ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 747 - (show annotations)
Tue Nov 15 17:35:10 2011 UTC (3 years, 10 months ago) by ph10
File size: 219192 byte(s)
Error occurred while calculating annotation data.
Fixed several items that were being incorrectly rejected as "not fixed length"
in lookbehinds. 
1 ChangeLog for PCRE
2 ------------------
4 Version 8.21
5 ------------
7 1. Updating the JIT compiler.
9 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
10 are added as well.
12 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
13 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
14 calling _pcre_jit_exec. Some extra comments are added.
16 4. Mark settings inside atomic groups that do not contain any capturing
17 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
18 was introduced by change 18 for 8.20.
20 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
21 ECMA-262 standard.
23 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
24 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
25 This bug was probably introduced by change 9 of 8.13.
27 7. While fixing 6 above, I noticed that a number of other items were being
28 incorrectly rejected as "not fixed length". This arose partly because newer
29 opcodes had not been added to the fixed-length checking code. I have (a)
30 corrected the bug and added tests for these items, and (b) arranged for an
31 error to occur if an unknown opcode is encountered while checking for fixed
32 length instead of just assuming "not fixed length". The items that were
33 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
34 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
35 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
38 Version 8.20 21-Oct-2011
39 ------------------------
41 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
42 a POSIX class. After further experiments with Perl, which convinced me that
43 Perl has bugs and confusions, a closing square bracket is no longer allowed
44 in a POSIX name. This bug also affected patterns with classes that started
45 with full stops.
47 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
48 captured substring, but while checking the failing first alternative,
49 substring 1 is temporarily captured. If the output vector supplied to
50 pcre_exec() was not big enough for this capture, the yield of the function
51 was still zero ("insufficient space for captured substrings"). This cannot
52 be totally fixed without adding another stack variable, which seems a lot
53 of expense for a edge case. However, I have improved the situation in cases
54 such as /(a)(b)x|abc/ matched against "abc", where the return code
55 indicates that fewer than the maximum number of slots in the ovector have
56 been set.
58 3. Related to (2) above: when there are more back references in a pattern than
59 slots in the output vector, pcre_exec() uses temporary memory during
60 matching, and copies in the captures as far as possible afterwards. It was
61 using the entire output vector, but this conflicts with the specification
62 that only 2/3 is used for passing back captured substrings. Now it uses
63 only the first 2/3, for compatibility. This is, of course, another edge
64 case.
66 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
67 main code base, and can be used by building with --enable-jit. When this is
68 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
69 runtime --no-jit option is given.
71 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
72 ovector, the return from the function was zero, implying that there were
73 other matches that did not fit. The correct "exactly full" value is now
74 returned.
76 6. If a subpattern that was called recursively or as a subroutine contained
77 (*PRUNE) or any other control that caused it to give a non-standard return,
78 invalid errors such as "Error -26 (nested recursion at the same subject
79 position)" or even infinite loops could occur.
81 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
82 computing the minimum length on reaching *ACCEPT, and so ended up with the
83 wrong value of 1 rather than 0. Further investigation indicates that
84 computing a minimum subject length in the presence of *ACCEPT is difficult
85 (think back references, subroutine calls), and so I have changed the code
86 so that no minimum is registered for a pattern that contains *ACCEPT.
88 8. If (*THEN) was present in the first (true) branch of a conditional group,
89 it was not handled as intended. [But see 16 below.]
91 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
92 Sheri Pierce.
94 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
95 the first byte in a match must be "a".
97 11. Change 17 for 8.13 increased the recursion depth for patterns like
98 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
99 pattern contains any instances of (*THEN). If it does not, the old
100 optimizations are restored. It would be nice to do this on a per-group
101 basis, but at the moment that is not feasible.
103 12. In some environments, the output of pcretest -C is CRLF terminated. This
104 broke RunTest's code that checks for the link size. A single white space
105 character after the value is now allowed for.
107 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
108 For "fr", it uses the Windows-specific input and output files.
110 14. If (*THEN) appeared in a group that was called recursively or as a
111 subroutine, it did not work as intended. [But see next item.]
113 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
114 pattern fragments (but not containing any | characters). If A and B are
115 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
116 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
117 D. In other words, Perl considers parentheses that do not contain any |
118 characters to be part of a surrounding alternative, whereas PCRE was
119 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
120 differently. PCRE now behaves in the same way as Perl, except in the case
121 of subroutine/recursion calls such as (?1) which have in any case always
122 been different (but PCRE had them first :-).
124 16. Related to 15 above: Perl does not treat the | in a conditional group as
125 creating alternatives. Such a group is treated in the same way as an
126 ordinary group without any | characters when processing (*THEN). PCRE has
127 been changed to match Perl's behaviour.
129 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
130 RunGrepTest script failed.
132 18. Change 22 for version 13 caused atomic groups to use more stack. This is
133 inevitable for groups that contain captures, but it can lead to a lot of
134 stack use in large patterns. The old behaviour has been restored for atomic
135 groups that do not contain any capturing parentheses.
137 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
138 suppress the check for a minimum subject length at run time. (If it was
139 given to pcre_exec() or pcre_dfa_exec() it did work.)
141 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
142 fail to work when decoding hex characters in data strings in EBCDIC
143 environments.
145 21. It appears that in at least one Mac OS environment, the isxdigit() function
146 is implemented as a macro that evaluates to its argument more than once,
147 contravening the C 90 Standard (I haven't checked a later standard). There
148 was an instance in pcretest which caused it to go wrong when processing
149 \x{...} escapes in subject strings. The has been rewritten to avoid using
150 things like p++ in the argument of isxdigit().
153 Version 8.13 16-Aug-2011
154 ------------------------
156 1. The Unicode data tables have been updated to Unicode 6.0.0.
158 2. Two minor typos in pcre_internal.h have been fixed.
160 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
161 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
162 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
164 4. There were a number of related bugs in the code for matching backrefences
165 caselessly in UTF-8 mode when codes for the characters concerned were
166 different numbers of bytes. For example, U+023A and U+2C65 are an upper
167 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
168 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
169 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
170 2-byte code at the end of the subject (it thought there wasn't enough data
171 left).
173 5. Comprehensive information about what went wrong is now returned by
174 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
175 as the output vector has at least 2 elements. The offset of the start of
176 the failing character and a reason code are placed in the vector.
178 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
179 now returned is for the first byte of the failing character, instead of the
180 last byte inspected. This is an incompatible change, but I hope it is small
181 enough not to be a problem. It makes the returned offset consistent with
182 pcre_exec() and pcre_dfa_exec().
184 7. pcretest now gives a text phrase as well as the error number when
185 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
186 failure, the offset and reason code are output.
188 8. When \R was used with a maximizing quantifier it failed to skip backwards
189 over a \r\n pair if the subsequent match failed. Instead, it just skipped
190 back over a single character (\n). This seems wrong (because it treated the
191 two characters as a single entity when going forwards), conflicts with the
192 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
193 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
194 has been changed.
196 9. Some internal refactoring has changed the processing so that the handling
197 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
198 time (the PCRE_DOTALL option was changed this way some time ago: version
199 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
200 which was always a bit of a fudge. It also means that there is one less
201 argument for the match() function, which reduces its stack requirements
202 slightly. This change also fixes an incompatibility with Perl: the pattern
203 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
205 10. More internal refactoring has drastically reduced the number of recursive
206 calls to match() for possessively repeated groups such as (abc)++ when
207 using pcre_exec().
209 11. While implementing 10, a number of bugs in the handling of groups were
210 discovered and fixed:
212 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
213 (a|)*(?1) gave a compile-time internal error.
214 ((a|)+)+ did not notice that the outer group could match an empty string.
215 (^a|^)+ was not marked as anchored.
216 (.*a|.*)+ was not marked as matching at start or after a newline.
218 12. Yet more internal refactoring has removed another argument from the match()
219 function. Special calls to this function are now indicated by setting a
220 value in a variable in the "match data" data block.
222 13. Be more explicit in pcre_study() instead of relying on "default" for
223 opcodes that mean there is no starting character; this means that when new
224 ones are added and accidentally left out of pcre_study(), testing should
225 pick them up.
227 14. The -s option of pcretest has been documented for ages as being an old
228 synonym of -m (show memory usage). I have changed it to mean "force study
229 for every regex", that is, assume /S for every regex. This is similar to -i
230 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
231 using it. It makes it easier to run collections of tests with and without
232 study enabled, and thereby test pcre_study() more easily. All the standard
233 tests are now run with and without -s (but some patterns can be marked as
234 "never study" - see 20 below).
236 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
237 restoration of the capturing data to the outer values was not happening
238 correctly.
240 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
241 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
242 pattern had matched an empty string, and so incorrectly returned a no
243 match.
245 17. There was optimizing code for the last branch of non-capturing parentheses,
246 and also for the obeyed branch of a conditional subexpression, which used
247 tail recursion to cut down on stack usage. Unfortunately, now that there is
248 the possibility of (*THEN) occurring in these branches, tail recursion is
249 no longer possible because the return has to be checked for (*THEN). These
250 two optimizations have therefore been removed. [But see 8.20/11 above.]
252 18. If a pattern containing \R was studied, it was assumed that \R always
253 matched two bytes, thus causing the minimum subject length to be
254 incorrectly computed because \R can also match just one byte.
256 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
257 was incorrectly computed.
259 20. If /S is present twice on a test pattern in pcretest input, it now
260 *disables* studying, thereby overriding the use of -s on the command line
261 (see 14 above). This is necessary for one or two tests to keep the output
262 identical in both cases.
264 21. When (*ACCEPT) was used in an assertion that matched an empty string and
265 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
267 22. When an atomic group that contained a capturing parenthesis was
268 successfully matched, but the branch in which it appeared failed, the
269 capturing was not being forgotten if a higher numbered group was later
270 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
271 group 1 to "a", when in fact it should be unset. This applied to multi-
272 branched capturing and non-capturing groups, repeated or not, and also to
273 positive assertions (capturing in negative assertions does not happen
274 in PCRE) and also to nested atomic groups.
276 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
277 subject after a captured substring, to make it easier to tell which of a
278 number of identical substrings has been captured.
280 24. The way atomic groups are processed by pcre_exec() has been changed so that
281 if they are repeated, backtracking one repetition now resets captured
282 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
283 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
284 "aaa". Previously, it would have been "a". As part of this code
285 refactoring, the way recursive calls are handled has also been changed.
287 25. If an assertion condition captured any substrings, they were not passed
288 back unless some other capturing happened later. For example, if
289 (?(?=(a))a) was matched against "a", no capturing was returned.
291 26. When studying a pattern that contained subroutine calls or assertions,
292 the code for finding the minimum length of a possible match was handling
293 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
294 group 1 called group 2 while simultaneously a separate group 2 called group
295 1). A stack overflow occurred in this case. I have fixed this by limiting
296 the recursion depth to 10.
298 27. Updated RunTest.bat in the distribution to the version supplied by Tom
299 Fortmann. This supports explicit test numbers on the command line, and has
300 argument validation and error reporting.
302 28. An instance of \X with an unlimited repeat could fail if at any point the
303 first character it looked at was a mark character.
305 29. Some minor code refactoring concerning Unicode properties and scripts
306 should reduce the stack requirement of match() slightly.
308 30. Added the '=' option to pcretest to check the setting of unused capturing
309 slots at the end of the pattern, which are documented as being -1, but are
310 not included in the return count.
312 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
313 compiled something random. Now it gives a compile-time error (as does
314 Perl).
316 32. A *MARK encountered during the processing of a positive assertion is now
317 recorded and passed back (compatible with Perl).
319 33. If --only-matching or --colour was set on a pcregrep call whose pattern
320 had alternative anchored branches, the search for a second match in a line
321 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
322 matched the line "0102" twice. The same bug affected patterns that started
323 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
324 twice.
326 34. Previously, PCRE did not allow quantification of assertions. However, Perl
327 does, and because of capturing effects, quantifying parenthesized
328 assertions may at times be useful. Quantifiers are now allowed for
329 parenthesized assertions.
331 35. A minor code tidy in pcre_compile() when checking options for \R usage.
333 36. \g was being checked for fancy things in a character class, when it should
334 just be a literal "g".
336 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
337 appearance of a nested POSIX class supersedes an apparent external class.
338 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
339 unescaped square brackets may also appear as part of class names. For
340 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
341 more like Perl. (But see 8.20/1 above.)
343 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
344 was because it thought it was \N{name}, which is not supported).
346 39. Add minix to OS list not supporting the -S option in pcretest.
348 40. PCRE tries to detect cases of infinite recursion at compile time, but it
349 cannot analyze patterns in sufficient detail to catch mutual recursions
350 such as ((?1))((?2)). There is now a runtime test that gives an error if a
351 subgroup is called recursively as a subpattern for a second time at the
352 same position in the subject string. In previous releases this might have
353 been caught by the recursion limit, or it might have run out of stack.
355 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
356 happen only once. PCRE was, however incorrectly giving a compile time error
357 "recursive call could loop indefinitely" because it cannot analyze the
358 pattern in sufficient detail. The compile time test no longer happens when
359 PCRE is compiling a conditional subpattern, but actual runaway loops are
360 now caught at runtime (see 40 above).
362 42. It seems that Perl allows any characters other than a closing parenthesis
363 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
364 has been changed to be the same.
366 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
367 as not to get warnings when autogen.sh is called. Also changed
368 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
370 44. To help people who use pcregrep to scan files containing exceedingly long
371 lines, the following changes have been made:
373 (a) The default value of the buffer size parameter has been increased from
374 8K to 20K. (The actual buffer used is three times this size.)
376 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
377 PCRE is built.
379 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
380 to be set at run time.
382 (d) Numerical values in pcregrep options can be followed by K or M, for
383 example --buffer-size=50K.
385 (e) If a line being scanned overflows pcregrep's buffer, an error is now
386 given and the return code is set to 2.
388 45. Add a pointer to the latest mark to the callout data block.
390 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
391 partial match of an empty string instead of no match. This was specific to
392 the use of ".".
394 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
395 complete match instead of a partial match. This bug was dependent on both
396 the PCRE_UTF8 and PCRE_DOTALL options being set.
398 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
399 starting byte set, because \b was not being ignored.
402 Version 8.12 15-Jan-2011
403 ------------------------
405 1. Fixed some typos in the markup of the man pages, and wrote a script that
406 checks for such things as part of the documentation building process.
408 2. On a big-endian 64-bit system, pcregrep did not correctly process the
409 --match-limit and --recursion-limit options (added for 8.11). In
410 particular, this made one of the standard tests fail. (The integer value
411 went into the wrong half of a long int.)
413 3. If the --colour option was given to pcregrep with -v (invert match), it
414 did strange things, either producing crazy output, or crashing. It should,
415 of course, ignore a request for colour when reporting lines that do not
416 match.
418 4. Another pcregrep bug caused similar problems if --colour was specified with
419 -M (multiline) and the pattern match finished with a line ending.
421 5. In pcregrep, when a pattern that ended with a literal newline sequence was
422 matched in multiline mode, the following line was shown as part of the
423 match. This seems wrong, so I have changed it.
425 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
426 the check for further matches in the same line (so they could be coloured)
427 to overrun the end of the current line. If another match was found, it was
428 incorrectly shown (and then shown again when found in the next line).
430 7. If pcregrep was compiled under Windows, there was a reference to the
431 function pcregrep_exit() before it was defined. I am assuming this was
432 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
433 reported by a user. I've moved the definition above the reference.
436 Version 8.11 10-Dec-2010
437 ------------------------
439 1. (*THEN) was not working properly if there were untried alternatives prior
440 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
441 backtracked to try for "b" instead of moving to the next alternative branch
442 at the same level (in this case, to look for "c"). The Perl documentation
443 is clear that when (*THEN) is backtracked onto, it goes to the "next
444 alternative in the innermost enclosing group".
446 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
447 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
448 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
449 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
450 (*THEN).
452 3. If \s appeared in a character class, it removed the VT character from
453 the class, even if it had been included by some previous item, for example
454 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
455 of \s, but is part of the POSIX "space" class.)
457 4. A partial match never returns an empty string (because you can always
458 match an empty string at the end of the subject); however the checking for
459 an empty string was starting at the "start of match" point. This has been
460 changed to the "earliest inspected character" point, because the returned
461 data for a partial match starts at this character. This means that, for
462 example, /(?<=abc)def/ gives a partial match for the subject "abc"
463 (previously it gave "no match").
465 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
466 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
467 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
468 has an implication that the given string is incomplete (because a partial
469 match is preferred over a full match). For this reason, these items now
470 give a partial match in this situation. [Aside: previously, the one case
471 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
472 match rather than a full match, which was wrong by the old rules, but is
473 now correct.]
475 6. There was a bug in the handling of #-introduced comments, recognized when
476 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
477 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
478 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
479 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
480 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
481 places in pcre_compile().
483 7. Related to (6) above, when pcre_compile() was skipping #-introduced
484 comments when looking ahead for named forward references to subpatterns,
485 the only newline sequence it recognized was NL. It now handles newlines
486 according to the set newline convention.
488 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
489 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
490 cater for a lack of strerror(). These oversights have been fixed.
492 9. Added --match-limit and --recursion-limit to pcregrep.
494 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
496 11. When the -o option was used, pcregrep was setting a return code of 1, even
497 when matches were found, and --line-buffered was not being honoured.
499 12. Added an optional parentheses number to the -o and --only-matching options
500 of pcregrep.
502 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
503 can match an empty string. The code to do it in pcretest and pcredemo
504 needed fixing:
506 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
507 only one byte after an empty string match just before CRLF (this case
508 just got forgotten; "any" and "anycrlf" were OK).
510 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
511 mode when an empty string match preceded an ASCII character followed by
512 a non-ASCII character. (The code for advancing by one character rather
513 than one byte was nonsense.)
515 (c) The pcredemo.c sample program did not have any code at all to handle
516 the cases when CRLF is a valid newline sequence.
518 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
519 as a starting offset was within the subject string. There is now a new
520 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
521 negative or greater than the length of the string. In order to test this,
522 pcretest is extended to allow the setting of negative starting offsets.
524 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
525 starting offset points to the beginning of a UTF-8 character was
526 unnecessarily clumsy. I tidied it up.
528 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
529 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
531 17. Nobody had reported that the --include_dir option, which was added in
532 release 7.7 should have been called --include-dir (hyphen, not underscore)
533 for compatibility with GNU grep. I have changed it to --include-dir, but
534 left --include_dir as an undocumented synonym, and the same for
535 --exclude-dir, though that is not available in GNU grep, at least as of
536 release 2.5.4.
538 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
539 characters from a string of bytes) have been redefined so as not to use
540 loops, in order to improve performance in some environments. At the same
541 time, I abstracted some of the common code into auxiliary macros to save
542 repetition (this should not affect the compiled code).
544 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
545 compile-time error is now given if \c is not followed by an ASCII
546 character, that is, a byte less than 128. (In EBCDIC mode, the code is
547 different, and any byte value is allowed.)
549 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
550 START_OPTIMIZE option, which is now allowed at compile time - but just
551 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
552 to pcregrep and other applications that have no direct access to PCRE
553 options. The new /Y option in pcretest sets this option when calling
554 pcre_compile().
556 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
557 back references. Groups containing recursive back references were forced to
558 be atomic by that change, but in the case of named groups, the amount of
559 memory required was incorrectly computed, leading to "Failed: internal
560 error: code overflow". This has been fixed.
562 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
563 pcretest.c, to avoid build problems in some Borland environments.
566 Version 8.10 25-Jun-2010
567 ------------------------
569 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
570 THEN.
572 2. (*ACCEPT) was not working when inside an atomic group.
574 3. Inside a character class, \B is treated as a literal by default, but
575 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
576 causes the error). The code is unchanged, but I tidied the documentation.
578 4. Inside a character class, PCRE always treated \R and \X as literals,
579 whereas Perl faults them if its -w option is set. I have changed PCRE so
580 that it faults them when PCRE_EXTRA is set.
582 5. Added support for \N, which always matches any character other than
583 newline. (It is the same as "." when PCRE_DOTALL is not set.)
585 6. When compiling pcregrep with newer versions of gcc which may have
586 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
587 declared with attribute warn_unused_result" were given. Just casting the
588 result to (void) does not stop the warnings; a more elaborate fudge is
589 needed. I've used a macro to implement this.
591 7. Minor change to pcretest.c to avoid a compiler warning.
593 8. Added four artifical Unicode properties to help with an option to make
594 \s etc use properties (see next item). The new properties are: Xan
595 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
597 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
598 use Unicode properties. (*UCP) at the start of a pattern can be used to set
599 this option. Modified pcretest to add /W to test this facility. Added
600 REG_UCP to make it available via the POSIX interface.
602 10. Added --line-buffered to pcregrep.
604 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
605 studied, and the match started with a letter with a code point greater than
606 127 whose first byte was different to the first byte of the other case of
607 the letter, the other case of this starting letter was not recognized
608 (#976).
610 12. If a pattern that was studied started with a repeated Unicode property
611 test, for example, \p{Nd}+, there was the theoretical possibility of
612 setting up an incorrect bitmap of starting bytes, but fortunately it could
613 not have actually happened in practice until change 8 above was made (it
614 added property types that matched character-matching opcodes).
616 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
617 possible starting bytes for non-anchored patterns.
619 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
620 \R, and also a number of cases that involve Unicode properties, both
621 explicit and implicit when PCRE_UCP is set.
623 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
624 input, it could crash or give wrong results if characters with values
625 greater than 0xc0 were present in the subject string. (Detail: it assumed
626 UTF-8 input when processing these items.)
628 16. Added a lot of (int) casts to avoid compiler warnings in systems where
629 size_t is 64-bit (#991).
631 17. Added a check for running out of memory when PCRE is compiled with
632 --disable-stack-for-recursion (#990).
634 18. If the last data line in a file for pcretest does not have a newline on
635 the end, a newline was missing in the output.
637 19. The default pcre_chartables.c file recognizes only ASCII characters (values
638 less than 128) in its various bitmaps. However, there is a facility for
639 generating tables according to the current locale when PCRE is compiled. It
640 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
641 characters, are recognized by isspace() and therefore were getting set in
642 these tables, and indeed these tables seem to approximate to ISO 8859. This
643 caused a problem in UTF-8 mode when pcre_study() was used to create a list
644 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
645 which of course cannot start UTF-8 characters. I have changed the code so
646 that only real ASCII characters (less than 128) and the correct starting
647 bytes for UTF-8 encodings are set for characters greater than 127 when in
648 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
649 altogether.)
651 20. Added the /T option to pcretest so as to be able to run tests with non-
652 standard character tables, thus making it possible to include the tests
653 used for 19 above in the standard set of tests.
655 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
656 reference to a subpattern the other side of a comment that contains an
657 opening parenthesis caused either an internal compiling error, or a
658 reference to the wrong subpattern.
661 Version 8.02 19-Mar-2010
662 ------------------------
664 1. The Unicode data tables have been updated to Unicode 5.2.0.
666 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
667 configured.
669 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
670 original author of that file, following a query about its status.
672 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
673 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
675 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
676 quantifier applied to a forward-referencing subroutine call, could compile
677 incorrect code or give the error "internal error: previously-checked
678 referenced subpattern not found".
680 6. Both MS Visual Studio and Symbian OS have problems with initializing
681 variables to point to external functions. For these systems, therefore,
682 pcre_malloc etc. are now initialized to local functions that call the
683 relevant global functions.
685 7. There were two entries missing in the vectors called coptable and poptable
686 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
687 I've fixed the data, and added a kludgy way of testing at compile time that
688 the lengths are correct (equal to the number of opcodes).
690 8. Following on from 7, I added a similar kludge to check the length of the
691 eint vector in pcreposix.c.
693 9. Error texts for pcre_compile() are held as one long string to avoid too
694 much relocation at load time. To find a text, the string is searched,
695 counting zeros. There was no check for running off the end of the string,
696 which could happen if a new error number was added without updating the
697 string.
699 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
701 11. \K was not working if it appeared in an atomic group or in a group that
702 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
703 \K is "not well defined" if used in an assertion. PCRE now accepts it if
704 the assertion is positive, but not if it is negative.
706 12. Change 11 fortuitously reduced the size of the stack frame used in the
707 "match()" function of pcre_exec.c by one pointer. Forthcoming
708 implementation of support for (*MARK) will need an extra pointer on the
709 stack; I have reserved it now, so that the stack frame size does not
710 decrease.
712 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
713 item in branch that calls a recursion is a subroutine call - as in the
714 second branch in the above example - was incorrectly given the compile-
715 time error "recursive call could loop indefinitely" because pcre_compile()
716 was not correctly checking the subroutine for matching a non-empty string.
718 14. The checks for overrunning compiling workspace could trigger after an
719 overrun had occurred. This is a "should never occur" error, but it can be
720 triggered by pathological patterns such as hundreds of nested parentheses.
721 The checks now trigger 100 bytes before the end of the workspace.
723 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
726 Version 8.01 19-Jan-2010
727 ------------------------
729 1. If a pattern contained a conditional subpattern with only one branch (in
730 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
731 computed the wrong minimum data length (which is of course zero for such
732 subpatterns). This could cause incorrect "no match" results.
734 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
735 the pattern is reset in the first branch, pcre_compile() failed with
736 "internal error: code overflow at offset...". This happened only when
737 the reset was to the original external option setting. (An optimization
738 abstracts leading options settings into an external setting, which was the
739 cause of this.)
741 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
742 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
743 assertion pattern did not match (meaning that the assertion was true), it
744 was incorrectly treated as false if the SKIP had been reached during the
745 matching. This also applied to assertions used as conditions.
747 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
748 assertion subpattern, including such a pattern used as a condition,
749 unpredictable results occurred, instead of the error return
752 5. The C++ GlobalReplace function was not working like Perl for the special
753 situation when an empty string is matched. It now does the fancy magic
754 stuff that is necessary.
756 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
757 removed. (These were left over from very, very early versions of PCRE.)
759 7. Some cosmetic changes to the code to make life easier when compiling it
760 as part of something else:
762 (a) Change DEBUG to PCRE_DEBUG.
764 (b) In pcre_compile(), rename the member of the "branch_chain" structure
765 called "current" as "current_branch", to prevent a collision with the
766 Linux macro when compiled as a kernel module.
768 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
769 prevent a collision with the Linux macro when compiled as a kernel
770 module.
772 8. In pcre_compile() there are some checks for integer overflows that used to
773 cast potentially large values to (double). This has been changed to that
774 when building, a check for int64_t is made, and if it is found, it is used
775 instead, thus avoiding the use of floating point arithmetic. (There is no
776 other use of FP in PCRE.) If int64_t is not found, the fallback is to
777 double.
779 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
780 2005 (difference between two addresses compared to an unsigned value).
782 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
783 custom one, because of the following reported problem in Windows:
785 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
786 under Win32.
787 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
788 therefore missing the function definition.
789 - The compiler thus generates a "C" signature for the test function.
790 - The linker fails to find the "C" function.
791 - PCRE fails to configure if asked to do so against libbz2.
793 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
794 messages were output:
796 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
797 rerunning libtoolize, to keep the correct libtool macros in-tree.
798 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
800 I have done both of these things.
802 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
803 most of the time, it *can* run out if it is given a pattern that contains a
804 runaway infinite recursion. I updated the discussion in the pcrestack man
805 page.
807 13. Now that we have gone to the x.xx style of version numbers, the minor
808 version may start with zero. Using 08 or 09 is a bad idea because users
809 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
810 interpreted as invalid octal numbers. I've updated the previous comment in
811 configure.ac, and also added a check that gives an error if 08 or 09 are
812 used.
814 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
815 causing partial matching to fail when the end of the subject matched \W
816 in a UTF-8 pattern where \W was quantified with a minimum of 3.
818 15. There were some discrepancies between the declarations in pcre_internal.h
819 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
820 their definitions. The declarations used "const uschar *" and the
821 definitions used USPTR. Even though USPTR is normally defined as "const
822 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
823 reported that: "This difference in casting confuses some C++ compilers, for
824 example, SunCC recognizes above declarations as different functions and
825 generates broken code for hbpcre." I have changed the declarations to use
826 USPTR.
828 16. GNU libtool is named differently on some systems. The autogen.sh script now
829 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
830 (FreeBSD).
832 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
833 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
834 comment: "Figure out how to create a longlong from a string: strtoll and
835 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
836 instance, but it only takes 2 args instead of 3!"
838 18. A subtle bug concerned with back references has been fixed by a change of
839 specification, with a corresponding code fix. A pattern such as
840 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
841 refers, was giving matches when it shouldn't. For example, xa=xaaa would
842 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
843 same bug. Such groups have to be quantified to be useful, or contained
844 inside another quantified group. (If there's no repetition, the reference
845 can never match.) The problem arises because, having left the group and
846 moved on to the rest of the pattern, a later failure that backtracks into
847 the group uses the captured value from the final iteration of the group
848 rather than the correct earlier one. I have fixed this in PCRE by forcing
849 any group that contains a reference to itself to be an atomic group; that
850 is, there cannot be any backtracking into it once it has completed. This is
851 similar to recursive and subroutine calls.
854 Version 8.00 19-Oct-09
855 ----------------------
857 1. The table for translating pcre_compile() error codes into POSIX error codes
858 was out-of-date, and there was no check on the pcre_compile() error code
859 being within the table. This could lead to an OK return being given in
860 error.
862 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
863 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
864 in a Windows environment.
866 3. The pcregrep --count option prints the count for each file even when it is
867 zero, as does GNU grep. However, pcregrep was also printing all files when
868 --files-with-matches was added. Now, when both options are given, it prints
869 counts only for those files that have at least one match. (GNU grep just
870 prints the file name in this circumstance, but including the count seems
871 more useful - otherwise, why use --count?) Also ensured that the
872 combination -clh just lists non-zero counts, with no names.
874 4. The long form of the pcregrep -F option was incorrectly implemented as
875 --fixed_strings instead of --fixed-strings. This is an incompatible change,
876 but it seems right to fix it, and I didn't think it was worth preserving
877 the old behaviour.
879 5. The command line items --regex=pattern and --regexp=pattern were not
880 recognized by pcregrep, which required --regex pattern or --regexp pattern
881 (with a space rather than an '='). The man page documented the '=' forms,
882 which are compatible with GNU grep; these now work.
884 6. No libpcreposix.pc file was created for pkg-config; there was just
885 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
887 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
888 when UCP support is not needed, by modifying the Python script that
889 generates it from Unicode data files. This should not matter if the module
890 is correctly used as a library, but I received one complaint about 50K of
891 unwanted data. My guess is that the person linked everything into his
892 program rather than using a library. Anyway, it does no harm.
894 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
895 was a minimum greater than 1 for a wide character in a possessive
896 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
897 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
898 character. Chaos in the form of incorrect output or a compiling loop could
899 result.
901 9. The restrictions on what a pattern can contain when partial matching is
902 requested for pcre_exec() have been removed. All patterns can now be
903 partially matched by this function. In addition, if there are at least two
904 slots in the offset vector, the offset of the earliest inspected character
905 for the match and the offset of the end of the subject are set in them when
906 PCRE_ERROR_PARTIAL is returned.
908 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
909 synonymous with PCRE_PARTIAL, for backwards compatibility, and
910 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
911 and may be more useful for multi-segment matching.
913 11. Partial matching with pcre_exec() is now more intuitive. A partial match
914 used to be given if ever the end of the subject was reached; now it is
915 given only if matching could not proceed because another character was
916 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
917 string "Z", which now yields "no match" instead of "partial match". In the
918 case of pcre_dfa_exec(), "no match" is given if every matching path for the
919 final character ended with (*FAIL).
921 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
922 if the pattern had a "must contain" character that was already found in the
923 earlier partial match, unless partial matching was again requested. For
924 example, with the pattern /dog.(body)?/, the "must contain" character is
925 "g". If the first part-match was for the string "dog", restarting with
926 "sbody" failed. This bug has been fixed.
928 13. The string returned by pcre_dfa_exec() after a partial match has been
929 changed so that it starts at the first inspected character rather than the
930 first character of the match. This makes a difference only if the pattern
931 starts with a lookbehind assertion or \b or \B (\K is not supported by
932 pcre_dfa_exec()). It's an incompatible change, but it makes the two
933 matching functions compatible, and I think it's the right thing to do.
935 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
936 so that the demonstration program is easily available in environments where
937 PCRE has not been installed from source.
939 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
940 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
941 library.
943 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
944 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
945 is not the first non-POSIX option to be added. Clearly some people find
946 these options useful.
948 17. If a caller to the POSIX matching function regexec() passes a non-zero
949 value for nmatch with a NULL value for pmatch, the value of
950 nmatch is forced to zero.
952 18. RunGrepTest did not have a test for the availability of the -u option of
953 the diff command, as RunTest does. It now checks in the same way as
954 RunTest, and also checks for the -b option.
956 19. If an odd number of negated classes containing just a single character
957 interposed, within parentheses, between a forward reference to a named
958 subpattern and the definition of the subpattern, compilation crashed with
959 an internal error, complaining that it could not find the referenced
960 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
961 [The bug was that it was starting one character too far in when skipping
962 over the character class, thus treating the ] as data rather than
963 terminating the class. This meant it could skip too much.]
965 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
966 /g option in pcretest when the pattern contains \K, which makes it possible
967 to have an empty string match not at the start, even when the pattern is
968 anchored. Updated pcretest and pcredemo to use this option.
970 21. If the maximum number of capturing subpatterns in a recursion was greater
971 than the maximum at the outer level, the higher number was returned, but
972 with unset values at the outer level. The correct (outer level) value is
973 now given.
975 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
976 PCRE did not set those parentheses (unlike Perl). I have now found a way to
977 make it do so. The string so far is captured, making this feature
978 compatible with Perl.
980 23. The tests have been re-organized, adding tests 11 and 12, to make it
981 possible to check the Perl 5.10 features against Perl 5.10.
983 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
984 pattern matches a fixed length string. PCRE did not allow this; now it
985 does. Neither allows recursion.
987 25. I finally figured out how to implement a request to provide the minimum
988 length of subject string that was needed in order to match a given pattern.
989 (It was back references and recursion that I had previously got hung up
990 on.) This code has now been added to pcre_study(); it finds a lower bound
991 to the length of subject needed. It is not necessarily the greatest lower
992 bound, but using it to avoid searching strings that are too short does give
993 some useful speed-ups. The value is available to calling programs via
994 pcre_fullinfo().
996 26. While implementing 25, I discovered to my embarrassment that pcretest had
997 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
998 study optimizations had never been tested with that matching function.
999 Oops. What is worse, even when it was passed study data, there was a bug in
1000 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1001 were also very few tests of studied patterns with pcre_dfa_exec().
1003 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1004 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1005 on the other side of the coin, they are no longer allowed to have different
1006 names, because these cannot be distinguished in PCRE, and this has caused
1007 confusion. (This is a difference from Perl.)
1009 28. When duplicate subpattern names are present (necessarily with different
1010 numbers, as required by 27 above), and a test is made by name in a
1011 conditional pattern, either for a subpattern having been matched, or for
1012 recursion in such a pattern, all the associated numbered subpatterns are
1013 tested, and the overall condition is true if the condition is true for any
1014 one of them. This is the way Perl works, and is also more like the way
1015 testing by number works.
1018 Version 7.9 11-Apr-09
1019 ---------------------
1021 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1022 (pcretest), all targets were linked against these libraries. This included
1023 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1024 libraries. This caused unwanted dependencies to be created. This problem
1025 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1026 pcretest is linked with readline.
1028 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1029 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1030 moved outside it again, because FALSE and TRUE are already defined in AIX,
1031 but BOOL is not.
1033 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1034 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1036 4. The pcregrep documentation said spaces were inserted as well as colons (or
1037 hyphens) following file names and line numbers when outputting matching
1038 lines. This is not true; no spaces are inserted. I have also clarified the
1039 wording for the --colour (or --color) option.
1041 5. In pcregrep, when --colour was used with -o, the list of matching strings
1042 was not coloured; this is different to GNU grep, so I have changed it to be
1043 the same.
1045 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1046 each matching line was coloured. Now it goes on to look for further matches
1047 of any of the test patterns, which is the same behaviour as GNU grep.
1049 7. A pattern that could match an empty string could cause pcregrep to loop; it
1050 doesn't make sense to accept an empty string match in pcregrep, so I have
1051 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1052 seems to be how GNU grep behaves.
1054 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1055 start or after a newline", because the conditional assertion was not being
1056 correctly handled. The rule now is that both the assertion and what follows
1057 in the first alternative must satisfy the test.
1059 9. If auto-callout was enabled in a pattern with a conditional group whose
1060 condition was an assertion, PCRE could crash during matching, both with
1061 pcre_exec() and pcre_dfa_exec().
1063 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1064 used for matching.
1066 11. Unicode property support in character classes was not working for
1067 characters (bytes) greater than 127 when not in UTF-8 mode.
1069 12. Added the -M command line option to pcretest.
1071 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1073 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1075 16. Added comments and documentation about mis-use of no_arg in the C++
1076 wrapper.
1078 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1079 from Martin Jerabek that uses macro names for all relevant character and
1080 string constants.
1082 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1083 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1084 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1085 these, but not everybody uses configure.
1087 19. A conditional group that had only one branch was not being correctly
1088 recognized as an item that could match an empty string. This meant that an
1089 enclosing group might also not be so recognized, causing infinite looping
1090 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1091 with the subject "ab", where knowledge that the repeated group can match
1092 nothing is needed in order to break the loop.
1094 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1095 exec(), but without supplying a callout function, matching went wrong.
1097 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1098 leak if the size of the offset vector was greater than 30. When the vector
1099 is smaller, the saved offsets during recursion go onto a local stack
1100 vector, but for larger vectors malloc() is used. It was failing to free
1101 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1102 error, in fact).
1104 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1105 heapframe that is used only when UTF-8 support is enabled. This caused no
1106 problem, but was untidy.
1108 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1109 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1110 included within another project.
1112 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1113 slightly modified by me:
1115 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1116 not building pcregrep.
1118 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1119 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1121 25. Forward references, both numeric and by name, in patterns that made use of
1122 duplicate group numbers, could behave incorrectly or give incorrect errors,
1123 because when scanning forward to find the reference group, PCRE was not
1124 taking into account the duplicate group numbers. A pattern such as
1125 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1127 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1128 the feature of a custom pointer more persuasive (as requested by a user).
1130 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1131 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1132 pre-defined.
1134 28. Added support for (*UTF8) at the start of a pattern.
1136 29. Arrange for flags added by the "release type" setting in CMake to be shown
1137 in the configuration summary.
1140 Version 7.8 05-Sep-08
1141 ---------------------
1143 1. Replaced UCP searching code with optimized version as implemented for Ad
1144 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1145 stage table and inline lookup instead of a function, giving speed ups of 2
1146 to 5 times on some simple patterns that I tested. Permission was given to
1147 distribute the MultiStage2.py script that generates the tables (it's not in
1148 the tarball, but is in the Subversion repository).
1150 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1151 scripts.
1153 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1154 a group with a zero qualifier. The result of the study could be incorrect,
1155 or the function might crash, depending on the pattern.
1157 4. Caseless matching was not working for non-ASCII characters in back
1158 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1159 It now works when Unicode Property Support is available.
1161 5. In pcretest, an escape such as \x{de} in the data was always generating
1162 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1163 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1164 truncation.
1166 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1168 7. Added two (int) casts to pcregrep when printing the difference of two
1169 pointers, in case they are 64-bit values.
1171 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1172 test 2 if it fails.
1174 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1175 and a #define of that name to empty if it is not externally set. This is to
1176 allow users of MSVC to set it if necessary.
1178 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1179 the convenience functions in the pcre_get.c source file.
1181 11. An option change at the start of a pattern that had top-level alternatives
1182 could cause overwriting and/or a crash. This command provoked a crash in
1183 some environments:
1185 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1187 This potential security problem was recorded as CVE-2008-2371.
1189 12. For a pattern where the match had to start at the beginning or immediately
1190 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1191 pcre_dfa_exec() could read past the end of the passed subject if there was
1192 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1193 pcretest so that it places the subject at the end of its malloc-ed buffer.
1195 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1196 exec() might read past the end of the data buffer in UTF-8 mode.
1198 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1199 the data contained the byte 0x85 as part of a UTF-8 character within its
1200 first line. This applied both to normal and DFA matching.
1202 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1203 /^[^d]*?$/8 failed to match "abc".
1205 16. Added a missing copyright notice to pcrecpp_internal.h.
1207 17. Make it more clear in the documentation that values returned from
1208 pcre_exec() in ovector are byte offsets, not character counts.
1210 18. Tidied a few places to stop certain compilers from issuing warnings.
1212 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1213 supplied by Stefan Weber. I made a further small update for 7.8 because
1214 there is a change of source arrangements: the pcre_searchfuncs.c module is
1215 replaced by pcre_ucd.c.
1218 Version 7.7 07-May-08
1219 ---------------------
1221 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1222 a string to a long long, pretend we don't even have a long long." This is
1223 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1225 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1226 pre-7.6 versions, which defined a global no_arg variable instead of putting
1227 it in the RE class. (See also #8 below.)
1229 3. Remove a line of dead code, identified by coverity and reported by Nuno
1230 Lopes.
1232 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1234 (1) The include/exclude patterns were being applied to the whole pathnames
1235 of files, instead of just to the final components.
1237 (2) If there was more than one level of directory, the subdirectories were
1238 skipped unless they satisfied the include/exclude conditions. This is
1239 inconsistent with GNU grep (and could even be seen as contrary to the
1240 pcregrep specification - which I improved to make it absolutely clear).
1241 The action now is always to scan all levels of directory, and just
1242 apply the include/exclude patterns to regular files.
1244 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1245 --exclude_dir in the tests to avoid scanning .svn directories.
1247 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1248 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1249 doesn't support NULs in patterns.
1251 7. Added some missing "const"s to declarations of static tables in
1252 pcre_compile.c and pcre_dfa_exec.c.
1254 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1255 caused by fix #2 above. (Subsequently also a second patch to fix the
1256 first patch. And a third patch - this was a messy problem.)
1258 9. Applied Craig's patch to remove the use of push_back().
1260 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1261 matching function regexec().
1263 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1264 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1265 references. PCRE supports relative numbers with this syntax (I don't think
1266 Oniguruma does).
1268 12. Previously, a group with a zero repeat such as (...){0} was completely
1269 omitted from the compiled regex. However, this means that if the group
1270 was called as a subroutine from elsewhere in the pattern, things went wrong
1271 (an internal error was given). Such groups are now left in the compiled
1272 pattern, with a new opcode that causes them to be skipped at execution
1273 time.
1275 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1276 to the way PCRE behaves:
1278 (a) A lone ] character is dis-allowed (Perl treats it as data).
1280 (b) A back reference to an unmatched subpattern matches an empty string
1281 (Perl fails the current match path).
1283 (c) A data ] in a character class must be notated as \] because if the
1284 first data character in a class is ], it defines an empty class. (In
1285 Perl it is not possible to have an empty class.) The empty class []
1286 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1287 The negative empty class [^] matches any one character, independently
1288 of the DOTALL setting.
1290 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1291 non-existent subpattern following a character class starting with ']' and
1292 containing () gave an internal compiling error instead of "reference to
1293 non-existent subpattern". Fortunately, when the pattern did exist, the
1294 compiled code was correct. (When scanning forwards to check for the
1295 existencd of the subpattern, it was treating the data ']' as terminating
1296 the class, so got the count wrong. When actually compiling, the reference
1297 was subsequently set up correctly.)
1299 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1300 it was being rejected as not supported by pcre_dfa_exec(), even though
1301 other assertions are supported. I have made pcre_dfa_exec() support
1302 (*FAIL).
1304 16. The implementation of 13c above involved the invention of a new opcode,
1305 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1306 cannot be changed at match time, I realized I could make a small
1307 improvement to matching performance by compiling OP_ALLANY instead of
1308 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1309 on the OP_ANY path.
1311 17. Compiling pcretest on Windows with readline support failed without the
1312 following two fixes: (1) Make the unistd.h include conditional on
1313 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1315 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1316 ncurses library to be included for pcretest when ReadLine support is
1317 requested, but also to allow for it to be overridden. This patch came from
1318 Daniel Bergström.
1320 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1321 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1322 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1323 spotting this.
1326 Version 7.6 28-Jan-08
1327 ---------------------
1329 1. A character class containing a very large number of characters with
1330 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1331 overflow.
1333 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1334 HAVE_LONG_LONG is not defined.
1336 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1337 bring them up to date and include new features. This patch includes:
1339 - Fixed PH's badly added libz and libbz2 support.
1340 - Fixed a problem with static linking.
1341 - Added pcredemo. [But later removed - see 7 below.]
1342 - Fixed dftables problem and added an option.
1343 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1345 - Added readline support for pcretest.
1346 - Added an listing of the option settings after cmake has run.
1348 4. A user submitted a patch to Makefile that makes it easy to create
1349 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1350 Makefile.am that cause it to include this special target, without
1351 affecting anything else. Note that the same mingw target plus all
1352 the other distribution libraries and programs are now supported
1353 when configuring with CMake (see 6 below) instead of with
1354 Configure/Make.
1356 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1357 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1358 exported in the Windows port". It has not yet been confirmed that the patch
1359 solves the problem, but it does no harm.
1361 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1362 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1363 with CMake, and also correct the comment about stack recursion.
1365 7. Remove the automatic building of pcredemo from the ./configure system and
1366 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1367 of a program that users should build themselves after PCRE is installed, so
1368 building it automatically is not really right. What is more, it gave
1369 trouble in some build environments.
1371 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1374 Version 7.5 10-Jan-08
1375 ---------------------
1377 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1378 values in parens when parsing an RE using the C++ wrapper."
1380 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1381 Characters greater than 255 were excluded from the class instead of being
1382 included.
1384 3. The same bug as (2) above applied to negated POSIX classes such as
1385 [:^space:].
1387 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1388 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1389 I have changed it.
1391 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1392 first named subpattern) and a construct such as (?&a) would reference the
1393 first named subpattern whose name started with "a" (in other words, the
1394 length check was missing). Both these problems are fixed. "Subpattern name
1395 expected" is now given for (?&) (a zero-length name), and this patch also
1396 makes it give the same error for \k'' (previously it complained that that
1397 was a reference to a non-existent subpattern).
1399 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1400 this is right because (?- can be followed by option settings as well as by
1401 digits. I have, however, made the messages clearer.
1403 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1404 than the number used in the conditional) now cause a compile-time error.
1405 This is actually not compatible with Perl, which accepts such patterns, but
1406 treats the conditional as always being FALSE (as PCRE used to), but it
1407 seems to me that giving a diagnostic is better.
1409 8. Change "alphameric" to the more common word "alphanumeric" in comments
1410 and messages.
1412 9. Fix two occurrences of "backslash" in comments that should have been
1413 "backspace".
1415 10. Remove two redundant lines of code that can never be obeyed (their function
1416 was moved elsewhere).
1418 11. The program that makes PCRE's Unicode character property table had a bug
1419 which caused it to generate incorrect table entries for sequences of
1420 characters that have the same character type, but are in different scripts.
1421 It amalgamated them into a single range, with the script of the first of
1422 them. In other words, some characters were in the wrong script. There were
1423 thirteen such cases, affecting characters in the following ranges:
1425 U+002b0 - U+002c1
1426 U+0060c - U+0060d
1427 U+0061e - U+00612
1428 U+0064b - U+0065e
1429 U+0074d - U+0076d
1430 U+01800 - U+01805
1431 U+01d00 - U+01d77
1432 U+01d9b - U+01dbf
1433 U+0200b - U+0200f
1434 U+030fc - U+030fe
1435 U+03260 - U+0327f
1436 U+0fb46 - U+0fbb1
1437 U+10450 - U+1049d
1439 12. The -o option (show only the matching part of a line) for pcregrep was not
1440 compatible with GNU grep in that, if there was more than one match in a
1441 line, it showed only the first of them. It now behaves in the same way as
1442 GNU grep.
1444 13. If the -o and -v options were combined for pcregrep, it printed a blank
1445 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1446 does the same. The return code can be used to tell if there were any
1447 non-matching lines.
1449 14. Added --file-offsets and --line-offsets to pcregrep.
1451 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1452 infinitely looping recursion. The bug was that positive lookaheads were not
1453 being skipped when checking for a possible empty match (negative lookaheads
1454 and both kinds of lookbehind were skipped).
1456 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1457 inclusion of <windows.h> to before rather than after the definition of
1458 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1460 17. Specifying a possessive quantifier with a specific limit for a Unicode
1461 character property caused pcre_compile() to compile bad code, which led at
1462 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1463 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1464 caused the error; without that there was no problem.
1466 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1468 19. Added --enable-pcretest-libreadline.
1470 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1471 RE::GlobalReplace(). As a result, the number of replacements returned was
1472 double what it should be. I removed one of the increments, but Craig sent a
1473 later patch that removed the other one (the right fix) and added unit tests
1474 that check the return values (which was not done before).
1476 21. Several CMake things:
1478 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1479 the names libpcre and libpcreposix, not just pcre and pcreposix.
1481 (2) The above change means that pcretest and pcregrep are now correctly
1482 linked with the newly-built libraries, not previously installed ones.
1486 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1487 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1488 UTF-8 newline character). The key issue is that the pattern starts .*;
1489 this means that the match must be either at the beginning, or after a
1490 newline. The bug was in the code for advancing after a failed match and
1491 checking that the new position followed a newline. It was not taking
1492 account of UTF-8 characters correctly.
1494 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1495 character classes. PCRE was not treating the sequence [:...:] as a
1496 character class unless the ... were all letters. Perl, however, seems to
1497 allow any characters between [: and :], though of course it rejects as
1498 unknown any "names" that contain non-letters, because all the known class
1499 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1500 for example, whereas PCRE did not - it did not recognize a POSIX character
1501 class. This seemed a bit dangerous, so the code has been changed to be
1502 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1503 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1504 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1505 Perl does, and where it didn't before.
1507 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1508 Windows environments %n is disabled by default.
1511 Version 7.4 21-Sep-07
1512 ---------------------
1514 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1515 means that a class such as [\s] counted as "explicit reference to CR or
1516 LF". That isn't really right - the whole point of the change was to try to
1517 help when there was an actual mention of one of the two characters. So now
1518 the change happens only if \r or \n (or a literal CR or LF) character is
1519 encountered.
1521 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1522 of both had grown to the point where there were only 3 bits left.
1523 Fortunately, there was spare space in the data structure, and so I have
1524 moved the internal flags into a new 16-bit field to free up more option
1525 bits.
1527 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1528 but did not set the internal JCHANGED flag - either of these is enough to
1529 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1530 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1531 start sets both bits.
1533 4. Added options (at build time, compile time, exec time) to change \R from
1534 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1536 5. doc/pcresyntax.html was missing from the distribution.
1538 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1539 compatibility, even though it is no longer used.
1541 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1542 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1543 windows.h file is present (where different names are used). [This was
1544 reversed later after testing - see 16 below.]
1546 8. Changed all #include <config.h> to #include "config.h". There were also
1547 some further <pcre.h> cases that I changed to "pcre.h".
1549 9. When pcregrep was used with the --colour option, it missed the line ending
1550 sequence off the lines that it output.
1552 10. It was pointed out to me that arrays of string pointers cause lots of
1553 relocations when a shared library is dynamically loaded. A technique of
1554 using a single long string with a table of offsets can drastically reduce
1555 these. I have refactored PCRE in four places to do this. The result is
1556 dramatic:
1558 Originally: 290
1559 After changing UCP table: 187
1560 After changing error message table: 43
1561 After changing table of "verbs" 36
1562 After changing table of Posix names 22
1564 Thanks to the folks working on Gregex for glib for this insight.
1566 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1567 unicode-properties was also set.
1569 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1571 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1572 checked only for CRLF.
1574 14. Added casts to pcretest.c to avoid compiler warnings.
1576 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1578 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1579 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1580 entirely. This removes changes made in 7 above.
1582 17. The CMake files have been updated, and there is now more information about
1583 building with CMake in the NON-UNIX-USE document.
1586 Version 7.3 28-Aug-07
1587 ---------------------
1589 1. In the rejigging of the build system that eventually resulted in 7.1, the
1590 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1591 brackets there is not right, since it causes compilers to look for an
1592 installed pcre.h, not the version that is in the source that is being
1593 compiled (which of course may be different). I have changed it back to:
1595 #include "pcre.h"
1597 I have a vague recollection that the change was concerned with compiling in
1598 different directories, but in the new build system, that is taken care of
1599 by the VPATH setting the Makefile.
1601 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1602 when the subject happened to end in the byte 0x85 (e.g. if the last
1603 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1604 characters but of course it shouldn't be taken as a newline when it is part
1605 of another character. The bug was that, for an unlimited repeat of . in
1606 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1607 characters when looking for a newline.
1609 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1611 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1612 in debug output.
1614 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1615 long printing in the pcrecpp unittest when running under MinGW.
1617 6. ESC_K was left out of the EBCDIC table.
1619 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1620 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1621 limit also applies to "virtual nesting" when a pattern is recursive, and in
1622 this case 1000 isn't so big. I have been able to remove this limit at the
1623 expense of backing off one optimization in certain circumstances. Normally,
1624 when pcre_exec() would call its internal match() function recursively and
1625 immediately return the result unconditionally, it uses a "tail recursion"
1626 feature to save stack. However, when a subpattern that can match an empty
1627 string has an unlimited repetition quantifier, it no longer makes this
1628 optimization. That gives it a stack frame in which to save the data for
1629 checking that an empty string has been matched. Previously this was taken
1630 from the 1000-entry workspace that had been reserved. So now there is no
1631 explicit limit, but more stack is used.
1633 8. Applied Daniel's patches to solve problems with the import/export magic
1634 syntax that is required for Windows, and which was going wrong for the
1635 pcreposix and pcrecpp parts of the library. These were overlooked when this
1636 problem was solved for the main library.
1638 9. There were some crude static tests to avoid integer overflow when computing
1639 the size of patterns that contain repeated groups with explicit upper
1640 limits. As the maximum quantifier is 65535, the maximum group length was
1641 set at 30,000 so that the product of these two numbers did not overflow a
1642 32-bit integer. However, it turns out that people want to use groups that
1643 are longer than 30,000 bytes (though not repeat them that many times).
1644 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1645 made it possible to implement the integer overflow checks in a much more
1646 dynamic way, which I have now done. The artificial limitation on group
1647 length has been removed - we now have only the limit on the total length of
1648 the compiled pattern, which depends on the LINK_SIZE setting.
1650 10. Fixed a bug in the documentation for get/copy named substring when
1651 duplicate names are permitted. If none of the named substrings are set, the
1652 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1653 empty string.
1655 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1656 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1657 because the ] is interpreted as the first data character and the
1658 terminating ] is not found. PCRE has been made compatible with Perl in this
1659 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1660 cause memory overwriting.
1662 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1663 string has been matched (to stop an infinite loop). It was not recognizing
1664 a conditional subpattern that could match an empty string if that
1665 subpattern was within another subpattern. For example, it looped when
1666 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1667 condition was not nested. This bug has been fixed.
1669 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1670 past the start of the subject in the presence of bytes with the top bit
1671 set, for example "\x8aBCD".
1673 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1674 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1676 14. Optimized (?!) to (*FAIL).
1678 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1679 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1680 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1681 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1682 does: it's just the validity check that is more restrictive.
1684 16. Inserted checks for integer overflows during escape sequence (backslash)
1685 processing, and also fixed erroneous offset values for syntax errors during
1686 backslash processing.
1688 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1689 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1691 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1692 caused an overrun.
1694 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1695 something other than just ASCII characters) inside a group that had an
1696 unlimited repeat caused a loop at compile time (while checking to see
1697 whether the group could match an empty string).
1699 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1700 [\P{Any}] did so. (Error in the code for printing property names.)
1702 21. An orphan \E inside a character class could cause a crash.
1704 22. A repeated capturing bracket such as (A)? could cause a wild memory
1705 reference during compilation.
1707 23. There are several functions in pcre_compile() that scan along a compiled
1708 expression for various reasons (e.g. to see if it's fixed length for look
1709 behind). There were bugs in these functions when a repeated \p or \P was
1710 present in the pattern. These operators have additional parameters compared
1711 with \d, etc, and these were not being taken into account when moving along
1712 the compiled data. Specifically:
1714 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1715 length.
1717 (b) An item such as \pL+ within a repeated group could cause crashes or
1718 loops.
1720 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1721 "reference to non-existent subpattern" error.
1723 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1725 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1726 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1728 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1730 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1731 character were causing crashes (broken optimization).
1733 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1734 \p or \P) caused a compile-time loop.
1736 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1737 break. For example, the unstudied pattern [\r\n]A does not match the string
1738 "\r\nA" because change 7.0/46 below moves the current point on by two
1739 characters after failing to match at the start. However, the pattern \nA
1740 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1741 the same is true. There doesn't seem any very clean way out of this, but
1742 what I have chosen to do makes the common cases work: PCRE now takes note
1743 of whether there can be an explicit match for \r or \n anywhere in the
1744 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1745 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1746 pattern has explicit CR or LF references.
1748 29. Added (*CR) etc for changing newline setting at start of pattern.
1751 Version 7.2 19-Jun-07
1752 ---------------------
1754 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1755 which is apparently normally available under Windows.
1757 2. Re-jig the pcregrep tests with different newline settings in an attempt
1758 to make them independent of the local environment's newline setting.
1760 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1762 4. Some of the "internals" tests were previously cut out when the link size
1763 was not 2, because the output contained actual offsets. The recent new
1764 "Z" feature of pcretest means that these can be cut out, making the tests
1765 usable with all link sizes.
1767 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1768 stack recursion. This gives a massive performance boost under BSD, but just
1769 a small improvement under Linux. However, it saves one field in the frame
1770 in all cases.
1772 6. Added more features from the forthcoming Perl 5.10:
1774 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1775 recursion call. It refers to the nth most recently opened parentheses.
1777 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1778 to be opened parentheses.
1780 (c) Conditions that refer to capturing parentheses can be specified
1781 relatively, for example, (?(-2)... or (?(+3)...
1783 (d) \K resets the start of the current match so that everything before
1784 is not part of it.
1786 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1788 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1789 reference syntax.
1791 (g) (?| introduces a group in which the numbering of parentheses in each
1792 alternative starts with the same number.
1794 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1796 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1799 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1800 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1801 for detecting groups that can match an empty string.
1803 9. A pattern with a very large number of alternatives (more than several
1804 hundred) was running out of internal workspace during the pre-compile
1805 phase, where pcre_compile() figures out how much memory will be needed. A
1806 bit of new cunning has reduced the workspace needed for groups with
1807 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1808 workspace instead of running out of the 4096 that are available.
1810 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1812 11. Applied patch from Google to remove an optimization that didn't quite work.
1813 The report of the bug said:
1815 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1816 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1817 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1819 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1820 it matched the wrong number of bytes.
1823 Version 7.1 24-Apr-07
1824 ---------------------
1826 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1827 that is more "standard", making use of automake and other Autotools. There
1828 is some re-arrangement of the files and adjustment of comments consequent
1829 on this.
1831 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1832 for recursive directory scanning broke on some systems because the files
1833 are not scanned in any specific order and on different systems the order
1834 was different. A call to "sort" has been inserted into RunGrepTest for the
1835 approprate test as a short-term fix. In the longer term there may be an
1836 alternative.
1838 3. I had an email from Eric Raymond about problems translating some of PCRE's
1839 man pages to HTML (despite the fact that I distribute HTML pages, some
1840 people do their own conversions for various reasons). The problems
1841 concerned the use of low-level troff macros .br and .in. I have therefore
1842 removed all such uses from the man pages (some were redundant, some could
1843 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1844 HTML has been updated to handle .nf/.fi and to complain if it encounters
1845 .br or .in.
1847 4. Updated comments in configure.ac that get placed in config.h.in and also
1848 arranged for config.h to be included in the distribution, with the name
1849 config.h.generic, for the benefit of those who have to compile without
1850 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1852 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1853 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1854 makevp.bat for the current PCRE, using the additional files
1855 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1857 6. A Windows user reported a minor discrepancy with test 2, which turned out
1858 to be caused by a trailing space on an input line that had got lost in his
1859 copy. The trailing space was an accident, so I've just removed it.
1861 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1862 that is needed.
1864 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1865 as "const" (a) because they are and (b) because it helps the PHP
1866 maintainers who have recently made a script to detect big data structures
1867 in the php code that should be moved to the .rodata section. I remembered
1868 to update Builducptable as well, so it won't revert if ucptable.h is ever
1869 re-created.
1871 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1872 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1873 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1874 support is not required. This saves 1.5-2K of code, which is important in
1875 some applications.
1877 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1878 so as not to refer to the tables, even though these functions will never be
1879 called when UTF-8 support is disabled. Otherwise there are problems with a
1880 shared library.
1882 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1884 (a) It was defining its arguments as char * instead of void *.
1886 (b) It was assuming that all moves were upwards in memory; this was true
1887 a long time ago when I wrote it, but is no longer the case.
1889 The emulated memove() is provided for those environments that have neither
1890 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1891 is clearly not the case, as these two bugs were recently reported.
1893 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1894 and Detrail to create the HTML documentation, the .txt form of the man
1895 pages, and it removes trailing spaces from listed files. It also creates
1896 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1897 case, it wraps all the #defines with #ifndefs. This script should be run
1898 before "make dist".
1900 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1901 with Unicode property support.
1903 (a) For a maximizing quantifier, if the two different cases of the
1904 character were of different lengths in their UTF-8 codings (there are
1905 some cases like this - I found 11), and the matching function had to
1906 back up over a mixture of the two cases, it incorrectly assumed they
1907 were both the same length.
1909 (b) When PCRE was configured to use the heap rather than the stack for
1910 recursion during matching, it was not correctly preserving the data for
1911 the other case of a UTF-8 character when checking ahead for a match
1912 while processing a minimizing repeat. If the check also involved
1913 matching a wide character, but failed, corruption could cause an
1914 erroneous result when trying to check for a repeat of the original
1915 character.
1917 13. Some tidying changes to the testing mechanism:
1919 (a) The RunTest script now detects the internal link size and whether there
1920 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1921 values substituted by "configure". (The RunGrepTest script already did
1922 this for UTF-8.) The configure.ac script no longer substitutes the
1923 relevant variables.
1925 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1926 with length and offset values. This means that the output is different
1927 for different internal link sizes. Test 2 is skipped for link sizes
1928 other than 2 because of this, bypassing the problem. Unfortunately,
1929 there was also a test in test 3 (the locale tests) that used /B and
1930 failed for link sizes other than 2. Rather than cut the whole test out,
1931 I have added a new /Z option to pcretest that replaces the length and
1932 offset values with spaces. This is now used to make test 3 independent
1933 of link size. (Test 2 will be tidied up later.)
1935 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1936 segmentation fault instead of returning the appropriate error message.
1938 15. In multiline mode when the newline sequence was set to "any", the pattern
1939 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1940 This doesn't seem right; it now treats the CRLF combination as the line
1941 ending, and so does not match in that case. It's only a pattern such as ^$
1942 that would hit this one: something like ^ABC$ would have failed after \r
1943 and then tried again after \r\n.
1945 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1946 in an attempt to make files that differ only in their line terminators
1947 compare equal. This works on Linux.
1949 17. Under certain error circumstances pcregrep might try to free random memory
1950 as it exited. This is now fixed, thanks to valgrind.
1952 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1953 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1954 was because its rules for how to advance for /g after matching an empty
1955 string at the end of a line did not allow for this case. They now check for
1956 it specially.
1958 20. pcretest is supposed to handle patterns and data of any length, by
1959 extending its buffers when necessary. It was getting this wrong when the
1960 buffer for a data line had to be extended.
1962 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1963 CRLF as a newline sequence.
1965 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1966 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1967 I have nevertheless tidied it up.
1969 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1971 24. Added a man page for pcre-config.
1974 Version 7.0 19-Dec-06
1975 ---------------------
1977 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1978 moving to gcc 4.1.1.
1980 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1981 sys/time.h, which is documented as needed for this function. It doesn't
1982 seem to matter on Linux, but it showed up on some releases of OS X.
1984 3. It seems that there are systems where bytes whose values are greater than
1985 127 match isprint() in the "C" locale. The "C" locale should be the
1986 default when a C program starts up. In most systems, only ASCII printing
1987 characters match isprint(). This difference caused the output from pcretest
1988 to vary, making some of the tests fail. I have changed pcretest so that:
1990 (a) When it is outputting text in the compiled version of a pattern, bytes
1991 other than 32-126 are always shown as hex escapes.
1993 (b) When it is outputting text that is a matched part of a subject string,
1994 it does the same, unless a different locale has been set for the match
1995 (using the /L modifier). In this case, it uses isprint() to decide.
1997 4. Fixed a major bug that caused incorrect computation of the amount of memory
1998 required for a compiled pattern when options that changed within the
1999 pattern affected the logic of the preliminary scan that determines the
2000 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2001 that the computed length was too small. The symptoms of this bug were
2002 either the PCRE error "internal error: code overflow" from pcre_compile(),
2003 or a glibc crash with a message such as "pcretest: free(): invalid next
2004 size (fast)". Examples of patterns that provoked this bug (shown in
2005 pcretest format) are:
2007 /(?-x: )/x
2008 /(?x)(?-x: \s*#\s*)/
2009 /((?i)[\x{c0}])/8
2010 /(?i:[\x{c0}])/8
2012 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2013 is now done differently.
2015 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2016 wrapper classes; (b) implement a new function in the C++ scanner that is
2017 more efficient than the old way of doing things because it avoids levels of
2018 recursion in the regex matching; (c) add a paragraph to the documentation
2019 for the FullMatch() function.
2021 6. The escape sequence \n was being treated as whatever was defined as
2022 "newline". Not only was this contrary to the documentation, which states
2023 that \n is character 10 (hex 0A), but it also went horribly wrong when
2024 "newline" was defined as CRLF. This has been fixed.
2026 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2027 was being set to -1 for the "end of line" case (supposedly a value that no
2028 character can have). Though this value is never used (the check for end of
2029 line is "zero bytes in current character"), it caused compiler complaints.
2030 I've changed it to 0xffffffff.
2032 8. In pcre_version.c, the version string was being built by a sequence of
2033 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2034 string (as it is for production releases) called a macro with an empty
2035 argument. The C standard says the result of this is undefined. The gcc
2036 compiler treats it as an empty string (which was what was wanted) but it is
2037 reported that Visual C gives an error. The source has been hacked around to
2038 avoid this problem.
2040 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2041 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2042 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2043 of them did).
2045 10. Originally, pcretest opened its input and output without "b"; then I was
2046 told that "b" was needed in some environments, so it was added for release
2047 5.0 to both the input and output. (It makes no difference on Unix-like
2048 systems.) Later I was told that it is wrong for the input on Windows. I've
2049 now abstracted the modes into two macros, to make it easier to fiddle with
2050 them, and removed "b" from the input mode under Windows.
2052 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2054 12. Added -help and --help to pcretest as an official way of being reminded
2055 of the options.
2057 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2058 and pcrecpp.cc because they annoy compilers at high warning levels.
2060 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2062 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2063 scripts are not C programs :-) and which was not noticed because it works
2064 on Linux.
2066 16. pcretest is supposed to handle any length of pattern and data line (as one
2067 line or as a continued sequence of lines) by extending its input buffer if
2068 necessary. This feature was broken for very long pattern lines, leading to
2069 a string of junk being passed to pcre_compile() if the pattern was longer
2070 than about 50K.
2072 17. I have done a major re-factoring of the way pcre_compile() computes the
2073 amount of memory needed for a compiled pattern. Previously, there was code
2074 that made a preliminary scan of the pattern in order to do this. That was
2075 OK when PCRE was new, but as the facilities have expanded, it has become
2076 harder and harder to keep it in step with the real compile phase, and there
2077 have been a number of bugs (see for example, 4 above). I have now found a
2078 cunning way of running the real compile function in a "fake" mode that
2079 enables it to compute how much memory it would need, while actually only
2080 ever using a few hundred bytes of working memory and without too many
2081 tests of the mode. This should make future maintenance and development
2082 easier. A side effect of this work is that the limit of 200 on the nesting
2083 depth of parentheses has been removed (though this was never a serious
2084 limitation, I suspect). However, there is a downside: pcre_compile() now
2085 runs more slowly than before (30% or more, depending on the pattern). I
2086 hope this isn't a big issue. There is no effect on runtime performance.
2088 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2089 newline (only possible for the last line of a file) and it was a
2090 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2092 19. Added additional timing features to pcretest. (1) The -tm option now times
2093 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2094 separate command line item, by a number that specifies the number of
2095 repeats to use when timing. The default is 50000; this gives better
2096 precision, but takes uncomfortably long for very large patterns.
2098 20. Extended pcre_study() to be more clever in cases where a branch of a
2099 subpattern has no definite first character. For example, (a*|b*)[cd] would
2100 previously give no result from pcre_study(). Now it recognizes that the
2101 first character must be a, b, c, or d.
2103 21. There was an incorrect error "recursive call could loop indefinitely" if
2104 a subpattern (or the entire pattern) that was being tested for matching an
2105 empty string contained only one non-empty item after a nested subpattern.
2106 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2107 incorrectly, because the \d was being skipped in the check.
2109 22. The pcretest program now has a new pattern option /B and a command line
2110 option -b, which is equivalent to adding /B to every pattern. This causes
2111 it to show the compiled bytecode, without the additional information that
2112 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2113 is the same as /B/I).
2115 23. A new optimization is now able automatically to treat some sequences such
2116 as a*b as a*+b. More specifically, if something simple (such as a character
2117 or a simple class like \d) has an unlimited quantifier, and is followed by
2118 something that cannot possibly match the quantified thing, the quantifier
2119 is automatically "possessified".
2121 24. A recursive reference to a subpattern whose number was greater than 39
2122 went wrong under certain circumstances in UTF-8 mode. This bug could also
2123 have affected the operation of pcre_study().
2125 25. Realized that a little bit of performance could be had by replacing
2126 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2128 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2130 27. Possessive quantifiers such as a++ were previously implemented by turning
2131 them into atomic groups such as ($>a+). Now they have their own opcodes,
2132 which improves performance. This includes the automatically created ones
2133 from 23 above.
2135 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2136 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2137 the first matched character to be a colon. This applied both to named and
2138 numbered groups.
2140 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2142 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2143 building PCRE on Windows easier, so I have included it in the distribution.
2145 31. There is now a check in pcretest against a ridiculously large number being
2146 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2147 loop, the loop is abandoned.
2149 32. Forward references to subpatterns in conditions such as (?(2)...) where
2150 subpattern 2 is defined later cause pcre_compile() to search forwards in
2151 the pattern for the relevant set of parentheses. This search went wrong
2152 when there were unescaped parentheses in a character class, parentheses
2153 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2155 33. "Subroutine" calls and backreferences were previously restricted to
2156 referencing subpatterns earlier in the regex. This restriction has now
2157 been removed.
2159 34. Added a number of extra features that are going to be in Perl 5.10. On the
2160 whole, these are just syntactic alternatives for features that PCRE had
2161 previously implemented using the Python syntax or my own invention. The
2162 other formats are all retained for compatibility.
2164 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2165 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2166 also .NET compatible.
2168 (b) A recursion or subroutine call to a named group can now be defined as
2169 (?&name) as well as (?P>name).
2171 (c) A backreference to a named group can now be defined as \k<name> or
2172 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2173 5.10, are also .NET compatible.
2175 (d) A conditional reference to a named group can now use the syntax
2176 (?(<name>) or (?('name') as well as (?(name).
2178 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2179 groups (named and numbered) that are never evaluated inline, but can be
2180 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2181 is always false. There may be only one alternative in such a group.
2183 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2184 as the simple (?(R). The condition is true only if the most recent
2185 recursion is that of the given number or name. It does not search out
2186 through the entire recursion stack.
2188 (g) The escape \gN or \g{N} has been added, where N is a positive or
2189 negative number, specifying an absolute or relative reference.
2191 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2192 some "unreachable code" warnings.
2194 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2195 things, this adds five new scripts.
2197 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2198 There were also incompatibilities regarding the handling of \Q..\E inside
2199 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2200 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2202 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2203 matches an empty string, and forcibly breaks the loop. There were bugs in
2204 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2205 against aaaa the result was just "a" rather than "aaaa", for example. Two
2206 separate and independent bugs (that affected different cases) have been
2207 fixed.
2209 39. Refactored the code to abolish the use of different opcodes for small
2210 capturing bracket numbers. This is a tidy that I avoided doing when I
2211 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2212 The new approach is not only tidier, it makes it possible to reduce the
2213 memory needed to fix the previous bug (38).
2215 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2216 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2217 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2218 mode.
2220 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2221 report.
2223 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2224 copy construction and assignment in the C++ wrapper.
2226 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2227 couple of bugs that could have given wrong results in the "--newline=crlf"
2228 case.
2230 44. Added a number of casts and did some reorganization of signed/unsigned int
2231 variables following suggestions from Dair Grant. Also renamed the variable
2232 "this" as "item" because it is a C++ keyword.
2234 45. Arranged for dftables to add
2236 #include "pcre_internal.h"
2238 to pcre_chartables.c because without it, gcc 4.x may remove the array
2239 definition from the final binary if PCRE is built into a static library and
2240 dead code stripping is activated.
2242 46. For an unanchored pattern, if a match attempt fails at the start of a
2243 newline sequence, and the newline setting is CRLF or ANY, and the next two
2244 characters are CRLF, advance by two characters instead of one.
2247 Version 6.7 04-Jul-06
2248 ---------------------
2250 1. In order to handle tests when input lines are enormously long, pcretest has
2251 been re-factored so that it automatically extends its buffers when
2252 necessary. The code is crude, but this _is_ just a test program. The
2253 default size has been increased from 32K to 50K.
2255 2. The code in pcre_study() was using the value of the re argument before
2256 testing it for NULL. (Of course, in any sensible call of the function, it
2257 won't be NULL.)
2259 3. The memmove() emulation function in pcre_internal.h, which is used on
2260 systems that lack both memmove() and bcopy() - that is, hardly ever -
2261 was missing a "static" storage class specifier.
2263 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2264 containing an extended class (one that cannot be represented by a bitmap
2265 because it contains high-valued characters or Unicode property items, e.g.
2266 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2267 pattern, but PCRE should not loop if you do not (it no longer does).
2268 [Detail: two cases were found: (a) a repeated subpattern containing an
2269 extended class; (b) a recursive reference to a subpattern that followed a
2270 previous extended class. It wasn't skipping over the extended class
2271 correctly when UTF-8 mode was not set.]
2273 5. A negated single-character class was not being recognized as fixed-length
2274 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2275 compile error "lookbehind assertion is not fixed length".
2277 6. The RunPerlTest auxiliary script was showing an unexpected difference
2278 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2279 write a Perl script that can interpret lines of an input file either as
2280 byte characters or as UTF-8, which is what "perltest" was being required to
2281 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2282 can't do is switch easily at run time between having the "use utf8;" pragma
2283 or not. In the end, I fudged it by using the RunPerlTest script to insert
2284 "use utf8;" explicitly for the UTF-8 tests.
2286 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2287 the end of the subject string, contrary to the documentation and to what
2288 Perl does. This was true of both matching functions. Now it matches only at
2289 the start of the subject and immediately after *internal* newlines.
2291 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2292 a pointer to an int instead of a pointer to an unsigned long int. This
2293 caused problems on 64-bit systems.
2295 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2296 instance of the 'standard' template library not being so standard".
2298 10. There was no check on the number of named subpatterns nor the maximum
2299 length of a subpattern name. The product of these values is used to compute
2300 the size of the memory block for a compiled pattern. By supplying a very
2301 long subpattern name and a large number of named subpatterns, the size
2302 computation could be caused to overflow. This is now prevented by limiting
2303 the length of names to 32 characters, and the number of named subpatterns
2304 to 10,000.
2306 11. Subpatterns that are repeated with specific counts have to be replicated in
2307 the compiled pattern. The size of memory for this was computed from the
2308 length of the subpattern and the repeat count. The latter is limited to
2309 65535, but there was no limit on the former, meaning that integer overflow
2310 could in principle occur. The compiled length of a repeated subpattern is
2311 now limited to 30,000 bytes in order to prevent this.
2313 12. Added the optional facility to have named substrings with the same name.
2315 13. Added the ability to use a named substring as a condition, using the
2316 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2317 are numbers (not recommended). Forward references are permitted.
2319 14. Added forward references in named backreferences (if you see what I mean).
2321 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2322 pattern could run off the end of the subject. For example, the pattern
2323 "(?s)(.{1,5})"8 did this with the subject "ab".
2325 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2326 PCRE_CASELESS was set when matching characters that were quantified with ?
2327 or *.
2329 17. A character class other than a single negated character that had a minimum
2330 but no maximum quantifier - for example [ab]{6,} - was not handled
2331 correctly by pce_dfa_exec(). It would match only one character.
2333 18. A valid (though odd) pattern that looked like a POSIX character
2334 class but used an invalid character after [ (for example [[,abc,]]) caused
2335 pcre_compile() to give the error "Failed: internal error: code overflow" or
2336 in some cases to crash with a glibc free() error. This could even happen if
2337 the pattern terminated after [[ but there just happened to be a sequence of
2338 letters, a binary zero, and a closing ] in the memory that followed.
2340 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2341 over the years. Originally (before any Unicode support), just the bottom 8
2342 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2343 output from "man perlunicode" includes this:
2345 The regular expression compiler produces polymorphic opcodes. That
2346 is, the pattern adapts to the data and automatically switches to
2347 the Unicode character scheme when presented with Unicode data--or
2348 instead uses a traditional byte scheme when presented with byte
2349 data.
2351 Sadly, a wide octal escape does not cause a switch, and in a string with
2352 no other multibyte characters, these octal escapes are treated as before.
2353 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2354 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2355 Unicode string.
2357 I have not perpetrated such confusion in PCRE. Up till now, it took just
2358 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2359 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2360 translate to the appropriate multibyte character.
2362 29. Applied some refactoring to reduce the number of warnings from Microsoft
2363 and Borland compilers. This has included removing the fudge introduced
2364 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2365 a warning about an unused variable.
2367 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2368 characters since release 4.0, because Perl (from release 5.004) does not.
2369 [Or at least, is documented not to: some releases seem to be in conflict
2370 with the documentation.] However, when a pattern was studied with
2371 pcre_study() and all its branches started with \s, PCRE still included VT
2372 as a possible starting character. Of course, this did no harm; it just
2373 caused an unnecessary match attempt.
2375 22. Removed a now-redundant internal flag bit that recorded the fact that case
2376 dependency changed within the pattern. This was once needed for "required
2377 byte" processing, but is no longer used. This recovers a now-scarce options
2378 bit. Also moved the least significant internal flag bit to the most-
2379 significant bit of the word, which was not previously used (hangover from
2380 the days when it was an int rather than a uint) to free up another bit for
2381 the future.
2383 23. Added support for CRLF line endings as well as CR and LF. As well as the
2384 default being selectable at build time, it can now be changed at runtime
2385 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2386 specify that it is scanning data with non-default line endings.
2388 24. Changed the definition of CXXLINK to make it agree with the definition of
2389 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2391 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2392 recursions. This makes a big different to stack usage for some patterns.
2394 26. If a subpattern containing a named recursion or subroutine reference such
2395 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2396 the space required for the compiled pattern went wrong and gave too small a
2397 value. Depending on the environment, this could lead to "Failed: internal
2398 error: code overflow at offset 49" or "glibc detected double free or
2399 corruption" errors.
2401 27. Applied patches from Google (a) to support the new newline modes and (b) to
2402 advance over multibyte UTF-8 characters in GlobalReplace.
2404 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2405 difference for some implementation of PCRE in some Windows version.
2407 29. Added some extra testing facilities to pcretest:
2409 \q<number> in a data line sets the "match limit" value
2410 \Q<number> in a data line sets the "match recursion limt" value
2411 -S <number> sets the stack size, where <number> is in megabytes
2413 The -S option isn't available for Windows.
2416 Version 6.6 06-Feb-06
2417 ---------------------
2419 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2420 in pcreposix.h. I have copied the definition from pcre.h.
2422 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2423 because pcre.h is no longer a built file.
2425 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2426 not normally included in the compiled code.
2429 Version 6.5 01-Feb-06
2430 ---------------------
2432 1. When using the partial match feature with pcre_dfa_exec(), it was not
2433 anchoring the second and subsequent partial matches at the new starting
2434 point. This could lead to incorrect results. For example, with the pattern
2435 /1234/, partially matching against "123" and then "a4" gave a match.
2437 2. Changes to pcregrep:
2439 (a) All non-match returns from pcre_exec() were being treated as failures
2440 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2441 error message is output. Some extra information is given for the
2443 probably the only errors that are likely to be caused by users (by
2444 specifying a regex that has nested indefinite repeats, for instance).
2445 If there are more than 20 of these errors, pcregrep is abandoned.
2447 (b) A binary zero was treated as data while matching, but terminated the
2448 output line if it was written out. This has been fixed: binary zeroes
2449 are now no different to any other data bytes.
2451 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2452 used to set a locale for matching. The --locale=xxxx long option has
2453 been added (no short equivalent) to specify a locale explicitly on the
2454 pcregrep command, overriding the environment variables.
2456 (d) When -B was used with -n, some line numbers in the output were one less
2457 than they should have been.
2459 (e) Added the -o (--only-matching) option.
2461 (f) If -A or -C was used with -c (count only), some lines of context were
2462 accidentally printed for the final match.
2464 (g) Added the -H (--with-filename) option.
2466 (h) The combination of options -rh failed to suppress file names for files
2467 that were found from directory arguments.
2469 (i) Added the -D (--devices) and -d (--directories) options.
2471 (j) Added the -F (--fixed-strings) option.
2473 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2475 (l) Added the --colo(u)r option.
2477 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2478 is not present by default.
2480 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2481 items such as (?R) or (?1), when the recursion could match a number of
2482 alternatives. If it matched one of the alternatives, but subsequently,
2483 outside the recursion, there was a failure, the code tried to back up into
2484 the recursion. However, because of the way PCRE is implemented, this is not
2485 possible, and the result was an incorrect result from the match.
2487 In order to prevent this happening, the specification of recursion has
2488 been changed so that all such subpatterns are automatically treated as
2489 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2491 4. I had overlooked the fact that, in some locales, there are characters for
2492 which isalpha() is true but neither isupper() nor islower() are true. In
2493 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2494 and ordfeminine) are like this. This affected the treatment of \w and \W
2495 when they appeared in character classes, but not when they appeared outside
2496 a character class. The bit map for "word" characters is now created
2497 separately from the results of isalnum() instead of just taking it from the
2498 upper, lower, and digit maps. (Plus the underscore character, of course.)
2500 5. The above bug also affected the handling of POSIX character classes such as
2501 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2502 permanent tables. Instead, the bit maps for such a class were previously
2503 created as the appropriate unions of the upper, lower, and digit bitmaps.
2504 Now they are created by subtraction from the [[:word:]] class, which has
2505 its own bitmap.
2507 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2508 It is created by subtracting the vertical space characters (\x09, \x0a,
2509 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2510 subtraction was done in the overall bitmap for a character class, meaning
2511 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2512 be recognized. This bug has been fixed.
2514 7. Patches from the folks at Google:
2516 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2517 real life, but is still worth protecting against".
2519 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2520 regular expressions".
2522 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2523 have it.
2525 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2526 "configure" and the latter not, in order to fix a problem somebody had
2527 with compiling the Arg class on HP-UX.
2529 (e) Improve the error-handling of the C++ wrapper a little bit.
2531 (f) New tests for checking recursion limiting.
2533 8. The pcre_memmove() function, which is used only if the environment does not
2534 have a standard memmove() function (and is therefore rarely compiled),
2535 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2536 returning a result (though PCRE never actually uses the result).
2538 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2539 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2540 returned instead of calling malloc() with an overflowing number that would
2541 most likely cause subsequent chaos.
2543 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2545 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2546 with this option is matched, the nmatch and pmatch options of regexec() are
2547 ignored.
2549 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2550 provided in case anyone wants to the the POSIX interface with UTF-8
2551 strings.
2553 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2554 C++ linking (needed for some HP-UX environments).
2556 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2557 (unused parameter) and in the pcre_printint() function (omitted "default"
2558 switch label when the default is to do nothing).
2560 15. Added some code to make it possible, when PCRE is compiled as a C++
2561 library, to replace subject pointers for pcre_exec() with a smart pointer
2562 class, thus making it possible to process discontinuous strings.
2564 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2565 much the same function. They were added by different people who were trying
2566 to make PCRE easy to compile on non-Unix systems. It has been suggested
2567 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2568 for compiling on Windows systems. I have therefore replaced it with
2569 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2570 defaults to "extern" for C or "extern C" for C++, which works fine on
2571 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2572 SCOPE with something explicit in config.h. In addition:
2574 (a) pcreposix.h still had just "extern" instead of either of these macros;
2575 I have replaced it with PCRE_DATA_SCOPE.
2577 (b) Functions such as _pcre_xclass(), which are internal to the library,
2578 but external in the C sense, all had PCRE_EXPORT in their definitions.
2579 This is apparently wrong for the Windows case, so I have removed it.
2580 (It makes no difference on Unix-like systems.)
2582 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2583 of recursive calls to match(). This is different to MATCH_LIMIT because
2584 that limits the total number of calls to match(), not all of which increase
2585 the depth of recursion. Limiting the recursion depth limits the amount of
2586 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2587 when PCRE is compiled, and changed at run time. A patch from Google adds
2588 this functionality to the C++ interface.
2590 18. Changes to the handling of Unicode character properties:
2592 (a) Updated the table to Unicode 4.1.0.
2594 (b) Recognize characters that are not in the table as "Cn" (undefined).
2596 (c) I revised the way the table is implemented to a much improved format
2597 which includes recognition of ranges. It now supports the ranges that
2598 are defined in UnicodeData.txt, and it also amalgamates other
2599 characters into ranges. This has reduced the number of entries in the
2600 table from around 16,000 to around 3,000, thus reducing its size
2601 considerably. I realized I did not need to use a tree structure after
2602 all - a binary chop search is just as efficient. Having reduced the
2603 number of entries, I extended their size from 6 bytes to 8 bytes to
2604 allow for more data.
2606 (d) Added support for Unicode script names via properties such as \p{Han}.
2608 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2609 matching that character.
2611 20. When matching a repeated Unicode property with a minimum greater than zero,
2612 (for example \pL{2,}), PCRE could look past the end of the subject if it
2613 reached it while seeking the minimum number of characters. This could
2614 happen only if some of the characters were more than one byte long, because
2615 there is a check for at least the minimum number of bytes.
2617 21. Refactored the implementation of \p and \P so as to be more general, to
2618 allow for more different types of property in future. This has changed the
2619 compiled form incompatibly. Anybody with saved compiled patterns that use
2620 \p or \P will have to recompile them.
2622 22. Added "Any" and "L&" to the supported property types.
2624 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2625 but give a compile time error if the value is greater than 0xff.
2627 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2628 accidentally not being installed or uninstalled.
2630 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2631 made were to insert the current release number. This seemed silly, because
2632 it made things harder for people building PCRE on systems that don't run
2633 "configure". I have turned pcre.h into a distributed file, no longer built
2634 by "configure", with the version identification directly included. There is
2635 no longer a pcre.h.in file.
2637 However, this change necessitated a change to the pcre-config script as
2638 well. It is built from pcre-config.in, and one of the substitutions was the
2639 release number. I have updated configure.ac so that ./configure now finds
2640 the release number by grepping pcre.h.
2642 26. Added the ability to run the tests under valgrind.
2645 Version 6.4 05-Sep-05
2646 ---------------------
2648 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2649 "--" to be printed when multiple files were scanned, even when none of the
2650 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2651 consider it to be a bug, and have restored the previous behaviour.
2653 2. A couple of code tidies to get rid of compiler warnings.
2655 3. The pcretest program used to cheat by referring to symbols in the library
2656 whose names begin with _pcre_. These are internal symbols that are not
2657 really supposed to be visible externally, and in some environments it is
2658 possible to suppress them. The cheating is now confined to including
2659 certain files from the library's source, which is a bit cleaner.
2661 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2662 file's purpose clearer.
2664 5. Reorganized pcre_ucp_findchar().
2667 Version 6.3 15-Aug-05
2668 ---------------------
2670 1. The file libpcre.pc.in did not have general read permission in the tarball.
2672 2. There were some problems when building without C++ support:
2674 (a) If C++ support was not built, "make install" and "make test" still
2675 tried to test it.
2677 (b) There were problems when the value of CXX was explicitly set. Some
2678 changes have been made to try to fix these, and ...
2680 (c) --disable-cpp can now be used to explicitly disable C++ support.
2682 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2683 backslash in a target when C++ was disabled. This confuses some
2684 versions of "make", apparently. Using an intermediate variable solves
2685 this. (Same for CPP_LOBJ.)
2687 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2688 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2689 necessary on certain architectures.
2691 4. Added a setting of -export-symbols-regex to the link command to remove
2692 those symbols that are exported in the C sense, but actually are local
2693 within the library, and not documented. Their names all begin with
2694 "_pcre_". This is not a perfect job, because (a) we have to except some
2695 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2696 available (and never for static libraries). I have made a note to try to
2697 find a way round (a) in the future.
2700 Version 6.2 01-Aug-05
2701 ---------------------
2703 1. There was no test for integer overflow of quantifier values. A construction
2704 such as {1111111111111111} would give undefined results. What is worse, if
2705 a minimum quantifier for a parenthesized subpattern overflowed and became
2706 negative, the calculation of the memory size went wrong. This could have
2707 led to memory overwriting.
2709 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2711 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2712 operating environments where this matters.
2714 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2715 PCRE options from within the C++ wrapper.
2717 5. Named capturing subpatterns were not being correctly counted when a pattern
2718 was compiled. This caused two problems: (a) If there were more than 100
2719 such subpatterns, the calculation of the memory needed for the whole
2720 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2721 back references of the form \12, where the number was greater than 9, were
2722 not recognized as back references, even though there were sufficient
2723 previous subpatterns.
2725 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2726 versions of gcc, e.g. 2.95.4.
2729 Version 6.1 21-Jun-05
2730 ---------------------
2732 1. There was one reference to the variable "posix" in pcretest.c that was not
2733 surrounded by "#if !defined NOPOSIX".
2735 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2736 the cross-check on the old pcre_info() function, for the benefit of the
2737 cut-down version of PCRE that is currently imported into Exim.
2739 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2740 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2741 patterns that start (?i) but I don't think that matters. The use of (?i) is
2742 just an example; this all applies to the other options as well.
2744 4. Since libtool seems to echo the compile commands it is issuing, the output
2745 from "make" can be reduced a bit by putting "@" in front of each libtool
2746 compile command.
2748 5. Patch from the folks at Google for configure.in to be a bit more thorough
2749 in checking for a suitable C++ installation before trying to compile the
2750 C++ stuff. This should fix a reported problem when a compiler was present,
2751 but no suitable headers.
2753 6. The man pages all had just "PCRE" as their title. I have changed them to
2754 be the relevant file name. I have also arranged that these names are
2755 retained in the file doc/pcre.txt, which is a concatenation in text format
2756 of all the man pages except the little individual ones for each function.
2758 7. The NON-UNIX-USE file had not been updated for the different set of source
2759 files that come with release 6. I also added a few comments about the C++
2760 wrapper.
2763 Version 6.0 07-Jun-05
2764 ---------------------
2766 1. Some minor internal re-organization to help with my DFA experiments.
2768 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2769 didn't matter for the library itself when fully configured, but did matter
2770 when compiling without UCP support, or within Exim, where the ucp files are
2771 not imported.
2773 3. Refactoring of the library code to split up the various functions into
2774 different source modules. The addition of the new DFA matching code (see
2775 below) to a single monolithic source would have made it really too
2776 unwieldy, quite apart from causing all the code to be include in a
2777 statically linked application, when only some functions are used. This is
2778 relevant even without the DFA addition now that patterns can be compiled in
2779 one application and matched in another.
2781 The downside of splitting up is that there have to be some external
2782 functions and data tables that are used internally in different modules of
2783 the library but which are not part of the API. These have all had their
2784 names changed to start with "_pcre_" so that they are unlikely to clash
2785 with other external names.
2787 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2788 a different (DFA) algorithm. Although it is slower than the original
2789 function, it does have some advantages for certain types of matching
2790 problem.
2792 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2793 including restarting after a partial match.
2795 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2796 defined when compiling for Windows was sent to me. I have put it into the
2797 code, though I have no means of testing or verifying it.
2799 7. Added the pcre_refcount() auxiliary function.
2801 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2802 match before or at the first newline in the subject string. In pcretest,
2803 the /f option on a pattern can be used to set this.
2805 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2806 would behave wrongly. This has been present in PCRE since release 4.0.
2808 10. A number of changes to the pcregrep command:
2810 (a) Refactored how -x works; insert ^(...)$ instead of setting
2811 PCRE_ANCHORED and checking the length, in preparation for adding
2812 something similar for -w.
2814 (b) Added the -w (match as a word) option.
2816 (c) Refactored the way lines are read and buffered so as to have more
2817 than one at a time available.
2819 (d) Implemented a pcregrep test script.
2821 (e) Added the -M (multiline match) option. This allows patterns to match
2822 over several lines of the subject. The buffering ensures that at least
2823 8K, or the rest of the document (whichever is the shorter) is available
2824 for matching (and similarly the previous 8K for lookbehind assertions).
2826 (f) Changed the --help output so that it now says
2828 -w, --word-regex(p)
2830 instead of two lines, one with "regex" and the other with "regexp"
2831 because that confused at least one person since the short forms are the
2832 same. (This required a bit of code, as the output is generated
2833 automatically from a table. It wasn't just a text change.)
2835 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2836 option but starts with a hyphen. Could be a pattern or a path name
2837 starting with a hyphen, for instance.
2839 (h) "-" can be given as a file name to represent stdin.
2841 (i) When file names are being printed, "(standard input)" is used for
2842 the standard input, for compatibility with GNU grep. Previously
2843 "<stdin>" was used.
2845 (j) The option --label=xxx can be used to supply a name to be used for
2846 stdin when file names are being printed. There is no short form.
2848 (k) Re-factored the options decoding logic because we are going to add
2849 two more options that take data. Such options can now be given in four
2850 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2852 (l) Added the -A, -B, and -C options for requesting that lines of context
2853 around matches be printed.
2855 (m) Added the -L option to print the names of files that do not contain
2856 any matching lines, that is, the complement of -l.
2858 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2859 continue to scan other files.
2861 (o) The -s option was incorrectly implemented. For compatibility with other
2862 greps, it now suppresses the error message for a non-existent or non-
2863 accessible file (but not the return code). There is a new option called
2864 -q that suppresses the output of matching lines, which was what -s was
2865 previously doing.
2867 (p) Added --include and --exclude options to specify files for inclusion
2868 and exclusion when recursing.
2870 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2871 Hopefully, it now does.
2873 12. Missing cast in pcre_study().
2875 13. Added an "uninstall" target to the makefile.
2877 14. Replaced "extern" in the function prototypes in Makefile.in with
2878 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2879 world, but is set differently for Windows.
2881 15. Added a second compiling function called pcre_compile2(). The only
2882 difference is that it has an extra argument, which is a pointer to an
2883 integer error code. When there is a compile-time failure, this is set
2884 non-zero, in addition to the error test pointer being set to point to an
2885 error message. The new argument may be NULL if no error number is required
2886 (but then you may as well call pcre_compile(), which is now just a
2887 wrapper). This facility is provided because some applications need a
2888 numeric error indication, but it has also enabled me to tidy up the way
2889 compile-time errors are handled in the POSIX wrapper.
2891 16. Added VPATH=.libs to the makefile; this should help when building with one
2892 prefix path and installing with another. (Or so I'm told by someone who
2893 knows more about this stuff than I do.)
2895 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2896 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2897 match everything, including newlines. This is not POSIX-compatible, but
2898 somebody wanted the feature. From pcretest it can be activated by using
2899 both the P and the s flags.
2901 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2903 19. libpcre.pc was being incorrectly installed as executable.
2905 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2906 it now also looks for '\r' so that it will work unmodified on Windows.
2908 21. Added Google's contributed C++ wrapper to the distribution.
2910 22. Added some untidy missing memory free() calls in pcretest, to keep
2911 Electric Fence happy when testing.
2915 Version 5.0 13-Sep-04
2916 ---------------------
2918 1. Internal change: literal characters are no longer packed up into items
2919 containing multiple characters in a single byte-string. Each character
2920 is now matched using a separate opcode. However, there may be more than one
2921 byte in the character in UTF-8 mode.
2923 2. The pcre_callout_block structure has two new fields: pattern_position and
2924 next_item_length. These contain the offset in the pattern to the next match
2925 item, and its length, respectively.
2927 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2928 insertion of callouts before each pattern item. Added the /C option to
2929 pcretest to make use of this.
2931 4. On the advice of a Windows user, the lines
2933 #if defined(_WIN32) || defined(WIN32)
2934 _setmode( _fileno( stdout ), 0x8000 );
2935 #endif /* defined(_WIN32) || defined(WIN32) */
2937 have been added to the source of pcretest. This apparently does useful
2938 magic in relation to line terminators.
2940 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2941 for the benefit of those environments where the "b" makes a difference.
2943 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2944 to know about it. I have put a hack into configure.in that adds in code
2945 to set GCC=yes if CC=icc. This seems to end up at a point in the
2946 generated configure script that is early enough to affect the setting of
2947 compiler options, which is what is needed, but I have no means of testing
2948 whether it really works. (The user who reported this had patched the
2949 generated configure script, which of course I cannot do.)
2951 LATER: After change 22 below (new libtool files), the configure script
2952 seems to know about icc (and also ecc). Therefore, I have commented out
2953 this hack in configure.in.
2955 7. Added support for pkg-config (2 patches were sent in).
2957 8. Negated POSIX character classes that used a combination of internal tables
2958 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2959 [[:^ascii]]. Typically, they would match almost any characters. The other
2960 POSIX classes were not broken in this way.
2962 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2963 to find the match, as PCRE was deluded into thinking that the match had to
2964 start at the start point or following a newline. The same bug applied to
2965 patterns with negative forward assertions or any backward assertions
2966 preceding ".*" at the start, unless the pattern required a fixed first
2967 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2969 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2970 starting at the last subject character, bytes beyond the end of the subject
2971 string were read.
2973 11. Renamed the variable "class" as "classbits" to make life easier for C++
2974 users. (Previously there was a macro definition, but it apparently wasn't
2975 enough.)
2977 12. Added the new field "tables" to the extra data so that tables can be passed
2978 in at exec time, or the internal tables can be re-selected. This allows
2979 a compiled regex to be saved and re-used at a later time by a different
2980 program that might have everything at different addresses.
2982 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2983 -R library as well as a -L library.
2985 14. The debugging options of pcretest (-d on the command line or D on a
2986 pattern) showed incorrect output for anything following an extended class
2987 that contained multibyte characters and which was followed by a quantifier.
2989 15. Added optional support for general category Unicode character properties
2990 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2991 support. It adds about 90K to the size of the library. The meanings of the
2992 inbuilt class escapes such as \d and \s have NOT been changed.
2994 16. Updated pcredemo.c to include calls to free() to release the memory for the
2995 compiled pattern.
2997 17. The generated file chartables.c was being created in the source directory
2998 instead of in the building directory. This caused the build to fail if the
2999 source directory was different from the building directory, and was
3000 read-only.
3002 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3003 file. No doubt somebody will tell me if they don't make sense... Also added
3004 Dan Mooney's comments about building on OpenVMS.
3006 19. Added support for partial matching via the PCRE_PARTIAL option for
3007 pcre_exec() and the \P data escape in pcretest.
3009 20. Extended pcretest with 3 new pattern features:
3011 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3012 write the compiled pattern to the file whose name is "rest-of-line".
3013 This is a straight binary dump of the data, with the saved pointer to
3014 the character tables forced to be NULL. The study data, if any, is
3015 written too. After writing, pcretest reads a new pattern.
3017 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3018 compiled pattern from the given file. There must not be any
3019 occurrences of "<" in the file name (pretty unlikely); if there are,
3020 pcretest will instead treat the initial "<" as a pattern delimiter.
3021 After reading in the pattern, pcretest goes on to read data lines as
3022 usual.
3024 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3025 and 16-bit fields in a compiled pattern, to simulate a pattern that
3026 was compiled on a host of opposite endianness.
3028 21. The pcre-exec() function can now cope with patterns that were compiled on
3029 hosts of opposite endianness, with this restriction:
3031 As for any compiled expression that is saved and used later, the tables
3032 pointer field cannot be preserved; the extra_data field in the arguments
3033 to pcre_exec() should be used to pass in a tables address if a value
3034 other than the default internal tables were used at compile time.
3036 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3037 now diagnosed as an error. Previously, most of the time, a negative number
3038 would have been treated as zero, but if in addition "ovector" was passed as
3039 NULL, a crash could occur.
3041 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3042 new versions from the libtool 1.5 distribution (the last one is a copy of
3043 a file called libtool.m4). This seems to have fixed the need to patch
3044 "configure" to support Darwin 1.3 (which I used to do). However, I still
3045 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3046 workstation).
3048 24. Changed the PCRE licence to be the more standard "BSD" licence.
3051 Version 4.5 01-Dec-03
3052 ---------------------
3054 1. There has been some re-arrangement of the code for the match() function so
3055 that it can be compiled in a version that does not call itself recursively.
3056 Instead, it keeps those local variables that need separate instances for
3057 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3058 needs to "recurse". Keeping track of where control must go is done by means
3059 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3060 hide most of the details from the main code, and operates only if
3061 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3062 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3063 operating.
3065 To make it easier for callers to provide specially tailored get/free
3066 functions for this usage, two new functions, pcre_stack_malloc, and
3067 pcre_stack_free, are used. They are always called in strict stacking order,
3068 and the size of block requested is always the same.
3070 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3071 PCRE has been compiled to use the stack or the heap for recursion. The
3072 -C option of pcretest uses this to show which version is compiled.
3074 A new data escape \S, is added to pcretest; it causes the amounts of store
3075 obtained and freed by both kinds of malloc/free at match time to be added
3076 to the output.
3078 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3079 what's available on my current Linux desktop machine.
3081 3. When matching a UTF-8 string, the test for a valid string at the start has
3082 been extended. If start_offset is not zero, PCRE now checks that it points
3083 to a byte that is the start of a UTF-8 character. If not, it returns
3084 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3085 this is necessary because there may be backward assertions in the pattern.
3086 When matching the same subject several times, it may save resources to use
3087 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3089 4. The code for checking the validity of UTF-8 strings has been tightened so
3090 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3091 containing "overlong sequences".
3093 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3094 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3095 should have been "&", but it just so happened that all the cases this let
3096 through by mistake were picked up later in the function.
3098 6. I had used a variable called "isblank" - this is a C99 function, causing
3099 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3101 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3102 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3103 so the version is shown; (c) stop "make test" echoing "./RunTest".
3105 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3107 9. The prototype for memmove() for systems that don't have it was using
3108 size_t, but the inclusion of the header that defines size_t was later. I've
3109 moved the #includes for the C headers earlier to avoid this.
3111 10. Added some adjustments to the code to make it easier to compiler on certain
3112 special systems:
3114 (a) Some "const" qualifiers were missing.
3115 (b) Added the macro EXPORT before all exported functions; by default this
3116 is defined to be empty.
3117 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3118 that it reads its output file name as an argument instead of writing
3119 to the standard output and assuming this can be redirected.
3121 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3122 class containing characters with values greater than 255, PCRE compilation
3123 went into a loop.
3125 12. A recursive reference to a subpattern that was within another subpattern
3126 that had a minimum quantifier of zero caused PCRE to crash. For example,
3127 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3128 recursion. If the recursively-called subpattern itself had a zero repeat,
3129 that was OK.
3131 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3132 buffer into which it was copied (for escape processing) was still set at
3133 1024, so long lines caused crashes.
3135 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3136 "internal error: code overflow...". This applied to any character class
3137 that was followed by a possessive quantifier.
3139 15. Modified the Makefile to add libpcre.la as a prerequisite for
3140 libpcreposix.la because I was told this is needed for a parallel build to
3141 work.
3143 16. If a pattern that contained .* following optional items at the start was
3144 studied, the wrong optimizing data was generated, leading to matching
3145 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3146 matching string must start with a or b or c. The correct conclusion for
3147 this pattern is that a match can start with any character.
3150 Version 4.4 13-Aug-03
3151 ---------------------
3153 1. In UTF-8 mode, a character class containing characters with values between
3154 127 and 255 was not handled correctly if the compiled pattern was studied.
3155 In fixing this, I have also improved the studying algorithm for such
3156 classes (slightly).
3158 2. Three internal functions had redundant arguments passed to them. Removal
3159 might give a very teeny performance improvement.
3161 3. Documentation bug: the value of the capture_top field in a callout is *one
3162 more than* the number of the hightest numbered captured substring.
3164 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3165 in incorrectly linking with a previously installed version. They now link
3166 explicitly with libpcre.la.
3168 5. configure.in no longer needs to recognize Cygwin specially.
3170 6. A problem in pcre.in for Windows platforms is fixed.
3172 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3173 pcretest, it used to include the size of the study block as part of its
3174 output. Unfortunately, the structure contains a field that has a different
3175 size on different hardware architectures. This meant that the tests that
3176 showed this size failed. As the block is currently always of a fixed size,
3177 this information isn't actually particularly useful in pcretest output, so
3178 I have just removed it.
3180 8. Three pre-processor statements accidentally did not start in column 1.
3181 Sadly, there are *still* compilers around that complain, even though
3182 standard C has not required this for well over a decade. Sigh.
3184 9. In pcretest, the code for checking callouts passed small integers in the
3185 callout_data field, which is a void * field. However, some picky compilers
3186 complained about the casts involved for this on 64-bit systems. Now
3187 pcretest passes the address of the small integer instead, which should get
3188 rid of the warnings.
3190 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3191 both compile and run time, and gives an error if an invalid UTF-8 sequence
3192 is found. There is a option for disabling this check in cases where the
3193 string is known to be correct and/or the maximum performance is wanted.
3195 11. In response to a bug report, I changed one line in Makefile.in from
3197 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3198 to
3199 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3201 to look similar to other lines, but I have no way of telling whether this
3202 is the right thing to do, as I do not use Windows. No doubt I'll get told
3203 if it's wrong...
3206 Version 4.3 21-May-03
3207 ---------------------
3209 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3210 Makefile.
3212 2. Some refactoring to improve the quality of the code:
3214 (i) The utf8_table... variables are now declared "const".
3216 (ii) The code for \cx, which used the "case flipping" table to upper case
3217 lower case letters, now just substracts 32. This is ASCII-specific,
3218 but the whole concept of \cx is ASCII-specific, so it seems
3219 reasonable.
3221 (iii) PCRE was using its character types table to recognize decimal and
3222 hexadecimal digits in the pattern. This is silly, because it handles
3223 only 0-9, a-f, and A-F, but the character types table is locale-
3224 specific, which means strange things might happen. A private
3225 table is now used for this - though it costs 256 bytes, a table is
3226 much faster than multiple explicit tests. Of course, the standard
3227 character types table is still used for matching digits in subject
3228 strings against \d.
3230 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3231 ending in _t are). So I've renamed it as ESC_tee.
3233 3. The first argument for regexec() in the POSIX wrapper should have been
3234 defined as "const".
3236 4. Changed pcretest to use malloc() for its buffers so that they can be
3237 Electric Fenced for debugging.
3239 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3240 to read one or more bytes before the start of the subject string. Often this
3241 had no effect on PCRE's behaviour, but in some circumstances it could
3242 provoke a segmentation fault.
3244 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3245 to try to read one or more bytes before the start of the subject string.
3247 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3248 UTF-8 support could misbehave in various ways if the subject string
3249 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3250 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3251 back over UTF-8 characters.)
3254 Version 4.2 14-Apr-03
3255 ---------------------
3257 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3259 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3260 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3261 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3262 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3263 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3265 Note: automatic setting of the BUILD variables is not yet working
3266 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3267 compile-time but not at link-time
3268 [LINK]: use for linking executables only
3269 make different versions for Windows and non-Windows
3270 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3271 libraries
3272 [LINK_FOR_BUILD]: new variable
3273 [OBJEXT]: use throughout
3274 [EXEEXT]: use throughout
3275 <winshared>: new target
3276 <wininstall>: new target
3277 <dftables.o>: use native compiler
3278 <dftables>: use native linker
3279 <install>: handle Windows platform correctly
3280 <clean>: ditto
3281 <check>: ditto
3282 copy DLL to top builddir before testing
3284 As part of these changes, -no-undefined was removed again. This was reported
3285 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3286 in any case.
3288 3. Some tidies to get rid of compiler warnings:
3290 . In the match_data structure, match_limit was an unsigned long int, whereas
3291 match_call_count was an int. I've made them both unsigned long ints.
3293 . In pcretest the fact that a const uschar * doesn't automatically cast to
3294 a void * provoked a warning.
3296 . Turning on some more compiler warnings threw up some "shadow" variables
3297 and a few more missing casts.
3299 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3300 option, a class that contained a single character with a value between 128
3301 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3303 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3304 option, a class that contained several characters, but with at least one
3305 whose value was between 128 and 255 caused PCRE to crash.
3308 Version 4.1 12-Mar-03
3309 ---------------------
3311 1. Compiling with gcc -pedantic found a couple of places where casts were
3312 needed, and a string in dftables.c that was longer than standard compilers are
3313 required to support.
3315 2. Compiling with Sun's compiler found a few more places where the code could
3316 be tidied up in order to avoid warnings.
3318 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3319 first of these names is deprecated in the latest Autoconf in favour of the name
3320 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3321 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3322 analogy I have changed it to CFLAGS_FOR_BUILD.
3324 4. Added -no-undefined to the linking command in the Makefile, because this is
3325 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3326 linking step for the pcreposix library.
3328 5. PCRE was failing to diagnose the case of two named groups with the same
3329 name.
3331 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3332 literal character that is needed in the subject for a match, and scans along to
3333 ensure that it is present before embarking on the full matching process. This
3334 saves time in cases of nested unlimited repeats that are never going to match.
3335 Problem: the scan can take a lot of time if the subject is very long (e.g.
3336 megabytes), thus penalizing straightforward matches. It is now done only if the
3337 amount of subject to be scanned is less than 1000 bytes.
3339 7. A lesser problem with the same optimization is that it was recording the
3340 first character of an anchored pattern as "needed", thus provoking a search
3341 right along the subject, even when the first match of the pattern was going to
3342 fail. The "needed" character is now not set for anchored patterns, unless it
3343 follows something in the pattern that is of non-fixed length. Thus, it still
3344 fulfils its original purpose of finding quick non-matches in cases of nested
3345 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3348 Version 4.0 17-Feb-03
3349 ---------------------
3351 1. If a comment in an extended regex that started immediately after a meta-item
3352 extended to the end of string, PCRE compiled incorrect data. This could lead to
3353 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3355 2. Moved to autoconf 2.53 and libtool 1.4.2.
3357 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3358 the special perltest8 script is no longer needed - all the tests can be run
3359 from a single perltest script.
3361 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3362 by \s. It has now been removed in PCRE. This means it isn't recognized as
3363 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3364 class [:space:] *does* include VT, thereby creating a mess.
3366 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3367 space and tab.
3369 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3370 its new features into the main test script, reducing the number of scripts.
3372 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3373 were backward compatible, and made the (?i) apply to the whole pattern, as if
3374 /i were given. Now it behaves more logically, and applies the option setting
3375 only to what follows. PCRE has been changed to follow suit. However, if it
3376 finds options settings right at the start of the pattern, it extracts them into
3377 the global options, as before. Thus, they show up in the info data.
3379 8. Added support for the \Q...\E escape sequence. Characters in between are
3380 treated as literals. This is slightly different from Perl in that $ and @ are
3381 also handled as literals inside the quotes. In Perl, they will cause variable
3382 interpolation. Note the following examples:
3384 Pattern PCRE matches Perl matches
3386 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3387 \Qabc\$xyz\E abc\$xyz abc\$xyz
3388 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3390 For compatibility with Perl, \Q...\E sequences are recognized inside character
3391 classes as well as outside them.
3393 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3394 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3395 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3396 signed/unsigned warnings.
3398 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3399 option for pcretest, so I've replaced it by a simple function that does just
3400 that job.
3402 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3403 "pcregrep -".
3405 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3406 Java package. This provides some syntactic sugar for simple cases of what my
3407 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3408 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3409 item, you can use this simplified notation. Note that only makes sense with
3410 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3411 greediness, whatever the setting of the PCRE_UNGREEDY option.
3413 13. A change of greediness default within a pattern was not taking effect at
3414 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3415 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3416 was abstracted outside.
3418 14. PCRE now supports the \G assertion. It is true when the current matching
3419 position is at the start point of the match. This differs from \A when the
3420 starting offset is non-zero. Used with the /g option of pcretest (or similar
3421 code), it works in the same way as it does for Perl's /g option. If all
3422 alternatives of a regex begin with \G, the expression is anchored to the start
3423 match position, and the "anchored" flag is set in the compiled expression.
3425 15. Some bugs concerning the handling of certain option changes within patterns
3426 have been fixed. These applied to options other than (?ims). For example,
3427 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3428 been the other way round. Some of this was related to change 7 above.
3430 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3431 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3432 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3433 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3435 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3436 mode. Unlike ".", it always matches newline, whatever the setting of
3437 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3438 assertions. Perl allows it, but it doesn't (in general) work because it can't
3439 calculate the length of the lookbehind. At least, that's the case for Perl
3440 5.8.0 - I've been told they are going to document that it doesn't work in
3441 future.
3443 18. Added an error diagnosis for escapes that PCRE does not support: these are
3444 \L, \l, \N, \P, \p, \U, \u, and \X.
3446 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3447 reading past the end of the pattern in cases such as /[abcd/.
3449 20. PCRE was getting more memory than necessary for patterns with classes that
3450 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3452 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3453 compiling PCRE for use with Virtual Pascal.
3455 22. Small fix to the Makefile to make it work properly if the build is done
3456 outside the source tree.
3458 23. Added a new extension: a condition to go with recursion. If a conditional
3459 subpattern starts with (?(R) the "true" branch is used if recursion has
3460 happened, whereas the "false" branch is used only at the top level.
3462 24. When there was a very long string of literal characters (over 255 bytes
3463 without UTF support, over 250 bytes with UTF support), the computation of how
3464 much memory was required could be incorrect, leading to segfaults or other
3465 strange effects.
3467 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3468 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3469 there was a subsequent back reference to those brackets. This meant that, for
3470 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3471 possible to check for precisely this case. All we can do is abandon the
3472 optimization if .* occurs inside capturing brackets when there are any back
3473 references whatsoever. (See below for a better fix that came later.)
3475 26. The handling of the optimization for finding the first character of a
3476 non-anchored pattern, and for finding a character that is required later in the
3477 match were failing in some cases. This didn't break the matching; it just
3478 failed to optimize when it could. The way this is done has been re-implemented.
3480 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3482 28. Added a new feature that provides some of the functionality that Perl
3483 provides with (?{...}). The facility is termed a "callout". The way it is done
3484 in PCRE is for the caller to provide an optional function, by setting
3485 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3486 global variable. By default it is unset, which disables all calling out. To get
3487 the function called, the regex must include (?C) at appropriate points. This
3488 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3489 This provides a means of identifying different callout points. When PCRE
3490 reaches such a point in the regex, if pcre_callout has been set, the external
3491 function is called. It is provided with data in a structure called
3492 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3493 matching continues; if it returns a non-zero value, the match at the current
3494 point fails. However, backtracking will occur if possible. [This was changed
3495 later and other features added - see item 49 below.]
3497 29. pcretest is upgraded to test the callout functionality. It provides a
3498 callout function that displays information. By default, it shows the start of
3499 the match and the current position in the text. There are some new data escapes
3500 to vary what happens:
3502 \C+ in addition, show current contents of captured substrings
3503 \C- do not supply a callout function
3504 \C!n return 1 when callout number n is reached
3505 \C!n!m return 1 when callout number n is reached for the mth time
3507 30. If pcregrep was called with the -l option and just a single file name, it
3508 output "<stdin>" if a match was found, instead of the file name.
3510 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3511 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3512 pcre_exec(). This saves a malloc/free per call. The default value of
3513 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3514 when configuring.
3516 32. The default maximum size of a compiled pattern is 64K. There have been a
3517 few cases of people hitting this limit. The code now uses macros to handle the
3518 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3519 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3520 configuring. Tests 2 and 5 work only with 2-byte links because they output
3521 debugging information about compiled patterns.
3523 33. Internal code re-arrangements:
3525 (a) Moved the debugging function for printing out a compiled regex into
3526 its own source file (printint.c) and used #include to pull it into
3527 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3528 separate copies.
3530 (b) Defined the list of op-code names for debugging as a macro in
3531 internal.h so that it is next to the definition of the opcodes.
3533 (c) Defined a table of op-code lengths for simpler skipping along compiled
3534 code. This is again a macro in internal.h so that it is next to the
3535 definition of the opcodes.
3537 34. Added support for recursive calls to individual subpatterns, along the
3538 lines of Robin Houston's patch (but implemented somewhat differently).
3540 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3541 allow it to read and process whole directories in Win32. This code was
3542 contributed by Lionel Fourquaux; it has not been tested by me.
3544 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3545 used to name a group. Names consist of alphanumerics and underscores, and must
3546 be unique. Back references use the syntax (?P=name) and recursive calls use
3547 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3548 numbers. The function pcre_fullinfo() can be used after compilation to extract
3549 a name/number map. There are three relevant calls:
3551 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3552 PCRE_INFO_NAMECOUNT yields the number of entries
3553 PCRE_INFO_NAMETABLE yields a pointer to the map.
3555 The map is a vector of fixed-size entries. The size of each entry depends on
3556 the length of the longest name used. The first two bytes of each entry are the
3557 group number, most significant byte first. There follows the corresponding
3558 name, zero terminated. The names are in alphabetical order.
3560 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3561 case instead of 255. Making it the same both with and without UTF-8 support
3562 means that the same test output works with both.
3564 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3565 calling malloc() with a zero argument.
3567 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3568 optimization. I've improved things by keeping a bitmap of backreferences with
3569 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3570 fact referenced, the optimization can be applied. It is unlikely that a
3571 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3572 the match to follow \n) will appear inside brackets with a number greater than
3573 31, but if it does, any back reference > 31 suppresses the optimization.
3575 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3576 of disabling numbered capturing parentheses. Any opening parenthesis that is
3577 not followed by ? behaves as if it were followed by ?: but named parentheses
3578 can still be used for capturing (and they will acquire numbers in the usual
3579 way).
3581 41. Redesigned the return codes from the match() function into yes/no/error so
3582 that errors can be passed back from deep inside the nested calls. A malloc
3583 failure while inside a recursive subpattern call now causes the
3584 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3586 42. It is now possible to set a limit on the number of times the match()
3587 function is called in a call to pcre_exec(). This facility makes it possible to
3588 limit the amount of recursion and backtracking, though not in a directly
3589 obvious way, because the match() function is used in a number of different
3590 circumstances. The count starts from zero for each position in the subject
3591 string (for non-anchored patterns). The default limit is, for compatibility, a
3592 large number, namely 10 000 000. You can change this in two ways:
3594 (a) When configuring PCRE before making, you can use --with-match-limit=n
3595 to set a default value for the compiled library.
3597 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3598 a different value is set. See 45 below.
3600 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3602 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3603 of things that can be changed at compile time. The first argument specifies
3604 what is wanted and the second points to where the information is to be placed.
3605 The current list of available information is:
3609 The output is an integer that is set to one if UTF-8 support is available;
3610 otherwise it is set to zero.
3614 The output is an integer that it set to the value of the code that is used for
3615 newline. It is either LF (10) or CR (13).
3619 The output is an integer that contains the number of bytes used for internal
3620 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3624 The output is an integer that contains the threshold above which the POSIX
3625 interface uses malloc() for output vectors. See item 31 above.
3629 The output is an unsigned integer that contains the default limit of the number
3630 of match() calls in a pcre_exec() execution. See 42 above.
3632 44. pcretest has been upgraded by the addition of the -C option. This causes it
3633 to extract all the available output from the new pcre_config() function, and to
3634 output it. The program then exits immediately.
3636 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3637 order to support additional features. One way would have been to define
3638 pcre_exec2() (for example) with extra arguments, but this would not have been
3639 extensible, and would also have required all calls to the original function to
3640 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3641 is used for passing in "extra" data from pcre_study().
3643 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3644 contains the following fields:
3646 flags a bitmap indicating which of the following fields are set
3647 study_data opaque data from pcre_study()
3648 match_limit a way of specifying a limit on match() calls for a specific
3649 call to pcre_exec()
3650 callout_data data for callouts (see 49 below)
3652 The flag bits are also defined in pcre.h, and are
3658 The pcre_study() function now returns one of these new pcre_extra blocks, with
3659 the actual study data pointed to by the study_data field, and the
3660 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3661 before. That is, this change is entirely upwards-compatible and requires no
3662 change to existing code.
3664 If you want to pass in additional data to pcre_exec(), you can either place it
3665 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3666 block.
3668 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3669 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3670 times with different match limits, until it finds the minimum value needed for
3671 pcre_exec() to complete. The value is then output. This can be instructive; for
3672 most simple matches the number is quite small, but for pathological cases it
3673 gets very large very quickly.
3675 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3676 returns the size of the data block pointed to by the study_data field in a
3677 pcre_extra block, that is, the value that was passed as the argument to
3678 pcre_malloc() when PCRE was getting memory in which to place the information
3679 created by pcre_study(). The fourth argument should point to a size_t variable.
3680 pcretest has been extended so that this information is shown after a successful
3681 pcre_study() call when information about the compiled regex is being displayed.
3683 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3684 because what follows is always an absolute path. (Later: it turns out that this
3685 is more than cosmetic for MinGW, because it doesn't like empty path
3686 components.)
3688 49. Some changes have been made to the callout feature (see 28 above):
3690 (i) A callout function now has three choices for what it returns:
3692 0 => success, carry on matching
3693 > 0 => failure at this point, but backtrack if possible
3694 < 0 => serious error, return this value from pcre_exec()
3696 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3697 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3698 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3699 use by callout functions. It will never be used by PCRE itself.
3701 (ii) The pcre_extra structure (see 45 above) has a void * field called
3702 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3703 pcre_callout_block structure has a field of the same name. The contents of
3704 the field passed in the pcre_extra structure are passed to the callout
3705 function in the corresponding field in the callout block. This makes it
3706 easier to use the same callout-containing regex from multiple threads. For
3707 testing, the pcretest program has a new data escape
3709 \C*n pass the number n (may be negative) as callout_data
3711 If the callout function in pcretest receives a non-zero value as
3712 callout_data, it returns that value.
3714 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3715 there were some redundant $(CFLAGS) in commands that are now specified as
3716 $(LINK), which already includes $(CFLAGS).
3718 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3719 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3720 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3721 one-byte characters throughout. Note that case-insensitive matching applies
3722 only to characters whose values are less than 256. PCRE doesn't support the
3723 notion of cases for higher-valued characters.
3725 (i) A character class whose characters are all within 0-255 is handled as
3726 a bit map, and the map is inverted for negative classes. Previously, a
3727 character > 255 always failed to match such a class; however it should
3728 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3730 (ii) A negated character class with a single character < 255 is coded as
3731 "not this character" (OP_NOT). This wasn't working properly when the test
3732 character was multibyte, either singly or repeated.
3734 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3735 mode, for example: \x{100}{2,3}.
3737 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3738 singly or repeated) now correctly test multibyte characters. However,
3739 PCRE doesn't recognize any characters with values greater than 255 as
3740 digits, spaces, or word characters. Such characters always match \D, \S,
3741 and \W, and never match \d, \s, or \w.
3743 (v) Classes may now contain characters and character ranges with values
3744 greater than 255. For example: [ab\x{100}-\x{400}].
3746 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3747 PCRE in UTF-8 mode.
3749 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3750 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3751 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3752 value.)
3754 53. The single man page has become too large. I have therefore split it up into
3755 a number of separate man pages. These also give rise to individual HTML pages;
3756 these are now put in a separate directory, and there is an index.html page that
3757 lists them all. Some hyperlinking between the pages has been installed.
3759 54. Added convenience functions for handling named capturing parentheses.
3761 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3762 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3763 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3764 are faulted.
3766 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3767 calling configure. These values are used when compiling the dftables.c program
3768 which is run to generate the source of the default character tables. They
3769 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3770 you will need to set these values.
3772 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3775 Version 3.9 02-Jan-02
3776 ---------------------
3778 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3780 2. If --disable-static was given, the building process failed when trying to
3781 build pcretest and pcregrep. (For some reason it was using libtool to compile
3782 them, which is not right, as they aren't part of the library.)
3785 Version 3.8 18-Dec-01
3786 ---------------------
3788 1. The experimental UTF-8 code was completely screwed up. It was packing the
3789 bytes in the wrong order. How dumb can you get?
3792 Version 3.7 29-Oct-01
3793 ---------------------
3795 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3796 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3797 this didn't happen under Solaris 8, where I normally test things.
3799 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3800 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3801 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3802 this fix an hour or so after the initial 3.7 release.)
3805 Version 3.6 23-Oct-01
3806 ---------------------
3808 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3809 offsets passed as NULL with zero offset count.
3811 2. The config.guess and config.sub files had not been updated when I moved to
3812 the latest autoconf.
3815 Version 3.5 15-Aug-01
3816 ---------------------
3818 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3819 had been forgotten.
3821 2. By using declared but undefined structures, we can avoid using "void"
3822 definitions in pcre.h while keeping the internal definitions of the structures
3823 private.
3825 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3826 user point of view, this means that both static and shared libraries are built
3827 by default, but this can be individually controlled. More of the work of
3828 handling this static/shared cases is now inside libtool instead of PCRE's make
3829 file.
3831 4. The pcretest utility is now installed along with pcregrep because it is
3832 useful for users (to test regexs) and by doing this, it automatically gets
3833 relinked by libtool. The documentation has been turned into a man page, so
3834 there are now .1, .txt, and .html versions in /doc.
3836 5. Upgrades to pcregrep:
3837 (i) Added long-form option names like gnu grep.
3838 (ii) Added --help to list all options with an explanatory phrase.
3839 (iii) Added -r, --recursive to recurse into sub-directories.
3840 (iv) Added -f, --file to read patterns from a file.
3842 6. pcre_exec() was referring to its "code" argument before testing that
3843 argument for NULL (and giving an error if it was NULL).
3845 7. Upgraded Makefile.in to allow for compiling in a different directory from
3846 the source directory.
3848 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3849 options bits, the pointer it was passed was to an int instead of to an unsigned
3850 long int. This mattered only on 64-bit systems.
3852 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3853 generated) instead of pcre.in, which it its source. Also made the same change
3854 in several of the .c files.
3856 10. A new release of gcc defines printf() as a macro, which broke pcretest
3857 because it had an ifdef in the middle of a string argument for printf(). Fixed
3858 by using separate calls to printf().
3860 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3861 script, to force use of CR or LF instead of \n in the source. On non-Unix
3862 systems, the value can be set in config.h.
3864 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3865 absolute limit. Changed the text of the error message to make this clear, and
3866 likewise updated the man page.
3868 13. The limit of 99 on the number of capturing subpatterns has been removed.
3869 The new limit is 65535, which I hope will not be a "real" limit.
3872 Version 3.4 22-Aug-00
3873 ---------------------
3875 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3877 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3880 Version 3.3 01-Aug-00
3881 ---------------------
3883 1. If an octal character was given, but the value was greater than \377, it
3884 was not getting masked to the least significant bits, as documented. This could
3885 lead to crashes in some systems.
3887 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3888 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3890 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3891 These just pass their arguments on to (pcre_free)(), but they are provided
3892 because some uses of PCRE bind it to non-C systems that can call its functions,
3893 but cannot call free() or pcre_free() directly.
3895 4. Add "make test" as a synonym for "make check". Corrected some comments in
3896 the Makefile.
3898 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3899 Makefile.
3901 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3902 command called pgrep for grepping around the active processes.
3904 7. Added the beginnings of support for UTF-8 character strings.
3906 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3907 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3908 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3909 out for the ar command.)
3912 Version 3.2 12-May-00
3913 ---------------------
3915 This is purely a bug fixing release.
3917 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3918 of ZA. This was just one example of several cases that could provoke this bug,
3919 which was introduced by change 9 of version 2.00. The code for breaking
3920 infinite loops after an iteration that matches an empty string was't working
3921 correctly.
3923 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3924 when matched against abbab (for example). After matching an empty string, it
3925 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3926 caused it to match further down the string than it should.
3928 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3929 was there because it doesn't seem to be needed, and it causes trouble on some
3930 systems, as it is not a Standard C header. It has been removed.
3932 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3933 were reported on the Macintosh. The changes were from
3935 while ((c = *(++ptr)) != 0 && c != '\n');
3936 to
3937 while ((c = *(++ptr)) != 0 && c != '\n') ;
3939 Totally extraordinary, but if that's what it takes...
3941 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3942 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3943 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3944 assumes the way PCRE uses memmove() (always moving upwards).
3946 6. PCRE is being used in one environment where strchr() is not available. There
3947 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3948 faster code anyway.
3951 Version 3.1 09-Feb-00
3952 ---------------------
3954 The only change in this release is the fixing of some bugs in Makefile.in for
3955 the "install" target:
3957 (1) It was failing to install pcreposix.h.
3959 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3962 Version 3.0 01-Feb-00
3963 ---------------------
3965 1. Add support for the /+ modifier to perltest (to output $` like it does in
3966 pcretest).
3968 2. Add support for the /g modifier to perltest.
3970 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3971 matches null strings.
3973 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3974 pattern. Perl treats empty patterns specially - it reuses the most recent
3975 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3976 effect.
3978 5. The POSIX interface was broken in that it was just handing over the POSIX
3979 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3980 required a bigger vector, with some working space on the end. This means that
3981 the POSIX wrapper now has to get and free some memory, and copy the results.
3983 6. Added some simple autoconf support, placing the test data and the
3984 documentation in separate directories, re-organizing some of the
3985 information files, and making it build pcre-config (a GNU standard). Also added
3986 libtool support for building PCRE as a shared library, which is now the
3987 default.
3989 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3990 09 are not valid octal constants. Single digits will be used for minor values
3991 less than 10.
3993 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3994 existing programs that set these in the POSIX interface can use PCRE without
3995 modification.
3997 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3998 return all that pcre_info() returns, plus additional data. The pcre_info()
3999 function is retained for compatibility, but is considered to be obsolete.
4001 10. Added experimental recursion feature (?R) to handle one common case that
4002 Perl 5.6 will be able to do with (?p{...}).
4004 11. Added support for POSIX character classes like [:alpha:], which Perl is
4005 adopting.
4008 Version 2.08 31-Aug-99
4009 ----------------------
4011 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4012 trying to match at the startoffset position, but instead was moving forward to
4013 the next newline as if a previous match had failed.
4015 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4016 and could get into a loop if a null string was matched other than at the start
4017 of the subject.
4019 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4020 be distinguished at compile time, and for completeness also added PCRE_DATE.
4022 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4023 in GnuWin32 environments.
4026 Version 2.07 29-Jul-99
4027 ----------------------
4029 1. The documentation is now supplied in plain text form and HTML as well as in
4030 the form of man page sources.
4032 2. C++ compilers don't like assigning (void *) values to other pointer types.
4033 In particular this affects malloc(). Although there is no problem in Standard
4034 C, I've put in casts to keep C++ compilers happy.
4036 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4037 should be (const char *).
4039 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4040 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4041 However, I haven't made this a standard facility. The documentation doesn't
4042 mention it, and the Makefile doesn't support it.
4044 5. The Makefile now contains an "install" target, with editable destinations at
4045 the top of the file. The pcretest program is not installed.
4047 6. pgrep -V now gives the PCRE version number and date.
4049 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4050 causing the entire string to be ignored, instead of just the last character.
4052 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4053 non-matching string, it can take a very, very long time, even for strings of
4054 quite modest length, because of the nested recursion. PCRE now does better in
4055 some of these cases. It does this by remembering the last required literal
4056 character in the pattern, and pre-searching the subject to ensure it is present
4057 before running the real match. In other words, it applies a heuristic to detect
4058 some types of certain failure quickly, and in the above example, if presented
4059 with a string that has no trailing " it gives "no match" very quickly.
4061 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4062 other alternatives are tried instead.
4065 Version 2.06 09-Jun-99
4066 ----------------------
4068 1. Change pcretest's output for amount of store used to show just the code
4069 space, because the remainder (the data block) varies in size between 32-bit and
4070 64-bit systems.
4072 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4073 start matching at. This allows lookbehinds to work when searching for multiple
4074 occurrences in a string.
4076 3. Added additional options to pcretest for testing multiple occurrences:
4078 /+ outputs the rest of the string that follows a match
4079 /g loops for multiple occurrences, using the new startoffset argument
4080 /G loops for multiple occurrences by passing an incremented pointer
4082 4. PCRE wasn't doing the "first character" optimization for patterns starting
4083 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4084 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4085 the letter 'x'. On long subject strings, this gives a significant speed-up.
4088 Version 2.05 21-Apr-99
4089 ----------------------
4091 1. Changed the type of magic_number from int to long int so that it works
4092 properly on 16-bit systems.
4094 2. Fixed a bug which caused patterns starting with .* not to work correctly
4095 when the subject string contained newline characters. PCRE was assuming
4096 anchoring for such patterns in all cases, which is not correct because .* will
4097 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4098 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4099 must be retried after every newline in the subject.
4102 Version 2.04 18-Feb-99
4103 ----------------------
4105 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4106 computation of the store needed to hold the pattern was incorrect (too large).
4107 If such patterns were nested a few deep, this could multiply and become a real
4108 problem.
4110 2. Added /M option to pcretest to show the memory requirement of a specific
4111 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4113 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4114 compiled in such a way that the backtracking after subsequent failure was
4115 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4116 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4119 Version 2.03 02-Feb-99
4120 ----------------------
4122 1. Fixed typo and small mistake in man page.
4124 2. Added 4th condition (GPL supersedes if conflict) and created separate
4125 LICENCE file containing the conditions.
4127 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4128 Perl, that is the internal \ allows the delimiter to be included in the
4129 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4130 the final delimiter, add \ to the end of the pattern (to test the error).
4132 4. Added the convenience functions for extracting substrings after a successful
4133 match. Updated pcretest to make it able to test these functions.
4136 Version 2.02 14-Jan-99
4137 ----------------------
4139 1. Initialized the working variables associated with each extraction so that
4140 their saving and restoring doesn't refer to uninitialized store.
4142 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4143 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4144 fix the problem.
4146 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4147 calls, and wasn't printing the correct value for compiling calls. Increased the
4148 default value of LOOPREPEAT, and the number of significant figures in the
4149 times.
4151 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4153 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4154 a building problem on Windows NT with a FAT file system.
4157 Version 2.01 21-Oct-98
4158 ----------------------
4160 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4161 to character tables built by pcre_maketables() in the current locale. If NULL
4162 is passed, the default tables are used.
4165 Version 2.00 24-Sep-98
4166 ----------------------
4168 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4169 it any more.
4171 2. Allow quantification of (?>) groups, and make it work correctly.
4173 3. The first character computation wasn't working for (?>) groups.
4175 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4176 end of the subject) and add 5.005's \z, which really does match only at the
4177 very end of the subject.
4179 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4181 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4182 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4183 localized options. All options to pcre_study() were also removed.
4185 7. Add other new features from 5.005:
4187 $(?<= positive lookbehind
4188 $(?<! negative lookbehind
4189 (?imsx-imsx) added the unsetting capability
4190 such a setting is global if at outer level; local otherwise
4191 (?imsx-imsx:) non-capturing groups with option setting
4192 (?(cond)re|re) conditional pattern matching
4194 A backreference to itself in a repeated group matches the previous
4195 captured string.
4197 8. General tidying up of studying (both automatic and via "study")
4198 consequential on the addition of new assertions.
4200 9. As in 5.005, unlimited repeated groups that could match an empty substring
4201 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4202 runtime if any iteration does actually match an empty substring.
4204 10. Include the RunTest script in the distribution.
4206 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4207 discrepancies, some of which were old and were also with respect to 5.004. They
4208 have now been fixed.
4211 Version 1.09 28-Apr-98
4212 ----------------------
4214 1. A negated single character class followed by a quantifier with a minimum
4215 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4216 program crashes, or just wrong answers. This did not apply to negated classes
4217 containing more than one character, or to minima other than one.
4220 Version 1.08 27-Mar-98
4221 ----------------------
4223 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4225 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4226 latter must appear before anything that relies on it in the pattern.
4229 Version 1.07 16-Feb-98
4230 ----------------------
4232 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4233 repeat of a potentially empty string).
4236 Version 1.06 23-Jan-98
4237 ----------------------
4239 1. Added Markus Oberhumer's little patches for C++.
4241 2. Literal strings longer than 255 characters were broken.
4244 Version 1.05 23-Dec-97
4245 ----------------------
4247 1. Negated character classes containing more than one character were failing if
4248 PCRE_CASELESS was set at run time.
4251 Version 1.04 19-Dec-97
4252 ----------------------
4254 1. Corrected the man page, where some "const" qualifiers had been omitted.
4256 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4257 input syntax.
4259 3. Fixed memory leak which occurred when a regex with back references was
4260 matched with an offsets vector that wasn't big enough. The temporary memory
4261 that is used in this case wasn't being freed if the match failed.
4263 4. Tidied pcretest to ensure it frees memory that it gets.
4265 5. Temporary memory was being obtained in the case where the passed offsets
4266 vector was exactly big enough.
4268 6. Corrected definition of offsetof() from change 5 below.
4270 7. I had screwed up change 6 below and broken the rules for the use of
4271 setjmp(). Now fixed.
4274 Version 1.03 18-Dec-97
4275 ----------------------
4277 1. A erroneous regex with a missing opening parenthesis was correctly
4278 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4279 on some systems.
4281 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4282 it was reported that one broken compiler failed on the former because "code" is
4283 also an independent variable.
4285 3. The erroneous regex a[]b caused an array overrun reference.
4287 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4288 fail on data ending with that character. (It was going on too far, and checking
4289 the next character, typically a binary zero.) This was specific to the
4290 optimized code for single-character negative classes.
4292 5. Added a contributed patch from the TIN world which does the following:
4294 + Add an undef for memmove, in case the the system defines a macro for it.
4296 + Add a definition of offsetof(), in case there isn't one. (I don't know
4297 the reason behind this - offsetof() is part of the ANSI standard - but
4298 it does no harm).
4300 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4301 most of the places where whitespace preceded '#'. I have given up and
4302 allowed the remaining 2 cases to be at the margin.
4304 + Rename some variables in pcre to eliminate shadowing. This seems very
4305 pedantic, but does no harm, of course.
4307 6. Moved the call to setjmp() into its own function, to get rid of warnings
4308 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4310 7. Constructs such as \d{8,} were compiling into the equivalent of
4311 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4312 outcome, but in this particular case used more store than had been allocated,
4313 which caused the bug to be discovered because it threw up an internal error.
4315 8. The debugging code in both pcre and pcretest for outputting the compiled
4316 form of a regex was going wrong in the case of back references followed by
4317 curly-bracketed repeats.
4320 Version 1.02 12-Dec-97
4321 ----------------------
4323 1. Typos in pcre.3 and comments in the source fixed.
4325 2. Applied a contributed patch to get rid of places where it used to remove
4326 'const' from variables, and fixed some signed/unsigned and uninitialized
4327 variable warnings.
4329 3. Added the "runtest" target to Makefile.
4331 4. Set default compiler flag to -O2 rather than just -O.
4334 Version 1.01 19-Nov-97
4335 ----------------------
4337 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4338 like /([ab]*)*/, that is, for classes with more than one character in them.
4340 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4341 as /((?>a*))*/ (a PCRE_EXTRA facility).
4344 Version 1.00 18-Nov-97
4345 ----------------------
4347 1. Added compile-time macros to support systems such as SunOS4 which don't have
4348 memmove() or strerror() but have other things that can be used instead.
4350 2. Arranged that "make clean" removes the executables.
4353 Version 0.99 27-Oct-97
4354 ----------------------
4356 1. Fixed bug in code for optimizing classes with only one character. It was
4357 initializing a 32-byte map regardless, which could cause it to run off the end
4358 of the memory it had got.
4360 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4363 Version 0.98 22-Oct-97
4364 ----------------------
4366 1. Fixed bug in code for handling temporary memory usage when there are more
4367 back references than supplied space in the ovector. This could cause segfaults.
4370 Version 0.97 21-Oct-97
4371 ----------------------
4373 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4375 2. Optimized negated single characters not to use a bit map.
4377 3. Brought error texts together as macro definitions; clarified some of them;
4378 fixed one that was wrong - it said "range out of order" when it meant "invalid
4379 escape sequence".
4381 4. Changed some char * arguments to const char *.
4383 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4385 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4386 pcretest.
4389 Version 0.96 16-Oct-97
4390 ----------------------
4392 1. Added a simple "pgrep" utility to the distribution.
4394 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4395 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4396 where "ddd" means "one or more decimal digits".
4398 3. Fixed serious bug. If a pattern had a back reference, but the call to
4399 pcre_exec() didn't supply a large enough ovector to record the related
4400 identifying subpattern, the match always failed. PCRE now remembers the number
4401 of the largest back reference, and gets some temporary memory in which to save
4402 the offsets during matching if necessary, in order to ensure that
4403 backreferences always work.
4405 4. Increased the compatibility with Perl in a number of ways:
4407 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4408 to request this handling. The option can be set at compile or exec time.
4410 (b) $ matches before a terminating newline by default; an option
4411 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4412 mode). The option can be set at compile or exec time.
4414 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4415 the same as Perl's. If the decimal number it represents is less than 10
4416 or there aren't that many previous left capturing parentheses, an octal
4417 escape is read. Inside a character class, it's always an octal escape,
4418 even if it is a single digit.
4420 (d) An escaped but undefined alphabetic character is taken as a literal,
4421 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4422 escapes.
4424 (e) {0} is now permitted. (The previous item is removed from the compiled
4425 pattern).
4427 5. Changed all the names of code files so that the basic parts are no longer
4428 than 10 characters, and abolished the teeny "globals.c" file.
4430 6. Changed the handling of character classes; they are now done with a 32-byte
4431 bit map always.
4433 7. Added the -d and /D options to pcretest to make it possible to look at the
4434 internals of compilation without having to recompile pcre.
4437 Version 0.95 23-Sep-97
4438 ----------------------
4440 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4441 \x20 at the start of a run of normal characters. These were being treated as
4442 real characters, instead of the source characters being re-checked.
4445 Version 0.94 18-Sep-97
4446 ----------------------
4448 1. The functions are now thread-safe, with the caveat that the global variables
4449 containing pointers to malloc() and free() or alternative functions are the
4450 same for all threads.
4452 2. Get pcre_study() to generate a bitmap of initial characters for non-
4453 anchored patterns when this is possible, and use it if passed to pcre_exec().
4456 Version 0.93 15-Sep-97
4457 ----------------------
4459 1. /(b)|(:+)/ was computing an incorrect first character.
4461 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4462 but not actually doing anything yet.
4464 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4465 as Perl does (e.g. [-az] or [az-]).
4467 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4468 all possible positions.
4470 5. Split up into different modules to avoid including unneeded functions in a
4471 compiled binary. However, compile and exec are still in one module. The "study"
4472 function is split off.
4474 6. The character tables are now in a separate module whose source is generated
4475 by an auxiliary program - but can then be edited by hand if required. There are
4476 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4477 toupper() in the code.
4479 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4480 make them global. Abolish the function for setting them, as the caller can now
4481 set them directly.
4484 Version 0.92 11-Sep-97
4485 ----------------------
4487 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4488 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4490 2. Caseless matching was not working in character classes if the characters in
4491 the pattern were in upper case.
4493 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4495 4. Make PCRE_ANCHORED public and accept as a compile option.
4497 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4498 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4499 pass them.
4501 6. Give an error if bad option bits passed at compile or run time.
4503 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4504 pcretest to cause it to pass that flag.
4506 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4507 options, and the first character, if set.
4509 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4512 Version 0.91 10-Sep-97
4513 ----------------------
4515 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4516 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4518 2. PCRE was looping on encountering an indefinitely repeated back reference to
4519 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4520 Perl does - treats the match as successful.
4522 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5