( _fileno( stdout ), 0x8000 );
2782 #endif /* defined(_WIN32) || defined(WIN32) */
2784 have been added to the source of pcretest. This apparently does useful
2785 magic in relation to line terminators.
2787 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2788 for the benefit of those environments where the "b" makes a difference.
2790 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2791 to know about it. I have put a hack into configure.in that adds in code
2792 to set GCC=yes if CC=icc. This seems to end up at a point in the
2793 generated configure script that is early enough to affect the setting of
2794 compiler options, which is what is needed, but I have no means of testing
2795 whether it really works. (The user who reported this had patched the
2796 generated configure script, which of course I cannot do.)
2798 LATER: After change 22 below (new libtool files), the configure script
2799 seems to know about icc (and also ecc). Therefore, I have commented out
2800 this hack in configure.in.
2802 7. Added support for pkg-config (2 patches were sent in).
2804 8. Negated POSIX character classes that used a combination of internal tables
2805 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2806 [[:^ascii]]. Typically, they would match almost any characters. The other
2807 POSIX classes were not broken in this way.
2809 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2810 to find the match, as PCRE was deluded into thinking that the match had to
2811 start at the start point or following a newline. The same bug applied to
2812 patterns with negative forward assertions or any backward assertions
2813 preceding ".*" at the start, unless the pattern required a fixed first
2814 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2816 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2817 starting at the last subject character, bytes beyond the end of the subject
2818 string were read.
2820 11. Renamed the variable "class" as "classbits" to make life easier for C++
2821 users. (Previously there was a macro definition, but it apparently wasn't
2822 enough.)
2824 12. Added the new field "tables" to the extra data so that tables can be passed
2825 in at exec time, or the internal tables can be re-selected. This allows
2826 a compiled regex to be saved and re-used at a later time by a different
2827 program that might have everything at different addresses.
2829 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2830 -R library as well as a -L library.
2832 14. The debugging options of pcretest (-d on the command line or D on a
2833 pattern) showed incorrect output for anything following an extended class
2834 that contained multibyte characters and which was followed by a quantifier.
2836 15. Added optional support for general category Unicode character properties
2837 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2838 support. It adds about 90K to the size of the library. The meanings of the
2839 inbuilt class escapes such as \d and \s have NOT been changed.
2841 16. Updated pcredemo.c to include calls to free() to release the memory for the
2842 compiled pattern.
2844 17. The generated file chartables.c was being created in the source directory
2845 instead of in the building directory. This caused the build to fail if the
2846 source directory was different from the building directory, and was
2847 read-only.
2849 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2850 file. No doubt somebody will tell me if they don't make sense... Also added
2851 Dan Mooney's comments about building on OpenVMS.
2853 19. Added support for partial matching via the PCRE_PARTIAL option for
2854 pcre_exec() and the \P data escape in pcretest.
2856 20. Extended pcretest with 3 new pattern features:
2858 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2859 write the compiled pattern to the file whose name is "rest-of-line".
2860 This is a straight binary dump of the data, with the saved pointer to
2861 the character tables forced to be NULL. The study data, if any, is
2862 written too. After writing, pcretest reads a new pattern.
2864 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2865 compiled pattern from the given file. There must not be any
2866 occurrences of "<" in the file name (pretty unlikely); if there are,
2867 pcretest will instead treat the initial "<" as a pattern delimiter.
2868 After reading in the pattern, pcretest goes on to read data lines as
2869 usual.
2871 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2872 and 16-bit fields in a compiled pattern, to simulate a pattern that
2873 was compiled on a host of opposite endianness.
2875 21. The pcre-exec() function can now cope with patterns that were compiled on
2876 hosts of opposite endianness, with this restriction:
2878 As for any compiled expression that is saved and used later, the tables
2879 pointer field cannot be preserved; the extra_data field in the arguments
2880 to pcre_exec() should be used to pass in a tables address if a value
2881 other than the default internal tables were used at compile time.
2883 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2884 now diagnosed as an error. Previously, most of the time, a negative number
2885 would have been treated as zero, but if in addition "ovector" was passed as
2886 NULL, a crash could occur.
2888 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2889 new versions from the libtool 1.5 distribution (the last one is a copy of
2890 a file called libtool.m4). This seems to have fixed the need to patch
2891 "configure" to support Darwin 1.3 (which I used to do). However, I still
2892 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2893 workstation).
2895 24. Changed the PCRE licence to be the more standard "BSD" licence.
2898 Version 4.5 01-Dec-03
2899 ---------------------
2901 1. There has been some re-arrangement of the code for the match() function so
2902 that it can be compiled in a version that does not call itself recursively.
2903 Instead, it keeps those local variables that need separate instances for
2904 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2905 needs to "recurse". Keeping track of where control must go is done by means
2906 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2907 hide most of the details from the main code, and operates only if
2908 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2909 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2910 operating.
2912 To make it easier for callers to provide specially tailored get/free
2913 functions for this usage, two new functions, pcre_stack_malloc, and
2914 pcre_stack_free, are used. They are always called in strict stacking order,
2915 and the size of block requested is always the same.
2917 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2918 PCRE has been compiled to use the stack or the heap for recursion. The
2919 -C option of pcretest uses this to show which version is compiled.
2921 A new data escape \S, is added to pcretest; it causes the amounts of store
2922 obtained and freed by both kinds of malloc/free at match time to be added
2923 to the output.
2925 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2926 what's available on my current Linux desktop machine.
2928 3. When matching a UTF-8 string, the test for a valid string at the start has
2929 been extended. If start_offset is not zero, PCRE now checks that it points
2930 to a byte that is the start of a UTF-8 character. If not, it returns
2931 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2932 this is necessary because there may be backward assertions in the pattern.
2933 When matching the same subject several times, it may save resources to use
2934 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2936 4. The code for checking the validity of UTF-8 strings has been tightened so
2937 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2938 containing "overlong sequences".
2940 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2941 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2942 should have been "&", but it just so happened that all the cases this let
2943 through by mistake were picked up later in the function.
2945 6. I had used a variable called "isblank" - this is a C99 function, causing
2946 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2948 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2949 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2950 so the version is shown; (c) stop "make test" echoing "./RunTest".
2952 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2954 9. The prototype for memmove() for systems that don't have it was using
2955 size_t, but the inclusion of the header that defines size_t was later. I've
2956 moved the #includes for the C headers earlier to avoid this.
2958 10. Added some adjustments to the code to make it easier to compiler on certain
2959 special systems:
2961 (a) Some "const" qualifiers were missing.
2962 (b) Added the macro EXPORT before all exported functions; by default this
2963 is defined to be empty.
2964 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2965 that it reads its output file name as an argument instead of writing
2966 to the standard output and assuming this can be redirected.
2968 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2969 class containing characters with values greater than 255, PCRE compilation
2970 went into a loop.
2972 12. A recursive reference to a subpattern that was within another subpattern
2973 that had a minimum quantifier of zero caused PCRE to crash. For example,
2974 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2975 recursion. If the recursively-called subpattern itself had a zero repeat,
2976 that was OK.
2978 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2979 buffer into which it was copied (for escape processing) was still set at
2980 1024, so long lines caused crashes.
2982 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2983 "internal error: code overflow...". This applied to any character class
2984 that was followed by a possessive quantifier.
2986 15. Modified the Makefile to add libpcre.la as a prerequisite for
2987 libpcreposix.la because I was told this is needed for a parallel build to
2988 work.
2990 16. If a pattern that contained .* following optional items at the start was
2991 studied, the wrong optimizing data was generated, leading to matching
2992 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2993 matching string must start with a or b or c. The correct conclusion for
2994 this pattern is that a match can start with any character.
2997 Version 4.4 13-Aug-03
2998 ---------------------
3000 1. In UTF-8 mode, a character class containing characters with values between
3001 127 and 255 was not handled correctly if the compiled pattern was studied.
3002 In fixing this, I have also improved the studying algorithm for such
3003 classes (slightly).
3005 2. Three internal functions had redundant arguments passed to them. Removal
3006 might give a very teeny performance improvement.
3008 3. Documentation bug: the value of the capture_top field in a callout is *one
3009 more than* the number of the hightest numbered captured substring.
3011 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3012 in incorrectly linking with a previously installed version. They now link
3013 explicitly with libpcre.la.
3015 5. configure.in no longer needs to recognize Cygwin specially.
3017 6. A problem in pcre.in for Windows platforms is fixed.
3019 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3020 pcretest, it used to include the size of the study block as part of its
3021 output. Unfortunately, the structure contains a field that has a different
3022 size on different hardware architectures. This meant that the tests that
3023 showed this size failed. As the block is currently always of a fixed size,
3024 this information isn't actually particularly useful in pcretest output, so
3025 I have just removed it.
3027 8. Three pre-processor statements accidentally did not start in column 1.
3028 Sadly, there are *still* compilers around that complain, even though
3029 standard C has not required this for well over a decade. Sigh.
3031 9. In pcretest, the code for checking callouts passed small integers in the
3032 callout_data field, which is a void * field. However, some picky compilers
3033 complained about the casts involved for this on 64-bit systems. Now
3034 pcretest passes the address of the small integer instead, which should get
3035 rid of the warnings.
3037 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3038 both compile and run time, and gives an error if an invalid UTF-8 sequence
3039 is found. There is a option for disabling this check in cases where the
3040 string is known to be correct and/or the maximum performance is wanted.
3042 11. In response to a bug report, I changed one line in Makefile.in from
3044 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3045 to
3046 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3048 to look similar to other lines, but I have no way of telling whether this
3049 is the right thing to do, as I do not use Windows. No doubt I'll get told
3050 if it's wrong...
3053 Version 4.3 21-May-03
3054 ---------------------
3056 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3057 Makefile.
3059 2. Some refactoring to improve the quality of the code:
3061 (i) The utf8_table... variables are now declared "const".
3063 (ii) The code for \cx, which used the "case flipping" table to upper case
3064 lower case letters, now just substracts 32. This is ASCII-specific,
3065 but the whole concept of \cx is ASCII-specific, so it seems
3066 reasonable.
3068 (iii) PCRE was using its character types table to recognize decimal and
3069 hexadecimal digits in the pattern. This is silly, because it handles
3070 only 0-9, a-f, and A-F, but the character types table is locale-
3071 specific, which means strange things might happen. A private
3072 table is now used for this - though it costs 256 bytes, a table is
3073 much faster than multiple explicit tests. Of course, the standard
3074 character types table is still used for matching digits in subject
3075 strings against \d.
3077 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3078 ending in _t are). So I've renamed it as ESC_tee.
3080 3. The first argument for regexec() in the POSIX wrapper should have been
3081 defined as "const".
3083 4. Changed pcretest to use malloc() for its buffers so that they can be
3084 Electric Fenced for debugging.
3086 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3087 to read one or more bytes before the start of the subject string. Often this
3088 had no effect on PCRE's behaviour, but in some circumstances it could
3089 provoke a segmentation fault.
3091 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3092 to try to read one or more bytes before the start of the subject string.
3094 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3095 UTF-8 support could misbehave in various ways if the subject string
3096 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3097 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3098 back over UTF-8 characters.)
3101 Version 4.2 14-Apr-03
3102 ---------------------
3104 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3106 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3107 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3108 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3109 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3110 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3112 Note: automatic setting of the BUILD variables is not yet working
3113 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3114 compile-time but not at link-time
3115 [LINK]: use for linking executables only
3116 make different versions for Windows and non-Windows
3117 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3118 libraries
3119 [LINK_FOR_BUILD]: new variable
3120 [OBJEXT]: use throughout
3121 [EXEEXT]: use throughout
3122 <winshared>: new target
3123 <wininstall>: new target
3124 <dftables.o>: use native compiler
3125 <dftables>: use native linker
3126 <install>: handle Windows platform correctly
3127 <clean>: ditto
3128 <check>: ditto
3129 copy DLL to top builddir before testing
3131 As part of these changes, -no-undefined was removed again. This was reported
3132 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3133 in any case.
3135 3. Some tidies to get rid of compiler warnings:
3137 . In the match_data structure, match_limit was an unsigned long int, whereas
3138 match_call_count was an int. I've made them both unsigned long ints.
3140 . In pcretest the fact that a const uschar * doesn't automatically cast to
3141 a void * provoked a warning.
3143 . Turning on some more compiler warnings threw up some "shadow" variables
3144 and a few more missing casts.
3146 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3147 option, a class that contained a single character with a value between 128
3148 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3150 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3151 option, a class that contained several characters, but with at least one
3152 whose value was between 128 and 255 caused PCRE to crash.
3155 Version 4.1 12-Mar-03
3156 ---------------------
3158 1. Compiling with gcc -pedantic found a couple of places where casts were
3159 needed, and a string in dftables.c that was longer than standard compilers are
3160 required to support.
3162 2. Compiling with Sun's compiler found a few more places where the code could
3163 be tidied up in order to avoid warnings.
3165 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3166 first of these names is deprecated in the latest Autoconf in favour of the name
3167 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3168 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3169 analogy I have changed it to CFLAGS_FOR_BUILD.
3171 4. Added -no-undefined to the linking command in the Makefile, because this is
3172 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3173 linking step for the pcreposix library.
3175 5. PCRE was failing to diagnose the case of two named groups with the same
3176 name.
3178 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3179 literal character that is needed in the subject for a match, and scans along to
3180 ensure that it is present before embarking on the full matching process. This
3181 saves time in cases of nested unlimited repeats that are never going to match.
3182 Problem: the scan can take a lot of time if the subject is very long (e.g.
3183 megabytes), thus penalizing straightforward matches. It is now done only if the
3184 amount of subject to be scanned is less than 1000 bytes.
3186 7. A lesser problem with the same optimization is that it was recording the
3187 first character of an anchored pattern as "needed", thus provoking a search
3188 right along the subject, even when the first match of the pattern was going to
3189 fail. The "needed" character is now not set for anchored patterns, unless it
3190 follows something in the pattern that is of non-fixed length. Thus, it still
3191 fulfils its original purpose of finding quick non-matches in cases of nested
3192 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3195 Version 4.0 17-Feb-03
3196 ---------------------
3198 1. If a comment in an extended regex that started immediately after a meta-item
3199 extended to the end of string, PCRE compiled incorrect data. This could lead to
3200 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3202 2. Moved to autoconf 2.53 and libtool 1.4.2.
3204 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3205 the special perltest8 script is no longer needed - all the tests can be run
3206 from a single perltest script.
3208 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3209 by \s. It has now been removed in PCRE. This means it isn't recognized as
3210 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3211 class [:space:] *does* include VT, thereby creating a mess.
3213 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3214 space and tab.
3216 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3217 its new features into the main test script, reducing the number of scripts.
3219 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3220 were backward compatible, and made the (?i) apply to the whole pattern, as if
3221 /i were given. Now it behaves more logically, and applies the option setting
3222 only to what follows. PCRE has been changed to follow suit. However, if it
3223 finds options settings right at the start of the pattern, it extracts them into
3224 the global options, as before. Thus, they show up in the info data.
3226 8. Added support for the \Q...\E escape sequence. Characters in between are
3227 treated as literals. This is slightly different from Perl in that $ and @ are
3228 also handled as literals inside the quotes. In Perl, they will cause variable
3229 interpolation. Note the following examples:
3231 Pattern PCRE matches Perl matches
3233 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3234 \Qabc\$xyz\E abc\$xyz abc\$xyz
3235 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3237 For compatibility with Perl, \Q...\E sequences are recognized inside character
3238 classes as well as outside them.
3240 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3241 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3242 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3243 signed/unsigned warnings.
3245 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3246 option for pcretest, so I've replaced it by a simple function that does just
3247 that job.
3249 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3250 "pcregrep -".
3252 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3253 Java package. This provides some syntactic sugar for simple cases of what my
3254 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3255 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3256 item, you can use this simplified notation. Note that only makes sense with
3257 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3258 greediness, whatever the setting of the PCRE_UNGREEDY option.
3260 13. A change of greediness default within a pattern was not taking effect at
3261 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3262 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3263 was abstracted outside.
3265 14. PCRE now supports the \G assertion. It is true when the current matching
3266 position is at the start point of the match. This differs from \A when the
3267 starting offset is non-zero. Used with the /g option of pcretest (or similar
3268 code), it works in the same way as it does for Perl's /g option. If all
3269 alternatives of a regex begin with \G, the expression is anchored to the start
3270 match position, and the "anchored" flag is set in the compiled expression.
3272 15. Some bugs concerning the handling of certain option changes within patterns
3273 have been fixed. These applied to options other than (?ims). For example,
3274 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3275 been the other way round. Some of this was related to change 7 above.
3277 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3278 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3279 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3280 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3282 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3283 mode. Unlike ".", it always matches newline, whatever the setting of
3284 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3285 assertions. Perl allows it, but it doesn't (in general) work because it can't
3286 calculate the length of the lookbehind. At least, that's the case for Perl
3287 5.8.0 - I've been told they are going to document that it doesn't work in
3288 future.
3290 18. Added an error diagnosis for escapes that PCRE does not support: these are
3291 \L, \l, \N, \P, \p, \U, \u, and \X.
3293 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3294 reading past the end of the pattern in cases such as /[abcd/.
3296 20. PCRE was getting more memory than necessary for patterns with classes that
3297 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3299 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3300 compiling PCRE for use with Virtual Pascal.
3302 22. Small fix to the Makefile to make it work properly if the build is done
3303 outside the source tree.
3305 23. Added a new extension: a condition to go with recursion. If a conditional
3306 subpattern starts with (?(R) the "true" branch is used if recursion has
3307 happened, whereas the "false" branch is used only at the top level.
3309 24. When there was a very long string of literal characters (over 255 bytes
3310 without UTF support, over 250 bytes with UTF support), the computation of how
3311 much memory was required could be incorrect, leading to segfaults or other
3312 strange effects.
3314 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3315 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3316 there was a subsequent back reference to those brackets. This meant that, for
3317 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3318 possible to check for precisely this case. All we can do is abandon the
3319 optimization if .* occurs inside capturing brackets when there are any back
3320 references whatsoever. (See below for a better fix that came later.)
3322 26. The handling of the optimization for finding the first character of a
3323 non-anchored pattern, and for finding a character that is required later in the
3324 match were failing in some cases. This didn't break the matching; it just
3325 failed to optimize when it could. The way this is done has been re-implemented.
3327 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3329 28. Added a new feature that provides some of the functionality that Perl
3330 provides with (?{...}). The facility is termed a "callout". The way it is done
3331 in PCRE is for the caller to provide an optional function, by setting
3332 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3333 global variable. By default it is unset, which disables all calling out. To get
3334 the function called, the regex must include (?C) at appropriate points. This
3335 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3336 This provides a means of identifying different callout points. When PCRE
3337 reaches such a point in the regex, if pcre_callout has been set, the external
3338 function is called. It is provided with data in a structure called
3339 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3340 matching continues; if it returns a non-zero value, the match at the current
3341 point fails. However, backtracking will occur if possible. [This was changed
3342 later and other features added - see item 49 below.]
3344 29. pcretest is upgraded to test the callout functionality. It provides a
3345 callout function that displays information. By default, it shows the start of
3346 the match and the current position in the text. There are some new data escapes
3347 to vary what happens:
3349 \C+ in addition, show current contents of captured substrings
3350 \C- do not supply a callout function
3351 \C!n return 1 when callout number n is reached
3352 \C!n!m return 1 when callout number n is reached for the mth time
3354 30. If pcregrep was called with the -l option and just a single file name, it
3355 output "<stdin>" if a match was found, instead of the file name.
3357 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3358 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3359 pcre_exec(). This saves a malloc/free per call. The default value of
3360 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3361 when configuring.
3363 32. The default maximum size of a compiled pattern is 64K. There have been a
3364 few cases of people hitting this limit. The code now uses macros to handle the
3365 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3366 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3367 configuring. Tests 2 and 5 work only with 2-byte links because they output
3368 debugging information about compiled patterns.
3370 33. Internal code re-arrangements:
3372 (a) Moved the debugging function for printing out a compiled regex into
3373 its own source file (printint.c) and used #include to pull it into
3374 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3375 separate copies.
3377 (b) Defined the list of op-code names for debugging as a macro in
3378 internal.h so that it is next to the definition of the opcodes.
3380 (c) Defined a table of op-code lengths for simpler skipping along compiled
3381 code. This is again a macro in internal.h so that it is next to the
3382 definition of the opcodes.
3384 34. Added support for recursive calls to individual subpatterns, along the
3385 lines of Robin Houston's patch (but implemented somewhat differently).
3387 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3388 allow it to read and process whole directories in Win32. This code was
3389 contributed by Lionel Fourquaux; it has not been tested by me.
3391 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3392 used to name a group. Names consist of alphanumerics and underscores, and must
3393 be unique. Back references use the syntax (?P=name) and recursive calls use
3394 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3395 numbers. The function pcre_fullinfo() can be used after compilation to extract
3396 a name/number map. There are three relevant calls:
3398 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3399 PCRE_INFO_NAMECOUNT yields the number of entries
3400 PCRE_INFO_NAMETABLE yields a pointer to the map.
3402 The map is a vector of fixed-size entries. The size of each entry depends on
3403 the length of the longest name used. The first two bytes of each entry are the
3404 group number, most significant byte first. There follows the corresponding
3405 name, zero terminated. The names are in alphabetical order.
3407 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3408 case instead of 255. Making it the same both with and without UTF-8 support
3409 means that the same test output works with both.
3411 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3412 calling malloc() with a zero argument.
3414 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3415 optimization. I've improved things by keeping a bitmap of backreferences with
3416 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3417 fact referenced, the optimization can be applied. It is unlikely that a
3418 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3419 the match to follow \n) will appear inside brackets with a number greater than
3420 31, but if it does, any back reference > 31 suppresses the optimization.
3422 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3423 of disabling numbered capturing parentheses. Any opening parenthesis that is
3424 not followed by ? behaves as if it were followed by ?: but named parentheses
3425 can still be used for capturing (and they will acquire numbers in the usual
3426 way).
3428 41. Redesigned the return codes from the match() function into yes/no/error so
3429 that errors can be passed back from deep inside the nested calls. A malloc
3430 failure while inside a recursive subpattern call now causes the
3431 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3433 42. It is now possible to set a limit on the number of times the match()
3434 function is called in a call to pcre_exec(). This facility makes it possible to
3435 limit the amount of recursion and backtracking, though not in a directly
3436 obvious way, because the match() function is used in a number of different
3437 circumstances. The count starts from zero for each position in the subject
3438 string (for non-anchored patterns). The default limit is, for compatibility, a
3439 large number, namely 10 000 000. You can change this in two ways:
3441 (a) When configuring PCRE before making, you can use --with-match-limit=n
3442 to set a default value for the compiled library.
3444 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3445 a different value is set. See 45 below.
3447 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3449 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3450 of things that can be changed at compile time. The first argument specifies
3451 what is wanted and the second points to where the information is to be placed.
3452 The current list of available information is:
3456 The output is an integer that is set to one if UTF-8 support is available;
3457 otherwise it is set to zero.
3461 The output is an integer that it set to the value of the code that is used for
3462 newline. It is either LF (10) or CR (13).
3466 The output is an integer that contains the number of bytes used for internal
3467 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3471 The output is an integer that contains the threshold above which the POSIX
3472 interface uses malloc() for output vectors. See item 31 above.
3476 The output is an unsigned integer that contains the default limit of the number
3477 of match() calls in a pcre_exec() execution. See 42 above.
3479 44. pcretest has been upgraded by the addition of the -C option. This causes it
3480 to extract all the available output from the new pcre_config() function, and to
3481 output it. The program then exits immediately.
3483 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3484 order to support additional features. One way would have been to define
3485 pcre_exec2() (for example) with extra arguments, but this would not have been
3486 extensible, and would also have required all calls to the original function to
3487 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3488 is used for passing in "extra" data from pcre_study().
3490 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3491 contains the following fields:
3493 flags a bitmap indicating which of the following fields are set
3494 study_data opaque data from pcre_study()
3495 match_limit a way of specifying a limit on match() calls for a specific
3496 call to pcre_exec()
3497 callout_data data for callouts (see 49 below)
3499 The flag bits are also defined in pcre.h, and are
3505 The pcre_study() function now returns one of these new pcre_extra blocks, with
3506 the actual study data pointed to by the study_data field, and the
3507 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3508 before. That is, this change is entirely upwards-compatible and requires no
3509 change to existing code.
3511 If you want to pass in additional data to pcre_exec(), you can either place it
3512 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3513 block.
3515 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3516 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3517 times with different match limits, until it finds the minimum value needed for
3518 pcre_exec() to complete. The value is then output. This can be instructive; for
3519 most simple matches the number is quite small, but for pathological cases it
3520 gets very large very quickly.
3522 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3523 returns the size of the data block pointed to by the study_data field in a
3524 pcre_extra block, that is, the value that was passed as the argument to
3525 pcre_malloc() when PCRE was getting memory in which to place the information
3526 created by pcre_study(). The fourth argument should point to a size_t variable.
3527 pcretest has been extended so that this information is shown after a successful
3528 pcre_study() call when information about the compiled regex is being displayed.
3530 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3531 because what follows is always an absolute path. (Later: it turns out that this
3532 is more than cosmetic for MinGW, because it doesn't like empty path
3533 components.)
3535 49. Some changes have been made to the callout feature (see 28 above):
3537 (i) A callout function now has three choices for what it returns:
3539 0 => success, carry on matching
3540 > 0 => failure at this point, but backtrack if possible
3541 < 0 => serious error, return this value from pcre_exec()
3543 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3544 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3545 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3546 use by callout functions. It will never be used by PCRE itself.
3548 (ii) The pcre_extra structure (see 45 above) has a void * field called
3549 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3550 pcre_callout_block structure has a field of the same name. The contents of
3551 the field passed in the pcre_extra structure are passed to the callout
3552 function in the corresponding field in the callout block. This makes it
3553 easier to use the same callout-containing regex from multiple threads. For
3554 testing, the pcretest program has a new data escape
3556 \C*n pass the number n (may be negative) as callout_data
3558 If the callout function in pcretest receives a non-zero value as
3559 callout_data, it returns that value.
3561 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3562 there were some redundant $(CFLAGS) in commands that are now specified as
3563 $(LINK), which already includes $(CFLAGS).
3565 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3566 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3567 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3568 one-byte characters throughout. Note that case-insensitive matching applies
3569 only to characters whose values are less than 256. PCRE doesn't support the
3570 notion of cases for higher-valued characters.
3572 (i) A character class whose characters are all within 0-255 is handled as
3573 a bit map, and the map is inverted for negative classes. Previously, a
3574 character > 255 always failed to match such a class; however it should
3575 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3577 (ii) A negated character class with a single character < 255 is coded as
3578 "not this character" (OP_NOT). This wasn't working properly when the test
3579 character was multibyte, either singly or repeated.
3581 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3582 mode, for example: \x{100}{2,3}.
3584 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3585 singly or repeated) now correctly test multibyte characters. However,
3586 PCRE doesn't recognize any characters with values greater than 255 as
3587 digits, spaces, or word characters. Such characters always match \D, \S,
3588 and \W, and never match \d, \s, or \w.
3590 (v) Classes may now contain characters and character ranges with values
3591 greater than 255. For example: [ab\x{100}-\x{400}].
3593 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3594 PCRE in UTF-8 mode.
3596 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3597 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3598 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3599 value.)
3601 53. The single man page has become too large. I have therefore split it up into
3602 a number of separate man pages. These also give rise to individual HTML pages;
3603 these are now put in a separate directory, and there is an index.html page that
3604 lists them all. Some hyperlinking between the pages has been installed.
3606 54. Added convenience functions for handling named capturing parentheses.
3608 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3609 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3610 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3611 are faulted.
3613 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3614 calling configure. These values are used when compiling the dftables.c program
3615 which is run to generate the source of the default character tables. They
3616 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3617 you will need to set these values.
3619 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3622 Version 3.9 02-Jan-02
3623 ---------------------
3625 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3627 2. If --disable-static was given, the building process failed when trying to
3628 build pcretest and pcregrep. (For some reason it was using libtool to compile
3629 them, which is not right, as they aren't part of the library.)
3632 Version 3.8 18-Dec-01
3633 ---------------------
3635 1. The experimental UTF-8 code was completely screwed up. It was packing the
3636 bytes in the wrong order. How dumb can you get?
3639 Version 3.7 29-Oct-01
3640 ---------------------
3642 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3643 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3644 this didn't happen under Solaris 8, where I normally test things.
3646 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3647 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3648 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3649 this fix an hour or so after the initial 3.7 release.)
3652 Version 3.6 23-Oct-01
3653 ---------------------
3655 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3656 offsets passed as NULL with zero offset count.
3658 2. The config.guess and config.sub files had not been updated when I moved to
3659 the latest autoconf.
3662 Version 3.5 15-Aug-01
3663 ---------------------
3665 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3666 had been forgotten.
3668 2. By using declared but undefined structures, we can avoid using "void"
3669 definitions in pcre.h while keeping the internal definitions of the structures
3670 private.
3672 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3673 user point of view, this means that both static and shared libraries are built
3674 by default, but this can be individually controlled. More of the work of
3675 handling this static/shared cases is now inside libtool instead of PCRE's make
3676 file.
3678 4. The pcretest utility is now installed along with pcregrep because it is
3679 useful for users (to test regexs) and by doing this, it automatically gets
3680 relinked by libtool. The documentation has been turned into a man page, so
3681 there are now .1, .txt, and .html versions in /doc.
3683 5. Upgrades to pcregrep:
3684 (i) Added long-form option names like gnu grep.
3685 (ii) Added --help to list all options with an explanatory phrase.
3686 (iii) Added -r, --recursive to recurse into sub-directories.
3687 (iv) Added -f, --file to read patterns from a file.
3689 6. pcre_exec() was referring to its "code" argument before testing that
3690 argument for NULL (and giving an error if it was NULL).
3692 7. Upgraded Makefile.in to allow for compiling in a different directory from
3693 the source directory.
3695 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3696 options bits, the pointer it was passed was to an int instead of to an unsigned
3697 long int. This mattered only on 64-bit systems.
3699 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3700 generated) instead of pcre.in, which it its source. Also made the same change
3701 in several of the .c files.
3703 10. A new release of gcc defines printf() as a macro, which broke pcretest
3704 because it had an ifdef in the middle of a string argument for printf(). Fixed
3705 by using separate calls to printf().
3707 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3708 script, to force use of CR or LF instead of \n in the source. On non-Unix
3709 systems, the value can be set in config.h.
3711 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3712 absolute limit. Changed the text of the error message to make this clear, and
3713 likewise updated the man page.
3715 13. The limit of 99 on the number of capturing subpatterns has been removed.
3716 The new limit is 65535, which I hope will not be a "real" limit.
3719 Version 3.4 22-Aug-00
3720 ---------------------
3722 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3724 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3727 Version 3.3 01-Aug-00
3728 ---------------------
3730 1. If an octal character was given, but the value was greater than \377, it
3731 was not getting masked to the least significant bits, as documented. This could
3732 lead to crashes in some systems.
3734 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3735 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3737 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3738 These just pass their arguments on to (pcre_free)(), but they are provided
3739 because some uses of PCRE bind it to non-C systems that can call its functions,
3740 but cannot call free() or pcre_free() directly.
3742 4. Add "make test" as a synonym for "make check". Corrected some comments in
3743 the Makefile.
3745 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3746 Makefile.
3748 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3749 command called pgrep for grepping around the active processes.
3751 7. Added the beginnings of support for UTF-8 character strings.
3753 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3754 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3755 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3756 out for the ar command.)
3759 Version 3.2 12-May-00
3760 ---------------------
3762 This is purely a bug fixing release.
3764 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3765 of ZA. This was just one example of several cases that could provoke this bug,
3766 which was introduced by change 9 of version 2.00. The code for breaking
3767 infinite loops after an iteration that matches an empty string was't working
3768 correctly.
3770 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3771 when matched against abbab (for example). After matching an empty string, it
3772 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3773 caused it to match further down the string than it should.
3775 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3776 was there because it doesn't seem to be needed, and it causes trouble on some
3777 systems, as it is not a Standard C header. It has been removed.
3779 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3780 were reported on the Macintosh. The changes were from
3782 while ((c = *(++ptr)) != 0 && c != '\n');
3783 to
3784 while ((c = *(++ptr)) != 0 && c != '\n') ;
3786 Totally extraordinary, but if that's what it takes...
3788 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3789 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3790 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3791 assumes the way PCRE uses memmove() (always moving upwards).
3793 6. PCRE is being used in one environment where strchr() is not available. There
3794 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3795 faster code anyway.
3798 Version 3.1 09-Feb-00
3799 ---------------------
3801 The only change in this release is the fixing of some bugs in Makefile.in for
3802 the "install" target:
3804 (1) It was failing to install pcreposix.h.
3806 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3809 Version 3.0 01-Feb-00
3810 ---------------------
3812 1. Add support for the /+ modifier to perltest (to output $` like it does in
3813 pcretest).
3815 2. Add support for the /g modifier to perltest.
3817 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3818 matches null strings.
3820 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3821 pattern. Perl treats empty patterns specially - it reuses the most recent
3822 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3823 effect.
3825 5. The POSIX interface was broken in that it was just handing over the POSIX
3826 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3827 required a bigger vector, with some working space on the end. This means that
3828 the POSIX wrapper now has to get and free some memory, and copy the results.
3830 6. Added some simple autoconf support, placing the test data and the
3831 documentation in separate directories, re-organizing some of the
3832 information files, and making it build pcre-config (a GNU standard). Also added
3833 libtool support for building PCRE as a shared library, which is now the
3834 default.
3836 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3837 09 are not valid octal constants. Single digits will be used for minor values
3838 less than 10.
3840 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3841 existing programs that set these in the POSIX interface can use PCRE without
3842 modification.
3844 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3845 return all that pcre_info() returns, plus additional data. The pcre_info()
3846 function is retained for compatibility, but is considered to be obsolete.
3848 10. Added experimental recursion feature (?R) to handle one common case that
3849 Perl 5.6 will be able to do with (?p{...}).
3851 11. Added support for POSIX character classes like [:alpha:], which Perl is
3852 adopting.
3855 Version 2.08 31-Aug-99
3856 ----------------------
3858 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3859 trying to match at the startoffset position, but instead was moving forward to
3860 the next newline as if a previous match had failed.
3862 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3863 and could get into a loop if a null string was matched other than at the start
3864 of the subject.
3866 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3867 be distinguished at compile time, and for completeness also added PCRE_DATE.
3869 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3870 in GnuWin32 environments.
3873 Version 2.07 29-Jul-99
3874 ----------------------
3876 1. The documentation is now supplied in plain text form and HTML as well as in
3877 the form of man page sources.
3879 2. C++ compilers don't like assigning (void *) values to other pointer types.
3880 In particular this affects malloc(). Although there is no problem in Standard
3881 C, I've put in casts to keep C++ compilers happy.
3883 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3884 should be (const char *).
3886 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3887 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3888 However, I haven't made this a standard facility. The documentation doesn't
3889 mention it, and the Makefile doesn't support it.
3891 5. The Makefile now contains an "install" target, with editable destinations at
3892 the top of the file. The pcretest program is not installed.
3894 6. pgrep -V now gives the PCRE version number and date.
3896 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3897 causing the entire string to be ignored, instead of just the last character.
3899 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3900 non-matching string, it can take a very, very long time, even for strings of
3901 quite modest length, because of the nested recursion. PCRE now does better in
3902 some of these cases. It does this by remembering the last required literal
3903 character in the pattern, and pre-searching the subject to ensure it is present
3904 before running the real match. In other words, it applies a heuristic to detect
3905 some types of certain failure quickly, and in the above example, if presented
3906 with a string that has no trailing " it gives "no match" very quickly.
3908 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3909 other alternatives are tried instead.
3912 Version 2.06 09-Jun-99
3913 ----------------------
3915 1. Change pcretest's output for amount of store used to show just the code
3916 space, because the remainder (the data block) varies in size between 32-bit and
3917 64-bit systems.
3919 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3920 start matching at. This allows lookbehinds to work when searching for multiple
3921 occurrences in a string.
3923 3. Added additional options to pcretest for testing multiple occurrences:
3925 /+ outputs the rest of the string that follows a match
3926 /g loops for multiple occurrences, using the new startoffset argument
3927 /G loops for multiple occurrences by passing an incremented pointer
3929 4. PCRE wasn't doing the "first character" optimization for patterns starting
3930 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3931 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3932 the letter 'x'. On long subject strings, this gives a significant speed-up.
3935 Version 2.05 21-Apr-99
3936 ----------------------
3938 1. Changed the type of magic_number from int to long int so that it works
3939 properly on 16-bit systems.
3941 2. Fixed a bug which caused patterns starting with .* not to work correctly
3942 when the subject string contained newline characters. PCRE was assuming
3943 anchoring for such patterns in all cases, which is not correct because .* will
3944 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3945 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3946 must be retried after every newline in the subject.
3949 Version 2.04 18-Feb-99
3950 ----------------------
3952 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3953 computation of the store needed to hold the pattern was incorrect (too large).
3954 If such patterns were nested a few deep, this could multiply and become a real
3955 problem.
3957 2. Added /M option to pcretest to show the memory requirement of a specific
3958 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3960 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3961 compiled in such a way that the backtracking after subsequent failure was
3962 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3963 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3966 Version 2.03 02-Feb-99
3967 ----------------------
3969 1. Fixed typo and small mistake in man page.
3971 2. Added 4th condition (GPL supersedes if conflict) and created separate
3972 LICENCE file containing the conditions.
3974 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3975 Perl, that is the internal \ allows the delimiter to be included in the
3976 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3977 the final delimiter, add \ to the end of the pattern (to test the error).
3979 4. Added the convenience functions for extracting substrings after a successful
3980 match. Updated pcretest to make it able to test these functions.
3983 Version 2.02 14-Jan-99
3984 ----------------------
3986 1. Initialized the working variables associated with each extraction so that
3987 their saving and restoring doesn't refer to uninitialized store.
3989 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3990 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3991 fix the problem.
3993 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3994 calls, and wasn't printing the correct value for compiling calls. Increased the
3995 default value of LOOPREPEAT, and the number of significant figures in the
3996 times.
3998 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4000 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4001 a building problem on Windows NT with a FAT file system.
4004 Version 2.01 21-Oct-98
4005 ----------------------
4007 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4008 to character tables built by pcre_maketables() in the current locale. If NULL
4009 is passed, the default tables are used.
4012 Version 2.00 24-Sep-98
4013 ----------------------
4015 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4016 it any more.
4018 2. Allow quantification of (?>) groups, and make it work correctly.
4020 3. The first character computation wasn't working for (?>) groups.
4022 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4023 end of the subject) and add 5.005's \z, which really does match only at the
4024 very end of the subject.
4026 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4028 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4029 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4030 localized options. All options to pcre_study() were also removed.
4032 7. Add other new features from 5.005:
4034 $(?<= positive lookbehind
4035 $(?<! negative lookbehind
4036 (?imsx-imsx) added the unsetting capability
4037 such a setting is global if at outer level; local otherwise
4038 (?imsx-imsx:) non-capturing groups with option setting
4039 (?(cond)re|re) conditional pattern matching
4041 A backreference to itself in a repeated group matches the previous
4042 captured string.
4044 8. General tidying up of studying (both automatic and via "study")
4045 consequential on the addition of new assertions.
4047 9. As in 5.005, unlimited repeated groups that could match an empty substring
4048 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4049 runtime if any iteration does actually match an empty substring.
4051 10. Include the RunTest script in the distribution.
4053 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4054 discrepancies, some of which were old and were also with respect to 5.004. They
4055 have now been fixed.
4058 Version 1.09 28-Apr-98
4059 ----------------------
4061 1. A negated single character class followed by a quantifier with a minimum
4062 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4063 program crashes, or just wrong answers. This did not apply to negated classes
4064 containing more than one character, or to minima other than one.
4067 Version 1.08 27-Mar-98
4068 ----------------------
4070 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4072 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4073 latter must appear before anything that relies on it in the pattern.
4076 Version 1.07 16-Feb-98
4077 ----------------------
4079 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4080 repeat of a potentially empty string).
4083 Version 1.06 23-Jan-98
4084 ----------------------
4086 1. Added Markus Oberhumer's little patches for C++.
4088 2. Literal strings longer than 255 characters were broken.
4091 Version 1.05 23-Dec-97
4092 ----------------------
4094 1. Negated character classes containing more than one character were failing if
4095 PCRE_CASELESS was set at run time.
4098 Version 1.04 19-Dec-97
4099 ----------------------
4101 1. Corrected the man page, where some "const" qualifiers had been omitted.
4103 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4104 input syntax.
4106 3. Fixed memory leak which occurred when a regex with back references was
4107 matched with an offsets vector that wasn't big enough. The temporary memory
4108 that is used in this case wasn't being freed if the match failed.
4110 4. Tidied pcretest to ensure it frees memory that it gets.
4112 5. Temporary memory was being obtained in the case where the passed offsets
4113 vector was exactly big enough.
4115 6. Corrected definition of offsetof() from change 5 below.
4117 7. I had screwed up change 6 below and broken the rules for the use of
4118 setjmp(). Now fixed.
4121 Version 1.03 18-Dec-97
4122 ----------------------
4124 1. A erroneous regex with a missing opening parenthesis was correctly
4125 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4126 on some systems.
4128 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4129 it was reported that one broken compiler failed on the former because "code" is
4130 also an independent variable.
4132 3. The erroneous regex a[]b caused an array overrun reference.
4134 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4135 fail on data ending with that character. (It was going on too far, and checking
4136 the next character, typically a binary zero.) This was specific to the
4137 optimized code for single-character negative classes.
4139 5. Added a contributed patch from the TIN world which does the following:
4141 + Add an undef for memmove, in case the the system defines a macro for it.
4143 + Add a definition of offsetof(), in case there isn't one. (I don't know
4144 the reason behind this - offsetof() is part of the ANSI standard - but
4145 it does no harm).
4147 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4148 most of the places where whitespace preceded '#'. I have given up and
4149 allowed the remaining 2 cases to be at the margin.
4151 + Rename some variables in pcre to eliminate shadowing. This seems very
4152 pedantic, but does no harm, of course.
4154 6. Moved the call to setjmp() into its own function, to get rid of warnings
4155 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4157 7. Constructs such as \d{8,} were compiling into the equivalent of
4158 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4159 outcome, but in this particular case used more store than had been allocated,
4160 which caused the bug to be discovered because it threw up an internal error.
4162 8. The debugging code in both pcre and pcretest for outputting the compiled
4163 form of a regex was going wrong in the case of back references followed by
4164 curly-bracketed repeats.
4167 Version 1.02 12-Dec-97
4168 ----------------------
4170 1. Typos in pcre.3 and comments in the source fixed.
4172 2. Applied a contributed patch to get rid of places where it used to remove
4173 'const' from variables, and fixed some signed/unsigned and uninitialized
4174 variable warnings.
4176 3. Added the "runtest" target to Makefile.
4178 4. Set default compiler flag to -O2 rather than just -O.
4181 Version 1.01 19-Nov-97
4182 ----------------------
4184 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4185 like /([ab]*)*/, that is, for classes with more than one character in them.
4187 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4188 as /((?>a*))*/ (a PCRE_EXTRA facility).
4191 Version 1.00 18-Nov-97
4192 ----------------------
4194 1. Added compile-time macros to support systems such as SunOS4 which don't have
4195 memmove() or strerror() but have other things that can be used instead.
4197 2. Arranged that "make clean" removes the executables.
4200 Version 0.99 27-Oct-97
4201 ----------------------
4203 1. Fixed bug in code for optimizing classes with only one character. It was
4204 initializing a 32-byte map regardless, which could cause it to run off the end
4205 of the memory it had got.
4207 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4210 Version 0.98 22-Oct-97
4211 ----------------------
4213 1. Fixed bug in code for handling temporary memory usage when there are more
4214 back references than supplied space in the ovector. This could cause segfaults.
4217 Version 0.97 21-Oct-97
4218 ----------------------
4220 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4222 2. Optimized negated single characters not to use a bit map.
4224 3. Brought error texts together as macro definitions; clarified some of them;
4225 fixed one that was wrong - it said "range out of order" when it meant "invalid
4226 escape sequence".
4228 4. Changed some char * arguments to const char *.
4230 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4232 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4233 pcretest.
4236 Version 0.96 16-Oct-97
4237 ----------------------
4239 1. Added a simple "pgrep" utility to the distribution.
4241 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4242 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4243 where "ddd" means "one or more decimal digits".
4245 3. Fixed serious bug. If a pattern had a back reference, but the call to
4246 pcre_exec() didn't supply a large enough ovector to record the related
4247 identifying subpattern, the match always failed. PCRE now remembers the number
4248 of the largest back reference, and gets some temporary memory in which to save
4249 the offsets during matching if necessary, in order to ensure that
4250 backreferences always work.
4252 4. Increased the compatibility with Perl in a number of ways:
4254 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4255 to request this handling. The option can be set at compile or exec time.
4257 (b) $ matches before a terminating newline by default; an option
4258 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4259 mode). The option can be set at compile or exec time.
4261 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4262 the same as Perl's. If the decimal number it represents is less than 10
4263 or there aren't that many previous left capturing parentheses, an octal
4264 escape is read. Inside a character class, it's always an octal escape,
4265 even if it is a single digit.
4267 (d) An escaped but undefined alphabetic character is taken as a literal,
4268 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4269 escapes.
4271 (e) {0} is now permitted. (The previous item is removed from the compiled
4272 pattern).
4274 5. Changed all the names of code files so that the basic parts are no longer
4275 than 10 characters, and abolished the teeny "globals.c" file.
4277 6. Changed the handling of character classes; they are now done with a 32-byte
4278 bit map always.
4280 7. Added the -d and /D options to pcretest to make it possible to look at the
4281 internals of compilation without having to recompile pcre.
4284 Version 0.95 23-Sep-97
4285 ----------------------
4287 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4288 \x20 at the start of a run of normal characters. These were being treated as
4289 real characters, instead of the source characters being re-checked.
4292 Version 0.94 18-Sep-97
4293 ----------------------
4295 1. The functions are now thread-safe, with the caveat that the global variables
4296 containing pointers to malloc() and free() or alternative functions are the
4297 same for all threads.
4299 2. Get pcre_study() to generate a bitmap of initial characters for non-
4300 anchored patterns when this is possible, and use it if passed to pcre_exec().
4303 Version 0.93 15-Sep-97
4304 ----------------------
4306 1. /(b)|(:+)/ was computing an incorrect first character.
4308 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4309 but not actually doing anything yet.
4311 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4312 as Perl does (e.g. [-az] or [az-]).
4314 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4315 all possible positions.
4317 5. Split up into different modules to avoid including unneeded functions in a
4318 compiled binary. However, compile and exec are still in one module. The "study"
4319 function is split off.
4321 6. The character tables are now in a separate module whose source is generated
4322 by an auxiliary program - but can then be edited by hand if required. There are
4323 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4324 toupper() in the code.
4326 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4327 make them global. Abolish the function for setting them, as the caller can now
4328 set them directly.
4331 Version 0.92 11-Sep-97
4332 ----------------------
4334 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4335 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4337 2. Caseless matching was not working in character classes if the characters in
4338 the pattern were in upper case.
4340 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4342 4. Make PCRE_ANCHORED public and accept as a compile option.
4344 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4345 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4346 pass them.
4348 6. Give an error if bad option bits passed at compile or run time.
4350 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4351 pcretest to cause it to pass that flag.
4353 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4354 options, and the first character, if set.
4356 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4359 Version 0.91 10-Sep-97
4360 ----------------------
4362 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4363 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4365 2. PCRE was looping on encountering an indefinitely repeated back reference to
4366 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4367 Perl does - treats the match as successful.
4369 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5