/[pcre]/code/trunk/ChangeLog
ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 719 - (show annotations) (download)
Fri Oct 7 11:41:05 2011 UTC (3 years ago) by ph10
File size: 216413 byte(s)
Another improved RunTest.bat and also CMakeLists.txt provided by Sheri P.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.20 23-Sep-2011
5 ------------------------
6
7 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
8 a POSIX class. After further experiments with Perl, which convinced me that
9 Perl has bugs and confusions, a closing square bracket is no longer allowed
10 in a POSIX name. This bug also affected patterns with classes that started
11 with full stops.
12
13 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
14 captured substring, but while checking the failing first alternative,
15 substring 1 is temporarily captured. If the output vector supplied to
16 pcre_exec() was not big enough for this capture, the yield of the function
17 was still zero ("insufficient space for captured substrings"). This cannot
18 be totally fixed without adding another stack variable, which seems a lot
19 of expense for a edge case. However, I have improved the situation in cases
20 such as /(a)(b)x|abc/ matched against "abc", where the return code
21 indicates that fewer than the maximum number of slots in the ovector have
22 been set.
23
24 3. Related to (2) above: when there are more back references in a pattern than
25 slots in the output vector, pcre_exec() uses temporary memory during
26 matching, and copies in the captures as far as possible afterwards. It was
27 using the entire output vector, but this conflicts with the specification
28 that only 2/3 is used for passing back captured substrings. Now it uses
29 only the first 2/3, for compatibility. This is, of course, another edge
30 case.
31
32 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
33 main code base, and can be used by building with --enable-jit. When this is
34 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
35 runtime --no-jit option is given.
36
37 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
38 ovector, the return from the function was zero, implying that there were
39 other matches that did not fit. The correct "exactly full" value is now
40 returned.
41
42 6. If a subpattern that was called recursively or as a subroutine contained
43 (*PRUNE) or any other control that caused it to give a non-standard return,
44 invalid errors such as "Error -26 (nested recursion at the same subject
45 position)" or even infinite loops could occur.
46
47 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
48 computing the minimum length on reaching *ACCEPT, and so ended up with the
49 wrong value of 1 rather than 0. Further investigation indicates that
50 computing a minimum subject length in the presence of *ACCEPT is difficult
51 (think back references, subroutine calls), and so I have changed the code
52 so that no minimum is registered for a pattern that contains *ACCEPT.
53
54 8. If (*THEN) was present in the first (true) branch of a conditional group,
55 it was not handled as intended. [But see 16 below.]
56
57 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
58 Sheri Pierce.
59
60 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
61 the first byte in a match must be "a".
62
63 11. Change 17 for 8.13 increased the recursion depth for patterns like
64 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
65 pattern contains any instances of (*THEN). If it does not, the old
66 optimizations are restored. It would be nice to do this on a per-group
67 basis, but at the moment that is not feasible.
68
69 12. In some environments, the output of pcretest -C is CRLF terminated. This
70 broke RunTest's code that checks for the link size. A single white space
71 after the value is now allowed for.
72
73 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
74 For "fr", it uses the Windows-specific input and output files.
75
76 14. If (*THEN) appeared in a group that was called recursively or as a
77 subroutine, it did not work as intended. [But see next item.]
78
79 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
80 pattern fragments (but not containing any | characters). If A and B are
81 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
82 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
83 D. In other words, Perl considers parentheses that do not contain any |
84 characters to be part of a surrounding alternative, whereas PCRE was
85 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
86 differently. PCRE now behaves in the same way as Perl, except in the case
87 of subroutine/recursion calls such as (?1) which have in any case always
88 been different (but PCRE had them first :-).
89
90 16. Related to 15 above: Perl does not treat the | in a conditional group as
91 creating alternatives. Such a group is treated in the same way as an
92 ordinary group without any | characters when processing (*THEN). PCRE has
93 been changed to match Perl's behaviour.
94
95
96 Version 8.13 16-Aug-2011
97 ------------------------
98
99 1. The Unicode data tables have been updated to Unicode 6.0.0.
100
101 2. Two minor typos in pcre_internal.h have been fixed.
102
103 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
104 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
105 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
106
107 4. There were a number of related bugs in the code for matching backrefences
108 caselessly in UTF-8 mode when codes for the characters concerned were
109 different numbers of bytes. For example, U+023A and U+2C65 are an upper
110 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
111 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
112 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
113 2-byte code at the end of the subject (it thought there wasn't enough data
114 left).
115
116 5. Comprehensive information about what went wrong is now returned by
117 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
118 as the output vector has at least 2 elements. The offset of the start of
119 the failing character and a reason code are placed in the vector.
120
121 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
122 now returned is for the first byte of the failing character, instead of the
123 last byte inspected. This is an incompatible change, but I hope it is small
124 enough not to be a problem. It makes the returned offset consistent with
125 pcre_exec() and pcre_dfa_exec().
126
127 7. pcretest now gives a text phrase as well as the error number when
128 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
129 failure, the offset and reason code are output.
130
131 8. When \R was used with a maximizing quantifier it failed to skip backwards
132 over a \r\n pair if the subsequent match failed. Instead, it just skipped
133 back over a single character (\n). This seems wrong (because it treated the
134 two characters as a single entity when going forwards), conflicts with the
135 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
136 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
137 has been changed.
138
139 9. Some internal refactoring has changed the processing so that the handling
140 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
141 time (the PCRE_DOTALL option was changed this way some time ago: version
142 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
143 which was always a bit of a fudge. It also means that there is one less
144 argument for the match() function, which reduces its stack requirements
145 slightly. This change also fixes an incompatibility with Perl: the pattern
146 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
147
148 10. More internal refactoring has drastically reduced the number of recursive
149 calls to match() for possessively repeated groups such as (abc)++ when
150 using pcre_exec().
151
152 11. While implementing 10, a number of bugs in the handling of groups were
153 discovered and fixed:
154
155 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
156 (a|)*(?1) gave a compile-time internal error.
157 ((a|)+)+ did not notice that the outer group could match an empty string.
158 (^a|^)+ was not marked as anchored.
159 (.*a|.*)+ was not marked as matching at start or after a newline.
160
161 12. Yet more internal refactoring has removed another argument from the match()
162 function. Special calls to this function are now indicated by setting a
163 value in a variable in the "match data" data block.
164
165 13. Be more explicit in pcre_study() instead of relying on "default" for
166 opcodes that mean there is no starting character; this means that when new
167 ones are added and accidentally left out of pcre_study(), testing should
168 pick them up.
169
170 14. The -s option of pcretest has been documented for ages as being an old
171 synonym of -m (show memory usage). I have changed it to mean "force study
172 for every regex", that is, assume /S for every regex. This is similar to -i
173 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
174 using it. It makes it easier to run collections of tests with and without
175 study enabled, and thereby test pcre_study() more easily. All the standard
176 tests are now run with and without -s (but some patterns can be marked as
177 "never study" - see 20 below).
178
179 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
180 restoration of the capturing data to the outer values was not happening
181 correctly.
182
183 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
184 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
185 pattern had matched an empty string, and so incorrectly returned a no
186 match.
187
188 17. There was optimizing code for the last branch of non-capturing parentheses,
189 and also for the obeyed branch of a conditional subexpression, which used
190 tail recursion to cut down on stack usage. Unfortunately, now that there is
191 the possibility of (*THEN) occurring in these branches, tail recursion is
192 no longer possible because the return has to be checked for (*THEN). These
193 two optimizations have therefore been removed. [But see 8.20/11 above.]
194
195 18. If a pattern containing \R was studied, it was assumed that \R always
196 matched two bytes, thus causing the minimum subject length to be
197 incorrectly computed because \R can also match just one byte.
198
199 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
200 was incorrectly computed.
201
202 20. If /S is present twice on a test pattern in pcretest input, it now
203 *disables* studying, thereby overriding the use of -s on the command line
204 (see 14 above). This is necessary for one or two tests to keep the output
205 identical in both cases.
206
207 21. When (*ACCEPT) was used in an assertion that matched an empty string and
208 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
209
210 22. When an atomic group that contained a capturing parenthesis was
211 successfully matched, but the branch in which it appeared failed, the
212 capturing was not being forgotten if a higher numbered group was later
213 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
214 group 1 to "a", when in fact it should be unset. This applied to multi-
215 branched capturing and non-capturing groups, repeated or not, and also to
216 positive assertions (capturing in negative assertions does not happen
217 in PCRE) and also to nested atomic groups.
218
219 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
220 subject after a captured substring, to make it easier to tell which of a
221 number of identical substrings has been captured.
222
223 24. The way atomic groups are processed by pcre_exec() has been changed so that
224 if they are repeated, backtracking one repetition now resets captured
225 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
226 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
227 "aaa". Previously, it would have been "a". As part of this code
228 refactoring, the way recursive calls are handled has also been changed.
229
230 25. If an assertion condition captured any substrings, they were not passed
231 back unless some other capturing happened later. For example, if
232 (?(?=(a))a) was matched against "a", no capturing was returned.
233
234 26. When studying a pattern that contained subroutine calls or assertions,
235 the code for finding the minimum length of a possible match was handling
236 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
237 group 1 called group 2 while simultaneously a separate group 2 called group
238 1). A stack overflow occurred in this case. I have fixed this by limiting
239 the recursion depth to 10.
240
241 27. Updated RunTest.bat in the distribution to the version supplied by Tom
242 Fortmann. This supports explicit test numbers on the command line, and has
243 argument validation and error reporting.
244
245 28. An instance of \X with an unlimited repeat could fail if at any point the
246 first character it looked at was a mark character.
247
248 29. Some minor code refactoring concerning Unicode properties and scripts
249 should reduce the stack requirement of match() slightly.
250
251 30. Added the '=' option to pcretest to check the setting of unused capturing
252 slots at the end of the pattern, which are documented as being -1, but are
253 not included in the return count.
254
255 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
256 compiled something random. Now it gives a compile-time error (as does
257 Perl).
258
259 32. A *MARK encountered during the processing of a positive assertion is now
260 recorded and passed back (compatible with Perl).
261
262 33. If --only-matching or --colour was set on a pcregrep call whose pattern
263 had alternative anchored branches, the search for a second match in a line
264 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
265 matched the line "0102" twice. The same bug affected patterns that started
266 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
267 twice.
268
269 34. Previously, PCRE did not allow quantification of assertions. However, Perl
270 does, and because of capturing effects, quantifying parenthesized
271 assertions may at times be useful. Quantifiers are now allowed for
272 parenthesized assertions.
273
274 35. A minor code tidy in pcre_compile() when checking options for \R usage.
275
276 36. \g was being checked for fancy things in a character class, when it should
277 just be a literal "g".
278
279 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
280 appearance of a nested POSIX class supersedes an apparent external class.
281 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
282 unescaped square brackets may also appear as part of class names. For
283 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
284 more like Perl. (But see 8.20/1 above.)
285
286 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
287 was because it thought it was \N{name}, which is not supported).
288
289 39. Add minix to OS list not supporting the -S option in pcretest.
290
291 40. PCRE tries to detect cases of infinite recursion at compile time, but it
292 cannot analyze patterns in sufficient detail to catch mutual recursions
293 such as ((?1))((?2)). There is now a runtime test that gives an error if a
294 subgroup is called recursively as a subpattern for a second time at the
295 same position in the subject string. In previous releases this might have
296 been caught by the recursion limit, or it might have run out of stack.
297
298 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
299 happen only once. PCRE was, however incorrectly giving a compile time error
300 "recursive call could loop indefinitely" because it cannot analyze the
301 pattern in sufficient detail. The compile time test no longer happens when
302 PCRE is compiling a conditional subpattern, but actual runaway loops are
303 now caught at runtime (see 40 above).
304
305 42. It seems that Perl allows any characters other than a closing parenthesis
306 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
307 has been changed to be the same.
308
309 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
310 as not to get warnings when autogen.sh is called. Also changed
311 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
312
313 44. To help people who use pcregrep to scan files containing exceedingly long
314 lines, the following changes have been made:
315
316 (a) The default value of the buffer size parameter has been increased from
317 8K to 20K. (The actual buffer used is three times this size.)
318
319 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
320 PCRE is built.
321
322 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
323 to be set at run time.
324
325 (d) Numerical values in pcregrep options can be followed by K or M, for
326 example --buffer-size=50K.
327
328 (e) If a line being scanned overflows pcregrep's buffer, an error is now
329 given and the return code is set to 2.
330
331 45. Add a pointer to the latest mark to the callout data block.
332
333 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
334 partial match of an empty string instead of no match. This was specific to
335 the use of ".".
336
337 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
338 complete match instead of a partial match. This bug was dependent on both
339 the PCRE_UTF8 and PCRE_DOTALL options being set.
340
341 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
342 starting byte set, because \b was not being ignored.
343
344
345 Version 8.12 15-Jan-2011
346 ------------------------
347
348 1. Fixed some typos in the markup of the man pages, and wrote a script that
349 checks for such things as part of the documentation building process.
350
351 2. On a big-endian 64-bit system, pcregrep did not correctly process the
352 --match-limit and --recursion-limit options (added for 8.11). In
353 particular, this made one of the standard tests fail. (The integer value
354 went into the wrong half of a long int.)
355
356 3. If the --colour option was given to pcregrep with -v (invert match), it
357 did strange things, either producing crazy output, or crashing. It should,
358 of course, ignore a request for colour when reporting lines that do not
359 match.
360
361 4. Another pcregrep bug caused similar problems if --colour was specified with
362 -M (multiline) and the pattern match finished with a line ending.
363
364 5. In pcregrep, when a pattern that ended with a literal newline sequence was
365 matched in multiline mode, the following line was shown as part of the
366 match. This seems wrong, so I have changed it.
367
368 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
369 the check for further matches in the same line (so they could be coloured)
370 to overrun the end of the current line. If another match was found, it was
371 incorrectly shown (and then shown again when found in the next line).
372
373 7. If pcregrep was compiled under Windows, there was a reference to the
374 function pcregrep_exit() before it was defined. I am assuming this was
375 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
376 reported by a user. I've moved the definition above the reference.
377
378
379 Version 8.11 10-Dec-2010
380 ------------------------
381
382 1. (*THEN) was not working properly if there were untried alternatives prior
383 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
384 backtracked to try for "b" instead of moving to the next alternative branch
385 at the same level (in this case, to look for "c"). The Perl documentation
386 is clear that when (*THEN) is backtracked onto, it goes to the "next
387 alternative in the innermost enclosing group".
388
389 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
390 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
391 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
392 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
393 (*THEN).
394
395 3. If \s appeared in a character class, it removed the VT character from
396 the class, even if it had been included by some previous item, for example
397 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
398 of \s, but is part of the POSIX "space" class.)
399
400 4. A partial match never returns an empty string (because you can always
401 match an empty string at the end of the subject); however the checking for
402 an empty string was starting at the "start of match" point. This has been
403 changed to the "earliest inspected character" point, because the returned
404 data for a partial match starts at this character. This means that, for
405 example, /(?<=abc)def/ gives a partial match for the subject "abc"
406 (previously it gave "no match").
407
408 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
409 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
410 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
411 has an implication that the given string is incomplete (because a partial
412 match is preferred over a full match). For this reason, these items now
413 give a partial match in this situation. [Aside: previously, the one case
414 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
415 match rather than a full match, which was wrong by the old rules, but is
416 now correct.]
417
418 6. There was a bug in the handling of #-introduced comments, recognized when
419 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
420 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
421 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
422 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
423 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
424 places in pcre_compile().
425
426 7. Related to (6) above, when pcre_compile() was skipping #-introduced
427 comments when looking ahead for named forward references to subpatterns,
428 the only newline sequence it recognized was NL. It now handles newlines
429 according to the set newline convention.
430
431 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
432 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
433 cater for a lack of strerror(). These oversights have been fixed.
434
435 9. Added --match-limit and --recursion-limit to pcregrep.
436
437 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
438
439 11. When the -o option was used, pcregrep was setting a return code of 1, even
440 when matches were found, and --line-buffered was not being honoured.
441
442 12. Added an optional parentheses number to the -o and --only-matching options
443 of pcregrep.
444
445 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
446 can match an empty string. The code to do it in pcretest and pcredemo
447 needed fixing:
448
449 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
450 only one byte after an empty string match just before CRLF (this case
451 just got forgotten; "any" and "anycrlf" were OK).
452
453 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
454 mode when an empty string match preceded an ASCII character followed by
455 a non-ASCII character. (The code for advancing by one character rather
456 than one byte was nonsense.)
457
458 (c) The pcredemo.c sample program did not have any code at all to handle
459 the cases when CRLF is a valid newline sequence.
460
461 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
462 as a starting offset was within the subject string. There is now a new
463 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
464 negative or greater than the length of the string. In order to test this,
465 pcretest is extended to allow the setting of negative starting offsets.
466
467 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
468 starting offset points to the beginning of a UTF-8 character was
469 unnecessarily clumsy. I tidied it up.
470
471 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
472 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
473
474 17. Nobody had reported that the --include_dir option, which was added in
475 release 7.7 should have been called --include-dir (hyphen, not underscore)
476 for compatibility with GNU grep. I have changed it to --include-dir, but
477 left --include_dir as an undocumented synonym, and the same for
478 --exclude-dir, though that is not available in GNU grep, at least as of
479 release 2.5.4.
480
481 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
482 characters from a string of bytes) have been redefined so as not to use
483 loops, in order to improve performance in some environments. At the same
484 time, I abstracted some of the common code into auxiliary macros to save
485 repetition (this should not affect the compiled code).
486
487 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
488 compile-time error is now given if \c is not followed by an ASCII
489 character, that is, a byte less than 128. (In EBCDIC mode, the code is
490 different, and any byte value is allowed.)
491
492 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
493 START_OPTIMIZE option, which is now allowed at compile time - but just
494 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
495 to pcregrep and other applications that have no direct access to PCRE
496 options. The new /Y option in pcretest sets this option when calling
497 pcre_compile().
498
499 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
500 back references. Groups containing recursive back references were forced to
501 be atomic by that change, but in the case of named groups, the amount of
502 memory required was incorrectly computed, leading to "Failed: internal
503 error: code overflow". This has been fixed.
504
505 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
506 pcretest.c, to avoid build problems in some Borland environments.
507
508
509 Version 8.10 25-Jun-2010
510 ------------------------
511
512 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
513 THEN.
514
515 2. (*ACCEPT) was not working when inside an atomic group.
516
517 3. Inside a character class, \B is treated as a literal by default, but
518 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
519 causes the error). The code is unchanged, but I tidied the documentation.
520
521 4. Inside a character class, PCRE always treated \R and \X as literals,
522 whereas Perl faults them if its -w option is set. I have changed PCRE so
523 that it faults them when PCRE_EXTRA is set.
524
525 5. Added support for \N, which always matches any character other than
526 newline. (It is the same as "." when PCRE_DOTALL is not set.)
527
528 6. When compiling pcregrep with newer versions of gcc which may have
529 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
530 declared with attribute warn_unused_result" were given. Just casting the
531 result to (void) does not stop the warnings; a more elaborate fudge is
532 needed. I've used a macro to implement this.
533
534 7. Minor change to pcretest.c to avoid a compiler warning.
535
536 8. Added four artifical Unicode properties to help with an option to make
537 \s etc use properties (see next item). The new properties are: Xan
538 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
539
540 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
541 use Unicode properties. (*UCP) at the start of a pattern can be used to set
542 this option. Modified pcretest to add /W to test this facility. Added
543 REG_UCP to make it available via the POSIX interface.
544
545 10. Added --line-buffered to pcregrep.
546
547 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
548 studied, and the match started with a letter with a code point greater than
549 127 whose first byte was different to the first byte of the other case of
550 the letter, the other case of this starting letter was not recognized
551 (#976).
552
553 12. If a pattern that was studied started with a repeated Unicode property
554 test, for example, \p{Nd}+, there was the theoretical possibility of
555 setting up an incorrect bitmap of starting bytes, but fortunately it could
556 not have actually happened in practice until change 8 above was made (it
557 added property types that matched character-matching opcodes).
558
559 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
560 possible starting bytes for non-anchored patterns.
561
562 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
563 \R, and also a number of cases that involve Unicode properties, both
564 explicit and implicit when PCRE_UCP is set.
565
566 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
567 input, it could crash or give wrong results if characters with values
568 greater than 0xc0 were present in the subject string. (Detail: it assumed
569 UTF-8 input when processing these items.)
570
571 16. Added a lot of (int) casts to avoid compiler warnings in systems where
572 size_t is 64-bit (#991).
573
574 17. Added a check for running out of memory when PCRE is compiled with
575 --disable-stack-for-recursion (#990).
576
577 18. If the last data line in a file for pcretest does not have a newline on
578 the end, a newline was missing in the output.
579
580 19. The default pcre_chartables.c file recognizes only ASCII characters (values
581 less than 128) in its various bitmaps. However, there is a facility for
582 generating tables according to the current locale when PCRE is compiled. It
583 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
584 characters, are recognized by isspace() and therefore were getting set in
585 these tables, and indeed these tables seem to approximate to ISO 8859. This
586 caused a problem in UTF-8 mode when pcre_study() was used to create a list
587 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
588 which of course cannot start UTF-8 characters. I have changed the code so
589 that only real ASCII characters (less than 128) and the correct starting
590 bytes for UTF-8 encodings are set for characters greater than 127 when in
591 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
592 altogether.)
593
594 20. Added the /T option to pcretest so as to be able to run tests with non-
595 standard character tables, thus making it possible to include the tests
596 used for 19 above in the standard set of tests.
597
598 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
599 reference to a subpattern the other side of a comment that contains an
600 opening parenthesis caused either an internal compiling error, or a
601 reference to the wrong subpattern.
602
603
604 Version 8.02 19-Mar-2010
605 ------------------------
606
607 1. The Unicode data tables have been updated to Unicode 5.2.0.
608
609 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
610 configured.
611
612 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
613 original author of that file, following a query about its status.
614
615 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
616 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
617
618 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
619 quantifier applied to a forward-referencing subroutine call, could compile
620 incorrect code or give the error "internal error: previously-checked
621 referenced subpattern not found".
622
623 6. Both MS Visual Studio and Symbian OS have problems with initializing
624 variables to point to external functions. For these systems, therefore,
625 pcre_malloc etc. are now initialized to local functions that call the
626 relevant global functions.
627
628 7. There were two entries missing in the vectors called coptable and poptable
629 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
630 I've fixed the data, and added a kludgy way of testing at compile time that
631 the lengths are correct (equal to the number of opcodes).
632
633 8. Following on from 7, I added a similar kludge to check the length of the
634 eint vector in pcreposix.c.
635
636 9. Error texts for pcre_compile() are held as one long string to avoid too
637 much relocation at load time. To find a text, the string is searched,
638 counting zeros. There was no check for running off the end of the string,
639 which could happen if a new error number was added without updating the
640 string.
641
642 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
643
644 11. \K was not working if it appeared in an atomic group or in a group that
645 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
646 \K is "not well defined" if used in an assertion. PCRE now accepts it if
647 the assertion is positive, but not if it is negative.
648
649 12. Change 11 fortuitously reduced the size of the stack frame used in the
650 "match()" function of pcre_exec.c by one pointer. Forthcoming
651 implementation of support for (*MARK) will need an extra pointer on the
652 stack; I have reserved it now, so that the stack frame size does not
653 decrease.
654
655 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
656 item in branch that calls a recursion is a subroutine call - as in the
657 second branch in the above example - was incorrectly given the compile-
658 time error "recursive call could loop indefinitely" because pcre_compile()
659 was not correctly checking the subroutine for matching a non-empty string.
660
661 14. The checks for overrunning compiling workspace could trigger after an
662 overrun had occurred. This is a "should never occur" error, but it can be
663 triggered by pathological patterns such as hundreds of nested parentheses.
664 The checks now trigger 100 bytes before the end of the workspace.
665
666 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
667
668
669 Version 8.01 19-Jan-2010
670 ------------------------
671
672 1. If a pattern contained a conditional subpattern with only one branch (in
673 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
674 computed the wrong minimum data length (which is of course zero for such
675 subpatterns). This could cause incorrect "no match" results.
676
677 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
678 the pattern is reset in the first branch, pcre_compile() failed with
679 "internal error: code overflow at offset...". This happened only when
680 the reset was to the original external option setting. (An optimization
681 abstracts leading options settings into an external setting, which was the
682 cause of this.)
683
684 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
685 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
686 assertion pattern did not match (meaning that the assertion was true), it
687 was incorrectly treated as false if the SKIP had been reached during the
688 matching. This also applied to assertions used as conditions.
689
690 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
691 assertion subpattern, including such a pattern used as a condition,
692 unpredictable results occurred, instead of the error return
693 PCRE_ERROR_DFA_UITEM.
694
695 5. The C++ GlobalReplace function was not working like Perl for the special
696 situation when an empty string is matched. It now does the fancy magic
697 stuff that is necessary.
698
699 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
700 removed. (These were left over from very, very early versions of PCRE.)
701
702 7. Some cosmetic changes to the code to make life easier when compiling it
703 as part of something else:
704
705 (a) Change DEBUG to PCRE_DEBUG.
706
707 (b) In pcre_compile(), rename the member of the "branch_chain" structure
708 called "current" as "current_branch", to prevent a collision with the
709 Linux macro when compiled as a kernel module.
710
711 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
712 prevent a collision with the Linux macro when compiled as a kernel
713 module.
714
715 8. In pcre_compile() there are some checks for integer overflows that used to
716 cast potentially large values to (double). This has been changed to that
717 when building, a check for int64_t is made, and if it is found, it is used
718 instead, thus avoiding the use of floating point arithmetic. (There is no
719 other use of FP in PCRE.) If int64_t is not found, the fallback is to
720 double.
721
722 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
723 2005 (difference between two addresses compared to an unsigned value).
724
725 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
726 custom one, because of the following reported problem in Windows:
727
728 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
729 under Win32.
730 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
731 therefore missing the function definition.
732 - The compiler thus generates a "C" signature for the test function.
733 - The linker fails to find the "C" function.
734 - PCRE fails to configure if asked to do so against libbz2.
735
736 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
737 messages were output:
738
739 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
740 rerunning libtoolize, to keep the correct libtool macros in-tree.
741 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
742
743 I have done both of these things.
744
745 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
746 most of the time, it *can* run out if it is given a pattern that contains a
747 runaway infinite recursion. I updated the discussion in the pcrestack man
748 page.
749
750 13. Now that we have gone to the x.xx style of version numbers, the minor
751 version may start with zero. Using 08 or 09 is a bad idea because users
752 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
753 interpreted as invalid octal numbers. I've updated the previous comment in
754 configure.ac, and also added a check that gives an error if 08 or 09 are
755 used.
756
757 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
758 causing partial matching to fail when the end of the subject matched \W
759 in a UTF-8 pattern where \W was quantified with a minimum of 3.
760
761 15. There were some discrepancies between the declarations in pcre_internal.h
762 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
763 their definitions. The declarations used "const uschar *" and the
764 definitions used USPTR. Even though USPTR is normally defined as "const
765 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
766 reported that: "This difference in casting confuses some C++ compilers, for
767 example, SunCC recognizes above declarations as different functions and
768 generates broken code for hbpcre." I have changed the declarations to use
769 USPTR.
770
771 16. GNU libtool is named differently on some systems. The autogen.sh script now
772 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
773 (FreeBSD).
774
775 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
776 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
777 comment: "Figure out how to create a longlong from a string: strtoll and
778 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
779 instance, but it only takes 2 args instead of 3!"
780
781 18. A subtle bug concerned with back references has been fixed by a change of
782 specification, with a corresponding code fix. A pattern such as
783 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
784 refers, was giving matches when it shouldn't. For example, xa=xaaa would
785 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
786 same bug. Such groups have to be quantified to be useful, or contained
787 inside another quantified group. (If there's no repetition, the reference
788 can never match.) The problem arises because, having left the group and
789 moved on to the rest of the pattern, a later failure that backtracks into
790 the group uses the captured value from the final iteration of the group
791 rather than the correct earlier one. I have fixed this in PCRE by forcing
792 any group that contains a reference to itself to be an atomic group; that
793 is, there cannot be any backtracking into it once it has completed. This is
794 similar to recursive and subroutine calls.
795
796
797 Version 8.00 19-Oct-09
798 ----------------------
799
800 1. The table for translating pcre_compile() error codes into POSIX error codes
801 was out-of-date, and there was no check on the pcre_compile() error code
802 being within the table. This could lead to an OK return being given in
803 error.
804
805 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
806 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
807 in a Windows environment.
808
809 3. The pcregrep --count option prints the count for each file even when it is
810 zero, as does GNU grep. However, pcregrep was also printing all files when
811 --files-with-matches was added. Now, when both options are given, it prints
812 counts only for those files that have at least one match. (GNU grep just
813 prints the file name in this circumstance, but including the count seems
814 more useful - otherwise, why use --count?) Also ensured that the
815 combination -clh just lists non-zero counts, with no names.
816
817 4. The long form of the pcregrep -F option was incorrectly implemented as
818 --fixed_strings instead of --fixed-strings. This is an incompatible change,
819 but it seems right to fix it, and I didn't think it was worth preserving
820 the old behaviour.
821
822 5. The command line items --regex=pattern and --regexp=pattern were not
823 recognized by pcregrep, which required --regex pattern or --regexp pattern
824 (with a space rather than an '='). The man page documented the '=' forms,
825 which are compatible with GNU grep; these now work.
826
827 6. No libpcreposix.pc file was created for pkg-config; there was just
828 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
829
830 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
831 when UCP support is not needed, by modifying the Python script that
832 generates it from Unicode data files. This should not matter if the module
833 is correctly used as a library, but I received one complaint about 50K of
834 unwanted data. My guess is that the person linked everything into his
835 program rather than using a library. Anyway, it does no harm.
836
837 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
838 was a minimum greater than 1 for a wide character in a possessive
839 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
840 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
841 character. Chaos in the form of incorrect output or a compiling loop could
842 result.
843
844 9. The restrictions on what a pattern can contain when partial matching is
845 requested for pcre_exec() have been removed. All patterns can now be
846 partially matched by this function. In addition, if there are at least two
847 slots in the offset vector, the offset of the earliest inspected character
848 for the match and the offset of the end of the subject are set in them when
849 PCRE_ERROR_PARTIAL is returned.
850
851 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
852 synonymous with PCRE_PARTIAL, for backwards compatibility, and
853 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
854 and may be more useful for multi-segment matching.
855
856 11. Partial matching with pcre_exec() is now more intuitive. A partial match
857 used to be given if ever the end of the subject was reached; now it is
858 given only if matching could not proceed because another character was
859 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
860 string "Z", which now yields "no match" instead of "partial match". In the
861 case of pcre_dfa_exec(), "no match" is given if every matching path for the
862 final character ended with (*FAIL).
863
864 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
865 if the pattern had a "must contain" character that was already found in the
866 earlier partial match, unless partial matching was again requested. For
867 example, with the pattern /dog.(body)?/, the "must contain" character is
868 "g". If the first part-match was for the string "dog", restarting with
869 "sbody" failed. This bug has been fixed.
870
871 13. The string returned by pcre_dfa_exec() after a partial match has been
872 changed so that it starts at the first inspected character rather than the
873 first character of the match. This makes a difference only if the pattern
874 starts with a lookbehind assertion or \b or \B (\K is not supported by
875 pcre_dfa_exec()). It's an incompatible change, but it makes the two
876 matching functions compatible, and I think it's the right thing to do.
877
878 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
879 so that the demonstration program is easily available in environments where
880 PCRE has not been installed from source.
881
882 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
883 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
884 library.
885
886 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
887 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
888 is not the first non-POSIX option to be added. Clearly some people find
889 these options useful.
890
891 17. If a caller to the POSIX matching function regexec() passes a non-zero
892 value for nmatch with a NULL value for pmatch, the value of
893 nmatch is forced to zero.
894
895 18. RunGrepTest did not have a test for the availability of the -u option of
896 the diff command, as RunTest does. It now checks in the same way as
897 RunTest, and also checks for the -b option.
898
899 19. If an odd number of negated classes containing just a single character
900 interposed, within parentheses, between a forward reference to a named
901 subpattern and the definition of the subpattern, compilation crashed with
902 an internal error, complaining that it could not find the referenced
903 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
904 [The bug was that it was starting one character too far in when skipping
905 over the character class, thus treating the ] as data rather than
906 terminating the class. This meant it could skip too much.]
907
908 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
909 /g option in pcretest when the pattern contains \K, which makes it possible
910 to have an empty string match not at the start, even when the pattern is
911 anchored. Updated pcretest and pcredemo to use this option.
912
913 21. If the maximum number of capturing subpatterns in a recursion was greater
914 than the maximum at the outer level, the higher number was returned, but
915 with unset values at the outer level. The correct (outer level) value is
916 now given.
917
918 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
919 PCRE did not set those parentheses (unlike Perl). I have now found a way to
920 make it do so. The string so far is captured, making this feature
921 compatible with Perl.
922
923 23. The tests have been re-organized, adding tests 11 and 12, to make it
924 possible to check the Perl 5.10 features against Perl 5.10.
925
926 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
927 pattern matches a fixed length string. PCRE did not allow this; now it
928 does. Neither allows recursion.
929
930 25. I finally figured out how to implement a request to provide the minimum
931 length of subject string that was needed in order to match a given pattern.
932 (It was back references and recursion that I had previously got hung up
933 on.) This code has now been added to pcre_study(); it finds a lower bound
934 to the length of subject needed. It is not necessarily the greatest lower
935 bound, but using it to avoid searching strings that are too short does give
936 some useful speed-ups. The value is available to calling programs via
937 pcre_fullinfo().
938
939 26. While implementing 25, I discovered to my embarrassment that pcretest had
940 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
941 study optimizations had never been tested with that matching function.
942 Oops. What is worse, even when it was passed study data, there was a bug in
943 pcre_dfa_exec() that meant it never actually used it. Double oops. There
944 were also very few tests of studied patterns with pcre_dfa_exec().
945
946 27. If (?| is used to create subpatterns with duplicate numbers, they are now
947 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
948 on the other side of the coin, they are no longer allowed to have different
949 names, because these cannot be distinguished in PCRE, and this has caused
950 confusion. (This is a difference from Perl.)
951
952 28. When duplicate subpattern names are present (necessarily with different
953 numbers, as required by 27 above), and a test is made by name in a
954 conditional pattern, either for a subpattern having been matched, or for
955 recursion in such a pattern, all the associated numbered subpatterns are
956 tested, and the overall condition is true if the condition is true for any
957 one of them. This is the way Perl works, and is also more like the way
958 testing by number works.
959
960
961 Version 7.9 11-Apr-09
962 ---------------------
963
964 1. When building with support for bzlib/zlib (pcregrep) and/or readline
965 (pcretest), all targets were linked against these libraries. This included
966 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
967 libraries. This caused unwanted dependencies to be created. This problem
968 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
969 pcretest is linked with readline.
970
971 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
972 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
973 moved outside it again, because FALSE and TRUE are already defined in AIX,
974 but BOOL is not.
975
976 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
977 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
978
979 4. The pcregrep documentation said spaces were inserted as well as colons (or
980 hyphens) following file names and line numbers when outputting matching
981 lines. This is not true; no spaces are inserted. I have also clarified the
982 wording for the --colour (or --color) option.
983
984 5. In pcregrep, when --colour was used with -o, the list of matching strings
985 was not coloured; this is different to GNU grep, so I have changed it to be
986 the same.
987
988 6. When --colo(u)r was used in pcregrep, only the first matching substring in
989 each matching line was coloured. Now it goes on to look for further matches
990 of any of the test patterns, which is the same behaviour as GNU grep.
991
992 7. A pattern that could match an empty string could cause pcregrep to loop; it
993 doesn't make sense to accept an empty string match in pcregrep, so I have
994 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
995 seems to be how GNU grep behaves.
996
997 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
998 start or after a newline", because the conditional assertion was not being
999 correctly handled. The rule now is that both the assertion and what follows
1000 in the first alternative must satisfy the test.
1001
1002 9. If auto-callout was enabled in a pattern with a conditional group whose
1003 condition was an assertion, PCRE could crash during matching, both with
1004 pcre_exec() and pcre_dfa_exec().
1005
1006 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1007 used for matching.
1008
1009 11. Unicode property support in character classes was not working for
1010 characters (bytes) greater than 127 when not in UTF-8 mode.
1011
1012 12. Added the -M command line option to pcretest.
1013
1014 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1015
1016 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1017
1018 16. Added comments and documentation about mis-use of no_arg in the C++
1019 wrapper.
1020
1021 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1022 from Martin Jerabek that uses macro names for all relevant character and
1023 string constants.
1024
1025 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1026 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1027 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1028 these, but not everybody uses configure.
1029
1030 19. A conditional group that had only one branch was not being correctly
1031 recognized as an item that could match an empty string. This meant that an
1032 enclosing group might also not be so recognized, causing infinite looping
1033 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1034 with the subject "ab", where knowledge that the repeated group can match
1035 nothing is needed in order to break the loop.
1036
1037 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1038 exec(), but without supplying a callout function, matching went wrong.
1039
1040 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1041 leak if the size of the offset vector was greater than 30. When the vector
1042 is smaller, the saved offsets during recursion go onto a local stack
1043 vector, but for larger vectors malloc() is used. It was failing to free
1044 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1045 error, in fact).
1046
1047 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1048 heapframe that is used only when UTF-8 support is enabled. This caused no
1049 problem, but was untidy.
1050
1051 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1052 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1053 included within another project.
1054
1055 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1056 slightly modified by me:
1057
1058 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1059 not building pcregrep.
1060
1061 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1062 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1063
1064 25. Forward references, both numeric and by name, in patterns that made use of
1065 duplicate group numbers, could behave incorrectly or give incorrect errors,
1066 because when scanning forward to find the reference group, PCRE was not
1067 taking into account the duplicate group numbers. A pattern such as
1068 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1069
1070 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1071 the feature of a custom pointer more persuasive (as requested by a user).
1072
1073 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1074 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1075 pre-defined.
1076
1077 28. Added support for (*UTF8) at the start of a pattern.
1078
1079 29. Arrange for flags added by the "release type" setting in CMake to be shown
1080 in the configuration summary.
1081
1082
1083 Version 7.8 05-Sep-08
1084 ---------------------
1085
1086 1. Replaced UCP searching code with optimized version as implemented for Ad
1087 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1088 stage table and inline lookup instead of a function, giving speed ups of 2
1089 to 5 times on some simple patterns that I tested. Permission was given to
1090 distribute the MultiStage2.py script that generates the tables (it's not in
1091 the tarball, but is in the Subversion repository).
1092
1093 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1094 scripts.
1095
1096 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1097 a group with a zero qualifier. The result of the study could be incorrect,
1098 or the function might crash, depending on the pattern.
1099
1100 4. Caseless matching was not working for non-ASCII characters in back
1101 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1102 It now works when Unicode Property Support is available.
1103
1104 5. In pcretest, an escape such as \x{de} in the data was always generating
1105 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1106 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1107 truncation.
1108
1109 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1110
1111 7. Added two (int) casts to pcregrep when printing the difference of two
1112 pointers, in case they are 64-bit values.
1113
1114 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1115 test 2 if it fails.
1116
1117 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1118 and a #define of that name to empty if it is not externally set. This is to
1119 allow users of MSVC to set it if necessary.
1120
1121 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1122 the convenience functions in the pcre_get.c source file.
1123
1124 11. An option change at the start of a pattern that had top-level alternatives
1125 could cause overwriting and/or a crash. This command provoked a crash in
1126 some environments:
1127
1128 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1129
1130 This potential security problem was recorded as CVE-2008-2371.
1131
1132 12. For a pattern where the match had to start at the beginning or immediately
1133 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1134 pcre_dfa_exec() could read past the end of the passed subject if there was
1135 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1136 pcretest so that it places the subject at the end of its malloc-ed buffer.
1137
1138 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1139 exec() might read past the end of the data buffer in UTF-8 mode.
1140
1141 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1142 the data contained the byte 0x85 as part of a UTF-8 character within its
1143 first line. This applied both to normal and DFA matching.
1144
1145 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1146 /^[^d]*?$/8 failed to match "abc".
1147
1148 16. Added a missing copyright notice to pcrecpp_internal.h.
1149
1150 17. Make it more clear in the documentation that values returned from
1151 pcre_exec() in ovector are byte offsets, not character counts.
1152
1153 18. Tidied a few places to stop certain compilers from issuing warnings.
1154
1155 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1156 supplied by Stefan Weber. I made a further small update for 7.8 because
1157 there is a change of source arrangements: the pcre_searchfuncs.c module is
1158 replaced by pcre_ucd.c.
1159
1160
1161 Version 7.7 07-May-08
1162 ---------------------
1163
1164 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1165 a string to a long long, pretend we don't even have a long long." This is
1166 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1167
1168 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1169 pre-7.6 versions, which defined a global no_arg variable instead of putting
1170 it in the RE class. (See also #8 below.)
1171
1172 3. Remove a line of dead code, identified by coverity and reported by Nuno
1173 Lopes.
1174
1175 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1176
1177 (1) The include/exclude patterns were being applied to the whole pathnames
1178 of files, instead of just to the final components.
1179
1180 (2) If there was more than one level of directory, the subdirectories were
1181 skipped unless they satisfied the include/exclude conditions. This is
1182 inconsistent with GNU grep (and could even be seen as contrary to the
1183 pcregrep specification - which I improved to make it absolutely clear).
1184 The action now is always to scan all levels of directory, and just
1185 apply the include/exclude patterns to regular files.
1186
1187 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1188 --exclude_dir in the tests to avoid scanning .svn directories.
1189
1190 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1191 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1192 doesn't support NULs in patterns.
1193
1194 7. Added some missing "const"s to declarations of static tables in
1195 pcre_compile.c and pcre_dfa_exec.c.
1196
1197 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1198 caused by fix #2 above. (Subsequently also a second patch to fix the
1199 first patch. And a third patch - this was a messy problem.)
1200
1201 9. Applied Craig's patch to remove the use of push_back().
1202
1203 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1204 matching function regexec().
1205
1206 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1207 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1208 references. PCRE supports relative numbers with this syntax (I don't think
1209 Oniguruma does).
1210
1211 12. Previously, a group with a zero repeat such as (...){0} was completely
1212 omitted from the compiled regex. However, this means that if the group
1213 was called as a subroutine from elsewhere in the pattern, things went wrong
1214 (an internal error was given). Such groups are now left in the compiled
1215 pattern, with a new opcode that causes them to be skipped at execution
1216 time.
1217
1218 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1219 to the way PCRE behaves:
1220
1221 (a) A lone ] character is dis-allowed (Perl treats it as data).
1222
1223 (b) A back reference to an unmatched subpattern matches an empty string
1224 (Perl fails the current match path).
1225
1226 (c) A data ] in a character class must be notated as \] because if the
1227 first data character in a class is ], it defines an empty class. (In
1228 Perl it is not possible to have an empty class.) The empty class []
1229 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1230 The negative empty class [^] matches any one character, independently
1231 of the DOTALL setting.
1232
1233 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1234 non-existent subpattern following a character class starting with ']' and
1235 containing () gave an internal compiling error instead of "reference to
1236 non-existent subpattern". Fortunately, when the pattern did exist, the
1237 compiled code was correct. (When scanning forwards to check for the
1238 existencd of the subpattern, it was treating the data ']' as terminating
1239 the class, so got the count wrong. When actually compiling, the reference
1240 was subsequently set up correctly.)
1241
1242 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1243 it was being rejected as not supported by pcre_dfa_exec(), even though
1244 other assertions are supported. I have made pcre_dfa_exec() support
1245 (*FAIL).
1246
1247 16. The implementation of 13c above involved the invention of a new opcode,
1248 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1249 cannot be changed at match time, I realized I could make a small
1250 improvement to matching performance by compiling OP_ALLANY instead of
1251 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1252 on the OP_ANY path.
1253
1254 17. Compiling pcretest on Windows with readline support failed without the
1255 following two fixes: (1) Make the unistd.h include conditional on
1256 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1257
1258 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1259 ncurses library to be included for pcretest when ReadLine support is
1260 requested, but also to allow for it to be overridden. This patch came from
1261 Daniel Bergström.
1262
1263 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1264 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1265 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1266 spotting this.
1267
1268
1269 Version 7.6 28-Jan-08
1270 ---------------------
1271
1272 1. A character class containing a very large number of characters with
1273 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1274 overflow.
1275
1276 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1277 HAVE_LONG_LONG is not defined.
1278
1279 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1280 bring them up to date and include new features. This patch includes:
1281
1282 - Fixed PH's badly added libz and libbz2 support.
1283 - Fixed a problem with static linking.
1284 - Added pcredemo. [But later removed - see 7 below.]
1285 - Fixed dftables problem and added an option.
1286 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1287 HAVE_LONG_LONG.
1288 - Added readline support for pcretest.
1289 - Added an listing of the option settings after cmake has run.
1290
1291 4. A user submitted a patch to Makefile that makes it easy to create
1292 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1293 Makefile.am that cause it to include this special target, without
1294 affecting anything else. Note that the same mingw target plus all
1295 the other distribution libraries and programs are now supported
1296 when configuring with CMake (see 6 below) instead of with
1297 Configure/Make.
1298
1299 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1300 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1301 exported in the Windows port". It has not yet been confirmed that the patch
1302 solves the problem, but it does no harm.
1303
1304 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1305 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1306 with CMake, and also correct the comment about stack recursion.
1307
1308 7. Remove the automatic building of pcredemo from the ./configure system and
1309 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1310 of a program that users should build themselves after PCRE is installed, so
1311 building it automatically is not really right. What is more, it gave
1312 trouble in some build environments.
1313
1314 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1315
1316
1317 Version 7.5 10-Jan-08
1318 ---------------------
1319
1320 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1321 values in parens when parsing an RE using the C++ wrapper."
1322
1323 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1324 Characters greater than 255 were excluded from the class instead of being
1325 included.
1326
1327 3. The same bug as (2) above applied to negated POSIX classes such as
1328 [:^space:].
1329
1330 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1331 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1332 I have changed it.
1333
1334 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1335 first named subpattern) and a construct such as (?&a) would reference the
1336 first named subpattern whose name started with "a" (in other words, the
1337 length check was missing). Both these problems are fixed. "Subpattern name
1338 expected" is now given for (?&) (a zero-length name), and this patch also
1339 makes it give the same error for \k'' (previously it complained that that
1340 was a reference to a non-existent subpattern).
1341
1342 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1343 this is right because (?- can be followed by option settings as well as by
1344 digits. I have, however, made the messages clearer.
1345
1346 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1347 than the number used in the conditional) now cause a compile-time error.
1348 This is actually not compatible with Perl, which accepts such patterns, but
1349 treats the conditional as always being FALSE (as PCRE used to), but it
1350 seems to me that giving a diagnostic is better.
1351
1352 8. Change "alphameric" to the more common word "alphanumeric" in comments
1353 and messages.
1354
1355 9. Fix two occurrences of "backslash" in comments that should have been
1356 "backspace".
1357
1358 10. Remove two redundant lines of code that can never be obeyed (their function
1359 was moved elsewhere).
1360
1361 11. The program that makes PCRE's Unicode character property table had a bug
1362 which caused it to generate incorrect table entries for sequences of
1363 characters that have the same character type, but are in different scripts.
1364 It amalgamated them into a single range, with the script of the first of
1365 them. In other words, some characters were in the wrong script. There were
1366 thirteen such cases, affecting characters in the following ranges:
1367
1368 U+002b0 - U+002c1
1369 U+0060c - U+0060d
1370 U+0061e - U+00612
1371 U+0064b - U+0065e
1372 U+0074d - U+0076d
1373 U+01800 - U+01805
1374 U+01d00 - U+01d77
1375 U+01d9b - U+01dbf
1376 U+0200b - U+0200f
1377 U+030fc - U+030fe
1378 U+03260 - U+0327f
1379 U+0fb46 - U+0fbb1
1380 U+10450 - U+1049d
1381
1382 12. The -o option (show only the matching part of a line) for pcregrep was not
1383 compatible with GNU grep in that, if there was more than one match in a
1384 line, it showed only the first of them. It now behaves in the same way as
1385 GNU grep.
1386
1387 13. If the -o and -v options were combined for pcregrep, it printed a blank
1388 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1389 does the same. The return code can be used to tell if there were any
1390 non-matching lines.
1391
1392 14. Added --file-offsets and --line-offsets to pcregrep.
1393
1394 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1395 infinitely looping recursion. The bug was that positive lookaheads were not
1396 being skipped when checking for a possible empty match (negative lookaheads
1397 and both kinds of lookbehind were skipped).
1398
1399 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1400 inclusion of <windows.h> to before rather than after the definition of
1401 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1402
1403 17. Specifying a possessive quantifier with a specific limit for a Unicode
1404 character property caused pcre_compile() to compile bad code, which led at
1405 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1406 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1407 caused the error; without that there was no problem.
1408
1409 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1410
1411 19. Added --enable-pcretest-libreadline.
1412
1413 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1414 RE::GlobalReplace(). As a result, the number of replacements returned was
1415 double what it should be. I removed one of the increments, but Craig sent a
1416 later patch that removed the other one (the right fix) and added unit tests
1417 that check the return values (which was not done before).
1418
1419 21. Several CMake things:
1420
1421 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1422 the names libpcre and libpcreposix, not just pcre and pcreposix.
1423
1424 (2) The above change means that pcretest and pcregrep are now correctly
1425 linked with the newly-built libraries, not previously installed ones.
1426
1427 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1428
1429 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1430 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1431 UTF-8 newline character). The key issue is that the pattern starts .*;
1432 this means that the match must be either at the beginning, or after a
1433 newline. The bug was in the code for advancing after a failed match and
1434 checking that the new position followed a newline. It was not taking
1435 account of UTF-8 characters correctly.
1436
1437 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1438 character classes. PCRE was not treating the sequence [:...:] as a
1439 character class unless the ... were all letters. Perl, however, seems to
1440 allow any characters between [: and :], though of course it rejects as
1441 unknown any "names" that contain non-letters, because all the known class
1442 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1443 for example, whereas PCRE did not - it did not recognize a POSIX character
1444 class. This seemed a bit dangerous, so the code has been changed to be
1445 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1446 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1447 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1448 Perl does, and where it didn't before.
1449
1450 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1451 Windows environments %n is disabled by default.
1452
1453
1454 Version 7.4 21-Sep-07
1455 ---------------------
1456
1457 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1458 means that a class such as [\s] counted as "explicit reference to CR or
1459 LF". That isn't really right - the whole point of the change was to try to
1460 help when there was an actual mention of one of the two characters. So now
1461 the change happens only if \r or \n (or a literal CR or LF) character is
1462 encountered.
1463
1464 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1465 of both had grown to the point where there were only 3 bits left.
1466 Fortunately, there was spare space in the data structure, and so I have
1467 moved the internal flags into a new 16-bit field to free up more option
1468 bits.
1469
1470 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1471 but did not set the internal JCHANGED flag - either of these is enough to
1472 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1473 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1474 start sets both bits.
1475
1476 4. Added options (at build time, compile time, exec time) to change \R from
1477 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1478
1479 5. doc/pcresyntax.html was missing from the distribution.
1480
1481 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1482 compatibility, even though it is no longer used.
1483
1484 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1485 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1486 windows.h file is present (where different names are used). [This was
1487 reversed later after testing - see 16 below.]
1488
1489 8. Changed all #include <config.h> to #include "config.h". There were also
1490 some further <pcre.h> cases that I changed to "pcre.h".
1491
1492 9. When pcregrep was used with the --colour option, it missed the line ending
1493 sequence off the lines that it output.
1494
1495 10. It was pointed out to me that arrays of string pointers cause lots of
1496 relocations when a shared library is dynamically loaded. A technique of
1497 using a single long string with a table of offsets can drastically reduce
1498 these. I have refactored PCRE in four places to do this. The result is
1499 dramatic:
1500
1501 Originally: 290
1502 After changing UCP table: 187
1503 After changing error message table: 43
1504 After changing table of "verbs" 36
1505 After changing table of Posix names 22
1506
1507 Thanks to the folks working on Gregex for glib for this insight.
1508
1509 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1510 unicode-properties was also set.
1511
1512 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1513
1514 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1515 checked only for CRLF.
1516
1517 14. Added casts to pcretest.c to avoid compiler warnings.
1518
1519 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1520
1521 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1522 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1523 entirely. This removes changes made in 7 above.
1524
1525 17. The CMake files have been updated, and there is now more information about
1526 building with CMake in the NON-UNIX-USE document.
1527
1528
1529 Version 7.3 28-Aug-07
1530 ---------------------
1531
1532 1. In the rejigging of the build system that eventually resulted in 7.1, the
1533 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1534 brackets there is not right, since it causes compilers to look for an
1535 installed pcre.h, not the version that is in the source that is being
1536 compiled (which of course may be different). I have changed it back to:
1537
1538 #include "pcre.h"
1539
1540 I have a vague recollection that the change was concerned with compiling in
1541 different directories, but in the new build system, that is taken care of
1542 by the VPATH setting the Makefile.
1543
1544 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1545 when the subject happened to end in the byte 0x85 (e.g. if the last
1546 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1547 characters but of course it shouldn't be taken as a newline when it is part
1548 of another character. The bug was that, for an unlimited repeat of . in
1549 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1550 characters when looking for a newline.
1551
1552 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1553
1554 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1555 in debug output.
1556
1557 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1558 long printing in the pcrecpp unittest when running under MinGW.
1559
1560 6. ESC_K was left out of the EBCDIC table.
1561
1562 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1563 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1564 limit also applies to "virtual nesting" when a pattern is recursive, and in
1565 this case 1000 isn't so big. I have been able to remove this limit at the
1566 expense of backing off one optimization in certain circumstances. Normally,
1567 when pcre_exec() would call its internal match() function recursively and
1568 immediately return the result unconditionally, it uses a "tail recursion"
1569 feature to save stack. However, when a subpattern that can match an empty
1570 string has an unlimited repetition quantifier, it no longer makes this
1571 optimization. That gives it a stack frame in which to save the data for
1572 checking that an empty string has been matched. Previously this was taken
1573 from the 1000-entry workspace that had been reserved. So now there is no
1574 explicit limit, but more stack is used.
1575
1576 8. Applied Daniel's patches to solve problems with the import/export magic
1577 syntax that is required for Windows, and which was going wrong for the
1578 pcreposix and pcrecpp parts of the library. These were overlooked when this
1579 problem was solved for the main library.
1580
1581 9. There were some crude static tests to avoid integer overflow when computing
1582 the size of patterns that contain repeated groups with explicit upper
1583 limits. As the maximum quantifier is 65535, the maximum group length was
1584 set at 30,000 so that the product of these two numbers did not overflow a
1585 32-bit integer. However, it turns out that people want to use groups that
1586 are longer than 30,000 bytes (though not repeat them that many times).
1587 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1588 made it possible to implement the integer overflow checks in a much more
1589 dynamic way, which I have now done. The artificial limitation on group
1590 length has been removed - we now have only the limit on the total length of
1591 the compiled pattern, which depends on the LINK_SIZE setting.
1592
1593 10. Fixed a bug in the documentation for get/copy named substring when
1594 duplicate names are permitted. If none of the named substrings are set, the
1595 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1596 empty string.
1597
1598 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1599 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1600 because the ] is interpreted as the first data character and the
1601 terminating ] is not found. PCRE has been made compatible with Perl in this
1602 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1603 cause memory overwriting.
1604
1605 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1606 string has been matched (to stop an infinite loop). It was not recognizing
1607 a conditional subpattern that could match an empty string if that
1608 subpattern was within another subpattern. For example, it looped when
1609 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1610 condition was not nested. This bug has been fixed.
1611
1612 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1613 past the start of the subject in the presence of bytes with the top bit
1614 set, for example "\x8aBCD".
1615
1616 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1617 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1618
1619 14. Optimized (?!) to (*FAIL).
1620
1621 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1622 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1623 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1624 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1625 does: it's just the validity check that is more restrictive.
1626
1627 16. Inserted checks for integer overflows during escape sequence (backslash)
1628 processing, and also fixed erroneous offset values for syntax errors during
1629 backslash processing.
1630
1631 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1632 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1633
1634 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1635 caused an overrun.
1636
1637 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1638 something other than just ASCII characters) inside a group that had an
1639 unlimited repeat caused a loop at compile time (while checking to see
1640 whether the group could match an empty string).
1641
1642 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1643 [\P{Any}] did so. (Error in the code for printing property names.)
1644
1645 21. An orphan \E inside a character class could cause a crash.
1646
1647 22. A repeated capturing bracket such as (A)? could cause a wild memory
1648 reference during compilation.
1649
1650 23. There are several functions in pcre_compile() that scan along a compiled
1651 expression for various reasons (e.g. to see if it's fixed length for look
1652 behind). There were bugs in these functions when a repeated \p or \P was
1653 present in the pattern. These operators have additional parameters compared
1654 with \d, etc, and these were not being taken into account when moving along
1655 the compiled data. Specifically:
1656
1657 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1658 length.
1659
1660 (b) An item such as \pL+ within a repeated group could cause crashes or
1661 loops.
1662
1663 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1664 "reference to non-existent subpattern" error.
1665
1666 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1667
1668 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1669 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1670
1671 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1672
1673 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1674 character were causing crashes (broken optimization).
1675
1676 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1677 \p or \P) caused a compile-time loop.
1678
1679 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1680 break. For example, the unstudied pattern [\r\n]A does not match the string
1681 "\r\nA" because change 7.0/46 below moves the current point on by two
1682 characters after failing to match at the start. However, the pattern \nA
1683 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1684 the same is true. There doesn't seem any very clean way out of this, but
1685 what I have chosen to do makes the common cases work: PCRE now takes note
1686 of whether there can be an explicit match for \r or \n anywhere in the
1687 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1688 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1689 pattern has explicit CR or LF references.
1690
1691 29. Added (*CR) etc for changing newline setting at start of pattern.
1692
1693
1694 Version 7.2 19-Jun-07
1695 ---------------------
1696
1697 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1698 which is apparently normally available under Windows.
1699
1700 2. Re-jig the pcregrep tests with different newline settings in an attempt
1701 to make them independent of the local environment's newline setting.
1702
1703 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1704
1705 4. Some of the "internals" tests were previously cut out when the link size
1706 was not 2, because the output contained actual offsets. The recent new
1707 "Z" feature of pcretest means that these can be cut out, making the tests
1708 usable with all link sizes.
1709
1710 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1711 stack recursion. This gives a massive performance boost under BSD, but just
1712 a small improvement under Linux. However, it saves one field in the frame
1713 in all cases.
1714
1715 6. Added more features from the forthcoming Perl 5.10:
1716
1717 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1718 recursion call. It refers to the nth most recently opened parentheses.
1719
1720 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1721 to be opened parentheses.
1722
1723 (c) Conditions that refer to capturing parentheses can be specified
1724 relatively, for example, (?(-2)... or (?(+3)...
1725
1726 (d) \K resets the start of the current match so that everything before
1727 is not part of it.
1728
1729 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1730
1731 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1732 reference syntax.
1733
1734 (g) (?| introduces a group in which the numbering of parentheses in each
1735 alternative starts with the same number.
1736
1737 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1738
1739 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1740 PCRE_INFO_JCHANGED.
1741
1742 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1743 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1744 for detecting groups that can match an empty string.
1745
1746 9. A pattern with a very large number of alternatives (more than several
1747 hundred) was running out of internal workspace during the pre-compile
1748 phase, where pcre_compile() figures out how much memory will be needed. A
1749 bit of new cunning has reduced the workspace needed for groups with
1750 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1751 workspace instead of running out of the 4096 that are available.
1752
1753 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1754
1755 11. Applied patch from Google to remove an optimization that didn't quite work.
1756 The report of the bug said:
1757
1758 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1759 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1760 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1761
1762 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1763 it matched the wrong number of bytes.
1764
1765
1766 Version 7.1 24-Apr-07
1767 ---------------------
1768
1769 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1770 that is more "standard", making use of automake and other Autotools. There
1771 is some re-arrangement of the files and adjustment of comments consequent
1772 on this.
1773
1774 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1775 for recursive directory scanning broke on some systems because the files
1776 are not scanned in any specific order and on different systems the order
1777 was different. A call to "sort" has been inserted into RunGrepTest for the
1778 approprate test as a short-term fix. In the longer term there may be an
1779 alternative.
1780
1781 3. I had an email from Eric Raymond about problems translating some of PCRE's
1782 man pages to HTML (despite the fact that I distribute HTML pages, some
1783 people do their own conversions for various reasons). The problems
1784 concerned the use of low-level troff macros .br and .in. I have therefore
1785 removed all such uses from the man pages (some were redundant, some could
1786 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1787 HTML has been updated to handle .nf/.fi and to complain if it encounters
1788 .br or .in.
1789
1790 4. Updated comments in configure.ac that get placed in config.h.in and also
1791 arranged for config.h to be included in the distribution, with the name
1792 config.h.generic, for the benefit of those who have to compile without
1793 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1794
1795 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1796 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1797 makevp.bat for the current PCRE, using the additional files
1798 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1799
1800 6. A Windows user reported a minor discrepancy with test 2, which turned out
1801 to be caused by a trailing space on an input line that had got lost in his
1802 copy. The trailing space was an accident, so I've just removed it.
1803
1804 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1805 that is needed.
1806
1807 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1808 as "const" (a) because they are and (b) because it helps the PHP
1809 maintainers who have recently made a script to detect big data structures
1810 in the php code that should be moved to the .rodata section. I remembered
1811 to update Builducptable as well, so it won't revert if ucptable.h is ever
1812 re-created.
1813
1814 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1815 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1816 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1817 support is not required. This saves 1.5-2K of code, which is important in
1818 some applications.
1819
1820 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1821 so as not to refer to the tables, even though these functions will never be
1822 called when UTF-8 support is disabled. Otherwise there are problems with a
1823 shared library.
1824
1825 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1826
1827 (a) It was defining its arguments as char * instead of void *.
1828
1829 (b) It was assuming that all moves were upwards in memory; this was true
1830 a long time ago when I wrote it, but is no longer the case.
1831
1832 The emulated memove() is provided for those environments that have neither
1833 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1834 is clearly not the case, as these two bugs were recently reported.
1835
1836 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1837 and Detrail to create the HTML documentation, the .txt form of the man
1838 pages, and it removes trailing spaces from listed files. It also creates
1839 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1840 case, it wraps all the #defines with #ifndefs. This script should be run
1841 before "make dist".
1842
1843 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1844 with Unicode property support.
1845
1846 (a) For a maximizing quantifier, if the two different cases of the
1847 character were of different lengths in their UTF-8 codings (there are
1848 some cases like this - I found 11), and the matching function had to
1849 back up over a mixture of the two cases, it incorrectly assumed they
1850 were both the same length.
1851
1852 (b) When PCRE was configured to use the heap rather than the stack for
1853 recursion during matching, it was not correctly preserving the data for
1854 the other case of a UTF-8 character when checking ahead for a match
1855 while processing a minimizing repeat. If the check also involved
1856 matching a wide character, but failed, corruption could cause an
1857 erroneous result when trying to check for a repeat of the original
1858 character.
1859
1860 13. Some tidying changes to the testing mechanism:
1861
1862 (a) The RunTest script now detects the internal link size and whether there
1863 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1864 values substituted by "configure". (The RunGrepTest script already did
1865 this for UTF-8.) The configure.ac script no longer substitutes the
1866 relevant variables.
1867
1868 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1869 with length and offset values. This means that the output is different
1870 for different internal link sizes. Test 2 is skipped for link sizes
1871 other than 2 because of this, bypassing the problem. Unfortunately,
1872 there was also a test in test 3 (the locale tests) that used /B and
1873 failed for link sizes other than 2. Rather than cut the whole test out,
1874 I have added a new /Z option to pcretest that replaces the length and
1875 offset values with spaces. This is now used to make test 3 independent
1876 of link size. (Test 2 will be tidied up later.)
1877
1878 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1879 segmentation fault instead of returning the appropriate error message.
1880
1881 15. In multiline mode when the newline sequence was set to "any", the pattern
1882 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1883 This doesn't seem right; it now treats the CRLF combination as the line
1884 ending, and so does not match in that case. It's only a pattern such as ^$
1885 that would hit this one: something like ^ABC$ would have failed after \r
1886 and then tried again after \r\n.
1887
1888 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1889 in an attempt to make files that differ only in their line terminators
1890 compare equal. This works on Linux.
1891
1892 17. Under certain error circumstances pcregrep might try to free random memory
1893 as it exited. This is now fixed, thanks to valgrind.
1894
1895 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1896 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1897 was because its rules for how to advance for /g after matching an empty
1898 string at the end of a line did not allow for this case. They now check for
1899 it specially.
1900
1901 20. pcretest is supposed to handle patterns and data of any length, by
1902 extending its buffers when necessary. It was getting this wrong when the
1903 buffer for a data line had to be extended.
1904
1905 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1906 CRLF as a newline sequence.
1907
1908 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1909 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1910 I have nevertheless tidied it up.
1911
1912 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1913
1914 24. Added a man page for pcre-config.
1915
1916
1917 Version 7.0 19-Dec-06
1918 ---------------------
1919
1920 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1921 moving to gcc 4.1.1.
1922
1923 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1924 sys/time.h, which is documented as needed for this function. It doesn't
1925 seem to matter on Linux, but it showed up on some releases of OS X.
1926
1927 3. It seems that there are systems where bytes whose values are greater than
1928 127 match isprint() in the "C" locale. The "C" locale should be the
1929 default when a C program starts up. In most systems, only ASCII printing
1930 characters match isprint(). This difference caused the output from pcretest
1931 to vary, making some of the tests fail. I have changed pcretest so that:
1932
1933 (a) When it is outputting text in the compiled version of a pattern, bytes
1934 other than 32-126 are always shown as hex escapes.
1935
1936 (b) When it is outputting text that is a matched part of a subject string,
1937 it does the same, unless a different locale has been set for the match
1938 (using the /L modifier). In this case, it uses isprint() to decide.
1939
1940 4. Fixed a major bug that caused incorrect computation of the amount of memory
1941 required for a compiled pattern when options that changed within the
1942 pattern affected the logic of the preliminary scan that determines the
1943 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1944 that the computed length was too small. The symptoms of this bug were
1945 either the PCRE error "internal error: code overflow" from pcre_compile(),
1946 or a glibc crash with a message such as "pcretest: free(): invalid next
1947 size (fast)". Examples of patterns that provoked this bug (shown in
1948 pcretest format) are:
1949
1950 /(?-x: )/x
1951 /(?x)(?-x: \s*#\s*)/
1952 /((?i)[\x{c0}])/8
1953 /(?i:[\x{c0}])/8
1954
1955 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1956 is now done differently.
1957
1958 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1959 wrapper classes; (b) implement a new function in the C++ scanner that is
1960 more efficient than the old way of doing things because it avoids levels of
1961 recursion in the regex matching; (c) add a paragraph to the documentation
1962 for the FullMatch() function.
1963
1964 6. The escape sequence \n was being treated as whatever was defined as
1965 "newline". Not only was this contrary to the documentation, which states
1966 that \n is character 10 (hex 0A), but it also went horribly wrong when
1967 "newline" was defined as CRLF. This has been fixed.
1968
1969 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1970 was being set to -1 for the "end of line" case (supposedly a value that no
1971 character can have). Though this value is never used (the check for end of
1972 line is "zero bytes in current character"), it caused compiler complaints.
1973 I've changed it to 0xffffffff.
1974
1975 8. In pcre_version.c, the version string was being built by a sequence of
1976 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1977 string (as it is for production releases) called a macro with an empty
1978 argument. The C standard says the result of this is undefined. The gcc
1979 compiler treats it as an empty string (which was what was wanted) but it is
1980 reported that Visual C gives an error. The source has been hacked around to
1981 avoid this problem.
1982
1983 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1984 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1985 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1986 of them did).
1987
1988 10. Originally, pcretest opened its input and output without "b"; then I was
1989 told that "b" was needed in some environments, so it was added for release
1990 5.0 to both the input and output. (It makes no difference on Unix-like
1991 systems.) Later I was told that it is wrong for the input on Windows. I've
1992 now abstracted the modes into two macros, to make it easier to fiddle with
1993 them, and removed "b" from the input mode under Windows.
1994
1995 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1996
1997 12. Added -help and --help to pcretest as an official way of being reminded
1998 of the options.
1999
2000 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2001 and pcrecpp.cc because they annoy compilers at high warning levels.
2002
2003 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2004
2005 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2006 scripts are not C programs :-) and which was not noticed because it works
2007 on Linux.
2008
2009 16. pcretest is supposed to handle any length of pattern and data line (as one
2010 line or as a continued sequence of lines) by extending its input buffer if
2011 necessary. This feature was broken for very long pattern lines, leading to
2012 a string of junk being passed to pcre_compile() if the pattern was longer
2013 than about 50K.
2014
2015 17. I have done a major re-factoring of the way pcre_compile() computes the
2016 amount of memory needed for a compiled pattern. Previously, there was code
2017 that made a preliminary scan of the pattern in order to do this. That was
2018 OK when PCRE was new, but as the facilities have expanded, it has become
2019 harder and harder to keep it in step with the real compile phase, and there
2020 have been a number of bugs (see for example, 4 above). I have now found a
2021 cunning way of running the real compile function in a "fake" mode that
2022 enables it to compute how much memory it would need, while actually only
2023 ever using a few hundred bytes of working memory and without too many
2024 tests of the mode. This should make future maintenance and development
2025 easier. A side effect of this work is that the limit of 200 on the nesting
2026 depth of parentheses has been removed (though this was never a serious
2027 limitation, I suspect). However, there is a downside: pcre_compile() now
2028 runs more slowly than before (30% or more, depending on the pattern). I
2029 hope this isn't a big issue. There is no effect on runtime performance.
2030
2031 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2032 newline (only possible for the last line of a file) and it was a
2033 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2034
2035 19. Added additional timing features to pcretest. (1) The -tm option now times
2036 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2037 separate command line item, by a number that specifies the number of
2038 repeats to use when timing. The default is 50000; this gives better
2039 precision, but takes uncomfortably long for very large patterns.
2040
2041 20. Extended pcre_study() to be more clever in cases where a branch of a
2042 subpattern has no definite first character. For example, (a*|b*)[cd] would
2043 previously give no result from pcre_study(). Now it recognizes that the
2044 first character must be a, b, c, or d.
2045
2046 21. There was an incorrect error "recursive call could loop indefinitely" if
2047 a subpattern (or the entire pattern) that was being tested for matching an
2048 empty string contained only one non-empty item after a nested subpattern.
2049 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2050 incorrectly, because the \d was being skipped in the check.
2051
2052 22. The pcretest program now has a new pattern option /B and a command line
2053 option -b, which is equivalent to adding /B to every pattern. This causes
2054 it to show the compiled bytecode, without the additional information that
2055 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2056 is the same as /B/I).
2057
2058 23. A new optimization is now able automatically to treat some sequences such
2059 as a*b as a*+b. More specifically, if something simple (such as a character
2060 or a simple class like \d) has an unlimited quantifier, and is followed by
2061 something that cannot possibly match the quantified thing, the quantifier
2062 is automatically "possessified".
2063
2064 24. A recursive reference to a subpattern whose number was greater than 39
2065 went wrong under certain circumstances in UTF-8 mode. This bug could also
2066 have affected the operation of pcre_study().
2067
2068 25. Realized that a little bit of performance could be had by replacing
2069 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2070
2071 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2072
2073 27. Possessive quantifiers such as a++ were previously implemented by turning
2074 them into atomic groups such as ($>a+). Now they have their own opcodes,
2075 which improves performance. This includes the automatically created ones
2076 from 23 above.
2077
2078 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2079 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2080 the first matched character to be a colon. This applied both to named and
2081 numbered groups.
2082
2083 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2084
2085 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2086 building PCRE on Windows easier, so I have included it in the distribution.
2087
2088 31. There is now a check in pcretest against a ridiculously large number being
2089 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2090 loop, the loop is abandoned.
2091
2092 32. Forward references to subpatterns in conditions such as (?(2)...) where
2093 subpattern 2 is defined later cause pcre_compile() to search forwards in
2094 the pattern for the relevant set of parentheses. This search went wrong
2095 when there were unescaped parentheses in a character class, parentheses
2096 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2097
2098 33. "Subroutine" calls and backreferences were previously restricted to
2099 referencing subpatterns earlier in the regex. This restriction has now
2100 been removed.
2101
2102 34. Added a number of extra features that are going to be in Perl 5.10. On the
2103 whole, these are just syntactic alternatives for features that PCRE had
2104 previously implemented using the Python syntax or my own invention. The
2105 other formats are all retained for compatibility.
2106
2107 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2108 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2109 also .NET compatible.
2110
2111 (b) A recursion or subroutine call to a named group can now be defined as
2112 (?&name) as well as (?P>name).
2113
2114 (c) A backreference to a named group can now be defined as \k<name> or
2115 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2116 5.10, are also .NET compatible.
2117
2118 (d) A conditional reference to a named group can now use the syntax
2119 (?(<name>) or (?('name') as well as (?(name).
2120
2121 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2122 groups (named and numbered) that are never evaluated inline, but can be
2123 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2124 is always false. There may be only one alternative in such a group.
2125
2126 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2127 as the simple (?(R). The condition is true only if the most recent
2128 recursion is that of the given number or name. It does not search out
2129 through the entire recursion stack.
2130
2131 (g) The escape \gN or \g{N} has been added, where N is a positive or
2132 negative number, specifying an absolute or relative reference.
2133
2134 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2135 some "unreachable code" warnings.
2136
2137 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2138 things, this adds five new scripts.
2139
2140 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2141 There were also incompatibilities regarding the handling of \Q..\E inside
2142 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2143 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2144
2145 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2146 matches an empty string, and forcibly breaks the loop. There were bugs in
2147 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2148 against aaaa the result was just "a" rather than "aaaa", for example. Two
2149 separate and independent bugs (that affected different cases) have been
2150 fixed.
2151
2152 39. Refactored the code to abolish the use of different opcodes for small
2153 capturing bracket numbers. This is a tidy that I avoided doing when I
2154 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2155 The new approach is not only tidier, it makes it possible to reduce the
2156 memory needed to fix the previous bug (38).
2157
2158 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2159 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2160 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2161 mode.
2162
2163 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2164 report.
2165
2166 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2167 copy construction and assignment in the C++ wrapper.
2168
2169 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2170 couple of bugs that could have given wrong results in the "--newline=crlf"
2171 case.
2172
2173 44. Added a number of casts and did some reorganization of signed/unsigned int
2174 variables following suggestions from Dair Grant. Also renamed the variable
2175 "this" as "item" because it is a C++ keyword.
2176
2177 45. Arranged for dftables to add
2178
2179 #include "pcre_internal.h"
2180
2181 to pcre_chartables.c because without it, gcc 4.x may remove the array
2182 definition from the final binary if PCRE is built into a static library and
2183 dead code stripping is activated.
2184
2185 46. For an unanchored pattern, if a match attempt fails at the start of a
2186 newline sequence, and the newline setting is CRLF or ANY, and the next two
2187 characters are CRLF, advance by two characters instead of one.
2188
2189
2190 Version 6.7 04-Jul-06
2191 ---------------------
2192
2193 1. In order to handle tests when input lines are enormously long, pcretest has
2194 been re-factored so that it automatically extends its buffers when
2195 necessary. The code is crude, but this _is_ just a test program. The
2196 default size has been increased from 32K to 50K.
2197
2198 2. The code in pcre_study() was using the value of the re argument before
2199 testing it for NULL. (Of course, in any sensible call of the function, it
2200 won't be NULL.)
2201
2202 3. The memmove() emulation function in pcre_internal.h, which is used on
2203 systems that lack both memmove() and bcopy() - that is, hardly ever -
2204 was missing a "static" storage class specifier.
2205
2206 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2207 containing an extended class (one that cannot be represented by a bitmap
2208 because it contains high-valued characters or Unicode property items, e.g.
2209 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2210 pattern, but PCRE should not loop if you do not (it no longer does).
2211 [Detail: two cases were found: (a) a repeated subpattern containing an
2212 extended class; (b) a recursive reference to a subpattern that followed a
2213 previous extended class. It wasn't skipping over the extended class
2214 correctly when UTF-8 mode was not set.]
2215
2216 5. A negated single-character class was not being recognized as fixed-length
2217 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2218 compile error "lookbehind assertion is not fixed length".
2219
2220 6. The RunPerlTest auxiliary script was showing an unexpected difference
2221 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2222 write a Perl script that can interpret lines of an input file either as
2223 byte characters or as UTF-8, which is what "perltest" was being required to
2224 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2225 can't do is switch easily at run time between having the "use utf8;" pragma
2226 or not. In the end, I fudged it by using the RunPerlTest script to insert
2227 "use utf8;" explicitly for the UTF-8 tests.
2228
2229 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2230 the end of the subject string, contrary to the documentation and to what
2231 Perl does. This was true of both matching functions. Now it matches only at
2232 the start of the subject and immediately after *internal* newlines.
2233
2234 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2235 a pointer to an int instead of a pointer to an unsigned long int. This
2236 caused problems on 64-bit systems.
2237
2238 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2239 instance of the 'standard' template library not being so standard".
2240
2241 10. There was no check on the number of named subpatterns nor the maximum
2242 length of a subpattern name. The product of these values is used to compute
2243 the size of the memory block for a compiled pattern. By supplying a very
2244 long subpattern name and a large number of named subpatterns, the size
2245 computation could be caused to overflow. This is now prevented by limiting
2246 the length of names to 32 characters, and the number of named subpatterns
2247 to 10,000.
2248
2249 11. Subpatterns that are repeated with specific counts have to be replicated in
2250 the compiled pattern. The size of memory for this was computed from the
2251 length of the subpattern and the repeat count. The latter is limited to
2252 65535, but there was no limit on the former, meaning that integer overflow
2253 could in principle occur. The compiled length of a repeated subpattern is
2254 now limited to 30,000 bytes in order to prevent this.
2255
2256 12. Added the optional facility to have named substrings with the same name.
2257
2258 13. Added the ability to use a named substring as a condition, using the
2259 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2260 are numbers (not recommended). Forward references are permitted.
2261
2262 14. Added forward references in named backreferences (if you see what I mean).
2263
2264 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2265 pattern could run off the end of the subject. For example, the pattern
2266 "(?s)(.{1,5})"8 did this with the subject "ab".
2267
2268 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2269 PCRE_CASELESS was set when matching characters that were quantified with ?
2270 or *.
2271
2272 17. A character class other than a single negated character that had a minimum
2273 but no maximum quantifier - for example [ab]{6,} - was not handled
2274 correctly by pce_dfa_exec(). It would match only one character.
2275
2276 18. A valid (though odd) pattern that looked like a POSIX character
2277 class but used an invalid character after [ (for example [[,abc,]]) caused
2278 pcre_compile() to give the error "Failed: internal error: code overflow" or
2279 in some cases to crash with a glibc free() error. This could even happen if
2280 the pattern terminated after [[ but there just happened to be a sequence of
2281 letters, a binary zero, and a closing ] in the memory that followed.
2282
2283 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2284 over the years. Originally (before any Unicode support), just the bottom 8
2285 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2286 output from "man perlunicode" includes this:
2287
2288 The regular expression compiler produces polymorphic opcodes. That
2289 is, the pattern adapts to the data and automatically switches to
2290 the Unicode character scheme when presented with Unicode data--or
2291 instead uses a traditional byte scheme when presented with byte
2292 data.
2293
2294 Sadly, a wide octal escape does not cause a switch, and in a string with
2295 no other multibyte characters, these octal escapes are treated as before.
2296 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2297 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2298 Unicode string.
2299
2300 I have not perpetrated such confusion in PCRE. Up till now, it took just
2301 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2302 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2303 translate to the appropriate multibyte character.
2304
2305 29. Applied some refactoring to reduce the number of warnings from Microsoft
2306 and Borland compilers. This has included removing the fudge introduced
2307 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2308 a warning about an unused variable.
2309
2310 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2311 characters since release 4.0, because Perl (from release 5.004) does not.
2312 [Or at least, is documented not to: some releases seem to be in conflict
2313 with the documentation.] However, when a pattern was studied with
2314 pcre_study() and all its branches started with \s, PCRE still included VT
2315 as a possible starting character. Of course, this did no harm; it just
2316 caused an unnecessary match attempt.
2317
2318 22. Removed a now-redundant internal flag bit that recorded the fact that case
2319 dependency changed within the pattern. This was once needed for "required
2320 byte" processing, but is no longer used. This recovers a now-scarce options
2321 bit. Also moved the least significant internal flag bit to the most-
2322 significant bit of the word, which was not previously used (hangover from
2323 the days when it was an int rather than a uint) to free up another bit for
2324 the future.
2325
2326 23. Added support for CRLF line endings as well as CR and LF. As well as the
2327 default being selectable at build time, it can now be changed at runtime
2328 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2329 specify that it is scanning data with non-default line endings.
2330
2331 24. Changed the definition of CXXLINK to make it agree with the definition of
2332 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2333
2334 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2335 recursions. This makes a big different to stack usage for some patterns.
2336
2337 26. If a subpattern containing a named recursion or subroutine reference such
2338 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2339 the space required for the compiled pattern went wrong and gave too small a
2340 value. Depending on the environment, this could lead to "Failed: internal
2341 error: code overflow at offset 49" or "glibc detected double free or
2342 corruption" errors.
2343
2344 27. Applied patches from Google (a) to support the new newline modes and (b) to
2345 advance over multibyte UTF-8 characters in GlobalReplace.
2346
2347 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2348 difference for some implementation of PCRE in some Windows version.
2349
2350 29. Added some extra testing facilities to pcretest:
2351
2352 \q<number> in a data line sets the "match limit" value
2353 \Q<number> in a data line sets the "match recursion limt" value
2354 -S <number> sets the stack size, where <number> is in megabytes
2355
2356 The -S option isn't available for Windows.
2357
2358
2359 Version 6.6 06-Feb-06
2360 ---------------------
2361
2362 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2363 in pcreposix.h. I have copied the definition from pcre.h.
2364
2365 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2366 because pcre.h is no longer a built file.
2367
2368 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2369 not normally included in the compiled code.
2370
2371
2372 Version 6.5 01-Feb-06
2373 ---------------------
2374
2375 1. When using the partial match feature with pcre_dfa_exec(), it was not
2376 anchoring the second and subsequent partial matches at the new starting
2377 point. This could lead to incorrect results. For example, with the pattern
2378 /1234/, partially matching against "123" and then "a4" gave a match.
2379
2380 2. Changes to pcregrep:
2381
2382 (a) All non-match returns from pcre_exec() were being treated as failures
2383 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2384 error message is output. Some extra information is given for the
2385 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2386 probably the only errors that are likely to be caused by users (by
2387 specifying a regex that has nested indefinite repeats, for instance).
2388 If there are more than 20 of these errors, pcregrep is abandoned.
2389
2390 (b) A binary zero was treated as data while matching, but terminated the
2391 output line if it was written out. This has been fixed: binary zeroes
2392 are now no different to any other data bytes.
2393
2394 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2395 used to set a locale for matching. The --locale=xxxx long option has
2396 been added (no short equivalent) to specify a locale explicitly on the
2397 pcregrep command, overriding the environment variables.
2398
2399 (d) When -B was used with -n, some line numbers in the output were one less
2400 than they should have been.
2401
2402 (e) Added the -o (--only-matching) option.
2403
2404 (f) If -A or -C was used with -c (count only), some lines of context were
2405 accidentally printed for the final match.
2406
2407 (g) Added the -H (--with-filename) option.
2408
2409 (h) The combination of options -rh failed to suppress file names for files
2410 that were found from directory arguments.
2411
2412 (i) Added the -D (--devices) and -d (--directories) options.
2413
2414 (j) Added the -F (--fixed-strings) option.
2415
2416 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2417
2418 (l) Added the --colo(u)r option.
2419
2420 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2421 is not present by default.
2422
2423 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2424 items such as (?R) or (?1), when the recursion could match a number of
2425 alternatives. If it matched one of the alternatives, but subsequently,
2426 outside the recursion, there was a failure, the code tried to back up into
2427 the recursion. However, because of the way PCRE is implemented, this is not
2428 possible, and the result was an incorrect result from the match.
2429
2430 In order to prevent this happening, the specification of recursion has
2431 been changed so that all such subpatterns are automatically treated as
2432 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2433
2434 4. I had overlooked the fact that, in some locales, there are characters for
2435 which isalpha() is true but neither isupper() nor islower() are true. In
2436 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2437 and ordfeminine) are like this. This affected the treatment of \w and \W
2438 when they appeared in character classes, but not when they appeared outside
2439 a character class. The bit map for "word" characters is now created
2440 separately from the results of isalnum() instead of just taking it from the
2441 upper, lower, and digit maps. (Plus the underscore character, of course.)
2442
2443 5. The above bug also affected the handling of POSIX character classes such as
2444 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2445 permanent tables. Instead, the bit maps for such a class were previously
2446 created as the appropriate unions of the upper, lower, and digit bitmaps.
2447 Now they are created by subtraction from the [[:word:]] class, which has
2448 its own bitmap.
2449
2450 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2451 It is created by subtracting the vertical space characters (\x09, \x0a,
2452 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2453 subtraction was done in the overall bitmap for a character class, meaning
2454 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2455 be recognized. This bug has been fixed.
2456
2457 7. Patches from the folks at Google:
2458
2459 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2460 real life, but is still worth protecting against".
2461
2462 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2463 regular expressions".
2464
2465 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2466 have it.
2467
2468 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2469 "configure" and the latter not, in order to fix a problem somebody had
2470 with compiling the Arg class on HP-UX.
2471
2472 (e) Improve the error-handling of the C++ wrapper a little bit.
2473
2474 (f) New tests for checking recursion limiting.
2475
2476 8. The pcre_memmove() function, which is used only if the environment does not
2477 have a standard memmove() function (and is therefore rarely compiled),
2478 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2479 returning a result (though PCRE never actually uses the result).
2480
2481 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2482 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2483 returned instead of calling malloc() with an overflowing number that would
2484 most likely cause subsequent chaos.
2485
2486 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2487
2488 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2489 with this option is matched, the nmatch and pmatch options of regexec() are
2490 ignored.
2491
2492 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2493 provided in case anyone wants to the the POSIX interface with UTF-8
2494 strings.
2495
2496 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2497 C++ linking (needed for some HP-UX environments).
2498
2499 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2500 (unused parameter) and in the pcre_printint() function (omitted "default"
2501 switch label when the default is to do nothing).
2502
2503 15. Added some code to make it possible, when PCRE is compiled as a C++
2504 library, to replace subject pointers for pcre_exec() with a smart pointer
2505 class, thus making it possible to process discontinuous strings.
2506
2507 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2508 much the same function. They were added by different people who were trying
2509 to make PCRE easy to compile on non-Unix systems. It has been suggested
2510 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2511 for compiling on Windows systems. I have therefore replaced it with
2512 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2513 defaults to "extern" for C or "extern C" for C++, which works fine on
2514 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2515 SCOPE with something explicit in config.h. In addition:
2516
2517 (a) pcreposix.h still had just "extern" instead of either of these macros;
2518 I have replaced it with PCRE_DATA_SCOPE.
2519
2520 (b) Functions such as _pcre_xclass(), which are internal to the library,
2521 but external in the C sense, all had PCRE_EXPORT in their definitions.
2522 This is apparently wrong for the Windows case, so I have removed it.
2523 (It makes no difference on Unix-like systems.)
2524
2525 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2526 of recursive calls to match(). This is different to MATCH_LIMIT because
2527 that limits the total number of calls to match(), not all of which increase
2528 the depth of recursion. Limiting the recursion depth limits the amount of
2529 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2530 when PCRE is compiled, and changed at run time. A patch from Google adds
2531 this functionality to the C++ interface.
2532
2533 18. Changes to the handling of Unicode character properties:
2534
2535 (a) Updated the table to Unicode 4.1.0.
2536
2537 (b) Recognize characters that are not in the table as "Cn" (undefined).
2538
2539 (c) I revised the way the table is implemented to a much improved format
2540 which includes recognition of ranges. It now supports the ranges that
2541 are defined in UnicodeData.txt, and it also amalgamates other
2542 characters into ranges. This has reduced the number of entries in the
2543 table from around 16,000 to around 3,000, thus reducing its size
2544 considerably. I realized I did not need to use a tree structure after
2545 all - a binary chop search is just as efficient. Having reduced the
2546 number of entries, I extended their size from 6 bytes to 8 bytes to
2547 allow for more data.
2548
2549 (d) Added support for Unicode script names via properties such as \p{Han}.
2550
2551 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2552 matching that character.
2553
2554 20. When matching a repeated Unicode property with a minimum greater than zero,
2555 (for example \pL{2,}), PCRE could look past the end of the subject if it
2556 reached it while seeking the minimum number of characters. This could
2557 happen only if some of the characters were more than one byte long, because
2558 there is a check for at least the minimum number of bytes.
2559
2560 21. Refactored the implementation of \p and \P so as to be more general, to
2561 allow for more different types of property in future. This has changed the
2562 compiled form incompatibly. Anybody with saved compiled patterns that use
2563 \p or \P will have to recompile them.
2564
2565 22. Added "Any" and "L&" to the supported property types.
2566
2567 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2568 but give a compile time error if the value is greater than 0xff.
2569
2570 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2571 accidentally not being installed or uninstalled.
2572
2573 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2574 made were to insert the current release number. This seemed silly, because
2575 it made things harder for people building PCRE on systems that don't run
2576 "configure". I have turned pcre.h into a distributed file, no longer built
2577 by "configure", with the version identification directly included. There is
2578 no longer a pcre.h.in file.
2579
2580 However, this change necessitated a change to the pcre-config script as
2581 well. It is built from pcre-config.in, and one of the substitutions was the
2582 release number. I have updated configure.ac so that ./configure now finds
2583 the release number by grepping pcre.h.
2584
2585 26. Added the ability to run the tests under valgrind.
2586
2587
2588 Version 6.4 05-Sep-05
2589 ---------------------
2590
2591 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2592 "--" to be printed when multiple files were scanned, even when none of the
2593 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2594 consider it to be a bug, and have restored the previous behaviour.
2595
2596 2. A couple of code tidies to get rid of compiler warnings.
2597
2598 3. The pcretest program used to cheat by referring to symbols in the library
2599 whose names begin with _pcre_. These are internal symbols that are not
2600 really supposed to be visible externally, and in some environments it is
2601 possible to suppress them. The cheating is now confined to including
2602 certain files from the library's source, which is a bit cleaner.
2603
2604 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2605 file's purpose clearer.
2606
2607 5. Reorganized pcre_ucp_findchar().
2608
2609
2610 Version 6.3 15-Aug-05
2611 ---------------------
2612
2613 1. The file libpcre.pc.in did not have general read permission in the tarball.
2614
2615 2. There were some problems when building without C++ support:
2616
2617 (a) If C++ support was not built, "make install" and "make test" still
2618 tried to test it.
2619
2620 (b) There were problems when the value of CXX was explicitly set. Some
2621 changes have been made to try to fix these, and ...
2622
2623 (c) --disable-cpp can now be used to explicitly disable C++ support.
2624
2625 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2626 backslash in a target when C++ was disabled. This confuses some
2627 versions of "make", apparently. Using an intermediate variable solves
2628 this. (Same for CPP_LOBJ.)
2629
2630 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2631 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2632 necessary on certain architectures.
2633
2634 4. Added a setting of -export-symbols-regex to the link command to remove
2635 those symbols that are exported in the C sense, but actually are local
2636 within the library, and not documented. Their names all begin with
2637 "_pcre_". This is not a perfect job, because (a) we have to except some
2638 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2639 available (and never for static libraries). I have made a note to try to
2640 find a way round (a) in the future.
2641
2642
2643 Version 6.2 01-Aug-05
2644 ---------------------
2645
2646 1. There was no test for integer overflow of quantifier values. A construction
2647 such as {1111111111111111} would give undefined results. What is worse, if
2648 a minimum quantifier for a parenthesized subpattern overflowed and became
2649 negative, the calculation of the memory size went wrong. This could have
2650 led to memory overwriting.
2651
2652 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2653
2654 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2655 operating environments where this matters.
2656
2657 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2658 PCRE options from within the C++ wrapper.
2659
2660 5. Named capturing subpatterns were not being correctly counted when a pattern
2661 was compiled. This caused two problems: (a) If there were more than 100
2662 such subpatterns, the calculation of the memory needed for the whole
2663 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2664 back references of the form \12, where the number was greater than 9, were
2665 not recognized as back references, even though there were sufficient
2666 previous subpatterns.
2667
2668 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2669 versions of gcc, e.g. 2.95.4.
2670
2671
2672 Version 6.1 21-Jun-05
2673 ---------------------
2674
2675 1. There was one reference to the variable "posix" in pcretest.c that was not
2676 surrounded by "#if !defined NOPOSIX".
2677
2678 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2679 the cross-check on the old pcre_info() function, for the benefit of the
2680 cut-down version of PCRE that is currently imported into Exim.
2681
2682 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2683 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2684 patterns that start (?i) but I don't think that matters. The use of (?i) is
2685 just an example; this all applies to the other options as well.
2686
2687 4. Since libtool seems to echo the compile commands it is issuing, the output
2688 from "make" can be reduced a bit by putting "@" in front of each libtool
2689 compile command.
2690
2691 5. Patch from the folks at Google for configure.in to be a bit more thorough
2692 in checking for a suitable C++ installation before trying to compile the
2693 C++ stuff. This should fix a reported problem when a compiler was present,
2694 but no suitable headers.
2695
2696 6. The man pages all had just "PCRE" as their title. I have changed them to
2697 be the relevant file name. I have also arranged that these names are
2698 retained in the file doc/pcre.txt, which is a concatenation in text format
2699 of all the man pages except the little individual ones for each function.
2700
2701 7. The NON-UNIX-USE file had not been updated for the different set of source
2702 files that come with release 6. I also added a few comments about the C++
2703 wrapper.
2704
2705
2706 Version 6.0 07-Jun-05
2707 ---------------------
2708
2709 1. Some minor internal re-organization to help with my DFA experiments.
2710
2711 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2712 didn't matter for the library itself when fully configured, but did matter
2713 when compiling without UCP support, or within Exim, where the ucp files are
2714 not imported.
2715
2716 3. Refactoring of the library code to split up the various functions into
2717 different source modules. The addition of the new DFA matching code (see
2718 below) to a single monolithic source would have made it really too
2719 unwieldy, quite apart from causing all the code to be include in a
2720 statically linked application, when only some functions are used. This is
2721 relevant even without the DFA addition now that patterns can be compiled in
2722 one application and matched in another.
2723
2724 The downside of splitting up is that there have to be some external
2725 functions and data tables that are used internally in different modules of
2726 the library but which are not part of the API. These have all had their
2727 names changed to start with "_pcre_" so that they are unlikely to clash
2728 with other external names.
2729
2730 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2731 a different (DFA) algorithm. Although it is slower than the original
2732 function, it does have some advantages for certain types of matching
2733 problem.
2734
2735 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2736 including restarting after a partial match.
2737
2738 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2739 defined when compiling for Windows was sent to me. I have put it into the
2740 code, though I have no means of testing or verifying it.
2741
2742 7. Added the pcre_refcount() auxiliary function.
2743
2744 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2745 match before or at the first newline in the subject string. In pcretest,
2746 the /f option on a pattern can be used to set this.
2747
2748 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2749 would behave wrongly. This has been present in PCRE since release 4.0.
2750
2751 10. A number of changes to the pcregrep command:
2752
2753 (a) Refactored how -x works; insert ^(...)$ instead of setting
2754 PCRE_ANCHORED and checking the length, in preparation for adding
2755 something similar for -w.
2756
2757 (b) Added the -w (match as a word) option.
2758
2759 (c) Refactored the way lines are read and buffered so as to have more
2760 than one at a time available.
2761
2762 (d) Implemented a pcregrep test script.
2763
2764 (e) Added the -M (multiline match) option. This allows patterns to match
2765 over several lines of the subject. The buffering ensures that at least
2766 8K, or the rest of the document (whichever is the shorter) is available
2767 for matching (and similarly the previous 8K for lookbehind assertions).
2768
2769 (f) Changed the --help output so that it now says
2770
2771 -w, --word-regex(p)
2772
2773 instead of two lines, one with "regex" and the other with "regexp"
2774 because that confused at least one person since the short forms are the
2775 same. (This required a bit of code, as the output is generated
2776 automatically from a table. It wasn't just a text change.)
2777
2778 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2779 option but starts with a hyphen. Could be a pattern or a path name
2780 starting with a hyphen, for instance.
2781
2782 (h) "-" can be given as a file name to represent stdin.
2783
2784 (i) When file names are being printed, "(standard input)" is used for
2785 the standard input, for compatibility with GNU grep. Previously
2786 "<stdin>" was used.
2787
2788 (j) The option --label=xxx can be used to supply a name to be used for
2789 stdin when file names are being printed. There is no short form.
2790
2791 (k) Re-factored the options decoding logic because we are going to add
2792 two more options that take data. Such options can now be given in four
2793 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2794
2795 (l) Added the -A, -B, and -C options for requesting that lines of context
2796 around matches be printed.
2797
2798 (m) Added the -L option to print the names of files that do not contain
2799 any matching lines, that is, the complement of -l.
2800
2801 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2802 continue to scan other files.
2803
2804 (o) The -s option was incorrectly implemented. For compatibility with other
2805 greps, it now suppresses the error message for a non-existent or non-
2806 accessible file (but not the return code). There is a new option called
2807 -q that suppresses the output of matching lines, which was what -s was
2808 previously doing.
2809
2810 (p) Added --include and --exclude options to specify files for inclusion
2811 and exclusion when recursing.
2812
2813 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2814 Hopefully, it now does.
2815
2816 12. Missing cast in pcre_study().
2817
2818 13. Added an "uninstall" target to the makefile.
2819
2820 14. Replaced "extern" in the function prototypes in Makefile.in with
2821 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2822 world, but is set differently for Windows.
2823
2824 15. Added a second compiling function called pcre_compile2(). The only
2825 difference is that it has an extra argument, which is a pointer to an
2826 integer error code. When there is a compile-time failure, this is set
2827 non-zero, in addition to the error test pointer being set to point to an
2828 error message. The new argument may be NULL if no error number is required
2829 (but then you may as well call pcre_compile(), which is now just a
2830 wrapper). This facility is provided because some applications need a
2831 numeric error indication, but it has also enabled me to tidy up the way
2832 compile-time errors are handled in the POSIX wrapper.
2833
2834 16. Added VPATH=.libs to the makefile; this should help when building with one
2835 prefix path and installing with another. (Or so I'm told by someone who
2836 knows more about this stuff than I do.)
2837
2838 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2839 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2840 match everything, including newlines. This is not POSIX-compatible, but
2841 somebody wanted the feature. From pcretest it can be activated by using
2842 both the P and the s flags.
2843
2844 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2845
2846 19. libpcre.pc was being incorrectly installed as executable.
2847
2848 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2849 it now also looks for '\r' so that it will work unmodified on Windows.
2850
2851 21. Added Google's contributed C++ wrapper to the distribution.
2852
2853 22. Added some untidy missing memory free() calls in pcretest, to keep
2854 Electric Fence happy when testing.
2855
2856
2857
2858 Version 5.0 13-Sep-04
2859 ---------------------
2860
2861 1. Internal change: literal characters are no longer packed up into items
2862 containing multiple characters in a single byte-string. Each character
2863 is now matched using a separate opcode. However, there may be more than one
2864 byte in the character in UTF-8 mode.
2865
2866 2. The pcre_callout_block structure has two new fields: pattern_position and
2867 next_item_length. These contain the offset in the pattern to the next match
2868 item, and its length, respectively.
2869
2870 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2871 insertion of callouts before each pattern item. Added the /C option to
2872 pcretest to make use of this.
2873
2874 4. On the advice of a Windows user, the lines
2875
2876 #if defined(_WIN32) || defined(WIN32)
2877 _setmode( _fileno( stdout ), 0x8000 );
2878 #endif /* defined(_WIN32) || defined(WIN32) */
2879
2880 have been added to the source of pcretest. This apparently does useful
2881 magic in relation to line terminators.
2882
2883 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2884 for the benefit of those environments where the "b" makes a difference.
2885
2886 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2887 to know about it. I have put a hack into configure.in that adds in code
2888 to set GCC=yes if CC=icc. This seems to end up at a point in the
2889 generated configure script that is early enough to affect the setting of
2890 compiler options, which is what is needed, but I have no means of testing
2891 whether it really works. (The user who reported this had patched the
2892 generated configure script, which of course I cannot do.)
2893
2894 LATER: After change 22 below (new libtool files), the configure script
2895 seems to know about icc (and also ecc). Therefore, I have commented out
2896 this hack in configure.in.
2897
2898 7. Added support for pkg-config (2 patches were sent in).
2899
2900 8. Negated POSIX character classes that used a combination of internal tables
2901 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2902 [[:^ascii]]. Typically, they would match almost any characters. The other
2903 POSIX classes were not broken in this way.
2904
2905 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2906 to find the match, as PCRE was deluded into thinking that the match had to
2907 start at the start point or following a newline. The same bug applied to
2908 patterns with negative forward assertions or any backward assertions
2909 preceding ".*" at the start, unless the pattern required a fixed first
2910 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2911
2912 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2913 starting at the last subject character, bytes beyond the end of the subject
2914 string were read.
2915
2916 11. Renamed the variable "class" as "classbits" to make life easier for C++
2917 users. (Previously there was a macro definition, but it apparently wasn't
2918 enough.)
2919
2920 12. Added the new field "tables" to the extra data so that tables can be passed
2921 in at exec time, or the internal tables can be re-selected. This allows
2922 a compiled regex to be saved and re-used at a later time by a different
2923 program that might have everything at different addresses.
2924
2925 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2926 -R library as well as a -L library.
2927
2928 14. The debugging options of pcretest (-d on the command line or D on a
2929 pattern) showed incorrect output for anything following an extended class
2930 that contained multibyte characters and which was followed by a quantifier.
2931
2932 15. Added optional support for general category Unicode character properties
2933 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2934 support. It adds about 90K to the size of the library. The meanings of the
2935 inbuilt class escapes such as \d and \s have NOT been changed.
2936
2937 16. Updated pcredemo.c to include calls to free() to release the memory for the
2938 compiled pattern.
2939
2940 17. The generated file chartables.c was being created in the source directory
2941 instead of in the building directory. This caused the build to fail if the
2942 source directory was different from the building directory, and was
2943 read-only.
2944
2945 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2946 file. No doubt somebody will tell me if they don't make sense... Also added
2947 Dan Mooney's comments about building on OpenVMS.
2948
2949 19. Added support for partial matching via the PCRE_PARTIAL option for
2950 pcre_exec() and the \P data escape in pcretest.
2951
2952 20. Extended pcretest with 3 new pattern features:
2953
2954 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2955 write the compiled pattern to the file whose name is "rest-of-line".
2956 This is a straight binary dump of the data, with the saved pointer to
2957 the character tables forced to be NULL. The study data, if any, is
2958 written too. After writing, pcretest reads a new pattern.
2959
2960 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2961 compiled pattern from the given file. There must not be any
2962 occurrences of "<" in the file name (pretty unlikely); if there are,
2963 pcretest will instead treat the initial "<" as a pattern delimiter.
2964 After reading in the pattern, pcretest goes on to read data lines as
2965 usual.
2966
2967 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2968 and 16-bit fields in a compiled pattern, to simulate a pattern that
2969 was compiled on a host of opposite endianness.
2970
2971 21. The pcre-exec() function can now cope with patterns that were compiled on
2972 hosts of opposite endianness, with this restriction:
2973
2974 As for any compiled expression that is saved and used later, the tables
2975 pointer field cannot be preserved; the extra_data field in the arguments
2976 to pcre_exec() should be used to pass in a tables address if a value
2977 other than the default internal tables were used at compile time.
2978
2979 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2980 now diagnosed as an error. Previously, most of the time, a negative number
2981 would have been treated as zero, but if in addition "ovector" was passed as
2982 NULL, a crash could occur.
2983
2984 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2985 new versions from the libtool 1.5 distribution (the last one is a copy of
2986 a file called libtool.m4). This seems to have fixed the need to patch
2987 "configure" to support Darwin 1.3 (which I used to do). However, I still
2988 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2989 workstation).
2990
2991 24. Changed the PCRE licence to be the more standard "BSD" licence.
2992
2993
2994 Version 4.5 01-Dec-03
2995 ---------------------
2996
2997 1. There has been some re-arrangement of the code for the match() function so
2998 that it can be compiled in a version that does not call itself recursively.
2999 Instead, it keeps those local variables that need separate instances for
3000 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3001 needs to "recurse". Keeping track of where control must go is done by means
3002 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3003 hide most of the details from the main code, and operates only if
3004 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3005 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3006 operating.
3007
3008 To make it easier for callers to provide specially tailored get/free
3009 functions for this usage, two new functions, pcre_stack_malloc, and
3010 pcre_stack_free, are used. They are always called in strict stacking order,
3011 and the size of block requested is always the same.
3012
3013 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3014 PCRE has been compiled to use the stack or the heap for recursion. The
3015 -C option of pcretest uses this to show which version is compiled.
3016
3017 A new data escape \S, is added to pcretest; it causes the amounts of store
3018 obtained and freed by both kinds of malloc/free at match time to be added
3019 to the output.
3020
3021 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3022 what's available on my current Linux desktop machine.
3023
3024 3. When matching a UTF-8 string, the test for a valid string at the start has
3025 been extended. If start_offset is not zero, PCRE now checks that it points
3026 to a byte that is the start of a UTF-8 character. If not, it returns
3027 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3028 this is necessary because there may be backward assertions in the pattern.
3029 When matching the same subject several times, it may save resources to use
3030 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3031
3032 4. The code for checking the validity of UTF-8 strings has been tightened so
3033 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3034 containing "overlong sequences".
3035
3036 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3037 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3038 should have been "&", but it just so happened that all the cases this let
3039 through by mistake were picked up later in the function.
3040
3041 6. I had used a variable called "isblank" - this is a C99 function, causing
3042 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3043
3044 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3045 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3046 so the version is shown; (c) stop "make test" echoing "./RunTest".
3047
3048 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3049
3050 9. The prototype for memmove() for systems that don't have it was using
3051 size_t, but the inclusion of the header that defines size_t was later. I've
3052 moved the #includes for the C headers earlier to avoid this.
3053
3054 10. Added some adjustments to the code to make it easier to compiler on certain
3055 special systems:
3056
3057 (a) Some "const" qualifiers were missing.
3058 (b) Added the macro EXPORT before all exported functions; by default this
3059 is defined to be empty.
3060 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3061 that it reads its output file name as an argument instead of writing
3062 to the standard output and assuming this can be redirected.
3063
3064 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3065 class containing characters with values greater than 255, PCRE compilation
3066 went into a loop.
3067
3068 12. A recursive reference to a subpattern that was within another subpattern
3069 that had a minimum quantifier of zero caused PCRE to crash. For example,
3070 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3071 recursion. If the recursively-called subpattern itself had a zero repeat,
3072 that was OK.
3073
3074 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3075 buffer into which it was copied (for escape processing) was still set at
3076 1024, so long lines caused crashes.
3077
3078 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3079 "internal error: code overflow...". This applied to any character class
3080 that was followed by a possessive quantifier.
3081
3082 15. Modified the Makefile to add libpcre.la as a prerequisite for
3083 libpcreposix.la because I was told this is needed for a parallel build to
3084 work.
3085
3086 16. If a pattern that contained .* following optional items at the start was
3087 studied, the wrong optimizing data was generated, leading to matching
3088 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3089 matching string must start with a or b or c. The correct conclusion for
3090 this pattern is that a match can start with any character.
3091
3092
3093 Version 4.4 13-Aug-03
3094 ---------------------
3095
3096 1. In UTF-8 mode, a character class containing characters with values between
3097 127 and 255 was not handled correctly if the compiled pattern was studied.
3098 In fixing this, I have also improved the studying algorithm for such
3099 classes (slightly).
3100
3101 2. Three internal functions had redundant arguments passed to them. Removal
3102 might give a very teeny performance improvement.
3103
3104 3. Documentation bug: the value of the capture_top field in a callout is *one
3105 more than* the number of the hightest numbered captured substring.
3106
3107 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3108 in incorrectly linking with a previously installed version. They now link
3109 explicitly with libpcre.la.
3110
3111 5. configure.in no longer needs to recognize Cygwin specially.
3112
3113 6. A problem in pcre.in for Windows platforms is fixed.
3114
3115 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3116 pcretest, it used to include the size of the study block as part of its
3117 output. Unfortunately, the structure contains a field that has a different
3118 size on different hardware architectures. This meant that the tests that
3119 showed this size failed. As the block is currently always of a fixed size,
3120 this information isn't actually particularly useful in pcretest output, so
3121 I have just removed it.
3122
3123 8. Three pre-processor statements accidentally did not start in column 1.
3124 Sadly, there are *still* compilers around that complain, even though
3125 standard C has not required this for well over a decade. Sigh.
3126
3127 9. In pcretest, the code for checking callouts passed small integers in the
3128 callout_data field, which is a void * field. However, some picky compilers
3129 complained about the casts involved for this on 64-bit systems. Now
3130 pcretest passes the address of the small integer instead, which should get
3131 rid of the warnings.
3132
3133 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3134 both compile and run time, and gives an error if an invalid UTF-8 sequence
3135 is found. There is a option for disabling this check in cases where the
3136 string is known to be correct and/or the maximum performance is wanted.
3137
3138 11. In response to a bug report, I changed one line in Makefile.in from
3139
3140 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3141 to
3142 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3143
3144 to look similar to other lines, but I have no way of telling whether this
3145 is the right thing to do, as I do not use Windows. No doubt I'll get told
3146 if it's wrong...
3147
3148
3149 Version 4.3 21-May-03
3150 ---------------------
3151
3152 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3153 Makefile.
3154
3155 2. Some refactoring to improve the quality of the code:
3156
3157 (i) The utf8_table... variables are now declared "const".
3158
3159 (ii) The code for \cx, which used the "case flipping" table to upper case
3160 lower case letters, now just substracts 32. This is ASCII-specific,
3161 but the whole concept of \cx is ASCII-specific, so it seems
3162 reasonable.
3163
3164 (iii) PCRE was using its character types table to recognize decimal and
3165 hexadecimal digits in the pattern. This is silly, because it handles
3166 only 0-9, a-f, and A-F, but the character types table is locale-
3167 specific, which means strange things might happen. A private
3168 table is now used for this - though it costs 256 bytes, a table is
3169 much faster than multiple explicit tests. Of course, the standard
3170 character types table is still used for matching digits in subject
3171 strings against \d.
3172
3173 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3174 ending in _t are). So I've renamed it as ESC_tee.
3175
3176 3. The first argument for regexec() in the POSIX wrapper should have been
3177 defined as "const".
3178
3179 4. Changed pcretest to use malloc() for its buffers so that they can be
3180 Electric Fenced for debugging.
3181
3182 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3183 to read one or more bytes before the start of the subject string. Often this
3184 had no effect on PCRE's behaviour, but in some circumstances it could
3185 provoke a segmentation fault.
3186
3187 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3188 to try to read one or more bytes before the start of the subject string.
3189
3190 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3191 UTF-8 support could misbehave in various ways if the subject string
3192 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3193 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3194 back over UTF-8 characters.)
3195
3196
3197 Version 4.2 14-Apr-03
3198 ---------------------
3199
3200 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3201
3202 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3203 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3204 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3205 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3206 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3207 and BUILD_EXEEXT
3208 Note: automatic setting of the BUILD variables is not yet working
3209 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3210 compile-time but not at link-time
3211 [LINK]: use for linking executables only
3212 make different versions for Windows and non-Windows
3213 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3214 libraries
3215 [LINK_FOR_BUILD]: new variable
3216 [OBJEXT]: use throughout
3217 [EXEEXT]: use throughout
3218 <winshared>: new target
3219 <wininstall>: new target
3220 <dftables.o>: use native compiler
3221 <dftables>: use native linker
3222 <install>: handle Windows platform correctly
3223 <clean>: ditto
3224 <check>: ditto
3225 copy DLL to top builddir before testing
3226
3227 As part of these changes, -no-undefined was removed again. This was reported
3228 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3229 in any case.
3230
3231 3. Some tidies to get rid of compiler warnings:
3232
3233 . In the match_data structure, match_limit was an unsigned long int, whereas
3234 match_call_count was an int. I've made them both unsigned long ints.
3235
3236 . In pcretest the fact that a const uschar * doesn't automatically cast to
3237 a void * provoked a warning.
3238
3239 . Turning on some more compiler warnings threw up some "shadow" variables
3240 and a few more missing casts.
3241
3242 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3243 option, a class that contained a single character with a value between 128
3244 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3245
3246 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3247 option, a class that contained several characters, but with at least one
3248 whose value was between 128 and 255 caused PCRE to crash.
3249
3250
3251 Version 4.1 12-Mar-03
3252 ---------------------
3253
3254 1. Compiling with gcc -pedantic found a couple of places where casts were
3255 needed, and a string in dftables.c that was longer than standard compilers are
3256 required to support.
3257
3258 2. Compiling with Sun's compiler found a few more places where the code could
3259 be tidied up in order to avoid warnings.
3260
3261 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3262 first of these names is deprecated in the latest Autoconf in favour of the name
3263 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3264 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3265 analogy I have changed it to CFLAGS_FOR_BUILD.
3266
3267 4. Added -no-undefined to the linking command in the Makefile, because this is
3268 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3269 linking step for the pcreposix library.
3270
3271 5. PCRE was failing to diagnose the case of two named groups with the same
3272 name.
3273
3274 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3275 literal character that is needed in the subject for a match, and scans along to
3276 ensure that it is present before embarking on the full matching process. This
3277 saves time in cases of nested unlimited repeats that are never going to match.
3278 Problem: the scan can take a lot of time if the subject is very long (e.g.
3279 megabytes), thus penalizing straightforward matches. It is now done only if the
3280 amount of subject to be scanned is less than 1000 bytes.
3281
3282 7. A lesser problem with the same optimization is that it was recording the
3283 first character of an anchored pattern as "needed", thus provoking a search
3284 right along the subject, even when the first match of the pattern was going to
3285 fail. The "needed" character is now not set for anchored patterns, unless it
3286 follows something in the pattern that is of non-fixed length. Thus, it still
3287 fulfils its original purpose of finding quick non-matches in cases of nested
3288 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3289
3290
3291 Version 4.0 17-Feb-03
3292 ---------------------
3293
3294 1. If a comment in an extended regex that started immediately after a meta-item
3295 extended to the end of string, PCRE compiled incorrect data. This could lead to
3296 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3297
3298 2. Moved to autoconf 2.53 and libtool 1.4.2.
3299
3300 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3301 the special perltest8 script is no longer needed - all the tests can be run
3302 from a single perltest script.
3303
3304 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3305 by \s. It has now been removed in PCRE. This means it isn't recognized as
3306 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3307 class [:space:] *does* include VT, thereby creating a mess.
3308
3309 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3310 space and tab.
3311
3312 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3313 its new features into the main test script, reducing the number of scripts.
3314
3315 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3316 were backward compatible, and made the (?i) apply to the whole pattern, as if
3317 /i were given. Now it behaves more logically, and applies the option setting
3318 only to what follows. PCRE has been changed to follow suit. However, if it
3319 finds options settings right at the start of the pattern, it extracts them into
3320 the global options, as before. Thus, they show up in the info data.
3321
3322 8. Added support for the \Q...\E escape sequence. Characters in between are
3323 treated as literals. This is slightly different from Perl in that $ and @ are
3324 also handled as literals inside the quotes. In Perl, they will cause variable
3325 interpolation. Note the following examples:
3326
3327 Pattern PCRE matches Perl matches
3328
3329 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3330 \Qabc\$xyz\E abc\$xyz abc\$xyz
3331 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3332
3333 For compatibility with Perl, \Q...\E sequences are recognized inside character
3334 classes as well as outside them.
3335
3336 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3337 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3338 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3339 signed/unsigned warnings.
3340
3341 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3342 option for pcretest, so I've replaced it by a simple function that does just
3343 that job.
3344
3345 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3346 "pcregrep -".
3347
3348 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3349 Java package. This provides some syntactic sugar for simple cases of what my
3350 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3351 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3352 item, you can use this simplified notation. Note that only makes sense with
3353 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3354 greediness, whatever the setting of the PCRE_UNGREEDY option.
3355
3356 13. A change of greediness default within a pattern was not taking effect at
3357 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3358 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3359 was abstracted outside.
3360
3361 14. PCRE now supports the \G assertion. It is true when the current matching
3362 position is at the start point of the match. This differs from \A when the
3363 starting offset is non-zero. Used with the /g option of pcretest (or similar
3364 code), it works in the same way as it does for Perl's /g option. If all
3365 alternatives of a regex begin with \G, the expression is anchored to the start
3366 match position, and the "anchored" flag is set in the compiled expression.
3367
3368 15. Some bugs concerning the handling of certain option changes within patterns
3369 have been fixed. These applied to options other than (?ims). For example,
3370 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3371 been the other way round. Some of this was related to change 7 above.
3372
3373 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3374 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3375 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3376 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3377
3378 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3379 mode. Unlike ".", it always matches newline, whatever the setting of
3380 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3381 assertions. Perl allows it, but it doesn't (in general) work because it can't
3382 calculate the length of the lookbehind. At least, that's the case for Perl
3383 5.8.0 - I've been told they are going to document that it doesn't work in
3384 future.
3385
3386 18. Added an error diagnosis for escapes that PCRE does not support: these are
3387 \L, \l, \N, \P, \p, \U, \u, and \X.
3388
3389 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3390 reading past the end of the pattern in cases such as /[abcd/.
3391
3392 20. PCRE was getting more memory than necessary for patterns with classes that
3393 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3394
3395 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3396 compiling PCRE for use with Virtual Pascal.
3397
3398 22. Small fix to the Makefile to make it work properly if the build is done
3399 outside the source tree.
3400
3401 23. Added a new extension: a condition to go with recursion. If a conditional
3402 subpattern starts with (?(R) the "true" branch is used if recursion has
3403 happened, whereas the "false" branch is used only at the top level.
3404
3405 24. When there was a very long string of literal characters (over 255 bytes
3406 without UTF support, over 250 bytes with UTF support), the computation of how
3407 much memory was required could be incorrect, leading to segfaults or other
3408 strange effects.
3409
3410 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3411 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3412 there was a subsequent back reference to those brackets. This meant that, for
3413 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3414 possible to check for precisely this case. All we can do is abandon the
3415 optimization if .* occurs inside capturing brackets when there are any back
3416 references whatsoever. (See below for a better fix that came later.)
3417
3418 26. The handling of the optimization for finding the first character of a
3419 non-anchored pattern, and for finding a character that is required later in the
3420 match were failing in some cases. This didn't break the matching; it just
3421 failed to optimize when it could. The way this is done has been re-implemented.
3422
3423 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3424
3425 28. Added a new feature that provides some of the functionality that Perl
3426 provides with (?{...}). The facility is termed a "callout". The way it is done
3427 in PCRE is for the caller to provide an optional function, by setting
3428 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3429 global variable. By default it is unset, which disables all calling out. To get
3430 the function called, the regex must include (?C) at appropriate points. This
3431 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3432 This provides a means of identifying different callout points. When PCRE
3433 reaches such a point in the regex, if pcre_callout has been set, the external
3434 function is called. It is provided with data in a structure called
3435 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3436 matching continues; if it returns a non-zero value, the match at the current
3437 point fails. However, backtracking will occur if possible. [This was changed
3438 later and other features added - see item 49 below.]
3439
3440 29. pcretest is upgraded to test the callout functionality. It provides a
3441 callout function that displays information. By default, it shows the start of
3442 the match and the current position in the text. There are some new data escapes
3443 to vary what happens:
3444
3445 \C+ in addition, show current contents of captured substrings
3446 \C- do not supply a callout function
3447 \C!n return 1 when callout number n is reached
3448 \C!n!m return 1 when callout number n is reached for the mth time
3449
3450 30. If pcregrep was called with the -l option and just a single file name, it
3451 output "<stdin>" if a match was found, instead of the file name.
3452
3453 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3454 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3455 pcre_exec(). This saves a malloc/free per call. The default value of
3456 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3457 when configuring.
3458
3459 32. The default maximum size of a compiled pattern is 64K. There have been a
3460 few cases of people hitting this limit. The code now uses macros to handle the
3461 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3462 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3463 configuring. Tests 2 and 5 work only with 2-byte links because they output
3464 debugging information about compiled patterns.
3465
3466 33. Internal code re-arrangements:
3467
3468 (a) Moved the debugging function for printing out a compiled regex into
3469 its own source file (printint.c) and used #include to pull it into
3470 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3471 separate copies.
3472
3473 (b) Defined the list of op-code names for debugging as a macro in
3474 internal.h so that it is next to the definition of the opcodes.
3475
3476 (c) Defined a table of op-code lengths for simpler skipping along compiled
3477 code. This is again a macro in internal.h so that it is next to the
3478 definition of the opcodes.
3479
3480 34. Added support for recursive calls to individual subpatterns, along the
3481 lines of Robin Houston's patch (but implemented somewhat differently).
3482
3483 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3484 allow it to read and process whole directories in Win32. This code was
3485 contributed by Lionel Fourquaux; it has not been tested by me.
3486
3487 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3488 used to name a group. Names consist of alphanumerics and underscores, and must
3489 be unique. Back references use the syntax (?P=name) and recursive calls use
3490 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3491 numbers. The function pcre_fullinfo() can be used after compilation to extract
3492 a name/number map. There are three relevant calls:
3493
3494 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3495 PCRE_INFO_NAMECOUNT yields the number of entries
3496 PCRE_INFO_NAMETABLE yields a pointer to the map.
3497
3498 The map is a vector of fixed-size entries. The size of each entry depends on
3499 the length of the longest name used. The first two bytes of each entry are the
3500 group number, most significant byte first. There follows the corresponding
3501 name, zero terminated. The names are in alphabetical order.
3502
3503 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3504 case instead of 255. Making it the same both with and without UTF-8 support
3505 means that the same test output works with both.
3506
3507 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3508 calling malloc() with a zero argument.
3509
3510 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3511 optimization. I've improved things by keeping a bitmap of backreferences with
3512 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3513 fact referenced, the optimization can be applied. It is unlikely that a
3514 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3515 the match to follow \n) will appear inside brackets with a number greater than
3516 31, but if it does, any back reference > 31 suppresses the optimization.
3517
3518 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3519 of disabling numbered capturing parentheses. Any opening parenthesis that is
3520 not followed by ? behaves as if it were followed by ?: but named parentheses
3521 can still be used for capturing (and they will acquire numbers in the usual
3522 way).
3523
3524 41. Redesigned the return codes from the match() function into yes/no/error so
3525 that errors can be passed back from deep inside the nested calls. A malloc
3526 failure while inside a recursive subpattern call now causes the
3527 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3528
3529 42. It is now possible to set a limit on the number of times the match()
3530 function is called in a call to pcre_exec(). This facility makes it possible to
3531 limit the amount of recursion and backtracking, though not in a directly
3532 obvious way, because the match() function is used in a number of different
3533 circumstances. The count starts from zero for each position in the subject
3534 string (for non-anchored patterns). The default limit is, for compatibility, a
3535 large number, namely 10 000 000. You can change this in two ways:
3536
3537 (a) When configuring PCRE before making, you can use --with-match-limit=n
3538 to set a default value for the compiled library.
3539
3540 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3541 a different value is set. See 45 below.
3542
3543 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3544
3545 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3546 of things that can be changed at compile time. The first argument specifies
3547 what is wanted and the second points to where the information is to be placed.
3548 The current list of available information is:
3549
3550 PCRE_CONFIG_UTF8
3551
3552 The output is an integer that is set to one if UTF-8 support is available;
3553 otherwise it is set to zero.
3554
3555 PCRE_CONFIG_NEWLINE
3556
3557 The output is an integer that it set to the value of the code that is used for
3558 newline. It is either LF (10) or CR (13).
3559
3560 PCRE_CONFIG_LINK_SIZE
3561
3562 The output is an integer that contains the number of bytes used for internal
3563 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3564
3565 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3566
3567 The output is an integer that contains the threshold above which the POSIX
3568 interface uses malloc() for output vectors. See item 31 above.
3569
3570 PCRE_CONFIG_MATCH_LIMIT
3571
3572 The output is an unsigned integer that contains the default limit of the number
3573 of match() calls in a pcre_exec() execution. See 42 above.
3574
3575 44. pcretest has been upgraded by the addition of the -C option. This causes it
3576 to extract all the available output from the new pcre_config() function, and to
3577 output it. The program then exits immediately.
3578
3579 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3580 order to support additional features. One way would have been to define
3581 pcre_exec2() (for example) with extra arguments, but this would not have been
3582 extensible, and would also have required all calls to the original function to
3583 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3584 is used for passing in "extra" data from pcre_study().
3585
3586 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3587 contains the following fields:
3588
3589 flags a bitmap indicating which of the following fields are set
3590 study_data opaque data from pcre_study()
3591 match_limit a way of specifying a limit on match() calls for a specific
3592 call to pcre_exec()
3593 callout_data data for callouts (see 49 below)
3594
3595 The flag bits are also defined in pcre.h, and are
3596
3597 PCRE_EXTRA_STUDY_DATA
3598 PCRE_EXTRA_MATCH_LIMIT
3599 PCRE_EXTRA_CALLOUT_DATA
3600
3601 The pcre_study() function now returns one of these new pcre_extra blocks, with
3602 the actual study data pointed to by the study_data field, and the
3603 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3604 before. That is, this change is entirely upwards-compatible and requires no
3605 change to existing code.
3606
3607 If you want to pass in additional data to pcre_exec(), you can either place it
3608 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3609 block.
3610
3611 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3612 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3613 times with different match limits, until it finds the minimum value needed for
3614 pcre_exec() to complete. The value is then output. This can be instructive; for
3615 most simple matches the number is quite small, but for pathological cases it
3616 gets very large very quickly.
3617
3618 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3619 returns the size of the data block pointed to by the study_data field in a
3620 pcre_extra block, that is, the value that was passed as the argument to
3621 pcre_malloc() when PCRE was getting memory in which to place the information
3622 created by pcre_study(). The fourth argument should point to a size_t variable.
3623 pcretest has been extended so that this information is shown after a successful
3624 pcre_study() call when information about the compiled regex is being displayed.
3625
3626 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3627 because what follows is always an absolute path. (Later: it turns out that this
3628 is more than cosmetic for MinGW, because it doesn't like empty path
3629 components.)
3630
3631 49. Some changes have been made to the callout feature (see 28 above):
3632
3633 (i) A callout function now has three choices for what it returns:
3634
3635 0 => success, carry on matching
3636 > 0 => failure at this point, but backtrack if possible
3637 < 0 => serious error, return this value from pcre_exec()
3638
3639 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3640 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3641 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3642 use by callout functions. It will never be used by PCRE itself.
3643
3644 (ii) The pcre_extra structure (see 45 above) has a void * field called
3645 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3646 pcre_callout_block structure has a field of the same name. The contents of
3647 the field passed in the pcre_extra structure are passed to the callout
3648 function in the corresponding field in the callout block. This makes it
3649 easier to use the same callout-containing regex from multiple threads. For
3650 testing, the pcretest program has a new data escape
3651
3652 \C*n pass the number n (may be negative) as callout_data
3653
3654 If the callout function in pcretest receives a non-zero value as
3655 callout_data, it returns that value.
3656
3657 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3658 there were some redundant $(CFLAGS) in commands that are now specified as
3659 $(LINK), which already includes $(CFLAGS).
3660
3661 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3662 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3663 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3664 one-byte characters throughout. Note that case-insensitive matching applies
3665 only to characters whose values are less than 256. PCRE doesn't support the
3666 notion of cases for higher-valued characters.
3667
3668 (i) A character class whose characters are all within 0-255 is handled as
3669 a bit map, and the map is inverted for negative classes. Previously, a
3670 character > 255 always failed to match such a class; however it should
3671 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3672
3673 (ii) A negated character class with a single character < 255 is coded as
3674 "not this character" (OP_NOT). This wasn't working properly when the test
3675 character was multibyte, either singly or repeated.
3676
3677 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3678 mode, for example: \x{100}{2,3}.
3679
3680 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3681 singly or repeated) now correctly test multibyte characters. However,
3682 PCRE doesn't recognize any characters with values greater than 255 as
3683 digits, spaces, or word characters. Such characters always match \D, \S,
3684 and \W, and never match \d, \s, or \w.
3685
3686 (v) Classes may now contain characters and character ranges with values
3687 greater than 255. For example: [ab\x{100}-\x{400}].
3688
3689 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3690 PCRE in UTF-8 mode.
3691
3692 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3693 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3694 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3695 value.)
3696
3697 53. The single man page has become too large. I have therefore split it up into
3698 a number of separate man pages. These also give rise to individual HTML pages;
3699 these are now put in a separate directory, and there is an index.html page that
3700 lists them all. Some hyperlinking between the pages has been installed.
3701
3702 54. Added convenience functions for handling named capturing parentheses.
3703
3704 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3705 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3706 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3707 are faulted.
3708
3709 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3710 calling configure. These values are used when compiling the dftables.c program
3711 which is run to generate the source of the default character tables. They
3712 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3713 you will need to set these values.
3714
3715 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3716
3717
3718 Version 3.9 02-Jan-02
3719 ---------------------
3720
3721 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3722
3723 2. If --disable-static was given, the building process failed when trying to
3724 build pcretest and pcregrep. (For some reason it was using libtool to compile
3725 them, which is not right, as they aren't part of the library.)
3726
3727
3728 Version 3.8 18-Dec-01
3729 ---------------------
3730
3731 1. The experimental UTF-8 code was completely screwed up. It was packing the
3732 bytes in the wrong order. How dumb can you get?
3733
3734
3735 Version 3.7 29-Oct-01
3736 ---------------------
3737
3738 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3739 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3740 this didn't happen under Solaris 8, where I normally test things.
3741
3742 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3743 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3744 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3745 this fix an hour or so after the initial 3.7 release.)
3746
3747
3748 Version 3.6 23-Oct-01
3749 ---------------------
3750
3751 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3752 offsets passed as NULL with zero offset count.
3753
3754 2. The config.guess and config.sub files had not been updated when I moved to
3755 the latest autoconf.
3756
3757
3758 Version 3.5 15-Aug-01
3759 ---------------------
3760
3761 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3762 had been forgotten.
3763
3764 2. By using declared but undefined structures, we can avoid using "void"
3765 definitions in pcre.h while keeping the internal definitions of the structures
3766 private.
3767
3768 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3769 user point of view, this means that both static and shared libraries are built
3770 by default, but this can be individually controlled. More of the work of
3771 handling this static/shared cases is now inside libtool instead of PCRE's make
3772 file.
3773
3774 4. The pcretest utility is now installed along with pcregrep because it is
3775 useful for users (to test regexs) and by doing this, it automatically gets
3776 relinked by libtool. The documentation has been turned into a man page, so
3777 there are now .1, .txt, and .html versions in /doc.
3778
3779 5. Upgrades to pcregrep:
3780 (i) Added long-form option names like gnu grep.
3781 (ii) Added --help to list all options with an explanatory phrase.
3782 (iii) Added -r, --recursive to recurse into sub-directories.
3783 (iv) Added -f, --file to read patterns from a file.
3784
3785 6. pcre_exec() was referring to its "code" argument before testing that
3786 argument for NULL (and giving an error if it was NULL).
3787
3788 7. Upgraded Makefile.in to allow for compiling in a different directory from
3789 the source directory.
3790
3791 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3792 options bits, the pointer it was passed was to an int instead of to an unsigned
3793 long int. This mattered only on 64-bit systems.
3794
3795 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3796 generated) instead of pcre.in, which it its source. Also made the same change
3797 in several of the .c files.
3798
3799 10. A new release of gcc defines printf() as a macro, which broke pcretest
3800 because it had an ifdef in the middle of a string argument for printf(). Fixed
3801 by using separate calls to printf().
3802
3803 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3804 script, to force use of CR or LF instead of \n in the source. On non-Unix
3805 systems, the value can be set in config.h.
3806
3807 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3808 absolute limit. Changed the text of the error message to make this clear, and
3809 likewise updated the man page.
3810
3811 13. The limit of 99 on the number of capturing subpatterns has been removed.
3812 The new limit is 65535, which I hope will not be a "real" limit.
3813
3814
3815 Version 3.4 22-Aug-00
3816 ---------------------
3817
3818 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3819
3820 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3821
3822
3823 Version 3.3 01-Aug-00
3824 ---------------------
3825
3826 1. If an octal character was given, but the value was greater than \377, it
3827 was not getting masked to the least significant bits, as documented. This could
3828 lead to crashes in some systems.
3829
3830 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3831 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3832
3833 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3834 These just pass their arguments on to (pcre_free)(), but they are provided
3835 because some uses of PCRE bind it to non-C systems that can call its functions,
3836 but cannot call free() or pcre_free() directly.
3837
3838 4. Add "make test" as a synonym for "make check". Corrected some comments in
3839 the Makefile.
3840
3841 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3842 Makefile.
3843
3844 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3845 command called pgrep for grepping around the active processes.
3846
3847 7. Added the beginnings of support for UTF-8 character strings.
3848
3849 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3850 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3851 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3852 out for the ar command.)
3853
3854
3855 Version 3.2 12-May-00
3856 ---------------------
3857
3858 This is purely a bug fixing release.
3859
3860 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3861 of ZA. This was just one example of several cases that could provoke this bug,
3862 which was introduced by change 9 of version 2.00. The code for breaking
3863 infinite loops after an iteration that matches an empty string was't working
3864 correctly.
3865
3866 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3867 when matched against abbab (for example). After matching an empty string, it
3868 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3869 caused it to match further down the string than it should.
3870
3871 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3872 was there because it doesn't seem to be needed, and it causes trouble on some
3873 systems, as it is not a Standard C header. It has been removed.
3874
3875 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3876 were reported on the Macintosh. The changes were from
3877
3878 while ((c = *(++ptr)) != 0 && c != '\n');
3879 to
3880 while ((c = *(++ptr)) != 0 && c != '\n') ;
3881
3882 Totally extraordinary, but if that's what it takes...
3883
3884 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3885 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3886 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3887 assumes the way PCRE uses memmove() (always moving upwards).
3888
3889 6. PCRE is being used in one environment where strchr() is not available. There
3890 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3891 faster code anyway.
3892
3893
3894 Version 3.1 09-Feb-00
3895 ---------------------
3896
3897 The only change in this release is the fixing of some bugs in Makefile.in for
3898 the "install" target:
3899
3900 (1) It was failing to install pcreposix.h.
3901
3902 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3903
3904
3905 Version 3.0 01-Feb-00
3906 ---------------------
3907
3908 1. Add support for the /+ modifier to perltest (to output $` like it does in
3909 pcretest).
3910
3911 2. Add support for the /g modifier to perltest.
3912
3913 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3914 matches null strings.
3915
3916 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3917 pattern. Perl treats empty patterns specially - it reuses the most recent
3918 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3919 effect.
3920
3921 5. The POSIX interface was broken in that it was just handing over the POSIX
3922 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3923 required a bigger vector, with some working space on the end. This means that
3924 the POSIX wrapper now has to get and free some memory, and copy the results.
3925
3926 6. Added some simple autoconf support, placing the test data and the
3927 documentation in separate directories, re-organizing some of the
3928 information files, and making it build pcre-config (a GNU standard). Also added
3929 libtool support for building PCRE as a shared library, which is now the
3930 default.
3931
3932 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3933 09 are not valid octal constants. Single digits will be used for minor values
3934 less than 10.
3935
3936 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3937 existing programs that set these in the POSIX interface can use PCRE without
3938 modification.
3939
3940 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3941 return all that pcre_info() returns, plus additional data. The pcre_info()
3942 function is retained for compatibility, but is considered to be obsolete.
3943
3944 10. Added experimental recursion feature (?R) to handle one common case that
3945 Perl 5.6 will be able to do with (?p{...}).
3946
3947 11. Added support for POSIX character classes like [:alpha:], which Perl is
3948 adopting.
3949
3950
3951 Version 2.08 31-Aug-99
3952 ----------------------
3953
3954 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3955 trying to match at the startoffset position, but instead was moving forward to
3956 the next newline as if a previous match had failed.
3957
3958 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3959 and could get into a loop if a null string was matched other than at the start
3960 of the subject.
3961
3962 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3963 be distinguished at compile time, and for completeness also added PCRE_DATE.
3964
3965 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3966 in GnuWin32 environments.
3967
3968
3969 Version 2.07 29-Jul-99
3970 ----------------------
3971
3972 1. The documentation is now supplied in plain text form and HTML as well as in
3973 the form of man page sources.
3974
3975 2. C++ compilers don't like assigning (void *) values to other pointer types.
3976 In particular this affects malloc(). Although there is no problem in Standard
3977 C, I've put in casts to keep C++ compilers happy.
3978
3979 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3980 should be (const char *).
3981
3982 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3983 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3984 However, I haven't made this a standard facility. The documentation doesn't
3985 mention it, and the Makefile doesn't support it.
3986
3987 5. The Makefile now contains an "install" target, with editable destinations at
3988 the top of the file. The pcretest program is not installed.
3989
3990 6. pgrep -V now gives the PCRE version number and date.
3991
3992 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3993 causing the entire string to be ignored, instead of just the last character.
3994
3995 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3996 non-matching string, it can take a very, very long time, even for strings of
3997 quite modest length, because of the nested recursion. PCRE now does better in
3998 some of these cases. It does this by remembering the last required literal
3999 character in the pattern, and pre-searching the subject to ensure it is present
4000 before running the real match. In other words, it applies a heuristic to detect
4001 some types of certain failure quickly, and in the above example, if presented
4002 with a string that has no trailing " it gives "no match" very quickly.
4003
4004 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4005 other alternatives are tried instead.
4006
4007
4008 Version 2.06 09-Jun-99
4009 ----------------------
4010
4011 1. Change pcretest's output for amount of store used to show just the code
4012 space, because the remainder (the data block) varies in size between 32-bit and
4013 64-bit systems.
4014
4015 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4016 start matching at. This allows lookbehinds to work when searching for multiple
4017 occurrences in a string.
4018
4019 3. Added additional options to pcretest for testing multiple occurrences:
4020
4021 /+ outputs the rest of the string that follows a match
4022 /g loops for multiple occurrences, using the new startoffset argument
4023 /G loops for multiple occurrences by passing an incremented pointer
4024
4025 4. PCRE wasn't doing the "first character" optimization for patterns starting
4026 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4027 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4028 the letter 'x'. On long subject strings, this gives a significant speed-up.
4029
4030
4031 Version 2.05 21-Apr-99
4032 ----------------------
4033
4034 1. Changed the type of magic_number from int to long int so that it works
4035 properly on 16-bit systems.
4036
4037 2. Fixed a bug which caused patterns starting with .* not to work correctly
4038 when the subject string contained newline characters. PCRE was assuming
4039 anchoring for such patterns in all cases, which is not correct because .* will
4040 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4041 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4042 must be retried after every newline in the subject.
4043
4044
4045 Version 2.04 18-Feb-99
4046 ----------------------
4047
4048 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4049 computation of the store needed to hold the pattern was incorrect (too large).
4050 If such patterns were nested a few deep, this could multiply and become a real
4051 problem.
4052
4053 2. Added /M option to pcretest to show the memory requirement of a specific
4054 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4055
4056 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4057 compiled in such a way that the backtracking after subsequent failure was
4058 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4059 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4060
4061
4062 Version 2.03 02-Feb-99
4063 ----------------------
4064
4065 1. Fixed typo and small mistake in man page.
4066
4067 2. Added 4th condition (GPL supersedes if conflict) and created separate
4068 LICENCE file containing the conditions.
4069
4070 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4071 Perl, that is the internal \ allows the delimiter to be included in the
4072 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4073 the final delimiter, add \ to the end of the pattern (to test the error).
4074
4075 4. Added the convenience functions for extracting substrings after a successful
4076 match. Updated pcretest to make it able to test these functions.
4077
4078
4079 Version 2.02 14-Jan-99
4080 ----------------------
4081
4082 1. Initialized the working variables associated with each extraction so that
4083 their saving and restoring doesn't refer to uninitialized store.
4084
4085 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4086 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4087 fix the problem.
4088
4089 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4090 calls, and wasn't printing the correct value for compiling calls. Increased the
4091 default value of LOOPREPEAT, and the number of significant figures in the
4092 times.
4093
4094 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4095
4096 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4097 a building problem on Windows NT with a FAT file system.
4098
4099
4100 Version 2.01 21-Oct-98
4101 ----------------------
4102
4103 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4104 to character tables built by pcre_maketables() in the current locale. If NULL
4105 is passed, the default tables are used.
4106
4107
4108 Version 2.00 24-Sep-98
4109 ----------------------
4110
4111 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4112 it any more.
4113
4114 2. Allow quantification of (?>) groups, and make it work correctly.
4115
4116 3. The first character computation wasn't working for (?>) groups.
4117
4118 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4119 end of the subject) and add 5.005's \z, which really does match only at the
4120 very end of the subject.
4121
4122 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4123
4124 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4125 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4126 localized options. All options to pcre_study() were also removed.
4127
4128 7. Add other new features from 5.005:
4129
4130 $(?<= positive lookbehind
4131 $(?<! negative lookbehind
4132 (?imsx-imsx) added the unsetting capability
4133 such a setting is global if at outer level; local otherwise
4134 (?imsx-imsx:) non-capturing groups with option setting
4135 (?(cond)re|re) conditional pattern matching
4136
4137 A backreference to itself in a repeated group matches the previous
4138 captured string.
4139
4140 8. General tidying up of studying (both automatic and via "study")
4141 consequential on the addition of new assertions.
4142
4143 9. As in 5.005, unlimited repeated groups that could match an empty substring
4144 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4145 runtime if any iteration does actually match an empty substring.
4146
4147 10. Include the RunTest script in the distribution.
4148
4149 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4150 discrepancies, some of which were old and were also with respect to 5.004. They
4151 have now been fixed.
4152
4153
4154 Version 1.09 28-Apr-98
4155 ----------------------
4156
4157 1. A negated single character class followed by a quantifier with a minimum
4158 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4159 program crashes, or just wrong answers. This did not apply to negated classes
4160 containing more than one character, or to minima other than one.
4161
4162
4163 Version 1.08 27-Mar-98
4164 ----------------------
4165
4166 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4167
4168 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4169 latter must appear before anything that relies on it in the pattern.
4170
4171
4172 Version 1.07 16-Feb-98
4173 ----------------------
4174
4175 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4176 repeat of a potentially empty string).
4177
4178
4179 Version 1.06 23-Jan-98
4180 ----------------------
4181
4182 1. Added Markus Oberhumer's little patches for C++.
4183
4184 2. Literal strings longer than 255 characters were broken.
4185
4186
4187 Version 1.05 23-Dec-97
4188 ----------------------
4189
4190 1. Negated character classes containing more than one character were failing if
4191 PCRE_CASELESS was set at run time.
4192
4193
4194 Version 1.04 19-Dec-97
4195 ----------------------
4196
4197 1. Corrected the man page, where some "const" qualifiers had been omitted.
4198
4199 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4200 input syntax.
4201
4202 3. Fixed memory leak which occurred when a regex with back references was
4203 matched with an offsets vector that wasn't big enough. The temporary memory
4204 that is used in this case wasn't being freed if the match failed.
4205
4206 4. Tidied pcretest to ensure it frees memory that it gets.
4207
4208 5. Temporary memory was being obtained in the case where the passed offsets
4209 vector was exactly big enough.
4210
4211 6. Corrected definition of offsetof() from change 5 below.
4212
4213 7. I had screwed up change 6 below and broken the rules for the use of
4214 setjmp(). Now fixed.
4215
4216
4217 Version 1.03 18-Dec-97
4218 ----------------------
4219
4220 1. A erroneous regex with a missing opening parenthesis was correctly
4221 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4222 on some systems.
4223
4224 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4225 it was reported that one broken compiler failed on the former because "code" is
4226 also an independent variable.
4227
4228 3. The erroneous regex a[]b caused an array overrun reference.
4229
4230 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4231 fail on data ending with that character. (It was going on too far, and checking
4232 the next character, typically a binary zero.) This was specific to the
4233 optimized code for single-character negative classes.
4234
4235 5. Added a contributed patch from the TIN world which does the following:
4236
4237 + Add an undef for memmove, in case the the system defines a macro for it.
4238
4239 + Add a definition of offsetof(), in case there isn't one. (I don't know
4240 the reason behind this - offsetof() is part of the ANSI standard - but
4241 it does no harm).
4242
4243 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4244 most of the places where whitespace preceded '#'. I have given up and
4245 allowed the remaining 2 cases to be at the margin.
4246
4247 + Rename some variables in pcre to eliminate shadowing. This seems very
4248 pedantic, but does no harm, of course.
4249
4250 6. Moved the call to setjmp() into its own function, to get rid of warnings
4251 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4252
4253 7. Constructs such as \d{8,} were compiling into the equivalent of
4254 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4255 outcome, but in this particular case used more store than had been allocated,
4256 which caused the bug to be discovered because it threw up an internal error.
4257
4258 8. The debugging code in both pcre and pcretest for outputting the compiled
4259 form of a regex was going wrong in the case of back references followed by
4260 curly-bracketed repeats.
4261
4262
4263 Version 1.02 12-Dec-97
4264 ----------------------
4265
4266 1. Typos in pcre.3 and comments in the source fixed.
4267
4268 2. Applied a contributed patch to get rid of places where it used to remove
4269 'const' from variables, and fixed some signed/unsigned and uninitialized
4270 variable warnings.
4271
4272 3. Added the "runtest" target to Makefile.
4273
4274 4. Set default compiler flag to -O2 rather than just -O.
4275
4276
4277 Version 1.01 19-Nov-97
4278 ----------------------
4279
4280 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4281 like /([ab]*)*/, that is, for classes with more than one character in them.
4282
4283 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4284 as /((?>a*))*/ (a PCRE_EXTRA facility).
4285
4286
4287 Version 1.00 18-Nov-97
4288 ----------------------
4289
4290 1. Added compile-time macros to support systems such as SunOS4 which don't have
4291 memmove() or strerror() but have other things that can be used instead.
4292
4293 2. Arranged that "make clean" removes the executables.
4294
4295
4296 Version 0.99 27-Oct-97
4297 ----------------------
4298
4299 1. Fixed bug in code for optimizing classes with only one character. It was
4300 initializing a 32-byte map regardless, which could cause it to run off the end
4301 of the memory it had got.
4302
4303 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4304
4305
4306 Version 0.98 22-Oct-97
4307 ----------------------
4308
4309 1. Fixed bug in code for handling temporary memory usage when there are more
4310 back references than supplied space in the ovector. This could cause segfaults.
4311
4312
4313 Version 0.97 21-Oct-97
4314 ----------------------
4315
4316 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4317
4318 2. Optimized negated single characters not to use a bit map.
4319
4320 3. Brought error texts together as macro definitions; clarified some of them;
4321 fixed one that was wrong - it said "range out of order" when it meant "invalid
4322 escape sequence".
4323
4324 4. Changed some char * arguments to const char *.
4325
4326 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4327
4328 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4329 pcretest.
4330
4331
4332 Version 0.96 16-Oct-97
4333 ----------------------
4334
4335 1. Added a simple "pgrep" utility to the distribution.
4336
4337 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4338 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4339 where "ddd" means "one or more decimal digits".
4340
4341 3. Fixed serious bug. If a pattern had a back reference, but the call to
4342 pcre_exec() didn't supply a large enough ovector to record the related
4343 identifying subpattern, the match always failed. PCRE now remembers the number
4344 of the largest back reference, and gets some temporary memory in which to save
4345 the offsets during matching if necessary, in order to ensure that
4346 backreferences always work.
4347
4348 4. Increased the compatibility with Perl in a number of ways:
4349
4350 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4351 to request this handling. The option can be set at compile or exec time.
4352
4353 (b) $ matches before a terminating newline by default; an option
4354 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4355 mode). The option can be set at compile or exec time.
4356
4357 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4358 the same as Perl's. If the decimal number it represents is less than 10
4359 or there aren't that many previous left capturing parentheses, an octal
4360 escape is read. Inside a character class, it's always an octal escape,
4361 even if it is a single digit.
4362
4363 (d) An escaped but undefined alphabetic character is taken as a literal,
4364 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4365 escapes.
4366
4367 (e) {0} is now permitted. (The previous item is removed from the compiled
4368 pattern).
4369
4370 5. Changed all the names of code files so that the basic parts are no longer
4371 than 10 characters, and abolished the teeny "globals.c" file.
4372
4373 6. Changed the handling of character classes; they are now done with a 32-byte
4374 bit map always.
4375
4376 7. Added the -d and /D options to pcretest to make it possible to look at the
4377 internals of compilation without having to recompile pcre.
4378
4379
4380 Version 0.95 23-Sep-97
4381 ----------------------
4382
4383 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4384 \x20 at the start of a run of normal characters. These were being treated as
4385 real characters, instead of the source characters being re-checked.
4386
4387
4388 Version 0.94 18-Sep-97
4389 ----------------------
4390
4391 1. The functions are now thread-safe, with the caveat that the global variables
4392 containing pointers to malloc() and free() or alternative functions are the
4393 same for all threads.
4394
4395 2. Get pcre_study() to generate a bitmap of initial characters for non-
4396 anchored patterns when this is possible, and use it if passed to pcre_exec().
4397
4398
4399 Version 0.93 15-Sep-97
4400 ----------------------
4401
4402 1. /(b)|(:+)/ was computing an incorrect first character.
4403
4404 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4405 but not actually doing anything yet.
4406
4407 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4408 as Perl does (e.g. [-az] or [az-]).
4409
4410 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4411 all possible positions.
4412
4413 5. Split up into different modules to avoid including unneeded functions in a
4414 compiled binary. However, compile and exec are still in one module. The "study"
4415 function is split off.
4416
4417 6. The character tables are now in a separate module whose source is generated
4418 by an auxiliary program - but can then be edited by hand if required. There are
4419 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4420 toupper() in the code.
4421
4422 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4423 make them global. Abolish the function for setting them, as the caller can now
4424 set them directly.
4425
4426
4427 Version 0.92 11-Sep-97
4428 ----------------------
4429
4430 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4431 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4432
4433 2. Caseless matching was not working in character classes if the characters in
4434 the pattern were in upper case.
4435
4436 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4437
4438 4. Make PCRE_ANCHORED public and accept as a compile option.
4439
4440 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4441 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4442 pass them.
4443
4444 6. Give an error if bad option bits passed at compile or run time.
4445
4446 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4447 pcretest to cause it to pass that flag.
4448
4449 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4450 options, and the first character, if set.
4451
4452 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4453
4454
4455 Version 0.91 10-Sep-97
4456 ----------------------
4457
4458 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4459 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4460
4461 2. PCRE was looping on encountering an indefinitely repeated back reference to
4462 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4463 Perl does - treats the match as successful.
4464
4465 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12