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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 746 - (show annotations)
Tue Nov 15 15:07:02 2011 UTC (8 years ago) by ph10
File size: 218496 byte(s)
Error occurred while calculating annotation data.
Caseless matching of backreferences with fixed length repetitions was broken.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.21
5 ------------
6
7 1. Updating the JIT compiler.
8
9 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
10 are added as well.
11
12 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
13 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
14 calling _pcre_jit_exec. Some extra comments are added.
15
16 4. Mark settings inside atomic groups that do not contain any capturing
17 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
18 was introduced by change 18 for 8.20.
19
20 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
21 ECMA-262 standard.
22
23 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
24 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
25 This bug was probably introduced by change 9 of 8.13.
26
27
28 Version 8.20 21-Oct-2011
29 ------------------------
30
31 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
32 a POSIX class. After further experiments with Perl, which convinced me that
33 Perl has bugs and confusions, a closing square bracket is no longer allowed
34 in a POSIX name. This bug also affected patterns with classes that started
35 with full stops.
36
37 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
38 captured substring, but while checking the failing first alternative,
39 substring 1 is temporarily captured. If the output vector supplied to
40 pcre_exec() was not big enough for this capture, the yield of the function
41 was still zero ("insufficient space for captured substrings"). This cannot
42 be totally fixed without adding another stack variable, which seems a lot
43 of expense for a edge case. However, I have improved the situation in cases
44 such as /(a)(b)x|abc/ matched against "abc", where the return code
45 indicates that fewer than the maximum number of slots in the ovector have
46 been set.
47
48 3. Related to (2) above: when there are more back references in a pattern than
49 slots in the output vector, pcre_exec() uses temporary memory during
50 matching, and copies in the captures as far as possible afterwards. It was
51 using the entire output vector, but this conflicts with the specification
52 that only 2/3 is used for passing back captured substrings. Now it uses
53 only the first 2/3, for compatibility. This is, of course, another edge
54 case.
55
56 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
57 main code base, and can be used by building with --enable-jit. When this is
58 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
59 runtime --no-jit option is given.
60
61 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
62 ovector, the return from the function was zero, implying that there were
63 other matches that did not fit. The correct "exactly full" value is now
64 returned.
65
66 6. If a subpattern that was called recursively or as a subroutine contained
67 (*PRUNE) or any other control that caused it to give a non-standard return,
68 invalid errors such as "Error -26 (nested recursion at the same subject
69 position)" or even infinite loops could occur.
70
71 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
72 computing the minimum length on reaching *ACCEPT, and so ended up with the
73 wrong value of 1 rather than 0. Further investigation indicates that
74 computing a minimum subject length in the presence of *ACCEPT is difficult
75 (think back references, subroutine calls), and so I have changed the code
76 so that no minimum is registered for a pattern that contains *ACCEPT.
77
78 8. If (*THEN) was present in the first (true) branch of a conditional group,
79 it was not handled as intended. [But see 16 below.]
80
81 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
82 Sheri Pierce.
83
84 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
85 the first byte in a match must be "a".
86
87 11. Change 17 for 8.13 increased the recursion depth for patterns like
88 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
89 pattern contains any instances of (*THEN). If it does not, the old
90 optimizations are restored. It would be nice to do this on a per-group
91 basis, but at the moment that is not feasible.
92
93 12. In some environments, the output of pcretest -C is CRLF terminated. This
94 broke RunTest's code that checks for the link size. A single white space
95 character after the value is now allowed for.
96
97 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
98 For "fr", it uses the Windows-specific input and output files.
99
100 14. If (*THEN) appeared in a group that was called recursively or as a
101 subroutine, it did not work as intended. [But see next item.]
102
103 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
104 pattern fragments (but not containing any | characters). If A and B are
105 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
106 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
107 D. In other words, Perl considers parentheses that do not contain any |
108 characters to be part of a surrounding alternative, whereas PCRE was
109 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
110 differently. PCRE now behaves in the same way as Perl, except in the case
111 of subroutine/recursion calls such as (?1) which have in any case always
112 been different (but PCRE had them first :-).
113
114 16. Related to 15 above: Perl does not treat the | in a conditional group as
115 creating alternatives. Such a group is treated in the same way as an
116 ordinary group without any | characters when processing (*THEN). PCRE has
117 been changed to match Perl's behaviour.
118
119 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
120 RunGrepTest script failed.
121
122 18. Change 22 for version 13 caused atomic groups to use more stack. This is
123 inevitable for groups that contain captures, but it can lead to a lot of
124 stack use in large patterns. The old behaviour has been restored for atomic
125 groups that do not contain any capturing parentheses.
126
127 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
128 suppress the check for a minimum subject length at run time. (If it was
129 given to pcre_exec() or pcre_dfa_exec() it did work.)
130
131 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
132 fail to work when decoding hex characters in data strings in EBCDIC
133 environments.
134
135 21. It appears that in at least one Mac OS environment, the isxdigit() function
136 is implemented as a macro that evaluates to its argument more than once,
137 contravening the C 90 Standard (I haven't checked a later standard). There
138 was an instance in pcretest which caused it to go wrong when processing
139 \x{...} escapes in subject strings. The has been rewritten to avoid using
140 things like p++ in the argument of isxdigit().
141
142
143 Version 8.13 16-Aug-2011
144 ------------------------
145
146 1. The Unicode data tables have been updated to Unicode 6.0.0.
147
148 2. Two minor typos in pcre_internal.h have been fixed.
149
150 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
151 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
152 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
153
154 4. There were a number of related bugs in the code for matching backrefences
155 caselessly in UTF-8 mode when codes for the characters concerned were
156 different numbers of bytes. For example, U+023A and U+2C65 are an upper
157 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
158 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
159 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
160 2-byte code at the end of the subject (it thought there wasn't enough data
161 left).
162
163 5. Comprehensive information about what went wrong is now returned by
164 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
165 as the output vector has at least 2 elements. The offset of the start of
166 the failing character and a reason code are placed in the vector.
167
168 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
169 now returned is for the first byte of the failing character, instead of the
170 last byte inspected. This is an incompatible change, but I hope it is small
171 enough not to be a problem. It makes the returned offset consistent with
172 pcre_exec() and pcre_dfa_exec().
173
174 7. pcretest now gives a text phrase as well as the error number when
175 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
176 failure, the offset and reason code are output.
177
178 8. When \R was used with a maximizing quantifier it failed to skip backwards
179 over a \r\n pair if the subsequent match failed. Instead, it just skipped
180 back over a single character (\n). This seems wrong (because it treated the
181 two characters as a single entity when going forwards), conflicts with the
182 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
183 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
184 has been changed.
185
186 9. Some internal refactoring has changed the processing so that the handling
187 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
188 time (the PCRE_DOTALL option was changed this way some time ago: version
189 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
190 which was always a bit of a fudge. It also means that there is one less
191 argument for the match() function, which reduces its stack requirements
192 slightly. This change also fixes an incompatibility with Perl: the pattern
193 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
194
195 10. More internal refactoring has drastically reduced the number of recursive
196 calls to match() for possessively repeated groups such as (abc)++ when
197 using pcre_exec().
198
199 11. While implementing 10, a number of bugs in the handling of groups were
200 discovered and fixed:
201
202 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
203 (a|)*(?1) gave a compile-time internal error.
204 ((a|)+)+ did not notice that the outer group could match an empty string.
205 (^a|^)+ was not marked as anchored.
206 (.*a|.*)+ was not marked as matching at start or after a newline.
207
208 12. Yet more internal refactoring has removed another argument from the match()
209 function. Special calls to this function are now indicated by setting a
210 value in a variable in the "match data" data block.
211
212 13. Be more explicit in pcre_study() instead of relying on "default" for
213 opcodes that mean there is no starting character; this means that when new
214 ones are added and accidentally left out of pcre_study(), testing should
215 pick them up.
216
217 14. The -s option of pcretest has been documented for ages as being an old
218 synonym of -m (show memory usage). I have changed it to mean "force study
219 for every regex", that is, assume /S for every regex. This is similar to -i
220 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
221 using it. It makes it easier to run collections of tests with and without
222 study enabled, and thereby test pcre_study() more easily. All the standard
223 tests are now run with and without -s (but some patterns can be marked as
224 "never study" - see 20 below).
225
226 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
227 restoration of the capturing data to the outer values was not happening
228 correctly.
229
230 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
231 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
232 pattern had matched an empty string, and so incorrectly returned a no
233 match.
234
235 17. There was optimizing code for the last branch of non-capturing parentheses,
236 and also for the obeyed branch of a conditional subexpression, which used
237 tail recursion to cut down on stack usage. Unfortunately, now that there is
238 the possibility of (*THEN) occurring in these branches, tail recursion is
239 no longer possible because the return has to be checked for (*THEN). These
240 two optimizations have therefore been removed. [But see 8.20/11 above.]
241
242 18. If a pattern containing \R was studied, it was assumed that \R always
243 matched two bytes, thus causing the minimum subject length to be
244 incorrectly computed because \R can also match just one byte.
245
246 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
247 was incorrectly computed.
248
249 20. If /S is present twice on a test pattern in pcretest input, it now
250 *disables* studying, thereby overriding the use of -s on the command line
251 (see 14 above). This is necessary for one or two tests to keep the output
252 identical in both cases.
253
254 21. When (*ACCEPT) was used in an assertion that matched an empty string and
255 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
256
257 22. When an atomic group that contained a capturing parenthesis was
258 successfully matched, but the branch in which it appeared failed, the
259 capturing was not being forgotten if a higher numbered group was later
260 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
261 group 1 to "a", when in fact it should be unset. This applied to multi-
262 branched capturing and non-capturing groups, repeated or not, and also to
263 positive assertions (capturing in negative assertions does not happen
264 in PCRE) and also to nested atomic groups.
265
266 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
267 subject after a captured substring, to make it easier to tell which of a
268 number of identical substrings has been captured.
269
270 24. The way atomic groups are processed by pcre_exec() has been changed so that
271 if they are repeated, backtracking one repetition now resets captured
272 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
273 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
274 "aaa". Previously, it would have been "a". As part of this code
275 refactoring, the way recursive calls are handled has also been changed.
276
277 25. If an assertion condition captured any substrings, they were not passed
278 back unless some other capturing happened later. For example, if
279 (?(?=(a))a) was matched against "a", no capturing was returned.
280
281 26. When studying a pattern that contained subroutine calls or assertions,
282 the code for finding the minimum length of a possible match was handling
283 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
284 group 1 called group 2 while simultaneously a separate group 2 called group
285 1). A stack overflow occurred in this case. I have fixed this by limiting
286 the recursion depth to 10.
287
288 27. Updated RunTest.bat in the distribution to the version supplied by Tom
289 Fortmann. This supports explicit test numbers on the command line, and has
290 argument validation and error reporting.
291
292 28. An instance of \X with an unlimited repeat could fail if at any point the
293 first character it looked at was a mark character.
294
295 29. Some minor code refactoring concerning Unicode properties and scripts
296 should reduce the stack requirement of match() slightly.
297
298 30. Added the '=' option to pcretest to check the setting of unused capturing
299 slots at the end of the pattern, which are documented as being -1, but are
300 not included in the return count.
301
302 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
303 compiled something random. Now it gives a compile-time error (as does
304 Perl).
305
306 32. A *MARK encountered during the processing of a positive assertion is now
307 recorded and passed back (compatible with Perl).
308
309 33. If --only-matching or --colour was set on a pcregrep call whose pattern
310 had alternative anchored branches, the search for a second match in a line
311 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
312 matched the line "0102" twice. The same bug affected patterns that started
313 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
314 twice.
315
316 34. Previously, PCRE did not allow quantification of assertions. However, Perl
317 does, and because of capturing effects, quantifying parenthesized
318 assertions may at times be useful. Quantifiers are now allowed for
319 parenthesized assertions.
320
321 35. A minor code tidy in pcre_compile() when checking options for \R usage.
322
323 36. \g was being checked for fancy things in a character class, when it should
324 just be a literal "g".
325
326 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
327 appearance of a nested POSIX class supersedes an apparent external class.
328 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
329 unescaped square brackets may also appear as part of class names. For
330 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
331 more like Perl. (But see 8.20/1 above.)
332
333 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
334 was because it thought it was \N{name}, which is not supported).
335
336 39. Add minix to OS list not supporting the -S option in pcretest.
337
338 40. PCRE tries to detect cases of infinite recursion at compile time, but it
339 cannot analyze patterns in sufficient detail to catch mutual recursions
340 such as ((?1))((?2)). There is now a runtime test that gives an error if a
341 subgroup is called recursively as a subpattern for a second time at the
342 same position in the subject string. In previous releases this might have
343 been caught by the recursion limit, or it might have run out of stack.
344
345 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
346 happen only once. PCRE was, however incorrectly giving a compile time error
347 "recursive call could loop indefinitely" because it cannot analyze the
348 pattern in sufficient detail. The compile time test no longer happens when
349 PCRE is compiling a conditional subpattern, but actual runaway loops are
350 now caught at runtime (see 40 above).
351
352 42. It seems that Perl allows any characters other than a closing parenthesis
353 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
354 has been changed to be the same.
355
356 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
357 as not to get warnings when autogen.sh is called. Also changed
358 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
359
360 44. To help people who use pcregrep to scan files containing exceedingly long
361 lines, the following changes have been made:
362
363 (a) The default value of the buffer size parameter has been increased from
364 8K to 20K. (The actual buffer used is three times this size.)
365
366 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
367 PCRE is built.
368
369 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
370 to be set at run time.
371
372 (d) Numerical values in pcregrep options can be followed by K or M, for
373 example --buffer-size=50K.
374
375 (e) If a line being scanned overflows pcregrep's buffer, an error is now
376 given and the return code is set to 2.
377
378 45. Add a pointer to the latest mark to the callout data block.
379
380 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
381 partial match of an empty string instead of no match. This was specific to
382 the use of ".".
383
384 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
385 complete match instead of a partial match. This bug was dependent on both
386 the PCRE_UTF8 and PCRE_DOTALL options being set.
387
388 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
389 starting byte set, because \b was not being ignored.
390
391
392 Version 8.12 15-Jan-2011
393 ------------------------
394
395 1. Fixed some typos in the markup of the man pages, and wrote a script that
396 checks for such things as part of the documentation building process.
397
398 2. On a big-endian 64-bit system, pcregrep did not correctly process the
399 --match-limit and --recursion-limit options (added for 8.11). In
400 particular, this made one of the standard tests fail. (The integer value
401 went into the wrong half of a long int.)
402
403 3. If the --colour option was given to pcregrep with -v (invert match), it
404 did strange things, either producing crazy output, or crashing. It should,
405 of course, ignore a request for colour when reporting lines that do not
406 match.
407
408 4. Another pcregrep bug caused similar problems if --colour was specified with
409 -M (multiline) and the pattern match finished with a line ending.
410
411 5. In pcregrep, when a pattern that ended with a literal newline sequence was
412 matched in multiline mode, the following line was shown as part of the
413 match. This seems wrong, so I have changed it.
414
415 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
416 the check for further matches in the same line (so they could be coloured)
417 to overrun the end of the current line. If another match was found, it was
418 incorrectly shown (and then shown again when found in the next line).
419
420 7. If pcregrep was compiled under Windows, there was a reference to the
421 function pcregrep_exit() before it was defined. I am assuming this was
422 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
423 reported by a user. I've moved the definition above the reference.
424
425
426 Version 8.11 10-Dec-2010
427 ------------------------
428
429 1. (*THEN) was not working properly if there were untried alternatives prior
430 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
431 backtracked to try for "b" instead of moving to the next alternative branch
432 at the same level (in this case, to look for "c"). The Perl documentation
433 is clear that when (*THEN) is backtracked onto, it goes to the "next
434 alternative in the innermost enclosing group".
435
436 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
437 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
438 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
439 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
440 (*THEN).
441
442 3. If \s appeared in a character class, it removed the VT character from
443 the class, even if it had been included by some previous item, for example
444 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
445 of \s, but is part of the POSIX "space" class.)
446
447 4. A partial match never returns an empty string (because you can always
448 match an empty string at the end of the subject); however the checking for
449 an empty string was starting at the "start of match" point. This has been
450 changed to the "earliest inspected character" point, because the returned
451 data for a partial match starts at this character. This means that, for
452 example, /(?<=abc)def/ gives a partial match for the subject "abc"
453 (previously it gave "no match").
454
455 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
456 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
457 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
458 has an implication that the given string is incomplete (because a partial
459 match is preferred over a full match). For this reason, these items now
460 give a partial match in this situation. [Aside: previously, the one case
461 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
462 match rather than a full match, which was wrong by the old rules, but is
463 now correct.]
464
465 6. There was a bug in the handling of #-introduced comments, recognized when
466 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
467 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
468 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
469 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
470 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
471 places in pcre_compile().
472
473 7. Related to (6) above, when pcre_compile() was skipping #-introduced
474 comments when looking ahead for named forward references to subpatterns,
475 the only newline sequence it recognized was NL. It now handles newlines
476 according to the set newline convention.
477
478 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
479 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
480 cater for a lack of strerror(). These oversights have been fixed.
481
482 9. Added --match-limit and --recursion-limit to pcregrep.
483
484 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
485
486 11. When the -o option was used, pcregrep was setting a return code of 1, even
487 when matches were found, and --line-buffered was not being honoured.
488
489 12. Added an optional parentheses number to the -o and --only-matching options
490 of pcregrep.
491
492 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
493 can match an empty string. The code to do it in pcretest and pcredemo
494 needed fixing:
495
496 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
497 only one byte after an empty string match just before CRLF (this case
498 just got forgotten; "any" and "anycrlf" were OK).
499
500 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
501 mode when an empty string match preceded an ASCII character followed by
502 a non-ASCII character. (The code for advancing by one character rather
503 than one byte was nonsense.)
504
505 (c) The pcredemo.c sample program did not have any code at all to handle
506 the cases when CRLF is a valid newline sequence.
507
508 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
509 as a starting offset was within the subject string. There is now a new
510 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
511 negative or greater than the length of the string. In order to test this,
512 pcretest is extended to allow the setting of negative starting offsets.
513
514 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
515 starting offset points to the beginning of a UTF-8 character was
516 unnecessarily clumsy. I tidied it up.
517
518 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
519 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
520
521 17. Nobody had reported that the --include_dir option, which was added in
522 release 7.7 should have been called --include-dir (hyphen, not underscore)
523 for compatibility with GNU grep. I have changed it to --include-dir, but
524 left --include_dir as an undocumented synonym, and the same for
525 --exclude-dir, though that is not available in GNU grep, at least as of
526 release 2.5.4.
527
528 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
529 characters from a string of bytes) have been redefined so as not to use
530 loops, in order to improve performance in some environments. At the same
531 time, I abstracted some of the common code into auxiliary macros to save
532 repetition (this should not affect the compiled code).
533
534 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
535 compile-time error is now given if \c is not followed by an ASCII
536 character, that is, a byte less than 128. (In EBCDIC mode, the code is
537 different, and any byte value is allowed.)
538
539 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
540 START_OPTIMIZE option, which is now allowed at compile time - but just
541 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
542 to pcregrep and other applications that have no direct access to PCRE
543 options. The new /Y option in pcretest sets this option when calling
544 pcre_compile().
545
546 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
547 back references. Groups containing recursive back references were forced to
548 be atomic by that change, but in the case of named groups, the amount of
549 memory required was incorrectly computed, leading to "Failed: internal
550 error: code overflow". This has been fixed.
551
552 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
553 pcretest.c, to avoid build problems in some Borland environments.
554
555
556 Version 8.10 25-Jun-2010
557 ------------------------
558
559 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
560 THEN.
561
562 2. (*ACCEPT) was not working when inside an atomic group.
563
564 3. Inside a character class, \B is treated as a literal by default, but
565 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
566 causes the error). The code is unchanged, but I tidied the documentation.
567
568 4. Inside a character class, PCRE always treated \R and \X as literals,
569 whereas Perl faults them if its -w option is set. I have changed PCRE so
570 that it faults them when PCRE_EXTRA is set.
571
572 5. Added support for \N, which always matches any character other than
573 newline. (It is the same as "." when PCRE_DOTALL is not set.)
574
575 6. When compiling pcregrep with newer versions of gcc which may have
576 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
577 declared with attribute warn_unused_result" were given. Just casting the
578 result to (void) does not stop the warnings; a more elaborate fudge is
579 needed. I've used a macro to implement this.
580
581 7. Minor change to pcretest.c to avoid a compiler warning.
582
583 8. Added four artifical Unicode properties to help with an option to make
584 \s etc use properties (see next item). The new properties are: Xan
585 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
586
587 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
588 use Unicode properties. (*UCP) at the start of a pattern can be used to set
589 this option. Modified pcretest to add /W to test this facility. Added
590 REG_UCP to make it available via the POSIX interface.
591
592 10. Added --line-buffered to pcregrep.
593
594 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
595 studied, and the match started with a letter with a code point greater than
596 127 whose first byte was different to the first byte of the other case of
597 the letter, the other case of this starting letter was not recognized
598 (#976).
599
600 12. If a pattern that was studied started with a repeated Unicode property
601 test, for example, \p{Nd}+, there was the theoretical possibility of
602 setting up an incorrect bitmap of starting bytes, but fortunately it could
603 not have actually happened in practice until change 8 above was made (it
604 added property types that matched character-matching opcodes).
605
606 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
607 possible starting bytes for non-anchored patterns.
608
609 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
610 \R, and also a number of cases that involve Unicode properties, both
611 explicit and implicit when PCRE_UCP is set.
612
613 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
614 input, it could crash or give wrong results if characters with values
615 greater than 0xc0 were present in the subject string. (Detail: it assumed
616 UTF-8 input when processing these items.)
617
618 16. Added a lot of (int) casts to avoid compiler warnings in systems where
619 size_t is 64-bit (#991).
620
621 17. Added a check for running out of memory when PCRE is compiled with
622 --disable-stack-for-recursion (#990).
623
624 18. If the last data line in a file for pcretest does not have a newline on
625 the end, a newline was missing in the output.
626
627 19. The default pcre_chartables.c file recognizes only ASCII characters (values
628 less than 128) in its various bitmaps. However, there is a facility for
629 generating tables according to the current locale when PCRE is compiled. It
630 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
631 characters, are recognized by isspace() and therefore were getting set in
632 these tables, and indeed these tables seem to approximate to ISO 8859. This
633 caused a problem in UTF-8 mode when pcre_study() was used to create a list
634 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
635 which of course cannot start UTF-8 characters. I have changed the code so
636 that only real ASCII characters (less than 128) and the correct starting
637 bytes for UTF-8 encodings are set for characters greater than 127 when in
638 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
639 altogether.)
640
641 20. Added the /T option to pcretest so as to be able to run tests with non-
642 standard character tables, thus making it possible to include the tests
643 used for 19 above in the standard set of tests.
644
645 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
646 reference to a subpattern the other side of a comment that contains an
647 opening parenthesis caused either an internal compiling error, or a
648 reference to the wrong subpattern.
649
650
651 Version 8.02 19-Mar-2010
652 ------------------------
653
654 1. The Unicode data tables have been updated to Unicode 5.2.0.
655
656 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
657 configured.
658
659 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
660 original author of that file, following a query about its status.
661
662 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
663 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
664
665 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
666 quantifier applied to a forward-referencing subroutine call, could compile
667 incorrect code or give the error "internal error: previously-checked
668 referenced subpattern not found".
669
670 6. Both MS Visual Studio and Symbian OS have problems with initializing
671 variables to point to external functions. For these systems, therefore,
672 pcre_malloc etc. are now initialized to local functions that call the
673 relevant global functions.
674
675 7. There were two entries missing in the vectors called coptable and poptable
676 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
677 I've fixed the data, and added a kludgy way of testing at compile time that
678 the lengths are correct (equal to the number of opcodes).
679
680 8. Following on from 7, I added a similar kludge to check the length of the
681 eint vector in pcreposix.c.
682
683 9. Error texts for pcre_compile() are held as one long string to avoid too
684 much relocation at load time. To find a text, the string is searched,
685 counting zeros. There was no check for running off the end of the string,
686 which could happen if a new error number was added without updating the
687 string.
688
689 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
690
691 11. \K was not working if it appeared in an atomic group or in a group that
692 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
693 \K is "not well defined" if used in an assertion. PCRE now accepts it if
694 the assertion is positive, but not if it is negative.
695
696 12. Change 11 fortuitously reduced the size of the stack frame used in the
697 "match()" function of pcre_exec.c by one pointer. Forthcoming
698 implementation of support for (*MARK) will need an extra pointer on the
699 stack; I have reserved it now, so that the stack frame size does not
700 decrease.
701
702 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
703 item in branch that calls a recursion is a subroutine call - as in the
704 second branch in the above example - was incorrectly given the compile-
705 time error "recursive call could loop indefinitely" because pcre_compile()
706 was not correctly checking the subroutine for matching a non-empty string.
707
708 14. The checks for overrunning compiling workspace could trigger after an
709 overrun had occurred. This is a "should never occur" error, but it can be
710 triggered by pathological patterns such as hundreds of nested parentheses.
711 The checks now trigger 100 bytes before the end of the workspace.
712
713 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
714
715
716 Version 8.01 19-Jan-2010
717 ------------------------
718
719 1. If a pattern contained a conditional subpattern with only one branch (in
720 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
721 computed the wrong minimum data length (which is of course zero for such
722 subpatterns). This could cause incorrect "no match" results.
723
724 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
725 the pattern is reset in the first branch, pcre_compile() failed with
726 "internal error: code overflow at offset...". This happened only when
727 the reset was to the original external option setting. (An optimization
728 abstracts leading options settings into an external setting, which was the
729 cause of this.)
730
731 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
732 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
733 assertion pattern did not match (meaning that the assertion was true), it
734 was incorrectly treated as false if the SKIP had been reached during the
735 matching. This also applied to assertions used as conditions.
736
737 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
738 assertion subpattern, including such a pattern used as a condition,
739 unpredictable results occurred, instead of the error return
740 PCRE_ERROR_DFA_UITEM.
741
742 5. The C++ GlobalReplace function was not working like Perl for the special
743 situation when an empty string is matched. It now does the fancy magic
744 stuff that is necessary.
745
746 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
747 removed. (These were left over from very, very early versions of PCRE.)
748
749 7. Some cosmetic changes to the code to make life easier when compiling it
750 as part of something else:
751
752 (a) Change DEBUG to PCRE_DEBUG.
753
754 (b) In pcre_compile(), rename the member of the "branch_chain" structure
755 called "current" as "current_branch", to prevent a collision with the
756 Linux macro when compiled as a kernel module.
757
758 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
759 prevent a collision with the Linux macro when compiled as a kernel
760 module.
761
762 8. In pcre_compile() there are some checks for integer overflows that used to
763 cast potentially large values to (double). This has been changed to that
764 when building, a check for int64_t is made, and if it is found, it is used
765 instead, thus avoiding the use of floating point arithmetic. (There is no
766 other use of FP in PCRE.) If int64_t is not found, the fallback is to
767 double.
768
769 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
770 2005 (difference between two addresses compared to an unsigned value).
771
772 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
773 custom one, because of the following reported problem in Windows:
774
775 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
776 under Win32.
777 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
778 therefore missing the function definition.
779 - The compiler thus generates a "C" signature for the test function.
780 - The linker fails to find the "C" function.
781 - PCRE fails to configure if asked to do so against libbz2.
782
783 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
784 messages were output:
785
786 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
787 rerunning libtoolize, to keep the correct libtool macros in-tree.
788 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
789
790 I have done both of these things.
791
792 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
793 most of the time, it *can* run out if it is given a pattern that contains a
794 runaway infinite recursion. I updated the discussion in the pcrestack man
795 page.
796
797 13. Now that we have gone to the x.xx style of version numbers, the minor
798 version may start with zero. Using 08 or 09 is a bad idea because users
799 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
800 interpreted as invalid octal numbers. I've updated the previous comment in
801 configure.ac, and also added a check that gives an error if 08 or 09 are
802 used.
803
804 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
805 causing partial matching to fail when the end of the subject matched \W
806 in a UTF-8 pattern where \W was quantified with a minimum of 3.
807
808 15. There were some discrepancies between the declarations in pcre_internal.h
809 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
810 their definitions. The declarations used "const uschar *" and the
811 definitions used USPTR. Even though USPTR is normally defined as "const
812 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
813 reported that: "This difference in casting confuses some C++ compilers, for
814 example, SunCC recognizes above declarations as different functions and
815 generates broken code for hbpcre." I have changed the declarations to use
816 USPTR.
817
818 16. GNU libtool is named differently on some systems. The autogen.sh script now
819 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
820 (FreeBSD).
821
822 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
823 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
824 comment: "Figure out how to create a longlong from a string: strtoll and
825 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
826 instance, but it only takes 2 args instead of 3!"
827
828 18. A subtle bug concerned with back references has been fixed by a change of
829 specification, with a corresponding code fix. A pattern such as
830 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
831 refers, was giving matches when it shouldn't. For example, xa=xaaa would
832 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
833 same bug. Such groups have to be quantified to be useful, or contained
834 inside another quantified group. (If there's no repetition, the reference
835 can never match.) The problem arises because, having left the group and
836 moved on to the rest of the pattern, a later failure that backtracks into
837 the group uses the captured value from the final iteration of the group
838 rather than the correct earlier one. I have fixed this in PCRE by forcing
839 any group that contains a reference to itself to be an atomic group; that
840 is, there cannot be any backtracking into it once it has completed. This is
841 similar to recursive and subroutine calls.
842
843
844 Version 8.00 19-Oct-09
845 ----------------------
846
847 1. The table for translating pcre_compile() error codes into POSIX error codes
848 was out-of-date, and there was no check on the pcre_compile() error code
849 being within the table. This could lead to an OK return being given in
850 error.
851
852 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
853 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
854 in a Windows environment.
855
856 3. The pcregrep --count option prints the count for each file even when it is
857 zero, as does GNU grep. However, pcregrep was also printing all files when
858 --files-with-matches was added. Now, when both options are given, it prints
859 counts only for those files that have at least one match. (GNU grep just
860 prints the file name in this circumstance, but including the count seems
861 more useful - otherwise, why use --count?) Also ensured that the
862 combination -clh just lists non-zero counts, with no names.
863
864 4. The long form of the pcregrep -F option was incorrectly implemented as
865 --fixed_strings instead of --fixed-strings. This is an incompatible change,
866 but it seems right to fix it, and I didn't think it was worth preserving
867 the old behaviour.
868
869 5. The command line items --regex=pattern and --regexp=pattern were not
870 recognized by pcregrep, which required --regex pattern or --regexp pattern
871 (with a space rather than an '='). The man page documented the '=' forms,
872 which are compatible with GNU grep; these now work.
873
874 6. No libpcreposix.pc file was created for pkg-config; there was just
875 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
876
877 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
878 when UCP support is not needed, by modifying the Python script that
879 generates it from Unicode data files. This should not matter if the module
880 is correctly used as a library, but I received one complaint about 50K of
881 unwanted data. My guess is that the person linked everything into his
882 program rather than using a library. Anyway, it does no harm.
883
884 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
885 was a minimum greater than 1 for a wide character in a possessive
886 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
887 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
888 character. Chaos in the form of incorrect output or a compiling loop could
889 result.
890
891 9. The restrictions on what a pattern can contain when partial matching is
892 requested for pcre_exec() have been removed. All patterns can now be
893 partially matched by this function. In addition, if there are at least two
894 slots in the offset vector, the offset of the earliest inspected character
895 for the match and the offset of the end of the subject are set in them when
896 PCRE_ERROR_PARTIAL is returned.
897
898 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
899 synonymous with PCRE_PARTIAL, for backwards compatibility, and
900 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
901 and may be more useful for multi-segment matching.
902
903 11. Partial matching with pcre_exec() is now more intuitive. A partial match
904 used to be given if ever the end of the subject was reached; now it is
905 given only if matching could not proceed because another character was
906 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
907 string "Z", which now yields "no match" instead of "partial match". In the
908 case of pcre_dfa_exec(), "no match" is given if every matching path for the
909 final character ended with (*FAIL).
910
911 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
912 if the pattern had a "must contain" character that was already found in the
913 earlier partial match, unless partial matching was again requested. For
914 example, with the pattern /dog.(body)?/, the "must contain" character is
915 "g". If the first part-match was for the string "dog", restarting with
916 "sbody" failed. This bug has been fixed.
917
918 13. The string returned by pcre_dfa_exec() after a partial match has been
919 changed so that it starts at the first inspected character rather than the
920 first character of the match. This makes a difference only if the pattern
921 starts with a lookbehind assertion or \b or \B (\K is not supported by
922 pcre_dfa_exec()). It's an incompatible change, but it makes the two
923 matching functions compatible, and I think it's the right thing to do.
924
925 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
926 so that the demonstration program is easily available in environments where
927 PCRE has not been installed from source.
928
929 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
930 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
931 library.
932
933 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
934 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
935 is not the first non-POSIX option to be added. Clearly some people find
936 these options useful.
937
938 17. If a caller to the POSIX matching function regexec() passes a non-zero
939 value for nmatch with a NULL value for pmatch, the value of
940 nmatch is forced to zero.
941
942 18. RunGrepTest did not have a test for the availability of the -u option of
943 the diff command, as RunTest does. It now checks in the same way as
944 RunTest, and also checks for the -b option.
945
946 19. If an odd number of negated classes containing just a single character
947 interposed, within parentheses, between a forward reference to a named
948 subpattern and the definition of the subpattern, compilation crashed with
949 an internal error, complaining that it could not find the referenced
950 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
951 [The bug was that it was starting one character too far in when skipping
952 over the character class, thus treating the ] as data rather than
953 terminating the class. This meant it could skip too much.]
954
955 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
956 /g option in pcretest when the pattern contains \K, which makes it possible
957 to have an empty string match not at the start, even when the pattern is
958 anchored. Updated pcretest and pcredemo to use this option.
959
960 21. If the maximum number of capturing subpatterns in a recursion was greater
961 than the maximum at the outer level, the higher number was returned, but
962 with unset values at the outer level. The correct (outer level) value is
963 now given.
964
965 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
966 PCRE did not set those parentheses (unlike Perl). I have now found a way to
967 make it do so. The string so far is captured, making this feature
968 compatible with Perl.
969
970 23. The tests have been re-organized, adding tests 11 and 12, to make it
971 possible to check the Perl 5.10 features against Perl 5.10.
972
973 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
974 pattern matches a fixed length string. PCRE did not allow this; now it
975 does. Neither allows recursion.
976
977 25. I finally figured out how to implement a request to provide the minimum
978 length of subject string that was needed in order to match a given pattern.
979 (It was back references and recursion that I had previously got hung up
980 on.) This code has now been added to pcre_study(); it finds a lower bound
981 to the length of subject needed. It is not necessarily the greatest lower
982 bound, but using it to avoid searching strings that are too short does give
983 some useful speed-ups. The value is available to calling programs via
984 pcre_fullinfo().
985
986 26. While implementing 25, I discovered to my embarrassment that pcretest had
987 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
988 study optimizations had never been tested with that matching function.
989 Oops. What is worse, even when it was passed study data, there was a bug in
990 pcre_dfa_exec() that meant it never actually used it. Double oops. There
991 were also very few tests of studied patterns with pcre_dfa_exec().
992
993 27. If (?| is used to create subpatterns with duplicate numbers, they are now
994 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
995 on the other side of the coin, they are no longer allowed to have different
996 names, because these cannot be distinguished in PCRE, and this has caused
997 confusion. (This is a difference from Perl.)
998
999 28. When duplicate subpattern names are present (necessarily with different
1000 numbers, as required by 27 above), and a test is made by name in a
1001 conditional pattern, either for a subpattern having been matched, or for
1002 recursion in such a pattern, all the associated numbered subpatterns are
1003 tested, and the overall condition is true if the condition is true for any
1004 one of them. This is the way Perl works, and is also more like the way
1005 testing by number works.
1006
1007
1008 Version 7.9 11-Apr-09
1009 ---------------------
1010
1011 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1012 (pcretest), all targets were linked against these libraries. This included
1013 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1014 libraries. This caused unwanted dependencies to be created. This problem
1015 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1016 pcretest is linked with readline.
1017
1018 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1019 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1020 moved outside it again, because FALSE and TRUE are already defined in AIX,
1021 but BOOL is not.
1022
1023 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1024 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1025
1026 4. The pcregrep documentation said spaces were inserted as well as colons (or
1027 hyphens) following file names and line numbers when outputting matching
1028 lines. This is not true; no spaces are inserted. I have also clarified the
1029 wording for the --colour (or --color) option.
1030
1031 5. In pcregrep, when --colour was used with -o, the list of matching strings
1032 was not coloured; this is different to GNU grep, so I have changed it to be
1033 the same.
1034
1035 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1036 each matching line was coloured. Now it goes on to look for further matches
1037 of any of the test patterns, which is the same behaviour as GNU grep.
1038
1039 7. A pattern that could match an empty string could cause pcregrep to loop; it
1040 doesn't make sense to accept an empty string match in pcregrep, so I have
1041 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1042 seems to be how GNU grep behaves.
1043
1044 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1045 start or after a newline", because the conditional assertion was not being
1046 correctly handled. The rule now is that both the assertion and what follows
1047 in the first alternative must satisfy the test.
1048
1049 9. If auto-callout was enabled in a pattern with a conditional group whose
1050 condition was an assertion, PCRE could crash during matching, both with
1051 pcre_exec() and pcre_dfa_exec().
1052
1053 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1054 used for matching.
1055
1056 11. Unicode property support in character classes was not working for
1057 characters (bytes) greater than 127 when not in UTF-8 mode.
1058
1059 12. Added the -M command line option to pcretest.
1060
1061 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1062
1063 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1064
1065 16. Added comments and documentation about mis-use of no_arg in the C++
1066 wrapper.
1067
1068 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1069 from Martin Jerabek that uses macro names for all relevant character and
1070 string constants.
1071
1072 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1073 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1074 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1075 these, but not everybody uses configure.
1076
1077 19. A conditional group that had only one branch was not being correctly
1078 recognized as an item that could match an empty string. This meant that an
1079 enclosing group might also not be so recognized, causing infinite looping
1080 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1081 with the subject "ab", where knowledge that the repeated group can match
1082 nothing is needed in order to break the loop.
1083
1084 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1085 exec(), but without supplying a callout function, matching went wrong.
1086
1087 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1088 leak if the size of the offset vector was greater than 30. When the vector
1089 is smaller, the saved offsets during recursion go onto a local stack
1090 vector, but for larger vectors malloc() is used. It was failing to free
1091 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1092 error, in fact).
1093
1094 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1095 heapframe that is used only when UTF-8 support is enabled. This caused no
1096 problem, but was untidy.
1097
1098 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1099 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1100 included within another project.
1101
1102 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1103 slightly modified by me:
1104
1105 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1106 not building pcregrep.
1107
1108 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1109 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1110
1111 25. Forward references, both numeric and by name, in patterns that made use of
1112 duplicate group numbers, could behave incorrectly or give incorrect errors,
1113 because when scanning forward to find the reference group, PCRE was not
1114 taking into account the duplicate group numbers. A pattern such as
1115 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1116
1117 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1118 the feature of a custom pointer more persuasive (as requested by a user).
1119
1120 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1121 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1122 pre-defined.
1123
1124 28. Added support for (*UTF8) at the start of a pattern.
1125
1126 29. Arrange for flags added by the "release type" setting in CMake to be shown
1127 in the configuration summary.
1128
1129
1130 Version 7.8 05-Sep-08
1131 ---------------------
1132
1133 1. Replaced UCP searching code with optimized version as implemented for Ad
1134 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1135 stage table and inline lookup instead of a function, giving speed ups of 2
1136 to 5 times on some simple patterns that I tested. Permission was given to
1137 distribute the MultiStage2.py script that generates the tables (it's not in
1138 the tarball, but is in the Subversion repository).
1139
1140 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1141 scripts.
1142
1143 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1144 a group with a zero qualifier. The result of the study could be incorrect,
1145 or the function might crash, depending on the pattern.
1146
1147 4. Caseless matching was not working for non-ASCII characters in back
1148 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1149 It now works when Unicode Property Support is available.
1150
1151 5. In pcretest, an escape such as \x{de} in the data was always generating
1152 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1153 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1154 truncation.
1155
1156 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1157
1158 7. Added two (int) casts to pcregrep when printing the difference of two
1159 pointers, in case they are 64-bit values.
1160
1161 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1162 test 2 if it fails.
1163
1164 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1165 and a #define of that name to empty if it is not externally set. This is to
1166 allow users of MSVC to set it if necessary.
1167
1168 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1169 the convenience functions in the pcre_get.c source file.
1170
1171 11. An option change at the start of a pattern that had top-level alternatives
1172 could cause overwriting and/or a crash. This command provoked a crash in
1173 some environments:
1174
1175 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1176
1177 This potential security problem was recorded as CVE-2008-2371.
1178
1179 12. For a pattern where the match had to start at the beginning or immediately
1180 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1181 pcre_dfa_exec() could read past the end of the passed subject if there was
1182 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1183 pcretest so that it places the subject at the end of its malloc-ed buffer.
1184
1185 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1186 exec() might read past the end of the data buffer in UTF-8 mode.
1187
1188 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1189 the data contained the byte 0x85 as part of a UTF-8 character within its
1190 first line. This applied both to normal and DFA matching.
1191
1192 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1193 /^[^d]*?$/8 failed to match "abc".
1194
1195 16. Added a missing copyright notice to pcrecpp_internal.h.
1196
1197 17. Make it more clear in the documentation that values returned from
1198 pcre_exec() in ovector are byte offsets, not character counts.
1199
1200 18. Tidied a few places to stop certain compilers from issuing warnings.
1201
1202 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1203 supplied by Stefan Weber. I made a further small update for 7.8 because
1204 there is a change of source arrangements: the pcre_searchfuncs.c module is
1205 replaced by pcre_ucd.c.
1206
1207
1208 Version 7.7 07-May-08
1209 ---------------------
1210
1211 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1212 a string to a long long, pretend we don't even have a long long." This is
1213 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1214
1215 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1216 pre-7.6 versions, which defined a global no_arg variable instead of putting
1217 it in the RE class. (See also #8 below.)
1218
1219 3. Remove a line of dead code, identified by coverity and reported by Nuno
1220 Lopes.
1221
1222 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1223
1224 (1) The include/exclude patterns were being applied to the whole pathnames
1225 of files, instead of just to the final components.
1226
1227 (2) If there was more than one level of directory, the subdirectories were
1228 skipped unless they satisfied the include/exclude conditions. This is
1229 inconsistent with GNU grep (and could even be seen as contrary to the
1230 pcregrep specification - which I improved to make it absolutely clear).
1231 The action now is always to scan all levels of directory, and just
1232 apply the include/exclude patterns to regular files.
1233
1234 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1235 --exclude_dir in the tests to avoid scanning .svn directories.
1236
1237 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1238 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1239 doesn't support NULs in patterns.
1240
1241 7. Added some missing "const"s to declarations of static tables in
1242 pcre_compile.c and pcre_dfa_exec.c.
1243
1244 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1245 caused by fix #2 above. (Subsequently also a second patch to fix the
1246 first patch. And a third patch - this was a messy problem.)
1247
1248 9. Applied Craig's patch to remove the use of push_back().
1249
1250 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1251 matching function regexec().
1252
1253 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1254 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1255 references. PCRE supports relative numbers with this syntax (I don't think
1256 Oniguruma does).
1257
1258 12. Previously, a group with a zero repeat such as (...){0} was completely
1259 omitted from the compiled regex. However, this means that if the group
1260 was called as a subroutine from elsewhere in the pattern, things went wrong
1261 (an internal error was given). Such groups are now left in the compiled
1262 pattern, with a new opcode that causes them to be skipped at execution
1263 time.
1264
1265 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1266 to the way PCRE behaves:
1267
1268 (a) A lone ] character is dis-allowed (Perl treats it as data).
1269
1270 (b) A back reference to an unmatched subpattern matches an empty string
1271 (Perl fails the current match path).
1272
1273 (c) A data ] in a character class must be notated as \] because if the
1274 first data character in a class is ], it defines an empty class. (In
1275 Perl it is not possible to have an empty class.) The empty class []
1276 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1277 The negative empty class [^] matches any one character, independently
1278 of the DOTALL setting.
1279
1280 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1281 non-existent subpattern following a character class starting with ']' and
1282 containing () gave an internal compiling error instead of "reference to
1283 non-existent subpattern". Fortunately, when the pattern did exist, the
1284 compiled code was correct. (When scanning forwards to check for the
1285 existencd of the subpattern, it was treating the data ']' as terminating
1286 the class, so got the count wrong. When actually compiling, the reference
1287 was subsequently set up correctly.)
1288
1289 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1290 it was being rejected as not supported by pcre_dfa_exec(), even though
1291 other assertions are supported. I have made pcre_dfa_exec() support
1292 (*FAIL).
1293
1294 16. The implementation of 13c above involved the invention of a new opcode,
1295 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1296 cannot be changed at match time, I realized I could make a small
1297 improvement to matching performance by compiling OP_ALLANY instead of
1298 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1299 on the OP_ANY path.
1300
1301 17. Compiling pcretest on Windows with readline support failed without the
1302 following two fixes: (1) Make the unistd.h include conditional on
1303 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1304
1305 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1306 ncurses library to be included for pcretest when ReadLine support is
1307 requested, but also to allow for it to be overridden. This patch came from
1308 Daniel Bergström.
1309
1310 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1311 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1312 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1313 spotting this.
1314
1315
1316 Version 7.6 28-Jan-08
1317 ---------------------
1318
1319 1. A character class containing a very large number of characters with
1320 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1321 overflow.
1322
1323 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1324 HAVE_LONG_LONG is not defined.
1325
1326 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1327 bring them up to date and include new features. This patch includes:
1328
1329 - Fixed PH's badly added libz and libbz2 support.
1330 - Fixed a problem with static linking.
1331 - Added pcredemo. [But later removed - see 7 below.]
1332 - Fixed dftables problem and added an option.
1333 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1334 HAVE_LONG_LONG.
1335 - Added readline support for pcretest.
1336 - Added an listing of the option settings after cmake has run.
1337
1338 4. A user submitted a patch to Makefile that makes it easy to create
1339 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1340 Makefile.am that cause it to include this special target, without
1341 affecting anything else. Note that the same mingw target plus all
1342 the other distribution libraries and programs are now supported
1343 when configuring with CMake (see 6 below) instead of with
1344 Configure/Make.
1345
1346 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1347 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1348 exported in the Windows port". It has not yet been confirmed that the patch
1349 solves the problem, but it does no harm.
1350
1351 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1352 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1353 with CMake, and also correct the comment about stack recursion.
1354
1355 7. Remove the automatic building of pcredemo from the ./configure system and
1356 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1357 of a program that users should build themselves after PCRE is installed, so
1358 building it automatically is not really right. What is more, it gave
1359 trouble in some build environments.
1360
1361 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1362
1363
1364 Version 7.5 10-Jan-08
1365 ---------------------
1366
1367 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1368 values in parens when parsing an RE using the C++ wrapper."
1369
1370 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1371 Characters greater than 255 were excluded from the class instead of being
1372 included.
1373
1374 3. The same bug as (2) above applied to negated POSIX classes such as
1375 [:^space:].
1376
1377 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1378 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1379 I have changed it.
1380
1381 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1382 first named subpattern) and a construct such as (?&a) would reference the
1383 first named subpattern whose name started with "a" (in other words, the
1384 length check was missing). Both these problems are fixed. "Subpattern name
1385 expected" is now given for (?&) (a zero-length name), and this patch also
1386 makes it give the same error for \k'' (previously it complained that that
1387 was a reference to a non-existent subpattern).
1388
1389 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1390 this is right because (?- can be followed by option settings as well as by
1391 digits. I have, however, made the messages clearer.
1392
1393 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1394 than the number used in the conditional) now cause a compile-time error.
1395 This is actually not compatible with Perl, which accepts such patterns, but
1396 treats the conditional as always being FALSE (as PCRE used to), but it
1397 seems to me that giving a diagnostic is better.
1398
1399 8. Change "alphameric" to the more common word "alphanumeric" in comments
1400 and messages.
1401
1402 9. Fix two occurrences of "backslash" in comments that should have been
1403 "backspace".
1404
1405 10. Remove two redundant lines of code that can never be obeyed (their function
1406 was moved elsewhere).
1407
1408 11. The program that makes PCRE's Unicode character property table had a bug
1409 which caused it to generate incorrect table entries for sequences of
1410 characters that have the same character type, but are in different scripts.
1411 It amalgamated them into a single range, with the script of the first of
1412 them. In other words, some characters were in the wrong script. There were
1413 thirteen such cases, affecting characters in the following ranges:
1414
1415 U+002b0 - U+002c1
1416 U+0060c - U+0060d
1417 U+0061e - U+00612
1418 U+0064b - U+0065e
1419 U+0074d - U+0076d
1420 U+01800 - U+01805
1421 U+01d00 - U+01d77
1422 U+01d9b - U+01dbf
1423 U+0200b - U+0200f
1424 U+030fc - U+030fe
1425 U+03260 - U+0327f
1426 U+0fb46 - U+0fbb1
1427 U+10450 - U+1049d
1428
1429 12. The -o option (show only the matching part of a line) for pcregrep was not
1430 compatible with GNU grep in that, if there was more than one match in a
1431 line, it showed only the first of them. It now behaves in the same way as
1432 GNU grep.
1433
1434 13. If the -o and -v options were combined for pcregrep, it printed a blank
1435 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1436 does the same. The return code can be used to tell if there were any
1437 non-matching lines.
1438
1439 14. Added --file-offsets and --line-offsets to pcregrep.
1440
1441 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1442 infinitely looping recursion. The bug was that positive lookaheads were not
1443 being skipped when checking for a possible empty match (negative lookaheads
1444 and both kinds of lookbehind were skipped).
1445
1446 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1447 inclusion of <windows.h> to before rather than after the definition of
1448 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1449
1450 17. Specifying a possessive quantifier with a specific limit for a Unicode
1451 character property caused pcre_compile() to compile bad code, which led at
1452 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1453 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1454 caused the error; without that there was no problem.
1455
1456 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1457
1458 19. Added --enable-pcretest-libreadline.
1459
1460 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1461 RE::GlobalReplace(). As a result, the number of replacements returned was
1462 double what it should be. I removed one of the increments, but Craig sent a
1463 later patch that removed the other one (the right fix) and added unit tests
1464 that check the return values (which was not done before).
1465
1466 21. Several CMake things:
1467
1468 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1469 the names libpcre and libpcreposix, not just pcre and pcreposix.
1470
1471 (2) The above change means that pcretest and pcregrep are now correctly
1472 linked with the newly-built libraries, not previously installed ones.
1473
1474 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1475
1476 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1477 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1478 UTF-8 newline character). The key issue is that the pattern starts .*;
1479 this means that the match must be either at the beginning, or after a
1480 newline. The bug was in the code for advancing after a failed match and
1481 checking that the new position followed a newline. It was not taking
1482 account of UTF-8 characters correctly.
1483
1484 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1485 character classes. PCRE was not treating the sequence [:...:] as a
1486 character class unless the ... were all letters. Perl, however, seems to
1487 allow any characters between [: and :], though of course it rejects as
1488 unknown any "names" that contain non-letters, because all the known class
1489 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1490 for example, whereas PCRE did not - it did not recognize a POSIX character
1491 class. This seemed a bit dangerous, so the code has been changed to be
1492 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1493 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1494 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1495 Perl does, and where it didn't before.
1496
1497 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1498 Windows environments %n is disabled by default.
1499
1500
1501 Version 7.4 21-Sep-07
1502 ---------------------
1503
1504 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1505 means that a class such as [\s] counted as "explicit reference to CR or
1506 LF". That isn't really right - the whole point of the change was to try to
1507 help when there was an actual mention of one of the two characters. So now
1508 the change happens only if \r or \n (or a literal CR or LF) character is
1509 encountered.
1510
1511 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1512 of both had grown to the point where there were only 3 bits left.
1513 Fortunately, there was spare space in the data structure, and so I have
1514 moved the internal flags into a new 16-bit field to free up more option
1515 bits.
1516
1517 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1518 but did not set the internal JCHANGED flag - either of these is enough to
1519 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1520 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1521 start sets both bits.
1522
1523 4. Added options (at build time, compile time, exec time) to change \R from
1524 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1525
1526 5. doc/pcresyntax.html was missing from the distribution.
1527
1528 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1529 compatibility, even though it is no longer used.
1530
1531 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1532 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1533 windows.h file is present (where different names are used). [This was
1534 reversed later after testing - see 16 below.]
1535
1536 8. Changed all #include <config.h> to #include "config.h". There were also
1537 some further <pcre.h> cases that I changed to "pcre.h".
1538
1539 9. When pcregrep was used with the --colour option, it missed the line ending
1540 sequence off the lines that it output.
1541
1542 10. It was pointed out to me that arrays of string pointers cause lots of
1543 relocations when a shared library is dynamically loaded. A technique of
1544 using a single long string with a table of offsets can drastically reduce
1545 these. I have refactored PCRE in four places to do this. The result is
1546 dramatic:
1547
1548 Originally: 290
1549 After changing UCP table: 187
1550 After changing error message table: 43
1551 After changing table of "verbs" 36
1552 After changing table of Posix names 22
1553
1554 Thanks to the folks working on Gregex for glib for this insight.
1555
1556 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1557 unicode-properties was also set.
1558
1559 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1560
1561 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1562 checked only for CRLF.
1563
1564 14. Added casts to pcretest.c to avoid compiler warnings.
1565
1566 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1567
1568 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1569 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1570 entirely. This removes changes made in 7 above.
1571
1572 17. The CMake files have been updated, and there is now more information about
1573 building with CMake in the NON-UNIX-USE document.
1574
1575
1576 Version 7.3 28-Aug-07
1577 ---------------------
1578
1579 1. In the rejigging of the build system that eventually resulted in 7.1, the
1580 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1581 brackets there is not right, since it causes compilers to look for an
1582 installed pcre.h, not the version that is in the source that is being
1583 compiled (which of course may be different). I have changed it back to:
1584
1585 #include "pcre.h"
1586
1587 I have a vague recollection that the change was concerned with compiling in
1588 different directories, but in the new build system, that is taken care of
1589 by the VPATH setting the Makefile.
1590
1591 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1592 when the subject happened to end in the byte 0x85 (e.g. if the last
1593 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1594 characters but of course it shouldn't be taken as a newline when it is part
1595 of another character. The bug was that, for an unlimited repeat of . in
1596 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1597 characters when looking for a newline.
1598
1599 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1600
1601 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1602 in debug output.
1603
1604 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1605 long printing in the pcrecpp unittest when running under MinGW.
1606
1607 6. ESC_K was left out of the EBCDIC table.
1608
1609 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1610 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1611 limit also applies to "virtual nesting" when a pattern is recursive, and in
1612 this case 1000 isn't so big. I have been able to remove this limit at the
1613 expense of backing off one optimization in certain circumstances. Normally,
1614 when pcre_exec() would call its internal match() function recursively and
1615 immediately return the result unconditionally, it uses a "tail recursion"
1616 feature to save stack. However, when a subpattern that can match an empty
1617 string has an unlimited repetition quantifier, it no longer makes this
1618 optimization. That gives it a stack frame in which to save the data for
1619 checking that an empty string has been matched. Previously this was taken
1620 from the 1000-entry workspace that had been reserved. So now there is no
1621 explicit limit, but more stack is used.
1622
1623 8. Applied Daniel's patches to solve problems with the import/export magic
1624 syntax that is required for Windows, and which was going wrong for the
1625 pcreposix and pcrecpp parts of the library. These were overlooked when this
1626 problem was solved for the main library.
1627
1628 9. There were some crude static tests to avoid integer overflow when computing
1629 the size of patterns that contain repeated groups with explicit upper
1630 limits. As the maximum quantifier is 65535, the maximum group length was
1631 set at 30,000 so that the product of these two numbers did not overflow a
1632 32-bit integer. However, it turns out that people want to use groups that
1633 are longer than 30,000 bytes (though not repeat them that many times).
1634 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1635 made it possible to implement the integer overflow checks in a much more
1636 dynamic way, which I have now done. The artificial limitation on group
1637 length has been removed - we now have only the limit on the total length of
1638 the compiled pattern, which depends on the LINK_SIZE setting.
1639
1640 10. Fixed a bug in the documentation for get/copy named substring when
1641 duplicate names are permitted. If none of the named substrings are set, the
1642 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1643 empty string.
1644
1645 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1646 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1647 because the ] is interpreted as the first data character and the
1648 terminating ] is not found. PCRE has been made compatible with Perl in this
1649 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1650 cause memory overwriting.
1651
1652 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1653 string has been matched (to stop an infinite loop). It was not recognizing
1654 a conditional subpattern that could match an empty string if that
1655 subpattern was within another subpattern. For example, it looped when
1656 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1657 condition was not nested. This bug has been fixed.
1658
1659 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1660 past the start of the subject in the presence of bytes with the top bit
1661 set, for example "\x8aBCD".
1662
1663 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1664 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1665
1666 14. Optimized (?!) to (*FAIL).
1667
1668 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1669 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1670 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1671 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1672 does: it's just the validity check that is more restrictive.
1673
1674 16. Inserted checks for integer overflows during escape sequence (backslash)
1675 processing, and also fixed erroneous offset values for syntax errors during
1676 backslash processing.
1677
1678 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1679 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1680
1681 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1682 caused an overrun.
1683
1684 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1685 something other than just ASCII characters) inside a group that had an
1686 unlimited repeat caused a loop at compile time (while checking to see
1687 whether the group could match an empty string).
1688
1689 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1690 [\P{Any}] did so. (Error in the code for printing property names.)
1691
1692 21. An orphan \E inside a character class could cause a crash.
1693
1694 22. A repeated capturing bracket such as (A)? could cause a wild memory
1695 reference during compilation.
1696
1697 23. There are several functions in pcre_compile() that scan along a compiled
1698 expression for various reasons (e.g. to see if it's fixed length for look
1699 behind). There were bugs in these functions when a repeated \p or \P was
1700 present in the pattern. These operators have additional parameters compared
1701 with \d, etc, and these were not being taken into account when moving along
1702 the compiled data. Specifically:
1703
1704 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1705 length.
1706
1707 (b) An item such as \pL+ within a repeated group could cause crashes or
1708 loops.
1709
1710 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1711 "reference to non-existent subpattern" error.
1712
1713 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1714
1715 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1716 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1717
1718 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1719
1720 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1721 character were causing crashes (broken optimization).
1722
1723 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1724 \p or \P) caused a compile-time loop.
1725
1726 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1727 break. For example, the unstudied pattern [\r\n]A does not match the string
1728 "\r\nA" because change 7.0/46 below moves the current point on by two
1729 characters after failing to match at the start. However, the pattern \nA
1730 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1731 the same is true. There doesn't seem any very clean way out of this, but
1732 what I have chosen to do makes the common cases work: PCRE now takes note
1733 of whether there can be an explicit match for \r or \n anywhere in the
1734 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1735 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1736 pattern has explicit CR or LF references.
1737
1738 29. Added (*CR) etc for changing newline setting at start of pattern.
1739
1740
1741 Version 7.2 19-Jun-07
1742 ---------------------
1743
1744 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1745 which is apparently normally available under Windows.
1746
1747 2. Re-jig the pcregrep tests with different newline settings in an attempt
1748 to make them independent of the local environment's newline setting.
1749
1750 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1751
1752 4. Some of the "internals" tests were previously cut out when the link size
1753 was not 2, because the output contained actual offsets. The recent new
1754 "Z" feature of pcretest means that these can be cut out, making the tests
1755 usable with all link sizes.
1756
1757 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1758 stack recursion. This gives a massive performance boost under BSD, but just
1759 a small improvement under Linux. However, it saves one field in the frame
1760 in all cases.
1761
1762 6. Added more features from the forthcoming Perl 5.10:
1763
1764 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1765 recursion call. It refers to the nth most recently opened parentheses.
1766
1767 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1768 to be opened parentheses.
1769
1770 (c) Conditions that refer to capturing parentheses can be specified
1771 relatively, for example, (?(-2)... or (?(+3)...
1772
1773 (d) \K resets the start of the current match so that everything before
1774 is not part of it.
1775
1776 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1777
1778 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1779 reference syntax.
1780
1781 (g) (?| introduces a group in which the numbering of parentheses in each
1782 alternative starts with the same number.
1783
1784 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1785
1786 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1787 PCRE_INFO_JCHANGED.
1788
1789 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1790 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1791 for detecting groups that can match an empty string.
1792
1793 9. A pattern with a very large number of alternatives (more than several
1794 hundred) was running out of internal workspace during the pre-compile
1795 phase, where pcre_compile() figures out how much memory will be needed. A
1796 bit of new cunning has reduced the workspace needed for groups with
1797 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1798 workspace instead of running out of the 4096 that are available.
1799
1800 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1801
1802 11. Applied patch from Google to remove an optimization that didn't quite work.
1803 The report of the bug said:
1804
1805 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1806 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1807 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1808
1809 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1810 it matched the wrong number of bytes.
1811
1812
1813 Version 7.1 24-Apr-07
1814 ---------------------
1815
1816 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1817 that is more "standard", making use of automake and other Autotools. There
1818 is some re-arrangement of the files and adjustment of comments consequent
1819 on this.
1820
1821 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1822 for recursive directory scanning broke on some systems because the files
1823 are not scanned in any specific order and on different systems the order
1824 was different. A call to "sort" has been inserted into RunGrepTest for the
1825 approprate test as a short-term fix. In the longer term there may be an
1826 alternative.
1827
1828 3. I had an email from Eric Raymond about problems translating some of PCRE's
1829 man pages to HTML (despite the fact that I distribute HTML pages, some
1830 people do their own conversions for various reasons). The problems
1831 concerned the use of low-level troff macros .br and .in. I have therefore
1832 removed all such uses from the man pages (some were redundant, some could
1833 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1834 HTML has been updated to handle .nf/.fi and to complain if it encounters
1835 .br or .in.
1836
1837 4. Updated comments in configure.ac that get placed in config.h.in and also
1838 arranged for config.h to be included in the distribution, with the name
1839 config.h.generic, for the benefit of those who have to compile without
1840 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1841
1842 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1843 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1844 makevp.bat for the current PCRE, using the additional files
1845 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1846
1847 6. A Windows user reported a minor discrepancy with test 2, which turned out
1848 to be caused by a trailing space on an input line that had got lost in his
1849 copy. The trailing space was an accident, so I've just removed it.
1850
1851 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1852 that is needed.
1853
1854 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1855 as "const" (a) because they are and (b) because it helps the PHP
1856 maintainers who have recently made a script to detect big data structures
1857 in the php code that should be moved to the .rodata section. I remembered
1858 to update Builducptable as well, so it won't revert if ucptable.h is ever
1859 re-created.
1860
1861 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1862 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1863 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1864 support is not required. This saves 1.5-2K of code, which is important in
1865 some applications.
1866
1867 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1868 so as not to refer to the tables, even though these functions will never be
1869 called when UTF-8 support is disabled. Otherwise there are problems with a
1870 shared library.
1871
1872 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1873
1874 (a) It was defining its arguments as char * instead of void *.
1875
1876 (b) It was assuming that all moves were upwards in memory; this was true
1877 a long time ago when I wrote it, but is no longer the case.
1878
1879 The emulated memove() is provided for those environments that have neither
1880 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1881 is clearly not the case, as these two bugs were recently reported.
1882
1883 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1884 and Detrail to create the HTML documentation, the .txt form of the man
1885 pages, and it removes trailing spaces from listed files. It also creates
1886 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1887 case, it wraps all the #defines with #ifndefs. This script should be run
1888 before "make dist".
1889
1890 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1891 with Unicode property support.
1892
1893 (a) For a maximizing quantifier, if the two different cases of the
1894 character were of different lengths in their UTF-8 codings (there are
1895 some cases like this - I found 11), and the matching function had to
1896 back up over a mixture of the two cases, it incorrectly assumed they
1897 were both the same length.
1898
1899 (b) When PCRE was configured to use the heap rather than the stack for
1900 recursion during matching, it was not correctly preserving the data for
1901 the other case of a UTF-8 character when checking ahead for a match
1902 while processing a minimizing repeat. If the check also involved
1903 matching a wide character, but failed, corruption could cause an
1904 erroneous result when trying to check for a repeat of the original
1905 character.
1906
1907 13. Some tidying changes to the testing mechanism:
1908
1909 (a) The RunTest script now detects the internal link size and whether there
1910 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1911 values substituted by "configure". (The RunGrepTest script already did
1912 this for UTF-8.) The configure.ac script no longer substitutes the
1913 relevant variables.
1914
1915 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1916 with length and offset values. This means that the output is different
1917 for different internal link sizes. Test 2 is skipped for link sizes
1918 other than 2 because of this, bypassing the problem. Unfortunately,
1919 there was also a test in test 3 (the locale tests) that used /B and
1920 failed for link sizes other than 2. Rather than cut the whole test out,
1921 I have added a new /Z option to pcretest that replaces the length and
1922 offset values with spaces. This is now used to make test 3 independent
1923 of link size. (Test 2 will be tidied up later.)
1924
1925 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1926 segmentation fault instead of returning the appropriate error message.
1927
1928 15. In multiline mode when the newline sequence was set to "any", the pattern
1929 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1930 This doesn't seem right; it now treats the CRLF combination as the line
1931 ending, and so does not match in that case. It's only a pattern such as ^$
1932 that would hit this one: something like ^ABC$ would have failed after \r
1933 and then tried again after \r\n.
1934
1935 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1936 in an attempt to make files that differ only in their line terminators
1937 compare equal. This works on Linux.
1938
1939 17. Under certain error circumstances pcregrep might try to free random memory
1940 as it exited. This is now fixed, thanks to valgrind.
1941
1942 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1943 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1944 was because its rules for how to advance for /g after matching an empty
1945 string at the end of a line did not allow for this case. They now check for
1946 it specially.
1947
1948 20. pcretest is supposed to handle patterns and data of any length, by
1949 extending its buffers when necessary. It was getting this wrong when the
1950 buffer for a data line had to be extended.
1951
1952 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1953 CRLF as a newline sequence.
1954
1955 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1956 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1957 I have nevertheless tidied it up.
1958
1959 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1960
1961 24. Added a man page for pcre-config.
1962
1963
1964 Version 7.0 19-Dec-06
1965 ---------------------
1966
1967 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1968 moving to gcc 4.1.1.
1969
1970 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1971 sys/time.h, which is documented as needed for this function. It doesn't
1972 seem to matter on Linux, but it showed up on some releases of OS X.
1973
1974 3. It seems that there are systems where bytes whose values are greater than
1975 127 match isprint() in the "C" locale. The "C" locale should be the
1976 default when a C program starts up. In most systems, only ASCII printing
1977 characters match isprint(). This difference caused the output from pcretest
1978 to vary, making some of the tests fail. I have changed pcretest so that:
1979
1980 (a) When it is outputting text in the compiled version of a pattern, bytes
1981 other than 32-126 are always shown as hex escapes.
1982
1983 (b) When it is outputting text that is a matched part of a subject string,
1984 it does the same, unless a different locale has been set for the match
1985 (using the /L modifier). In this case, it uses isprint() to decide.
1986
1987 4. Fixed a major bug that caused incorrect computation of the amount of memory
1988 required for a compiled pattern when options that changed within the
1989 pattern affected the logic of the preliminary scan that determines the
1990 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1991 that the computed length was too small. The symptoms of this bug were
1992 either the PCRE error "internal error: code overflow" from pcre_compile(),
1993 or a glibc crash with a message such as "pcretest: free(): invalid next
1994 size (fast)". Examples of patterns that provoked this bug (shown in
1995 pcretest format) are:
1996
1997 /(?-x: )/x
1998 /(?x)(?-x: \s*#\s*)/
1999 /((?i)[\x{c0}])/8
2000 /(?i:[\x{c0}])/8
2001
2002 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2003 is now done differently.
2004
2005 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2006 wrapper classes; (b) implement a new function in the C++ scanner that is
2007 more efficient than the old way of doing things because it avoids levels of
2008 recursion in the regex matching; (c) add a paragraph to the documentation
2009 for the FullMatch() function.
2010
2011 6. The escape sequence \n was being treated as whatever was defined as
2012 "newline". Not only was this contrary to the documentation, which states
2013 that \n is character 10 (hex 0A), but it also went horribly wrong when
2014 "newline" was defined as CRLF. This has been fixed.
2015
2016 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2017 was being set to -1 for the "end of line" case (supposedly a value that no
2018 character can have). Though this value is never used (the check for end of
2019 line is "zero bytes in current character"), it caused compiler complaints.
2020 I've changed it to 0xffffffff.
2021
2022 8. In pcre_version.c, the version string was being built by a sequence of
2023 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2024 string (as it is for production releases) called a macro with an empty
2025 argument. The C standard says the result of this is undefined. The gcc
2026 compiler treats it as an empty string (which was what was wanted) but it is
2027 reported that Visual C gives an error. The source has been hacked around to
2028 avoid this problem.
2029
2030 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2031 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2032 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2033 of them did).
2034
2035 10. Originally, pcretest opened its input and output without "b"; then I was
2036 told that "b" was needed in some environments, so it was added for release
2037 5.0 to both the input and output. (It makes no difference on Unix-like
2038 systems.) Later I was told that it is wrong for the input on Windows. I've
2039 now abstracted the modes into two macros, to make it easier to fiddle with
2040 them, and removed "b" from the input mode under Windows.
2041
2042 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2043
2044 12. Added -help and --help to pcretest as an official way of being reminded
2045 of the options.
2046
2047 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2048 and pcrecpp.cc because they annoy compilers at high warning levels.
2049
2050 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2051
2052 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2053 scripts are not C programs :-) and which was not noticed because it works
2054 on Linux.
2055
2056 16. pcretest is supposed to handle any length of pattern and data line (as one
2057 line or as a continued sequence of lines) by extending its input buffer if
2058 necessary. This feature was broken for very long pattern lines, leading to
2059 a string of junk being passed to pcre_compile() if the pattern was longer
2060 than about 50K.
2061
2062 17. I have done a major re-factoring of the way pcre_compile() computes the
2063 amount of memory needed for a compiled pattern. Previously, there was code
2064 that made a preliminary scan of the pattern in order to do this. That was
2065 OK when PCRE was new, but as the facilities have expanded, it has become
2066 harder and harder to keep it in step with the real compile phase, and there
2067 have been a number of bugs (see for example, 4 above). I have now found a
2068 cunning way of running the real compile function in a "fake" mode that
2069 enables it to compute how much memory it would need, while actually only
2070 ever using a few hundred bytes of working memory and without too many
2071 tests of the mode. This should make future maintenance and development
2072 easier. A side effect of this work is that the limit of 200 on the nesting
2073 depth of parentheses has been removed (though this was never a serious
2074 limitation, I suspect). However, there is a downside: pcre_compile() now
2075 runs more slowly than before (30% or more, depending on the pattern). I
2076 hope this isn't a big issue. There is no effect on runtime performance.
2077
2078 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2079 newline (only possible for the last line of a file) and it was a
2080 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2081
2082 19. Added additional timing features to pcretest. (1) The -tm option now times
2083 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2084 separate command line item, by a number that specifies the number of
2085 repeats to use when timing. The default is 50000; this gives better
2086 precision, but takes uncomfortably long for very large patterns.
2087
2088 20. Extended pcre_study() to be more clever in cases where a branch of a
2089 subpattern has no definite first character. For example, (a*|b*)[cd] would
2090 previously give no result from pcre_study(). Now it recognizes that the
2091 first character must be a, b, c, or d.
2092
2093 21. There was an incorrect error "recursive call could loop indefinitely" if
2094 a subpattern (or the entire pattern) that was being tested for matching an
2095 empty string contained only one non-empty item after a nested subpattern.
2096 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2097 incorrectly, because the \d was being skipped in the check.
2098
2099 22. The pcretest program now has a new pattern option /B and a command line
2100 option -b, which is equivalent to adding /B to every pattern. This causes
2101 it to show the compiled bytecode, without the additional information that
2102 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2103 is the same as /B/I).
2104
2105 23. A new optimization is now able automatically to treat some sequences such
2106 as a*b as a*+b. More specifically, if something simple (such as a character
2107 or a simple class like \d) has an unlimited quantifier, and is followed by
2108 something that cannot possibly match the quantified thing, the quantifier
2109 is automatically "possessified".
2110
2111 24. A recursive reference to a subpattern whose number was greater than 39
2112 went wrong under certain circumstances in UTF-8 mode. This bug could also
2113 have affected the operation of pcre_study().
2114
2115 25. Realized that a little bit of performance could be had by replacing
2116 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2117
2118 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2119
2120 27. Possessive quantifiers such as a++ were previously implemented by turning
2121 them into atomic groups such as ($>a+). Now they have their own opcodes,
2122 which improves performance. This includes the automatically created ones
2123 from 23 above.
2124
2125 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2126 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2127 the first matched character to be a colon. This applied both to named and
2128 numbered groups.
2129
2130 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2131
2132 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2133 building PCRE on Windows easier, so I have included it in the distribution.
2134
2135 31. There is now a check in pcretest against a ridiculously large number being
2136 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2137 loop, the loop is abandoned.
2138
2139 32. Forward references to subpatterns in conditions such as (?(2)...) where
2140 subpattern 2 is defined later cause pcre_compile() to search forwards in
2141 the pattern for the relevant set of parentheses. This search went wrong
2142 when there were unescaped parentheses in a character class, parentheses
2143 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2144
2145 33. "Subroutine" calls and backreferences were previously restricted to
2146 referencing subpatterns earlier in the regex. This restriction has now
2147 been removed.
2148
2149 34. Added a number of extra features that are going to be in Perl 5.10. On the
2150 whole, these are just syntactic alternatives for features that PCRE had
2151 previously implemented using the Python syntax or my own invention. The
2152 other formats are all retained for compatibility.
2153
2154 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2155 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2156 also .NET compatible.
2157
2158 (b) A recursion or subroutine call to a named group can now be defined as
2159 (?&name) as well as (?P>name).
2160
2161 (c) A backreference to a named group can now be defined as \k<name> or
2162 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2163 5.10, are also .NET compatible.
2164
2165 (d) A conditional reference to a named group can now use the syntax
2166 (?(<name>) or (?('name') as well as (?(name).
2167
2168 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2169 groups (named and numbered) that are never evaluated inline, but can be
2170 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2171 is always false. There may be only one alternative in such a group.
2172
2173 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2174 as the simple (?(R). The condition is true only if the most recent
2175 recursion is that of the given number or name. It does not search out
2176 through the entire recursion stack.
2177
2178 (g) The escape \gN or \g{N} has been added, where N is a positive or
2179 negative number, specifying an absolute or relative reference.
2180
2181 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2182 some "unreachable code" warnings.
2183
2184 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2185 things, this adds five new scripts.
2186
2187 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2188 There were also incompatibilities regarding the handling of \Q..\E inside
2189 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2190 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2191
2192 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2193 matches an empty string, and forcibly breaks the loop. There were bugs in
2194 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2195 against aaaa the result was just "a" rather than "aaaa", for example. Two
2196 separate and independent bugs (that affected different cases) have been
2197 fixed.
2198
2199 39. Refactored the code to abolish the use of different opcodes for small
2200 capturing bracket numbers. This is a tidy that I avoided doing when I
2201 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2202 The new approach is not only tidier, it makes it possible to reduce the
2203 memory needed to fix the previous bug (38).
2204
2205 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2206 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2207 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2208 mode.
2209
2210 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2211 report.
2212
2213 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2214 copy construction and assignment in the C++ wrapper.
2215
2216 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2217 couple of bugs that could have given wrong results in the "--newline=crlf"
2218 case.
2219
2220 44. Added a number of casts and did some reorganization of signed/unsigned int
2221 variables following suggestions from Dair Grant. Also renamed the variable
2222 "this" as "item" because it is a C++ keyword.
2223
2224 45. Arranged for dftables to add
2225
2226 #include "pcre_internal.h"
2227
2228 to pcre_chartables.c because without it, gcc 4.x may remove the array
2229 definition from the final binary if PCRE is built into a static library and
2230 dead code stripping is activated.
2231
2232 46. For an unanchored pattern, if a match attempt fails at the start of a
2233 newline sequence, and the newline setting is CRLF or ANY, and the next two
2234 characters are CRLF, advance by two characters instead of one.
2235
2236
2237 Version 6.7 04-Jul-06
2238 ---------------------
2239
2240 1. In order to handle tests when input lines are enormously long, pcretest has
2241 been re-factored so that it automatically extends its buffers when
2242 necessary. The code is crude, but this _is_ just a test program. The
2243 default size has been increased from 32K to 50K.
2244
2245 2. The code in pcre_study() was using the value of the re argument before
2246 testing it for NULL. (Of course, in any sensible call of the function, it
2247 won't be NULL.)
2248
2249 3. The memmove() emulation function in pcre_internal.h, which is used on
2250 systems that lack both memmove() and bcopy() - that is, hardly ever -
2251 was missing a "static" storage class specifier.
2252
2253 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2254 containing an extended class (one that cannot be represented by a bitmap
2255 because it contains high-valued characters or Unicode property items, e.g.
2256 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2257 pattern, but PCRE should not loop if you do not (it no longer does).
2258 [Detail: two cases were found: (a) a repeated subpattern containing an
2259 extended class; (b) a recursive reference to a subpattern that followed a
2260 previous extended class. It wasn't skipping over the extended class
2261 correctly when UTF-8 mode was not set.]
2262
2263 5. A negated single-character class was not being recognized as fixed-length
2264 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2265 compile error "lookbehind assertion is not fixed length".
2266
2267 6. The RunPerlTest auxiliary script was showing an unexpected difference
2268 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2269 write a Perl script that can interpret lines of an input file either as
2270 byte characters or as UTF-8, which is what "perltest" was being required to
2271 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2272 can't do is switch easily at run time between having the "use utf8;" pragma
2273 or not. In the end, I fudged it by using the RunPerlTest script to insert
2274 "use utf8;" explicitly for the UTF-8 tests.
2275
2276 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2277 the end of the subject string, contrary to the documentation and to what
2278 Perl does. This was true of both matching functions. Now it matches only at
2279 the start of the subject and immediately after *internal* newlines.
2280
2281 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2282 a pointer to an int instead of a pointer to an unsigned long int. This
2283 caused problems on 64-bit systems.
2284
2285 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2286 instance of the 'standard' template library not being so standard".
2287
2288 10. There was no check on the number of named subpatterns nor the maximum
2289 length of a subpattern name. The product of these values is used to compute
2290 the size of the memory block for a compiled pattern. By supplying a very
2291 long subpattern name and a large number of named subpatterns, the size
2292 computation could be caused to overflow. This is now prevented by limiting
2293 the length of names to 32 characters, and the number of named subpatterns
2294 to 10,000.
2295
2296 11. Subpatterns that are repeated with specific counts have to be replicated in
2297 the compiled pattern. The size of memory for this was computed from the
2298 length of the subpattern and the repeat count. The latter is limited to
2299 65535, but there was no limit on the former, meaning that integer overflow
2300 could in principle occur. The compiled length of a repeated subpattern is
2301 now limited to 30,000 bytes in order to prevent this.
2302
2303 12. Added the optional facility to have named substrings with the same name.
2304
2305 13. Added the ability to use a named substring as a condition, using the
2306 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2307 are numbers (not recommended). Forward references are permitted.
2308
2309 14. Added forward references in named backreferences (if you see what I mean).
2310
2311 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2312 pattern could run off the end of the subject. For example, the pattern
2313 "(?s)(.{1,5})"8 did this with the subject "ab".
2314
2315 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2316 PCRE_CASELESS was set when matching characters that were quantified with ?
2317 or *.
2318
2319 17. A character class other than a single negated character that had a minimum
2320 but no maximum quantifier - for example [ab]{6,} - was not handled
2321 correctly by pce_dfa_exec(). It would match only one character.
2322
2323 18. A valid (though odd) pattern that looked like a POSIX character
2324 class but used an invalid character after [ (for example [[,abc,]]) caused
2325 pcre_compile() to give the error "Failed: internal error: code overflow" or
2326 in some cases to crash with a glibc free() error. This could even happen if
2327 the pattern terminated after [[ but there just happened to be a sequence of
2328 letters, a binary zero, and a closing ] in the memory that followed.
2329
2330 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2331 over the years. Originally (before any Unicode support), just the bottom 8
2332 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2333 output from "man perlunicode" includes this:
2334
2335 The regular expression compiler produces polymorphic opcodes. That
2336 is, the pattern adapts to the data and automatically switches to
2337 the Unicode character scheme when presented with Unicode data--or
2338 instead uses a traditional byte scheme when presented with byte
2339 data.
2340
2341 Sadly, a wide octal escape does not cause a switch, and in a string with
2342 no other multibyte characters, these octal escapes are treated as before.
2343 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2344 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2345 Unicode string.
2346
2347 I have not perpetrated such confusion in PCRE. Up till now, it took just
2348 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2349 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2350 translate to the appropriate multibyte character.
2351
2352 29. Applied some refactoring to reduce the number of warnings from Microsoft
2353 and Borland compilers. This has included removing the fudge introduced
2354 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2355 a warning about an unused variable.
2356
2357 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2358 characters since release 4.0, because Perl (from release 5.004) does not.
2359 [Or at least, is documented not to: some releases seem to be in conflict
2360 with the documentation.] However, when a pattern was studied with
2361 pcre_study() and all its branches started with \s, PCRE still included VT
2362 as a possible starting character. Of course, this did no harm; it just
2363 caused an unnecessary match attempt.
2364
2365 22. Removed a now-redundant internal flag bit that recorded the fact that case
2366 dependency changed within the pattern. This was once needed for "required
2367 byte" processing, but is no longer used. This recovers a now-scarce options
2368 bit. Also moved the least significant internal flag bit to the most-
2369 significant bit of the word, which was not previously used (hangover from
2370 the days when it was an int rather than a uint) to free up another bit for
2371 the future.
2372
2373 23. Added support for CRLF line endings as well as CR and LF. As well as the
2374 default being selectable at build time, it can now be changed at runtime
2375 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2376 specify that it is scanning data with non-default line endings.
2377
2378 24. Changed the definition of CXXLINK to make it agree with the definition of
2379 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2380
2381 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2382 recursions. This makes a big different to stack usage for some patterns.
2383
2384 26. If a subpattern containing a named recursion or subroutine reference such
2385 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2386 the space required for the compiled pattern went wrong and gave too small a
2387 value. Depending on the environment, this could lead to "Failed: internal
2388 error: code overflow at offset 49" or "glibc detected double free or
2389 corruption" errors.
2390
2391 27. Applied patches from Google (a) to support the new newline modes and (b) to
2392 advance over multibyte UTF-8 characters in GlobalReplace.
2393
2394 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2395 difference for some implementation of PCRE in some Windows version.
2396
2397 29. Added some extra testing facilities to pcretest:
2398
2399 \q<number> in a data line sets the "match limit" value
2400 \Q<number> in a data line sets the "match recursion limt" value
2401 -S <number> sets the stack size, where <number> is in megabytes
2402
2403 The -S option isn't available for Windows.
2404
2405
2406 Version 6.6 06-Feb-06
2407 ---------------------
2408
2409 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2410 in pcreposix.h. I have copied the definition from pcre.h.
2411
2412 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2413 because pcre.h is no longer a built file.
2414
2415 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2416 not normally included in the compiled code.
2417
2418
2419 Version 6.5 01-Feb-06
2420 ---------------------
2421
2422 1. When using the partial match feature with pcre_dfa_exec(), it was not
2423 anchoring the second and subsequent partial matches at the new starting
2424 point. This could lead to incorrect results. For example, with the pattern
2425 /1234/, partially matching against "123" and then "a4" gave a match.
2426
2427 2. Changes to pcregrep:
2428
2429 (a) All non-match returns from pcre_exec() were being treated as failures
2430 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2431 error message is output. Some extra information is given for the
2432 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2433 probably the only errors that are likely to be caused by users (by
2434 specifying a regex that has nested indefinite repeats, for instance).
2435 If there are more than 20 of these errors, pcregrep is abandoned.
2436
2437 (b) A binary zero was treated as data while matching, but terminated the
2438 output line if it was written out. This has been fixed: binary zeroes
2439 are now no different to any other data bytes.
2440
2441 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2442 used to set a locale for matching. The --locale=xxxx long option has
2443 been added (no short equivalent) to specify a locale explicitly on the
2444 pcregrep command, overriding the environment variables.
2445
2446 (d) When -B was used with -n, some line numbers in the output were one less
2447 than they should have been.
2448
2449 (e) Added the -o (--only-matching) option.
2450
2451 (f) If -A or -C was used with -c (count only), some lines of context were
2452 accidentally printed for the final match.
2453
2454 (g) Added the -H (--with-filename) option.
2455
2456 (h) The combination of options -rh failed to suppress file names for files
2457 that were found from directory arguments.
2458
2459 (i) Added the -D (--devices) and -d (--directories) options.
2460
2461 (j) Added the -F (--fixed-strings) option.
2462
2463 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2464
2465 (l) Added the --colo(u)r option.
2466
2467 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2468 is not present by default.
2469
2470 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2471 items such as (?R) or (?1), when the recursion could match a number of
2472 alternatives. If it matched one of the alternatives, but subsequently,
2473 outside the recursion, there was a failure, the code tried to back up into
2474 the recursion. However, because of the way PCRE is implemented, this is not
2475 possible, and the result was an incorrect result from the match.
2476
2477 In order to prevent this happening, the specification of recursion has
2478 been changed so that all such subpatterns are automatically treated as
2479 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2480
2481 4. I had overlooked the fact that, in some locales, there are characters for
2482 which isalpha() is true but neither isupper() nor islower() are true. In
2483 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2484 and ordfeminine) are like this. This affected the treatment of \w and \W
2485 when they appeared in character classes, but not when they appeared outside
2486 a character class. The bit map for "word" characters is now created
2487 separately from the results of isalnum() instead of just taking it from the
2488 upper, lower, and digit maps. (Plus the underscore character, of course.)
2489
2490 5. The above bug also affected the handling of POSIX character classes such as
2491 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2492 permanent tables. Instead, the bit maps for such a class were previously
2493 created as the appropriate unions of the upper, lower, and digit bitmaps.
2494 Now they are created by subtraction from the [[:word:]] class, which has
2495 its own bitmap.
2496
2497 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2498 It is created by subtracting the vertical space characters (\x09, \x0a,
2499 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2500 subtraction was done in the overall bitmap for a character class, meaning
2501 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2502 be recognized. This bug has been fixed.
2503
2504 7. Patches from the folks at Google:
2505
2506 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2507 real life, but is still worth protecting against".
2508
2509 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2510 regular expressions".
2511
2512 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2513 have it.
2514
2515 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2516 "configure" and the latter not, in order to fix a problem somebody had
2517 with compiling the Arg class on HP-UX.
2518
2519 (e) Improve the error-handling of the C++ wrapper a little bit.
2520
2521 (f) New tests for checking recursion limiting.
2522
2523 8. The pcre_memmove() function, which is used only if the environment does not
2524 have a standard memmove() function (and is therefore rarely compiled),
2525 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2526 returning a result (though PCRE never actually uses the result).
2527
2528 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2529 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2530 returned instead of calling malloc() with an overflowing number that would
2531 most likely cause subsequent chaos.
2532
2533 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2534
2535 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2536 with this option is matched, the nmatch and pmatch options of regexec() are
2537 ignored.
2538
2539 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2540 provided in case anyone wants to the the POSIX interface with UTF-8
2541 strings.
2542
2543 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2544 C++ linking (needed for some HP-UX environments).
2545
2546 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2547 (unused parameter) and in the pcre_printint() function (omitted "default"
2548 switch label when the default is to do nothing).
2549
2550 15. Added some code to make it possible, when PCRE is compiled as a C++
2551 library, to replace subject pointers for pcre_exec() with a smart pointer
2552 class, thus making it possible to process discontinuous strings.
2553
2554 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2555 much the same function. They were added by different people who were trying
2556 to make PCRE easy to compile on non-Unix systems. It has been suggested
2557 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2558 for compiling on Windows systems. I have therefore replaced it with
2559 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2560 defaults to "extern" for C or "extern C" for C++, which works fine on
2561 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2562 SCOPE with something explicit in config.h. In addition:
2563
2564 (a) pcreposix.h still had just "extern" instead of either of these macros;
2565 I have replaced it with PCRE_DATA_SCOPE.
2566
2567 (b) Functions such as _pcre_xclass(), which are internal to the library,
2568 but external in the C sense, all had PCRE_EXPORT in their definitions.
2569 This is apparently wrong for the Windows case, so I have removed it.
2570 (It makes no difference on Unix-like systems.)
2571
2572 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2573 of recursive calls to match(). This is different to MATCH_LIMIT because
2574 that limits the total number of calls to match(), not all of which increase
2575 the depth of recursion. Limiting the recursion depth limits the amount of
2576 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2577 when PCRE is compiled, and changed at run time. A patch from Google adds
2578 this functionality to the C++ interface.
2579
2580 18. Changes to the handling of Unicode character properties:
2581
2582 (a) Updated the table to Unicode 4.1.0.
2583
2584 (b) Recognize characters that are not in the table as "Cn" (undefined).
2585
2586 (c) I revised the way the table is implemented to a much improved format
2587 which includes recognition of ranges. It now supports the ranges that
2588 are defined in UnicodeData.txt, and it also amalgamates other
2589 characters into ranges. This has reduced the number of entries in the
2590 table from around 16,000 to around 3,000, thus reducing its size
2591 considerably. I realized I did not need to use a tree structure after
2592 all - a binary chop search is just as efficient. Having reduced the
2593 number of entries, I extended their size from 6 bytes to 8 bytes to
2594 allow for more data.
2595
2596 (d) Added support for Unicode script names via properties such as \p{Han}.
2597
2598 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2599 matching that character.
2600
2601 20. When matching a repeated Unicode property with a minimum greater than zero,
2602 (for example \pL{2,}), PCRE could look past the end of the subject if it
2603 reached it while seeking the minimum number of characters. This could
2604 happen only if some of the characters were more than one byte long, because
2605 there is a check for at least the minimum number of bytes.
2606
2607 21. Refactored the implementation of \p and \P so as to be more general, to
2608 allow for more different types of property in future. This has changed the
2609 compiled form incompatibly. Anybody with saved compiled patterns that use
2610 \p or \P will have to recompile them.
2611
2612 22. Added "Any" and "L&" to the supported property types.
2613
2614 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2615 but give a compile time error if the value is greater than 0xff.
2616
2617 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2618 accidentally not being installed or uninstalled.
2619
2620 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2621 made were to insert the current release number. This seemed silly, because
2622 it made things harder for people building PCRE on systems that don't run
2623 "configure". I have turned pcre.h into a distributed file, no longer built
2624 by "configure", with the version identification directly included. There is
2625 no longer a pcre.h.in file.
2626
2627 However, this change necessitated a change to the pcre-config script as
2628 well. It is built from pcre-config.in, and one of the substitutions was the
2629 release number. I have updated configure.ac so that ./configure now finds
2630 the release number by grepping pcre.h.
2631
2632 26. Added the ability to run the tests under valgrind.
2633
2634
2635 Version 6.4 05-Sep-05
2636 ---------------------
2637
2638 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2639 "--" to be printed when multiple files were scanned, even when none of the
2640 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2641 consider it to be a bug, and have restored the previous behaviour.
2642
2643 2. A couple of code tidies to get rid of compiler warnings.
2644
2645 3. The pcretest program used to cheat by referring to symbols in the library
2646 whose names begin with _pcre_. These are internal symbols that are not
2647 really supposed to be visible externally, and in some environments it is
2648 possible to suppress them. The cheating is now confined to including
2649 certain files from the library's source, which is a bit cleaner.
2650
2651 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2652 file's purpose clearer.
2653
2654 5. Reorganized pcre_ucp_findchar().
2655
2656
2657 Version 6.3 15-Aug-05
2658 ---------------------
2659
2660 1. The file libpcre.pc.in did not have general read permission in the tarball.
2661
2662 2. There were some problems when building without C++ support:
2663
2664 (a) If C++ support was not built, "make install" and "make test" still
2665 tried to test it.
2666
2667 (b) There were problems when the value of CXX was explicitly set. Some
2668 changes have been made to try to fix these, and ...
2669
2670 (c) --disable-cpp can now be used to explicitly disable C++ support.
2671
2672 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2673 backslash in a target when C++ was disabled. This confuses some
2674 versions of "make", apparently. Using an intermediate variable solves
2675 this. (Same for CPP_LOBJ.)
2676
2677 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2678 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2679 necessary on certain architectures.
2680
2681 4. Added a setting of -export-symbols-regex to the link command to remove
2682 those symbols that are exported in the C sense, but actually are local
2683 within the library, and not documented. Their names all begin with
2684 "_pcre_". This is not a perfect job, because (a) we have to except some
2685 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2686 available (and never for static libraries). I have made a note to try to
2687 find a way round (a) in the future.
2688
2689
2690 Version 6.2 01-Aug-05
2691 ---------------------
2692
2693 1. There was no test for integer overflow of quantifier values. A construction
2694 such as {1111111111111111} would give undefined results. What is worse, if
2695 a minimum quantifier for a parenthesized subpattern overflowed and became
2696 negative, the calculation of the memory size went wrong. This could have
2697 led to memory overwriting.
2698
2699 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2700
2701 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2702 operating environments where this matters.
2703
2704 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2705 PCRE options from within the C++ wrapper.
2706
2707 5. Named capturing subpatterns were not being correctly counted when a pattern
2708 was compiled. This caused two problems: (a) If there were more than 100
2709 such subpatterns, the calculation of the memory needed for the whole
2710 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2711 back references of the form \12, where the number was greater than 9, were
2712 not recognized as back references, even though there were sufficient
2713 previous subpatterns.
2714
2715 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2716 versions of gcc, e.g. 2.95.4.
2717
2718
2719 Version 6.1 21-Jun-05
2720 ---------------------
2721
2722 1. There was one reference to the variable "posix" in pcretest.c that was not
2723 surrounded by "#if !defined NOPOSIX".
2724
2725 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2726 the cross-check on the old pcre_info() function, for the benefit of the
2727 cut-down version of PCRE that is currently imported into Exim.
2728
2729 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2730 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2731 patterns that start (?i) but I don't think that matters. The use of (?i) is
2732 just an example; this all applies to the other options as well.
2733
2734 4. Since libtool seems to echo the compile commands it is issuing, the output
2735 from "make" can be reduced a bit by putting "@" in front of each libtool
2736 compile command.
2737
2738 5. Patch from the folks at Google for configure.in to be a bit more thorough
2739 in checking for a suitable C++ installation before trying to compile the
2740 C++ stuff. This should fix a reported problem when a compiler was present,
2741 but no suitable headers.
2742
2743 6. The man pages all had just "PCRE" as their title. I have changed them to
2744 be the relevant file name. I have also arranged that these names are
2745 retained in the file doc/pcre.txt, which is a concatenation in text format
2746 of all the man pages except the little individual ones for each function.
2747
2748 7. The NON-UNIX-USE file had not been updated for the different set of source
2749 files that come with release 6. I also added a few comments about the C++
2750 wrapper.
2751
2752
2753 Version 6.0 07-Jun-05
2754 ---------------------
2755
2756 1. Some minor internal re-organization to help with my DFA experiments.
2757
2758 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2759 didn't matter for the library itself when fully configured, but did matter
2760 when compiling without UCP support, or within Exim, where the ucp files are
2761 not imported.
2762
2763 3. Refactoring of the library code to split up the various functions into
2764 different source modules. The addition of the new DFA matching code (see
2765 below) to a single monolithic source would have made it really too
2766 unwieldy, quite apart from causing all the code to be include in a
2767 statically linked application, when only some functions are used. This is
2768 relevant even without the DFA addition now that patterns can be compiled in
2769 one application and matched in another.
2770
2771 The downside of splitting up is that there have to be some external
2772 functions and data tables that are used internally in different modules of
2773 the library but which are not part of the API. These have all had their
2774 names changed to start with "_pcre_" so that they are unlikely to clash
2775 with other external names.
2776
2777 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2778 a different (DFA) algorithm. Although it is slower than the original
2779 function, it does have some advantages for certain types of matching
2780 problem.
2781
2782 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2783 including restarting after a partial match.
2784
2785 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2786 defined when compiling for Windows was sent to me. I have put it into the
2787 code, though I have no means of testing or verifying it.
2788
2789 7. Added the pcre_refcount() auxiliary function.
2790
2791 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2792 match before or at the first newline in the subject string. In pcretest,
2793 the /f option on a pattern can be used to set this.
2794
2795 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2796 would behave wrongly. This has been present in PCRE since release 4.0.
2797
2798 10. A number of changes to the pcregrep command:
2799
2800 (a) Refactored how -x works; insert ^(...)$ instead of setting
2801 PCRE_ANCHORED and checking the length, in preparation for adding
2802 something similar for -w.
2803
2804 (b) Added the -w (match as a word) option.
2805
2806 (c) Refactored the way lines are read and buffered so as to have more
2807 than one at a time available.
2808
2809 (d) Implemented a pcregrep test script.
2810
2811 (e) Added the -M (multiline match) option. This allows patterns to match
2812 over several lines of the subject. The buffering ensures that at least
2813 8K, or the rest of the document (whichever is the shorter) is available
2814 for matching (and similarly the previous 8K for lookbehind assertions).
2815
2816 (f) Changed the --help output so that it now says
2817
2818 -w, --word-regex(p)
2819
2820 instead of two lines, one with "regex" and the other with "regexp"
2821 because that confused at least one person since the short forms are the
2822 same. (This required a bit of code, as the output is generated
2823 automatically from a table. It wasn't just a text change.)
2824
2825 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2826 option but starts with a hyphen. Could be a pattern or a path name
2827 starting with a hyphen, for instance.
2828
2829 (h) "-" can be given as a file name to represent stdin.
2830
2831 (i) When file names are being printed, "(standard input)" is used for
2832 the standard input, for compatibility with GNU grep. Previously
2833 "<stdin>" was used.
2834
2835 (j) The option --label=xxx can be used to supply a name to be used for
2836 stdin when file names are being printed. There is no short form.
2837
2838 (k) Re-factored the options decoding logic because we are going to add
2839 two more options that take data. Such options can now be given in four
2840 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2841
2842 (l) Added the -A, -B, and -C options for requesting that lines of context
2843 around matches be printed.
2844
2845 (m) Added the -L option to print the names of files that do not contain
2846 any matching lines, that is, the complement of -l.
2847
2848 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2849 continue to scan other files.
2850
2851 (o) The -s option was incorrectly implemented. For compatibility with other
2852 greps, it now suppresses the error message for a non-existent or non-
2853 accessible file (but not the return code). There is a new option called
2854 -q that suppresses the output of matching lines, which was what -s was
2855 previously doing.
2856
2857 (p) Added --include and --exclude options to specify files for inclusion
2858 and exclusion when recursing.
2859
2860 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2861 Hopefully, it now does.
2862
2863 12. Missing cast in pcre_study().
2864
2865 13. Added an "uninstall" target to the makefile.
2866
2867 14. Replaced "extern" in the function prototypes in Makefile.in with
2868 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2869 world, but is set differently for Windows.
2870
2871 15. Added a second compiling function called pcre_compile2(). The only
2872 difference is that it has an extra argument, which is a pointer to an
2873 integer error code. When there is a compile-time failure, this is set
2874 non-zero, in addition to the error test pointer being set to point to an
2875 error message. The new argument may be NULL if no error number is required
2876 (but then you may as well call pcre_compile(), which is now just a
2877 wrapper). This facility is provided because some applications need a
2878 numeric error indication, but it has also enabled me to tidy up the way
2879 compile-time errors are handled in the POSIX wrapper.
2880
2881 16. Added VPATH=.libs to the makefile; this should help when building with one
2882 prefix path and installing with another. (Or so I'm told by someone who
2883 knows more about this stuff than I do.)
2884
2885 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2886 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2887 match everything, including newlines. This is not POSIX-compatible, but
2888 somebody wanted the feature. From pcretest it can be activated by using
2889 both the P and the s flags.
2890
2891 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2892
2893 19. libpcre.pc was being incorrectly installed as executable.
2894
2895 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2896 it now also looks for '\r' so that it will work unmodified on Windows.
2897
2898 21. Added Google's contributed C++ wrapper to the distribution.
2899
2900 22. Added some untidy missing memory free() calls in pcretest, to keep
2901 Electric Fence happy when testing.
2902
2903
2904
2905 Version 5.0 13-Sep-04
2906 ---------------------
2907
2908 1. Internal change: literal characters are no longer packed up into items
2909 containing multiple characters in a single byte-string. Each character
2910 is now matched using a separate opcode. However, there may be more than one
2911 byte in the character in UTF-8 mode.
2912
2913 2. The pcre_callout_block structure has two new fields: pattern_position and
2914 next_item_length. These contain the offset in the pattern to the next match
2915 item, and its length, respectively.
2916
2917 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2918 insertion of callouts before each pattern item. Added the /C option to
2919 pcretest to make use of this.
2920
2921 4. On the advice of a Windows user, the lines
2922
2923 #if defined(_WIN32) || defined(WIN32)
2924 _setmode( _fileno( stdout ), 0x8000 );
2925 #endif /* defined(_WIN32) || defined(WIN32) */
2926
2927 have been added to the source of pcretest. This apparently does useful
2928 magic in relation to line terminators.
2929
2930 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2931 for the benefit of those environments where the "b" makes a difference.
2932
2933 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2934 to know about it. I have put a hack into configure.in that adds in code
2935 to set GCC=yes if CC=icc. This seems to end up at a point in the
2936 generated configure script that is early enough to affect the setting of
2937 compiler options, which is what is needed, but I have no means of testing
2938 whether it really works. (The user who reported this had patched the
2939 generated configure script, which of course I cannot do.)
2940
2941 LATER: After change 22 below (new libtool files), the configure script
2942 seems to know about icc (and also ecc). Therefore, I have commented out
2943 this hack in configure.in.
2944
2945 7. Added support for pkg-config (2 patches were sent in).
2946
2947 8. Negated POSIX character classes that used a combination of internal tables
2948 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2949 [[:^ascii]]. Typically, they would match almost any characters. The other
2950 POSIX classes were not broken in this way.
2951
2952 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2953 to find the match, as PCRE was deluded into thinking that the match had to
2954 start at the start point or following a newline. The same bug applied to
2955 patterns with negative forward assertions or any backward assertions
2956 preceding ".*" at the start, unless the pattern required a fixed first
2957 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2958
2959 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2960 starting at the last subject character, bytes beyond the end of the subject
2961 string were read.
2962
2963 11. Renamed the variable "class" as "classbits" to make life easier for C++
2964 users. (Previously there was a macro definition, but it apparently wasn't
2965 enough.)
2966
2967 12. Added the new field "tables" to the extra data so that tables can be passed
2968 in at exec time, or the internal tables can be re-selected. This allows
2969 a compiled regex to be saved and re-used at a later time by a different
2970 program that might have everything at different addresses.
2971
2972 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2973 -R library as well as a -L library.
2974
2975 14. The debugging options of pcretest (-d on the command line or D on a
2976 pattern) showed incorrect output for anything following an extended class
2977 that contained multibyte characters and which was followed by a quantifier.
2978
2979 15. Added optional support for general category Unicode character properties
2980 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2981 support. It adds about 90K to the size of the library. The meanings of the
2982 inbuilt class escapes such as \d and \s have NOT been changed.
2983
2984 16. Updated pcredemo.c to include calls to free() to release the memory for the
2985 compiled pattern.
2986
2987 17. The generated file chartables.c was being created in the source directory
2988 instead of in the building directory. This caused the build to fail if the
2989 source directory was different from the building directory, and was
2990 read-only.
2991
2992 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2993 file. No doubt somebody will tell me if they don't make sense... Also added
2994 Dan Mooney's comments about building on OpenVMS.
2995
2996 19. Added support for partial matching via the PCRE_PARTIAL option for
2997 pcre_exec() and the \P data escape in pcretest.
2998
2999 20. Extended pcretest with 3 new pattern features:
3000
3001 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3002 write the compiled pattern to the file whose name is "rest-of-line".
3003 This is a straight binary dump of the data, with the saved pointer to
3004 the character tables forced to be NULL. The study data, if any, is
3005 written too. After writing, pcretest reads a new pattern.
3006
3007 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3008 compiled pattern from the given file. There must not be any
3009 occurrences of "<" in the file name (pretty unlikely); if there are,
3010 pcretest will instead treat the initial "<" as a pattern delimiter.
3011 After reading in the pattern, pcretest goes on to read data lines as
3012 usual.
3013
3014 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3015 and 16-bit fields in a compiled pattern, to simulate a pattern that
3016 was compiled on a host of opposite endianness.
3017
3018 21. The pcre-exec() function can now cope with patterns that were compiled on
3019 hosts of opposite endianness, with this restriction:
3020
3021 As for any compiled expression that is saved and used later, the tables
3022 pointer field cannot be preserved; the extra_data field in the arguments
3023 to pcre_exec() should be used to pass in a tables address if a value
3024 other than the default internal tables were used at compile time.
3025
3026 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3027 now diagnosed as an error. Previously, most of the time, a negative number
3028 would have been treated as zero, but if in addition "ovector" was passed as
3029 NULL, a crash could occur.
3030
3031 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3032 new versions from the libtool 1.5 distribution (the last one is a copy of
3033 a file called libtool.m4). This seems to have fixed the need to patch
3034 "configure" to support Darwin 1.3 (which I used to do). However, I still
3035 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3036 workstation).
3037
3038 24. Changed the PCRE licence to be the more standard "BSD" licence.
3039
3040
3041 Version 4.5 01-Dec-03
3042 ---------------------
3043
3044 1. There has been some re-arrangement of the code for the match() function so
3045 that it can be compiled in a version that does not call itself recursively.
3046 Instead, it keeps those local variables that need separate instances for
3047 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3048 needs to "recurse". Keeping track of where control must go is done by means
3049 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3050 hide most of the details from the main code, and operates only if
3051 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3052 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3053 operating.
3054
3055 To make it easier for callers to provide specially tailored get/free
3056 functions for this usage, two new functions, pcre_stack_malloc, and
3057 pcre_stack_free, are used. They are always called in strict stacking order,
3058 and the size of block requested is always the same.
3059
3060 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3061 PCRE has been compiled to use the stack or the heap for recursion. The
3062 -C option of pcretest uses this to show which version is compiled.
3063
3064 A new data escape \S, is added to pcretest; it causes the amounts of store
3065 obtained and freed by both kinds of malloc/free at match time to be added
3066 to the output.
3067
3068 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3069 what's available on my current Linux desktop machine.
3070
3071 3. When matching a UTF-8 string, the test for a valid string at the start has
3072 been extended. If start_offset is not zero, PCRE now checks that it points
3073 to a byte that is the start of a UTF-8 character. If not, it returns
3074 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3075 this is necessary because there may be backward assertions in the pattern.
3076 When matching the same subject several times, it may save resources to use
3077 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3078
3079 4. The code for checking the validity of UTF-8 strings has been tightened so
3080 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3081 containing "overlong sequences".
3082
3083 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3084 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3085 should have been "&", but it just so happened that all the cases this let
3086 through by mistake were picked up later in the function.
3087
3088 6. I had used a variable called "isblank" - this is a C99 function, causing
3089 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3090
3091 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3092 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3093 so the version is shown; (c) stop "make test" echoing "./RunTest".
3094
3095 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3096
3097 9. The prototype for memmove() for systems that don't have it was using
3098 size_t, but the inclusion of the header that defines size_t was later. I've
3099 moved the #includes for the C headers earlier to avoid this.
3100
3101 10. Added some adjustments to the code to make it easier to compiler on certain
3102 special systems:
3103
3104 (a) Some "const" qualifiers were missing.
3105 (b) Added the macro EXPORT before all exported functions; by default this
3106 is defined to be empty.
3107 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3108 that it reads its output file name as an argument instead of writing
3109 to the standard output and assuming this can be redirected.
3110
3111 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3112 class containing characters with values greater than 255, PCRE compilation
3113 went into a loop.
3114
3115 12. A recursive reference to a subpattern that was within another subpattern
3116 that had a minimum quantifier of zero caused PCRE to crash. For example,
3117 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3118 recursion. If the recursively-called subpattern itself had a zero repeat,
3119 that was OK.
3120
3121 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3122 buffer into which it was copied (for escape processing) was still set at
3123 1024, so long lines caused crashes.
3124
3125 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3126 "internal error: code overflow...". This applied to any character class
3127 that was followed by a possessive quantifier.
3128
3129 15. Modified the Makefile to add libpcre.la as a prerequisite for
3130 libpcreposix.la because I was told this is needed for a parallel build to
3131 work.
3132
3133 16. If a pattern that contained .* following optional items at the start was
3134 studied, the wrong optimizing data was generated, leading to matching
3135 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3136 matching string must start with a or b or c. The correct conclusion for
3137 this pattern is that a match can start with any character.
3138
3139
3140 Version 4.4 13-Aug-03
3141 ---------------------
3142
3143 1. In UTF-8 mode, a character class containing characters with values between
3144 127 and 255 was not handled correctly if the compiled pattern was studied.
3145 In fixing this, I have also improved the studying algorithm for such
3146 classes (slightly).
3147
3148 2. Three internal functions had redundant arguments passed to them. Removal
3149 might give a very teeny performance improvement.
3150
3151 3. Documentation bug: the value of the capture_top field in a callout is *one
3152 more than* the number of the hightest numbered captured substring.
3153
3154 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3155 in incorrectly linking with a previously installed version. They now link
3156 explicitly with libpcre.la.
3157
3158 5. configure.in no longer needs to recognize Cygwin specially.
3159
3160 6. A problem in pcre.in for Windows platforms is fixed.
3161
3162 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3163 pcretest, it used to include the size of the study block as part of its
3164 output. Unfortunately, the structure contains a field that has a different
3165 size on different hardware architectures. This meant that the tests that
3166 showed this size failed. As the block is currently always of a fixed size,
3167 this information isn't actually particularly useful in pcretest output, so
3168 I have just removed it.
3169
3170 8. Three pre-processor statements accidentally did not start in column 1.
3171 Sadly, there are *still* compilers around that complain, even though
3172 standard C has not required this for well over a decade. Sigh.
3173
3174 9. In pcretest, the code for checking callouts passed small integers in the
3175 callout_data field, which is a void * field. However, some picky compilers
3176 complained about the casts involved for this on 64-bit systems. Now
3177 pcretest passes the address of the small integer instead, which should get
3178 rid of the warnings.
3179
3180 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3181 both compile and run time, and gives an error if an invalid UTF-8 sequence
3182 is found. There is a option for disabling this check in cases where the
3183 string is known to be correct and/or the maximum performance is wanted.
3184
3185 11. In response to a bug report, I changed one line in Makefile.in from
3186
3187 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3188 to
3189 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3190
3191 to look similar to other lines, but I have no way of telling whether this
3192 is the right thing to do, as I do not use Windows. No doubt I'll get told
3193 if it's wrong...
3194
3195
3196 Version 4.3 21-May-03
3197 ---------------------
3198
3199 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3200 Makefile.
3201
3202 2. Some refactoring to improve the quality of the code:
3203
3204 (i) The utf8_table... variables are now declared "const".
3205
3206 (ii) The code for \cx, which used the "case flipping" table to upper case
3207 lower case letters, now just substracts 32. This is ASCII-specific,
3208 but the whole concept of \cx is ASCII-specific, so it seems
3209 reasonable.
3210
3211 (iii) PCRE was using its character types table to recognize decimal and
3212 hexadecimal digits in the pattern. This is silly, because it handles
3213 only 0-9, a-f, and A-F, but the character types table is locale-
3214 specific, which means strange things might happen. A private
3215 table is now used for this - though it costs 256 bytes, a table is
3216 much faster than multiple explicit tests. Of course, the standard
3217 character types table is still used for matching digits in subject
3218 strings against \d.
3219
3220 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3221 ending in _t are). So I've renamed it as ESC_tee.
3222
3223 3. The first argument for regexec() in the POSIX wrapper should have been
3224 defined as "const".
3225
3226 4. Changed pcretest to use malloc() for its buffers so that they can be
3227 Electric Fenced for debugging.
3228
3229 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3230 to read one or more bytes before the start of the subject string. Often this
3231 had no effect on PCRE's behaviour, but in some circumstances it could
3232 provoke a segmentation fault.
3233
3234 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3235 to try to read one or more bytes before the start of the subject string.
3236
3237 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3238 UTF-8 support could misbehave in various ways if the subject string
3239 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3240 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3241 back over UTF-8 characters.)
3242
3243
3244 Version 4.2 14-Apr-03
3245 ---------------------
3246
3247 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3248
3249 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3250 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3251 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3252 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3253 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3254 and BUILD_EXEEXT
3255 Note: automatic setting of the BUILD variables is not yet working
3256 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3257 compile-time but not at link-time
3258 [LINK]: use for linking executables only
3259 make different versions for Windows and non-Windows
3260 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3261 libraries
3262 [LINK_FOR_BUILD]: new variable
3263 [OBJEXT]: use throughout
3264 [EXEEXT]: use throughout
3265 <winshared>: new target
3266 <wininstall>: new target
3267 <dftables.o>: use native compiler
3268 <dftables>: use native linker
3269 <install>: handle Windows platform correctly
3270 <clean>: ditto
3271 <check>: ditto
3272 copy DLL to top builddir before testing
3273
3274 As part of these changes, -no-undefined was removed again. This was reported
3275 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3276 in any case.
3277
3278 3. Some tidies to get rid of compiler warnings:
3279
3280 . In the match_data structure, match_limit was an unsigned long int, whereas
3281 match_call_count was an int. I've made them both unsigned long ints.
3282
3283 . In pcretest the fact that a const uschar * doesn't automatically cast to
3284 a void * provoked a warning.
3285
3286 . Turning on some more compiler warnings threw up some "shadow" variables
3287 and a few more missing casts.
3288
3289 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3290 option, a class that contained a single character with a value between 128
3291 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3292
3293 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3294 option, a class that contained several characters, but with at least one
3295 whose value was between 128 and 255 caused PCRE to crash.
3296
3297
3298 Version 4.1 12-Mar-03
3299 ---------------------
3300
3301 1. Compiling with gcc -pedantic found a couple of places where casts were
3302 needed, and a string in dftables.c that was longer than standard compilers are
3303 required to support.
3304
3305 2. Compiling with Sun's compiler found a few more places where the code could
3306 be tidied up in order to avoid warnings.
3307
3308 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3309 first of these names is deprecated in the latest Autoconf in favour of the name
3310 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3311 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3312 analogy I have changed it to CFLAGS_FOR_BUILD.
3313
3314 4. Added -no-undefined to the linking command in the Makefile, because this is
3315 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3316 linking step for the pcreposix library.
3317
3318 5. PCRE was failing to diagnose the case of two named groups with the same
3319 name.
3320
3321 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3322 literal character that is needed in the subject for a match, and scans along to
3323 ensure that it is present before embarking on the full matching process. This
3324 saves time in cases of nested unlimited repeats that are never going to match.
3325 Problem: the scan can take a lot of time if the subject is very long (e.g.
3326 megabytes), thus penalizing straightforward matches. It is now done only if the
3327 amount of subject to be scanned is less than 1000 bytes.
3328
3329 7. A lesser problem with the same optimization is that it was recording the
3330 first character of an anchored pattern as "needed", thus provoking a search
3331 right along the subject, even when the first match of the pattern was going to
3332 fail. The "needed" character is now not set for anchored patterns, unless it
3333 follows something in the pattern that is of non-fixed length. Thus, it still
3334 fulfils its original purpose of finding quick non-matches in cases of nested
3335 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3336
3337
3338 Version 4.0 17-Feb-03
3339 ---------------------
3340
3341 1. If a comment in an extended regex that started immediately after a meta-item
3342 extended to the end of string, PCRE compiled incorrect data. This could lead to
3343 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3344
3345 2. Moved to autoconf 2.53 and libtool 1.4.2.
3346
3347 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3348 the special perltest8 script is no longer needed - all the tests can be run
3349 from a single perltest script.
3350
3351 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3352 by \s. It has now been removed in PCRE. This means it isn't recognized as
3353 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3354 class [:space:] *does* include VT, thereby creating a mess.
3355
3356 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3357 space and tab.
3358
3359 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3360 its new features into the main test script, reducing the number of scripts.
3361
3362 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3363 were backward compatible, and made the (?i) apply to the whole pattern, as if
3364 /i were given. Now it behaves more logically, and applies the option setting
3365 only to what follows. PCRE has been changed to follow suit. However, if it
3366 finds options settings right at the start of the pattern, it extracts them into
3367 the global options, as before. Thus, they show up in the info data.
3368
3369 8. Added support for the \Q...\E escape sequence. Characters in between are
3370 treated as literals. This is slightly different from Perl in that $ and @ are
3371 also handled as literals inside the quotes. In Perl, they will cause variable
3372 interpolation. Note the following examples:
3373
3374 Pattern PCRE matches Perl matches
3375
3376 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3377 \Qabc\$xyz\E abc\$xyz abc\$xyz
3378 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3379
3380 For compatibility with Perl, \Q...\E sequences are recognized inside character
3381 classes as well as outside them.
3382
3383 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3384 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3385 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3386 signed/unsigned warnings.
3387
3388 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3389 option for pcretest, so I've replaced it by a simple function that does just
3390 that job.
3391
3392 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3393 "pcregrep -".
3394
3395 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3396 Java package. This provides some syntactic sugar for simple cases of what my
3397 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3398 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3399 item, you can use this simplified notation. Note that only makes sense with
3400 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3401 greediness, whatever the setting of the PCRE_UNGREEDY option.
3402
3403 13. A change of greediness default within a pattern was not taking effect at
3404 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3405 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3406 was abstracted outside.
3407
3408 14. PCRE now supports the \G assertion. It is true when the current matching
3409 position is at the start point of the match. This differs from \A when the
3410 starting offset is non-zero. Used with the /g option of pcretest (or similar
3411 code), it works in the same way as it does for Perl's /g option. If all
3412 alternatives of a regex begin with \G, the expression is anchored to the start
3413 match position, and the "anchored" flag is set in the compiled expression.
3414
3415 15. Some bugs concerning the handling of certain option changes within patterns
3416 have been fixed. These applied to options other than (?ims). For example,
3417 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3418 been the other way round. Some of this was related to change 7 above.
3419
3420 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3421 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3422 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3423 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3424
3425 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3426 mode. Unlike ".", it always matches newline, whatever the setting of
3427 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3428 assertions. Perl allows it, but it doesn't (in general) work because it can't
3429 calculate the length of the lookbehind. At least, that's the case for Perl
3430 5.8.0 - I've been told they are going to document that it doesn't work in
3431 future.
3432
3433 18. Added an error diagnosis for escapes that PCRE does not support: these are
3434 \L, \l, \N, \P, \p, \U, \u, and \X.
3435
3436 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3437 reading past the end of the pattern in cases such as /[abcd/.
3438
3439 20. PCRE was getting more memory than necessary for patterns with classes that
3440 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3441
3442 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3443 compiling PCRE for use with Virtual Pascal.
3444
3445 22. Small fix to the Makefile to make it work properly if the build is done
3446 outside the source tree.
3447
3448 23. Added a new extension: a condition to go with recursion. If a conditional
3449 subpattern starts with (?(R) the "true" branch is used if recursion has
3450 happened, whereas the "false" branch is used only at the top level.
3451
3452 24. When there was a very long string of literal characters (over 255 bytes
3453 without UTF support, over 250 bytes with UTF support), the computation of how
3454 much memory was required could be incorrect, leading to segfaults or other
3455 strange effects.
3456
3457 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3458 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3459 there was a subsequent back reference to those brackets. This meant that, for
3460 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3461 possible to check for precisely this case. All we can do is abandon the
3462 optimization if .* occurs inside capturing brackets when there are any back
3463 references whatsoever. (See below for a better fix that came later.)
3464
3465 26. The handling of the optimization for finding the first character of a
3466 non-anchored pattern, and for finding a character that is required later in the
3467 match were failing in some cases. This didn't break the matching; it just
3468 failed to optimize when it could. The way this is done has been re-implemented.
3469
3470 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3471
3472 28. Added a new feature that provides some of the functionality that Perl
3473 provides with (?{...}). The facility is termed a "callout". The way it is done
3474 in PCRE is for the caller to provide an optional function, by setting
3475 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3476 global variable. By default it is unset, which disables all calling out. To get
3477 the function called, the regex must include (?C) at appropriate points. This
3478 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3479 This provides a means of identifying different callout points. When PCRE
3480 reaches such a point in the regex, if pcre_callout has been set, the external
3481 function is called. It is provided with data in a structure called
3482 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3483 matching continues; if it returns a non-zero value, the match at the current
3484 point fails. However, backtracking will occur if possible. [This was changed
3485 later and other features added - see item 49 below.]
3486
3487 29. pcretest is upgraded to test the callout functionality. It provides a
3488 callout function that displays information. By default, it shows the start of
3489 the match and the current position in the text. There are some new data escapes
3490 to vary what happens:
3491
3492 \C+ in addition, show current contents of captured substrings
3493 \C- do not supply a callout function
3494 \C!n return 1 when callout number n is reached
3495 \C!n!m return 1 when callout number n is reached for the mth time
3496
3497 30. If pcregrep was called with the -l option and just a single file name, it
3498 output "<stdin>" if a match was found, instead of the file name.
3499
3500 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3501 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3502 pcre_exec(). This saves a malloc/free per call. The default value of
3503 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3504 when configuring.
3505
3506 32. The default maximum size of a compiled pattern is 64K. There have been a
3507 few cases of people hitting this limit. The code now uses macros to handle the
3508 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3509 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3510 configuring. Tests 2 and 5 work only with 2-byte links because they output
3511 debugging information about compiled patterns.
3512
3513 33. Internal code re-arrangements:
3514
3515 (a) Moved the debugging function for printing out a compiled regex into
3516 its own source file (printint.c) and used #include to pull it into
3517 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3518 separate copies.
3519
3520 (b) Defined the list of op-code names for debugging as a macro in
3521 internal.h so that it is next to the definition of the opcodes.
3522
3523 (c) Defined a table of op-code lengths for simpler skipping along compiled
3524 code. This is again a macro in internal.h so that it is next to the
3525 definition of the opcodes.
3526
3527 34. Added support for recursive calls to individual subpatterns, along the
3528 lines of Robin Houston's patch (but implemented somewhat differently).
3529
3530 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3531 allow it to read and process whole directories in Win32. This code was
3532 contributed by Lionel Fourquaux; it has not been tested by me.
3533
3534 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3535 used to name a group. Names consist of alphanumerics and underscores, and must
3536 be unique. Back references use the syntax (?P=name) and recursive calls use
3537 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3538 numbers. The function pcre_fullinfo() can be used after compilation to extract
3539 a name/number map. There are three relevant calls:
3540
3541 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3542 PCRE_INFO_NAMECOUNT yields the number of entries
3543 PCRE_INFO_NAMETABLE yields a pointer to the map.
3544
3545 The map is a vector of fixed-size entries. The size of each entry depends on
3546 the length of the longest name used. The first two bytes of each entry are the
3547 group number, most significant byte first. There follows the corresponding
3548 name, zero terminated. The names are in alphabetical order.
3549
3550 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3551 case instead of 255. Making it the same both with and without UTF-8 support
3552 means that the same test output works with both.
3553
3554 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3555 calling malloc() with a zero argument.
3556
3557 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3558 optimization. I've improved things by keeping a bitmap of backreferences with
3559 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3560 fact referenced, the optimization can be applied. It is unlikely that a
3561 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3562 the match to follow \n) will appear inside brackets with a number greater than
3563 31, but if it does, any back reference > 31 suppresses the optimization.
3564
3565 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3566 of disabling numbered capturing parentheses. Any opening parenthesis that is
3567 not followed by ? behaves as if it were followed by ?: but named parentheses
3568 can still be used for capturing (and they will acquire numbers in the usual
3569 way).
3570
3571 41. Redesigned the return codes from the match() function into yes/no/error so
3572 that errors can be passed back from deep inside the nested calls. A malloc
3573 failure while inside a recursive subpattern call now causes the
3574 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3575
3576 42. It is now possible to set a limit on the number of times the match()
3577 function is called in a call to pcre_exec(). This facility makes it possible to
3578 limit the amount of recursion and backtracking, though not in a directly
3579 obvious way, because the match() function is used in a number of different
3580 circumstances. The count starts from zero for each position in the subject
3581 string (for non-anchored patterns). The default limit is, for compatibility, a
3582 large number, namely 10 000 000. You can change this in two ways:
3583
3584 (a) When configuring PCRE before making, you can use --with-match-limit=n
3585 to set a default value for the compiled library.
3586
3587 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3588 a different value is set. See 45 below.
3589
3590 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3591
3592 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3593 of things that can be changed at compile time. The first argument specifies
3594 what is wanted and the second points to where the information is to be placed.
3595 The current list of available information is:
3596
3597 PCRE_CONFIG_UTF8
3598
3599 The output is an integer that is set to one if UTF-8 support is available;
3600 otherwise it is set to zero.
3601
3602 PCRE_CONFIG_NEWLINE
3603
3604 The output is an integer that it set to the value of the code that is used for
3605 newline. It is either LF (10) or CR (13).
3606
3607 PCRE_CONFIG_LINK_SIZE
3608
3609 The output is an integer that contains the number of bytes used for internal
3610 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3611
3612 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3613
3614 The output is an integer that contains the threshold above which the POSIX
3615 interface uses malloc() for output vectors. See item 31 above.
3616
3617 PCRE_CONFIG_MATCH_LIMIT
3618
3619 The output is an unsigned integer that contains the default limit of the number
3620 of match() calls in a pcre_exec() execution. See 42 above.
3621
3622 44. pcretest has been upgraded by the addition of the -C option. This causes it
3623 to extract all the available output from the new pcre_config() function, and to
3624 output it. The program then exits immediately.
3625
3626 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3627 order to support additional features. One way would have been to define
3628 pcre_exec2() (for example) with extra arguments, but this would not have been
3629 extensible, and would also have required all calls to the original function to
3630 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3631 is used for passing in "extra" data from pcre_study().
3632
3633 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3634 contains the following fields:
3635
3636 flags a bitmap indicating which of the following fields are set
3637 study_data opaque data from pcre_study()
3638 match_limit a way of specifying a limit on match() calls for a specific
3639 call to pcre_exec()
3640 callout_data data for callouts (see 49 below)
3641
3642 The flag bits are also defined in pcre.h, and are
3643
3644 PCRE_EXTRA_STUDY_DATA
3645 PCRE_EXTRA_MATCH_LIMIT
3646 PCRE_EXTRA_CALLOUT_DATA
3647
3648 The pcre_study() function now returns one of these new pcre_extra blocks, with
3649 the actual study data pointed to by the study_data field, and the
3650 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3651 before. That is, this change is entirely upwards-compatible and requires no
3652 change to existing code.
3653
3654 If you want to pass in additional data to pcre_exec(), you can either place it
3655 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3656 block.
3657
3658 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3659 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3660 times with different match limits, until it finds the minimum value needed for
3661 pcre_exec() to complete. The value is then output. This can be instructive; for
3662 most simple matches the number is quite small, but for pathological cases it
3663 gets very large very quickly.
3664
3665 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3666 returns the size of the data block pointed to by the study_data field in a
3667 pcre_extra block, that is, the value that was passed as the argument to
3668 pcre_malloc() when PCRE was getting memory in which to place the information
3669 created by pcre_study(). The fourth argument should point to a size_t variable.
3670 pcretest has been extended so that this information is shown after a successful
3671 pcre_study() call when information about the compiled regex is being displayed.
3672
3673 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3674 because what follows is always an absolute path. (Later: it turns out that this
3675 is more than cosmetic for MinGW, because it doesn't like empty path
3676 components.)
3677
3678 49. Some changes have been made to the callout feature (see 28 above):
3679
3680 (i) A callout function now has three choices for what it returns:
3681
3682 0 => success, carry on matching
3683 > 0 => failure at this point, but backtrack if possible
3684 < 0 => serious error, return this value from pcre_exec()
3685
3686 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3687 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3688 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3689 use by callout functions. It will never be used by PCRE itself.
3690
3691 (ii) The pcre_extra structure (see 45 above) has a void * field called
3692 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3693 pcre_callout_block structure has a field of the same name. The contents of
3694 the field passed in the pcre_extra structure are passed to the callout
3695 function in the corresponding field in the callout block. This makes it
3696 easier to use the same callout-containing regex from multiple threads. For
3697 testing, the pcretest program has a new data escape
3698
3699 \C*n pass the number n (may be negative) as callout_data
3700
3701 If the callout function in pcretest receives a non-zero value as
3702 callout_data, it returns that value.
3703
3704 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3705 there were some redundant $(CFLAGS) in commands that are now specified as
3706 $(LINK), which already includes $(CFLAGS).
3707
3708 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3709 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3710 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3711 one-byte characters throughout. Note that case-insensitive matching applies
3712 only to characters whose values are less than 256. PCRE doesn't support the
3713 notion of cases for higher-valued characters.
3714
3715 (i) A character class whose characters are all within 0-255 is handled as
3716 a bit map, and the map is inverted for negative classes. Previously, a
3717 character > 255 always failed to match such a class; however it should
3718 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3719
3720 (ii) A negated character class with a single character < 255 is coded as
3721 "not this character" (OP_NOT). This wasn't working properly when the test
3722 character was multibyte, either singly or repeated.
3723
3724 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3725 mode, for example: \x{100}{2,3}.
3726
3727 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3728 singly or repeated) now correctly test multibyte characters. However,
3729 PCRE doesn't recognize any characters with values greater than 255 as
3730 digits, spaces, or word characters. Such characters always match \D, \S,
3731 and \W, and never match \d, \s, or \w.
3732
3733 (v) Classes may now contain characters and character ranges with values
3734 greater than 255. For example: [ab\x{100}-\x{400}].
3735
3736 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3737 PCRE in UTF-8 mode.
3738
3739 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3740 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3741 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3742 value.)
3743
3744 53. The single man page has become too large. I have therefore split it up into
3745 a number of separate man pages. These also give rise to individual HTML pages;
3746 these are now put in a separate directory, and there is an index.html page that
3747 lists them all. Some hyperlinking between the pages has been installed.
3748
3749 54. Added convenience functions for handling named capturing parentheses.
3750
3751 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3752 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3753 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3754 are faulted.
3755
3756 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3757 calling configure. These values are used when compiling the dftables.c program
3758 which is run to generate the source of the default character tables. They
3759 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3760 you will need to set these values.
3761
3762 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3763
3764
3765 Version 3.9 02-Jan-02
3766 ---------------------
3767
3768 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3769
3770 2. If --disable-static was given, the building process failed when trying to
3771 build pcretest and pcregrep. (For some reason it was using libtool to compile
3772 them, which is not right, as they aren't part of the library.)
3773
3774
3775 Version 3.8 18-Dec-01
3776 ---------------------
3777
3778 1. The experimental UTF-8 code was completely screwed up. It was packing the
3779 bytes in the wrong order. How dumb can you get?
3780
3781
3782 Version 3.7 29-Oct-01
3783 ---------------------
3784
3785 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3786 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3787 this didn't happen under Solaris 8, where I normally test things.
3788
3789 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3790 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3791 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3792 this fix an hour or so after the initial 3.7 release.)
3793
3794
3795 Version 3.6 23-Oct-01
3796 ---------------------
3797
3798 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3799 offsets passed as NULL with zero offset count.
3800
3801 2. The config.guess and config.sub files had not been updated when I moved to
3802 the latest autoconf.
3803
3804
3805 Version 3.5 15-Aug-01
3806 ---------------------
3807
3808 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3809 had been forgotten.
3810
3811 2. By using declared but undefined structures, we can avoid using "void"
3812 definitions in pcre.h while keeping the internal definitions of the structures
3813 private.
3814
3815 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3816 user point of view, this means that both static and shared libraries are built
3817 by default, but this can be individually controlled. More of the work of
3818 handling this static/shared cases is now inside libtool instead of PCRE's make
3819 file.
3820
3821 4. The pcretest utility is now installed along with pcregrep because it is
3822 useful for users (to test regexs) and by doing this, it automatically gets
3823 relinked by libtool. The documentation has been turned into a man page, so
3824 there are now .1, .txt, and .html versions in /doc.
3825
3826 5. Upgrades to pcregrep:
3827 (i) Added long-form option names like gnu grep.
3828 (ii) Added --help to list all options with an explanatory phrase.
3829 (iii) Added -r, --recursive to recurse into sub-directories.
3830 (iv) Added -f, --file to read patterns from a file.
3831
3832 6. pcre_exec() was referring to its "code" argument before testing that
3833 argument for NULL (and giving an error if it was NULL).
3834
3835 7. Upgraded Makefile.in to allow for compiling in a different directory from
3836 the source directory.
3837
3838 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3839 options bits, the pointer it was passed was to an int instead of to an unsigned
3840 long int. This mattered only on 64-bit systems.
3841
3842 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3843 generated) instead of pcre.in, which it its source. Also made the same change
3844 in several of the .c files.
3845
3846 10. A new release of gcc defines printf() as a macro, which broke pcretest
3847 because it had an ifdef in the middle of a string argument for printf(). Fixed
3848 by using separate calls to printf().
3849
3850 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3851 script, to force use of CR or LF instead of \n in the source. On non-Unix
3852 systems, the value can be set in config.h.
3853
3854 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3855 absolute limit. Changed the text of the error message to make this clear, and
3856 likewise updated the man page.
3857
3858 13. The limit of 99 on the number of capturing subpatterns has been removed.
3859 The new limit is 65535, which I hope will not be a "real" limit.
3860
3861
3862 Version 3.4 22-Aug-00
3863 ---------------------
3864
3865 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3866
3867 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3868
3869
3870 Version 3.3 01-Aug-00
3871 ---------------------
3872
3873 1. If an octal character was given, but the value was greater than \377, it
3874 was not getting masked to the least significant bits, as documented. This could
3875 lead to crashes in some systems.
3876
3877 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3878 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3879
3880 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3881 These just pass their arguments on to (pcre_free)(), but they are provided
3882 because some uses of PCRE bind it to non-C systems that can call its functions,
3883 but cannot call free() or pcre_free() directly.
3884
3885 4. Add "make test" as a synonym for "make check". Corrected some comments in
3886 the Makefile.
3887
3888 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3889 Makefile.
3890
3891 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3892 command called pgrep for grepping around the active processes.
3893
3894 7. Added the beginnings of support for UTF-8 character strings.
3895
3896 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3897 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3898 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3899 out for the ar command.)
3900
3901
3902 Version 3.2 12-May-00
3903 ---------------------
3904
3905 This is purely a bug fixing release.
3906
3907 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3908 of ZA. This was just one example of several cases that could provoke this bug,
3909 which was introduced by change 9 of version 2.00. The code for breaking
3910 infinite loops after an iteration that matches an empty string was't working
3911 correctly.
3912
3913 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3914 when matched against abbab (for example). After matching an empty string, it
3915 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3916 caused it to match further down the string than it should.
3917
3918 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3919 was there because it doesn't seem to be needed, and it causes trouble on some
3920 systems, as it is not a Standard C header. It has been removed.
3921
3922 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3923 were reported on the Macintosh. The changes were from
3924
3925 while ((c = *(++ptr)) != 0 && c != '\n');
3926 to
3927 while ((c = *(++ptr)) != 0 && c != '\n') ;
3928
3929 Totally extraordinary, but if that's what it takes...
3930
3931 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3932 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3933 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3934 assumes the way PCRE uses memmove() (always moving upwards).
3935
3936 6. PCRE is being used in one environment where strchr() is not available. There
3937 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3938 faster code anyway.
3939
3940
3941 Version 3.1 09-Feb-00
3942 ---------------------
3943
3944 The only change in this release is the fixing of some bugs in Makefile.in for
3945 the "install" target:
3946
3947 (1) It was failing to install pcreposix.h.
3948
3949 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3950
3951
3952 Version 3.0 01-Feb-00
3953 ---------------------
3954
3955 1. Add support for the /+ modifier to perltest (to output $` like it does in
3956 pcretest).
3957
3958 2. Add support for the /g modifier to perltest.
3959
3960 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3961 matches null strings.
3962
3963 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3964 pattern. Perl treats empty patterns specially - it reuses the most recent
3965 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3966 effect.
3967
3968 5. The POSIX interface was broken in that it was just handing over the POSIX
3969 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3970 required a bigger vector, with some working space on the end. This means that
3971 the POSIX wrapper now has to get and free some memory, and copy the results.
3972
3973 6. Added some simple autoconf support, placing the test data and the
3974 documentation in separate directories, re-organizing some of the
3975 information files, and making it build pcre-config (a GNU standard). Also added
3976 libtool support for building PCRE as a shared library, which is now the
3977 default.
3978
3979 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3980 09 are not valid octal constants. Single digits will be used for minor values
3981 less than 10.
3982
3983 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3984 existing programs that set these in the POSIX interface can use PCRE without
3985 modification.
3986
3987 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3988 return all that pcre_info() returns, plus additional data. The pcre_info()
3989 function is retained for compatibility, but is considered to be obsolete.
3990
3991 10. Added experimental recursion feature (?R) to handle one common case that
3992 Perl 5.6 will be able to do with (?p{...}).
3993
3994 11. Added support for POSIX character classes like [:alpha:], which Perl is
3995 adopting.
3996
3997
3998 Version 2.08 31-Aug-99
3999 ----------------------
4000
4001 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4002 trying to match at the startoffset position, but instead was moving forward to
4003 the next newline as if a previous match had failed.
4004
4005 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4006 and could get into a loop if a null string was matched other than at the start
4007 of the subject.
4008
4009 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4010 be distinguished at compile time, and for completeness also added PCRE_DATE.
4011
4012 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4013 in GnuWin32 environments.
4014
4015
4016 Version 2.07 29-Jul-99
4017 ----------------------
4018
4019 1. The documentation is now supplied in plain text form and HTML as well as in
4020 the form of man page sources.
4021
4022 2. C++ compilers don't like assigning (void *) values to other pointer types.
4023 In particular this affects malloc(). Although there is no problem in Standard
4024 C, I've put in casts to keep C++ compilers happy.
4025
4026 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4027 should be (const char *).
4028
4029 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4030 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4031 However, I haven't made this a standard facility. The documentation doesn't
4032 mention it, and the Makefile doesn't support it.
4033
4034 5. The Makefile now contains an "install" target, with editable destinations at
4035 the top of the file. The pcretest program is not installed.
4036
4037 6. pgrep -V now gives the PCRE version number and date.
4038
4039 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4040 causing the entire string to be ignored, instead of just the last character.
4041
4042 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4043 non-matching string, it can take a very, very long time, even for strings of
4044 quite modest length, because of the nested recursion. PCRE now does better in
4045 some of these cases. It does this by remembering the last required literal
4046 character in the pattern, and pre-searching the subject to ensure it is present
4047 before running the real match. In other words, it applies a heuristic to detect
4048 some types of certain failure quickly, and in the above example, if presented
4049 with a string that has no trailing " it gives "no match" very quickly.
4050
4051 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4052 other alternatives are tried instead.
4053
4054
4055 Version 2.06 09-Jun-99
4056 ----------------------
4057
4058 1. Change pcretest's output for amount of store used to show just the code
4059 space, because the remainder (the data block) varies in size between 32-bit and
4060 64-bit systems.
4061
4062 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4063 start matching at. This allows lookbehinds to work when searching for multiple
4064 occurrences in a string.
4065
4066 3. Added additional options to pcretest for testing multiple occurrences:
4067
4068 /+ outputs the rest of the string that follows a match
4069 /g loops for multiple occurrences, using the new startoffset argument
4070 /G loops for multiple occurrences by passing an incremented pointer
4071
4072 4. PCRE wasn't doing the "first character" optimization for patterns starting
4073 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4074 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4075 the letter 'x'. On long subject strings, this gives a significant speed-up.
4076
4077
4078 Version 2.05 21-Apr-99
4079 ----------------------
4080
4081 1. Changed the type of magic_number from int to long int so that it works
4082 properly on 16-bit systems.
4083
4084 2. Fixed a bug which caused patterns starting with .* not to work correctly
4085 when the subject string contained newline characters. PCRE was assuming
4086 anchoring for such patterns in all cases, which is not correct because .* will
4087 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4088 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4089 must be retried after every newline in the subject.
4090
4091
4092 Version 2.04 18-Feb-99
4093 ----------------------
4094
4095 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4096 computation of the store needed to hold the pattern was incorrect (too large).
4097 If such patterns were nested a few deep, this could multiply and become a real
4098 problem.
4099
4100 2. Added /M option to pcretest to show the memory requirement of a specific
4101 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4102
4103 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4104 compiled in such a way that the backtracking after subsequent failure was
4105 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4106 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4107
4108
4109 Version 2.03 02-Feb-99
4110 ----------------------
4111
4112 1. Fixed typo and small mistake in man page.
4113
4114 2. Added 4th condition (GPL supersedes if conflict) and created separate
4115 LICENCE file containing the conditions.
4116
4117 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4118 Perl, that is the internal \ allows the delimiter to be included in the
4119 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4120 the final delimiter, add \ to the end of the pattern (to test the error).
4121
4122 4. Added the convenience functions for extracting substrings after a successful
4123 match. Updated pcretest to make it able to test these functions.
4124
4125
4126 Version 2.02 14-Jan-99
4127 ----------------------
4128
4129 1. Initialized the working variables associated with each extraction so that
4130 their saving and restoring doesn't refer to uninitialized store.
4131
4132 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4133 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4134 fix the problem.
4135
4136 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4137 calls, and wasn't printing the correct value for compiling calls. Increased the
4138 default value of LOOPREPEAT, and the number of significant figures in the
4139 times.
4140
4141 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4142
4143 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4144 a building problem on Windows NT with a FAT file system.
4145
4146
4147 Version 2.01 21-Oct-98
4148 ----------------------
4149
4150 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4151 to character tables built by pcre_maketables() in the current locale. If NULL
4152 is passed, the default tables are used.
4153
4154
4155 Version 2.00 24-Sep-98
4156 ----------------------
4157
4158 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4159 it any more.
4160
4161 2. Allow quantification of (?>) groups, and make it work correctly.
4162
4163 3. The first character computation wasn't working for (?>) groups.
4164
4165 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4166 end of the subject) and add 5.005's \z, which really does match only at the
4167 very end of the subject.
4168
4169 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4170
4171 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4172 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4173 localized options. All options to pcre_study() were also removed.
4174
4175 7. Add other new features from 5.005:
4176
4177 $(?<= positive lookbehind
4178 $(?<! negative lookbehind
4179 (?imsx-imsx) added the unsetting capability
4180 such a setting is global if at outer level; local otherwise
4181 (?imsx-imsx:) non-capturing groups with option setting
4182 (?(cond)re|re) conditional pattern matching
4183
4184 A backreference to itself in a repeated group matches the previous
4185 captured string.
4186
4187 8. General tidying up of studying (both automatic and via "study")
4188 consequential on the addition of new assertions.
4189
4190 9. As in 5.005, unlimited repeated groups that could match an empty substring
4191 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4192 runtime if any iteration does actually match an empty substring.
4193
4194 10. Include the RunTest script in the distribution.
4195
4196 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4197 discrepancies, some of which were old and were also with respect to 5.004. They
4198 have now been fixed.
4199
4200
4201 Version 1.09 28-Apr-98
4202 ----------------------
4203
4204 1. A negated single character class followed by a quantifier with a minimum
4205 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4206 program crashes, or just wrong answers. This did not apply to negated classes
4207 containing more than one character, or to minima other than one.
4208
4209
4210 Version 1.08 27-Mar-98
4211 ----------------------
4212
4213 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4214
4215 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4216 latter must appear before anything that relies on it in the pattern.
4217
4218
4219 Version 1.07 16-Feb-98
4220 ----------------------
4221
4222 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4223 repeat of a potentially empty string).
4224
4225
4226 Version 1.06 23-Jan-98
4227 ----------------------
4228
4229 1. Added Markus Oberhumer's little patches for C++.
4230
4231 2. Literal strings longer than 255 characters were broken.
4232
4233
4234 Version 1.05 23-Dec-97
4235 ----------------------
4236
4237 1. Negated character classes containing more than one character were failing if
4238 PCRE_CASELESS was set at run time.
4239
4240
4241 Version 1.04 19-Dec-97
4242 ----------------------
4243
4244 1. Corrected the man page, where some "const" qualifiers had been omitted.
4245
4246 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4247 input syntax.
4248
4249 3. Fixed memory leak which occurred when a regex with back references was
4250 matched with an offsets vector that wasn't big enough. The temporary memory
4251 that is used in this case wasn't being freed if the match failed.
4252
4253 4. Tidied pcretest to ensure it frees memory that it gets.
4254
4255 5. Temporary memory was being obtained in the case where the passed offsets
4256 vector was exactly big enough.
4257
4258 6. Corrected definition of offsetof() from change 5 below.
4259
4260 7. I had screwed up change 6 below and broken the rules for the use of
4261 setjmp(). Now fixed.
4262
4263
4264 Version 1.03 18-Dec-97
4265 ----------------------
4266
4267 1. A erroneous regex with a missing opening parenthesis was correctly
4268 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4269 on some systems.
4270
4271 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4272 it was reported that one broken compiler failed on the former because "code" is
4273 also an independent variable.
4274
4275 3. The erroneous regex a[]b caused an array overrun reference.
4276
4277 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4278 fail on data ending with that character. (It was going on too far, and checking
4279 the next character, typically a binary zero.) This was specific to the
4280 optimized code for single-character negative classes.
4281
4282 5. Added a contributed patch from the TIN world which does the following:
4283
4284 + Add an undef for memmove, in case the the system defines a macro for it.
4285
4286 + Add a definition of offsetof(), in case there isn't one. (I don't know
4287 the reason behind this - offsetof() is part of the ANSI standard - but
4288 it does no harm).
4289
4290 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4291 most of the places where whitespace preceded '#'. I have given up and
4292 allowed the remaining 2 cases to be at the margin.
4293
4294 + Rename some variables in pcre to eliminate shadowing. This seems very
4295 pedantic, but does no harm, of course.
4296
4297 6. Moved the call to setjmp() into its own function, to get rid of warnings
4298 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4299
4300 7. Constructs such as \d{8,} were compiling into the equivalent of
4301 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4302 outcome, but in this particular case used more store than had been allocated,
4303 which caused the bug to be discovered because it threw up an internal error.
4304
4305 8. The debugging code in both pcre and pcretest for outputting the compiled
4306 form of a regex was going wrong in the case of back references followed by
4307 curly-bracketed repeats.
4308
4309
4310 Version 1.02 12-Dec-97
4311 ----------------------
4312
4313 1. Typos in pcre.3 and comments in the source fixed.
4314
4315 2. Applied a contributed patch to get rid of places where it used to remove
4316 'const' from variables, and fixed some signed/unsigned and uninitialized
4317 variable warnings.
4318
4319 3. Added the "runtest" target to Makefile.
4320
4321 4. Set default compiler flag to -O2 rather than just -O.
4322
4323
4324 Version 1.01 19-Nov-97
4325 ----------------------
4326
4327 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4328 like /([ab]*)*/, that is, for classes with more than one character in them.
4329
4330 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4331 as /((?>a*))*/ (a PCRE_EXTRA facility).
4332
4333
4334 Version 1.00 18-Nov-97
4335 ----------------------
4336
4337 1. Added compile-time macros to support systems such as SunOS4 which don't have
4338 memmove() or strerror() but have other things that can be used instead.
4339
4340 2. Arranged that "make clean" removes the executables.
4341
4342
4343 Version 0.99 27-Oct-97
4344 ----------------------
4345
4346 1. Fixed bug in code for optimizing classes with only one character. It was
4347 initializing a 32-byte map regardless, which could cause it to run off the end
4348 of the memory it had got.
4349
4350 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4351
4352
4353 Version 0.98 22-Oct-97
4354 ----------------------
4355
4356 1. Fixed bug in code for handling temporary memory usage when there are more
4357 back references than supplied space in the ovector. This could cause segfaults.
4358
4359
4360 Version 0.97 21-Oct-97
4361 ----------------------
4362
4363 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4364
4365 2. Optimized negated single characters not to use a bit map.
4366
4367 3. Brought error texts together as macro definitions; clarified some of them;
4368 fixed one that was wrong - it said "range out of order" when it meant "invalid
4369 escape sequence".
4370
4371 4. Changed some char * arguments to const char *.
4372
4373 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4374
4375 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4376 pcretest.
4377
4378
4379 Version 0.96 16-Oct-97
4380 ----------------------
4381
4382 1. Added a simple "pgrep" utility to the distribution.
4383
4384 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4385 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4386 where "ddd" means "one or more decimal digits".
4387
4388 3. Fixed serious bug. If a pattern had a back reference, but the call to
4389 pcre_exec() didn't supply a large enough ovector to record the related
4390 identifying subpattern, the match always failed. PCRE now remembers the number
4391 of the largest back reference, and gets some temporary memory in which to save
4392 the offsets during matching if necessary, in order to ensure that
4393 backreferences always work.
4394
4395 4. Increased the compatibility with Perl in a number of ways:
4396
4397 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4398 to request this handling. The option can be set at compile or exec time.
4399
4400 (b) $ matches before a terminating newline by default; an option
4401 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4402 mode). The option can be set at compile or exec time.
4403
4404 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4405 the same as Perl's. If the decimal number it represents is less than 10
4406 or there aren't that many previous left capturing parentheses, an octal
4407 escape is read. Inside a character class, it's always an octal escape,
4408 even if it is a single digit.
4409
4410 (d) An escaped but undefined alphabetic character is taken as a literal,
4411 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4412 escapes.
4413
4414 (e) {0} is now permitted. (The previous item is removed from the compiled
4415 pattern).
4416
4417 5. Changed all the names of code files so that the basic parts are no longer
4418 than 10 characters, and abolished the teeny "globals.c" file.
4419
4420 6. Changed the handling of character classes; they are now done with a 32-byte
4421 bit map always.
4422
4423 7. Added the -d and /D options to pcretest to make it possible to look at the
4424 internals of compilation without having to recompile pcre.
4425
4426
4427 Version 0.95 23-Sep-97
4428 ----------------------
4429
4430 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4431 \x20 at the start of a run of normal characters. These were being treated as
4432 real characters, instead of the source characters being re-checked.
4433
4434
4435 Version 0.94 18-Sep-97
4436 ----------------------
4437
4438 1. The functions are now thread-safe, with the caveat that the global variables
4439 containing pointers to malloc() and free() or alternative functions are the
4440 same for all threads.
4441
4442 2. Get pcre_study() to generate a bitmap of initial characters for non-
4443 anchored patterns when this is possible, and use it if passed to pcre_exec().
4444
4445
4446 Version 0.93 15-Sep-97
4447 ----------------------
4448
4449 1. /(b)|(:+)/ was computing an incorrect first character.
4450
4451 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4452 but not actually doing anything yet.
4453
4454 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4455 as Perl does (e.g. [-az] or [az-]).
4456
4457 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4458 all possible positions.
4459
4460 5. Split up into different modules to avoid including unneeded functions in a
4461 compiled binary. However, compile and exec are still in one module. The "study"
4462 function is split off.
4463
4464 6. The character tables are now in a separate module whose source is generated
4465 by an auxiliary program - but can then be edited by hand if required. There are
4466 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4467 toupper() in the code.
4468
4469 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4470 make them global. Abolish the function for setting them, as the caller can now
4471 set them directly.
4472
4473
4474 Version 0.92 11-Sep-97
4475 ----------------------
4476
4477 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4478 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4479
4480 2. Caseless matching was not working in character classes if the characters in
4481 the pattern were in upper case.
4482
4483 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4484
4485 4. Make PCRE_ANCHORED public and accept as a compile option.
4486
4487 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4488 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4489 pass them.
4490
4491 6. Give an error if bad option bits passed at compile or run time.
4492
4493 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4494 pcretest to cause it to pass that flag.
4495
4496 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4497 options, and the first character, if set.
4498
4499 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4500
4501
4502 Version 0.91 10-Sep-97
4503 ----------------------
4504
4505 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4506 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4507
4508 2. PCRE was looping on encountering an indefinitely repeated back reference to
4509 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4510 Perl does - treats the match as successful.
4511
4512 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5