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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 752 - (show annotations)
Sat Nov 19 15:28:29 2011 UTC (8 years ago) by zherczeg
File size: 219758 byte(s)
Error occurred while calculating annotation data.
JIT should use pcre_malloc/pcre_free for allocation.
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 7. While fixing 6 above, I noticed that a number of other items were being
28 incorrectly rejected as "not fixed length". This arose partly because newer
29 opcodes had not been added to the fixed-length checking code. I have (a)
30 corrected the bug and added tests for these items, and (b) arranged for an
31 error to occur if an unknown opcode is encountered while checking for fixed
32 length instead of just assuming "not fixed length". The items that were
33 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
34 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
35 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
36
37 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
38 being incorrectly compiled and would have given unpredicatble results.
39
40 9. A possessively repeated subpattern with minimum repeat count greater than
41 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
42 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
43 the first (A) could occur when it should not.
44
45 10. Add a cast and remove a redundant test from the code.
46
47 11. JIT should use pcre_malloc/pcre_free for allocation.
48
49
50 Version 8.20 21-Oct-2011
51 ------------------------
52
53 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
54 a POSIX class. After further experiments with Perl, which convinced me that
55 Perl has bugs and confusions, a closing square bracket is no longer allowed
56 in a POSIX name. This bug also affected patterns with classes that started
57 with full stops.
58
59 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
60 captured substring, but while checking the failing first alternative,
61 substring 1 is temporarily captured. If the output vector supplied to
62 pcre_exec() was not big enough for this capture, the yield of the function
63 was still zero ("insufficient space for captured substrings"). This cannot
64 be totally fixed without adding another stack variable, which seems a lot
65 of expense for a edge case. However, I have improved the situation in cases
66 such as /(a)(b)x|abc/ matched against "abc", where the return code
67 indicates that fewer than the maximum number of slots in the ovector have
68 been set.
69
70 3. Related to (2) above: when there are more back references in a pattern than
71 slots in the output vector, pcre_exec() uses temporary memory during
72 matching, and copies in the captures as far as possible afterwards. It was
73 using the entire output vector, but this conflicts with the specification
74 that only 2/3 is used for passing back captured substrings. Now it uses
75 only the first 2/3, for compatibility. This is, of course, another edge
76 case.
77
78 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
79 main code base, and can be used by building with --enable-jit. When this is
80 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
81 runtime --no-jit option is given.
82
83 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
84 ovector, the return from the function was zero, implying that there were
85 other matches that did not fit. The correct "exactly full" value is now
86 returned.
87
88 6. If a subpattern that was called recursively or as a subroutine contained
89 (*PRUNE) or any other control that caused it to give a non-standard return,
90 invalid errors such as "Error -26 (nested recursion at the same subject
91 position)" or even infinite loops could occur.
92
93 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
94 computing the minimum length on reaching *ACCEPT, and so ended up with the
95 wrong value of 1 rather than 0. Further investigation indicates that
96 computing a minimum subject length in the presence of *ACCEPT is difficult
97 (think back references, subroutine calls), and so I have changed the code
98 so that no minimum is registered for a pattern that contains *ACCEPT.
99
100 8. If (*THEN) was present in the first (true) branch of a conditional group,
101 it was not handled as intended. [But see 16 below.]
102
103 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
104 Sheri Pierce.
105
106 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
107 the first byte in a match must be "a".
108
109 11. Change 17 for 8.13 increased the recursion depth for patterns like
110 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
111 pattern contains any instances of (*THEN). If it does not, the old
112 optimizations are restored. It would be nice to do this on a per-group
113 basis, but at the moment that is not feasible.
114
115 12. In some environments, the output of pcretest -C is CRLF terminated. This
116 broke RunTest's code that checks for the link size. A single white space
117 character after the value is now allowed for.
118
119 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
120 For "fr", it uses the Windows-specific input and output files.
121
122 14. If (*THEN) appeared in a group that was called recursively or as a
123 subroutine, it did not work as intended. [But see next item.]
124
125 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
126 pattern fragments (but not containing any | characters). If A and B are
127 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
128 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
129 D. In other words, Perl considers parentheses that do not contain any |
130 characters to be part of a surrounding alternative, whereas PCRE was
131 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
132 differently. PCRE now behaves in the same way as Perl, except in the case
133 of subroutine/recursion calls such as (?1) which have in any case always
134 been different (but PCRE had them first :-).
135
136 16. Related to 15 above: Perl does not treat the | in a conditional group as
137 creating alternatives. Such a group is treated in the same way as an
138 ordinary group without any | characters when processing (*THEN). PCRE has
139 been changed to match Perl's behaviour.
140
141 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
142 RunGrepTest script failed.
143
144 18. Change 22 for version 13 caused atomic groups to use more stack. This is
145 inevitable for groups that contain captures, but it can lead to a lot of
146 stack use in large patterns. The old behaviour has been restored for atomic
147 groups that do not contain any capturing parentheses.
148
149 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
150 suppress the check for a minimum subject length at run time. (If it was
151 given to pcre_exec() or pcre_dfa_exec() it did work.)
152
153 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
154 fail to work when decoding hex characters in data strings in EBCDIC
155 environments.
156
157 21. It appears that in at least one Mac OS environment, the isxdigit() function
158 is implemented as a macro that evaluates to its argument more than once,
159 contravening the C 90 Standard (I haven't checked a later standard). There
160 was an instance in pcretest which caused it to go wrong when processing
161 \x{...} escapes in subject strings. The has been rewritten to avoid using
162 things like p++ in the argument of isxdigit().
163
164
165 Version 8.13 16-Aug-2011
166 ------------------------
167
168 1. The Unicode data tables have been updated to Unicode 6.0.0.
169
170 2. Two minor typos in pcre_internal.h have been fixed.
171
172 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
173 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
174 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
175
176 4. There were a number of related bugs in the code for matching backrefences
177 caselessly in UTF-8 mode when codes for the characters concerned were
178 different numbers of bytes. For example, U+023A and U+2C65 are an upper
179 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
180 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
181 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
182 2-byte code at the end of the subject (it thought there wasn't enough data
183 left).
184
185 5. Comprehensive information about what went wrong is now returned by
186 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
187 as the output vector has at least 2 elements. The offset of the start of
188 the failing character and a reason code are placed in the vector.
189
190 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
191 now returned is for the first byte of the failing character, instead of the
192 last byte inspected. This is an incompatible change, but I hope it is small
193 enough not to be a problem. It makes the returned offset consistent with
194 pcre_exec() and pcre_dfa_exec().
195
196 7. pcretest now gives a text phrase as well as the error number when
197 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
198 failure, the offset and reason code are output.
199
200 8. When \R was used with a maximizing quantifier it failed to skip backwards
201 over a \r\n pair if the subsequent match failed. Instead, it just skipped
202 back over a single character (\n). This seems wrong (because it treated the
203 two characters as a single entity when going forwards), conflicts with the
204 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
205 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
206 has been changed.
207
208 9. Some internal refactoring has changed the processing so that the handling
209 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
210 time (the PCRE_DOTALL option was changed this way some time ago: version
211 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
212 which was always a bit of a fudge. It also means that there is one less
213 argument for the match() function, which reduces its stack requirements
214 slightly. This change also fixes an incompatibility with Perl: the pattern
215 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
216
217 10. More internal refactoring has drastically reduced the number of recursive
218 calls to match() for possessively repeated groups such as (abc)++ when
219 using pcre_exec().
220
221 11. While implementing 10, a number of bugs in the handling of groups were
222 discovered and fixed:
223
224 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
225 (a|)*(?1) gave a compile-time internal error.
226 ((a|)+)+ did not notice that the outer group could match an empty string.
227 (^a|^)+ was not marked as anchored.
228 (.*a|.*)+ was not marked as matching at start or after a newline.
229
230 12. Yet more internal refactoring has removed another argument from the match()
231 function. Special calls to this function are now indicated by setting a
232 value in a variable in the "match data" data block.
233
234 13. Be more explicit in pcre_study() instead of relying on "default" for
235 opcodes that mean there is no starting character; this means that when new
236 ones are added and accidentally left out of pcre_study(), testing should
237 pick them up.
238
239 14. The -s option of pcretest has been documented for ages as being an old
240 synonym of -m (show memory usage). I have changed it to mean "force study
241 for every regex", that is, assume /S for every regex. This is similar to -i
242 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
243 using it. It makes it easier to run collections of tests with and without
244 study enabled, and thereby test pcre_study() more easily. All the standard
245 tests are now run with and without -s (but some patterns can be marked as
246 "never study" - see 20 below).
247
248 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
249 restoration of the capturing data to the outer values was not happening
250 correctly.
251
252 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
253 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
254 pattern had matched an empty string, and so incorrectly returned a no
255 match.
256
257 17. There was optimizing code for the last branch of non-capturing parentheses,
258 and also for the obeyed branch of a conditional subexpression, which used
259 tail recursion to cut down on stack usage. Unfortunately, now that there is
260 the possibility of (*THEN) occurring in these branches, tail recursion is
261 no longer possible because the return has to be checked for (*THEN). These
262 two optimizations have therefore been removed. [But see 8.20/11 above.]
263
264 18. If a pattern containing \R was studied, it was assumed that \R always
265 matched two bytes, thus causing the minimum subject length to be
266 incorrectly computed because \R can also match just one byte.
267
268 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
269 was incorrectly computed.
270
271 20. If /S is present twice on a test pattern in pcretest input, it now
272 *disables* studying, thereby overriding the use of -s on the command line
273 (see 14 above). This is necessary for one or two tests to keep the output
274 identical in both cases.
275
276 21. When (*ACCEPT) was used in an assertion that matched an empty string and
277 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
278
279 22. When an atomic group that contained a capturing parenthesis was
280 successfully matched, but the branch in which it appeared failed, the
281 capturing was not being forgotten if a higher numbered group was later
282 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
283 group 1 to "a", when in fact it should be unset. This applied to multi-
284 branched capturing and non-capturing groups, repeated or not, and also to
285 positive assertions (capturing in negative assertions does not happen
286 in PCRE) and also to nested atomic groups.
287
288 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
289 subject after a captured substring, to make it easier to tell which of a
290 number of identical substrings has been captured.
291
292 24. The way atomic groups are processed by pcre_exec() has been changed so that
293 if they are repeated, backtracking one repetition now resets captured
294 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
295 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
296 "aaa". Previously, it would have been "a". As part of this code
297 refactoring, the way recursive calls are handled has also been changed.
298
299 25. If an assertion condition captured any substrings, they were not passed
300 back unless some other capturing happened later. For example, if
301 (?(?=(a))a) was matched against "a", no capturing was returned.
302
303 26. When studying a pattern that contained subroutine calls or assertions,
304 the code for finding the minimum length of a possible match was handling
305 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
306 group 1 called group 2 while simultaneously a separate group 2 called group
307 1). A stack overflow occurred in this case. I have fixed this by limiting
308 the recursion depth to 10.
309
310 27. Updated RunTest.bat in the distribution to the version supplied by Tom
311 Fortmann. This supports explicit test numbers on the command line, and has
312 argument validation and error reporting.
313
314 28. An instance of \X with an unlimited repeat could fail if at any point the
315 first character it looked at was a mark character.
316
317 29. Some minor code refactoring concerning Unicode properties and scripts
318 should reduce the stack requirement of match() slightly.
319
320 30. Added the '=' option to pcretest to check the setting of unused capturing
321 slots at the end of the pattern, which are documented as being -1, but are
322 not included in the return count.
323
324 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
325 compiled something random. Now it gives a compile-time error (as does
326 Perl).
327
328 32. A *MARK encountered during the processing of a positive assertion is now
329 recorded and passed back (compatible with Perl).
330
331 33. If --only-matching or --colour was set on a pcregrep call whose pattern
332 had alternative anchored branches, the search for a second match in a line
333 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
334 matched the line "0102" twice. The same bug affected patterns that started
335 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
336 twice.
337
338 34. Previously, PCRE did not allow quantification of assertions. However, Perl
339 does, and because of capturing effects, quantifying parenthesized
340 assertions may at times be useful. Quantifiers are now allowed for
341 parenthesized assertions.
342
343 35. A minor code tidy in pcre_compile() when checking options for \R usage.
344
345 36. \g was being checked for fancy things in a character class, when it should
346 just be a literal "g".
347
348 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
349 appearance of a nested POSIX class supersedes an apparent external class.
350 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
351 unescaped square brackets may also appear as part of class names. For
352 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
353 more like Perl. (But see 8.20/1 above.)
354
355 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
356 was because it thought it was \N{name}, which is not supported).
357
358 39. Add minix to OS list not supporting the -S option in pcretest.
359
360 40. PCRE tries to detect cases of infinite recursion at compile time, but it
361 cannot analyze patterns in sufficient detail to catch mutual recursions
362 such as ((?1))((?2)). There is now a runtime test that gives an error if a
363 subgroup is called recursively as a subpattern for a second time at the
364 same position in the subject string. In previous releases this might have
365 been caught by the recursion limit, or it might have run out of stack.
366
367 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
368 happen only once. PCRE was, however incorrectly giving a compile time error
369 "recursive call could loop indefinitely" because it cannot analyze the
370 pattern in sufficient detail. The compile time test no longer happens when
371 PCRE is compiling a conditional subpattern, but actual runaway loops are
372 now caught at runtime (see 40 above).
373
374 42. It seems that Perl allows any characters other than a closing parenthesis
375 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
376 has been changed to be the same.
377
378 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
379 as not to get warnings when autogen.sh is called. Also changed
380 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
381
382 44. To help people who use pcregrep to scan files containing exceedingly long
383 lines, the following changes have been made:
384
385 (a) The default value of the buffer size parameter has been increased from
386 8K to 20K. (The actual buffer used is three times this size.)
387
388 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
389 PCRE is built.
390
391 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
392 to be set at run time.
393
394 (d) Numerical values in pcregrep options can be followed by K or M, for
395 example --buffer-size=50K.
396
397 (e) If a line being scanned overflows pcregrep's buffer, an error is now
398 given and the return code is set to 2.
399
400 45. Add a pointer to the latest mark to the callout data block.
401
402 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
403 partial match of an empty string instead of no match. This was specific to
404 the use of ".".
405
406 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
407 complete match instead of a partial match. This bug was dependent on both
408 the PCRE_UTF8 and PCRE_DOTALL options being set.
409
410 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
411 starting byte set, because \b was not being ignored.
412
413
414 Version 8.12 15-Jan-2011
415 ------------------------
416
417 1. Fixed some typos in the markup of the man pages, and wrote a script that
418 checks for such things as part of the documentation building process.
419
420 2. On a big-endian 64-bit system, pcregrep did not correctly process the
421 --match-limit and --recursion-limit options (added for 8.11). In
422 particular, this made one of the standard tests fail. (The integer value
423 went into the wrong half of a long int.)
424
425 3. If the --colour option was given to pcregrep with -v (invert match), it
426 did strange things, either producing crazy output, or crashing. It should,
427 of course, ignore a request for colour when reporting lines that do not
428 match.
429
430 4. Another pcregrep bug caused similar problems if --colour was specified with
431 -M (multiline) and the pattern match finished with a line ending.
432
433 5. In pcregrep, when a pattern that ended with a literal newline sequence was
434 matched in multiline mode, the following line was shown as part of the
435 match. This seems wrong, so I have changed it.
436
437 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
438 the check for further matches in the same line (so they could be coloured)
439 to overrun the end of the current line. If another match was found, it was
440 incorrectly shown (and then shown again when found in the next line).
441
442 7. If pcregrep was compiled under Windows, there was a reference to the
443 function pcregrep_exit() before it was defined. I am assuming this was
444 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
445 reported by a user. I've moved the definition above the reference.
446
447
448 Version 8.11 10-Dec-2010
449 ------------------------
450
451 1. (*THEN) was not working properly if there were untried alternatives prior
452 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
453 backtracked to try for "b" instead of moving to the next alternative branch
454 at the same level (in this case, to look for "c"). The Perl documentation
455 is clear that when (*THEN) is backtracked onto, it goes to the "next
456 alternative in the innermost enclosing group".
457
458 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
459 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
460 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
461 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
462 (*THEN).
463
464 3. If \s appeared in a character class, it removed the VT character from
465 the class, even if it had been included by some previous item, for example
466 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
467 of \s, but is part of the POSIX "space" class.)
468
469 4. A partial match never returns an empty string (because you can always
470 match an empty string at the end of the subject); however the checking for
471 an empty string was starting at the "start of match" point. This has been
472 changed to the "earliest inspected character" point, because the returned
473 data for a partial match starts at this character. This means that, for
474 example, /(?<=abc)def/ gives a partial match for the subject "abc"
475 (previously it gave "no match").
476
477 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
478 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
479 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
480 has an implication that the given string is incomplete (because a partial
481 match is preferred over a full match). For this reason, these items now
482 give a partial match in this situation. [Aside: previously, the one case
483 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
484 match rather than a full match, which was wrong by the old rules, but is
485 now correct.]
486
487 6. There was a bug in the handling of #-introduced comments, recognized when
488 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
489 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
490 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
491 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
492 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
493 places in pcre_compile().
494
495 7. Related to (6) above, when pcre_compile() was skipping #-introduced
496 comments when looking ahead for named forward references to subpatterns,
497 the only newline sequence it recognized was NL. It now handles newlines
498 according to the set newline convention.
499
500 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
501 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
502 cater for a lack of strerror(). These oversights have been fixed.
503
504 9. Added --match-limit and --recursion-limit to pcregrep.
505
506 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
507
508 11. When the -o option was used, pcregrep was setting a return code of 1, even
509 when matches were found, and --line-buffered was not being honoured.
510
511 12. Added an optional parentheses number to the -o and --only-matching options
512 of pcregrep.
513
514 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
515 can match an empty string. The code to do it in pcretest and pcredemo
516 needed fixing:
517
518 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
519 only one byte after an empty string match just before CRLF (this case
520 just got forgotten; "any" and "anycrlf" were OK).
521
522 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
523 mode when an empty string match preceded an ASCII character followed by
524 a non-ASCII character. (The code for advancing by one character rather
525 than one byte was nonsense.)
526
527 (c) The pcredemo.c sample program did not have any code at all to handle
528 the cases when CRLF is a valid newline sequence.
529
530 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
531 as a starting offset was within the subject string. There is now a new
532 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
533 negative or greater than the length of the string. In order to test this,
534 pcretest is extended to allow the setting of negative starting offsets.
535
536 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
537 starting offset points to the beginning of a UTF-8 character was
538 unnecessarily clumsy. I tidied it up.
539
540 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
541 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
542
543 17. Nobody had reported that the --include_dir option, which was added in
544 release 7.7 should have been called --include-dir (hyphen, not underscore)
545 for compatibility with GNU grep. I have changed it to --include-dir, but
546 left --include_dir as an undocumented synonym, and the same for
547 --exclude-dir, though that is not available in GNU grep, at least as of
548 release 2.5.4.
549
550 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
551 characters from a string of bytes) have been redefined so as not to use
552 loops, in order to improve performance in some environments. At the same
553 time, I abstracted some of the common code into auxiliary macros to save
554 repetition (this should not affect the compiled code).
555
556 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
557 compile-time error is now given if \c is not followed by an ASCII
558 character, that is, a byte less than 128. (In EBCDIC mode, the code is
559 different, and any byte value is allowed.)
560
561 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
562 START_OPTIMIZE option, which is now allowed at compile time - but just
563 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
564 to pcregrep and other applications that have no direct access to PCRE
565 options. The new /Y option in pcretest sets this option when calling
566 pcre_compile().
567
568 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
569 back references. Groups containing recursive back references were forced to
570 be atomic by that change, but in the case of named groups, the amount of
571 memory required was incorrectly computed, leading to "Failed: internal
572 error: code overflow". This has been fixed.
573
574 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
575 pcretest.c, to avoid build problems in some Borland environments.
576
577
578 Version 8.10 25-Jun-2010
579 ------------------------
580
581 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
582 THEN.
583
584 2. (*ACCEPT) was not working when inside an atomic group.
585
586 3. Inside a character class, \B is treated as a literal by default, but
587 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
588 causes the error). The code is unchanged, but I tidied the documentation.
589
590 4. Inside a character class, PCRE always treated \R and \X as literals,
591 whereas Perl faults them if its -w option is set. I have changed PCRE so
592 that it faults them when PCRE_EXTRA is set.
593
594 5. Added support for \N, which always matches any character other than
595 newline. (It is the same as "." when PCRE_DOTALL is not set.)
596
597 6. When compiling pcregrep with newer versions of gcc which may have
598 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
599 declared with attribute warn_unused_result" were given. Just casting the
600 result to (void) does not stop the warnings; a more elaborate fudge is
601 needed. I've used a macro to implement this.
602
603 7. Minor change to pcretest.c to avoid a compiler warning.
604
605 8. Added four artifical Unicode properties to help with an option to make
606 \s etc use properties (see next item). The new properties are: Xan
607 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
608
609 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
610 use Unicode properties. (*UCP) at the start of a pattern can be used to set
611 this option. Modified pcretest to add /W to test this facility. Added
612 REG_UCP to make it available via the POSIX interface.
613
614 10. Added --line-buffered to pcregrep.
615
616 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
617 studied, and the match started with a letter with a code point greater than
618 127 whose first byte was different to the first byte of the other case of
619 the letter, the other case of this starting letter was not recognized
620 (#976).
621
622 12. If a pattern that was studied started with a repeated Unicode property
623 test, for example, \p{Nd}+, there was the theoretical possibility of
624 setting up an incorrect bitmap of starting bytes, but fortunately it could
625 not have actually happened in practice until change 8 above was made (it
626 added property types that matched character-matching opcodes).
627
628 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
629 possible starting bytes for non-anchored patterns.
630
631 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
632 \R, and also a number of cases that involve Unicode properties, both
633 explicit and implicit when PCRE_UCP is set.
634
635 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
636 input, it could crash or give wrong results if characters with values
637 greater than 0xc0 were present in the subject string. (Detail: it assumed
638 UTF-8 input when processing these items.)
639
640 16. Added a lot of (int) casts to avoid compiler warnings in systems where
641 size_t is 64-bit (#991).
642
643 17. Added a check for running out of memory when PCRE is compiled with
644 --disable-stack-for-recursion (#990).
645
646 18. If the last data line in a file for pcretest does not have a newline on
647 the end, a newline was missing in the output.
648
649 19. The default pcre_chartables.c file recognizes only ASCII characters (values
650 less than 128) in its various bitmaps. However, there is a facility for
651 generating tables according to the current locale when PCRE is compiled. It
652 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
653 characters, are recognized by isspace() and therefore were getting set in
654 these tables, and indeed these tables seem to approximate to ISO 8859. This
655 caused a problem in UTF-8 mode when pcre_study() was used to create a list
656 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
657 which of course cannot start UTF-8 characters. I have changed the code so
658 that only real ASCII characters (less than 128) and the correct starting
659 bytes for UTF-8 encodings are set for characters greater than 127 when in
660 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
661 altogether.)
662
663 20. Added the /T option to pcretest so as to be able to run tests with non-
664 standard character tables, thus making it possible to include the tests
665 used for 19 above in the standard set of tests.
666
667 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
668 reference to a subpattern the other side of a comment that contains an
669 opening parenthesis caused either an internal compiling error, or a
670 reference to the wrong subpattern.
671
672
673 Version 8.02 19-Mar-2010
674 ------------------------
675
676 1. The Unicode data tables have been updated to Unicode 5.2.0.
677
678 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
679 configured.
680
681 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
682 original author of that file, following a query about its status.
683
684 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
685 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
686
687 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
688 quantifier applied to a forward-referencing subroutine call, could compile
689 incorrect code or give the error "internal error: previously-checked
690 referenced subpattern not found".
691
692 6. Both MS Visual Studio and Symbian OS have problems with initializing
693 variables to point to external functions. For these systems, therefore,
694 pcre_malloc etc. are now initialized to local functions that call the
695 relevant global functions.
696
697 7. There were two entries missing in the vectors called coptable and poptable
698 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
699 I've fixed the data, and added a kludgy way of testing at compile time that
700 the lengths are correct (equal to the number of opcodes).
701
702 8. Following on from 7, I added a similar kludge to check the length of the
703 eint vector in pcreposix.c.
704
705 9. Error texts for pcre_compile() are held as one long string to avoid too
706 much relocation at load time. To find a text, the string is searched,
707 counting zeros. There was no check for running off the end of the string,
708 which could happen if a new error number was added without updating the
709 string.
710
711 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
712
713 11. \K was not working if it appeared in an atomic group or in a group that
714 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
715 \K is "not well defined" if used in an assertion. PCRE now accepts it if
716 the assertion is positive, but not if it is negative.
717
718 12. Change 11 fortuitously reduced the size of the stack frame used in the
719 "match()" function of pcre_exec.c by one pointer. Forthcoming
720 implementation of support for (*MARK) will need an extra pointer on the
721 stack; I have reserved it now, so that the stack frame size does not
722 decrease.
723
724 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
725 item in branch that calls a recursion is a subroutine call - as in the
726 second branch in the above example - was incorrectly given the compile-
727 time error "recursive call could loop indefinitely" because pcre_compile()
728 was not correctly checking the subroutine for matching a non-empty string.
729
730 14. The checks for overrunning compiling workspace could trigger after an
731 overrun had occurred. This is a "should never occur" error, but it can be
732 triggered by pathological patterns such as hundreds of nested parentheses.
733 The checks now trigger 100 bytes before the end of the workspace.
734
735 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
736
737
738 Version 8.01 19-Jan-2010
739 ------------------------
740
741 1. If a pattern contained a conditional subpattern with only one branch (in
742 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
743 computed the wrong minimum data length (which is of course zero for such
744 subpatterns). This could cause incorrect "no match" results.
745
746 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
747 the pattern is reset in the first branch, pcre_compile() failed with
748 "internal error: code overflow at offset...". This happened only when
749 the reset was to the original external option setting. (An optimization
750 abstracts leading options settings into an external setting, which was the
751 cause of this.)
752
753 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
754 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
755 assertion pattern did not match (meaning that the assertion was true), it
756 was incorrectly treated as false if the SKIP had been reached during the
757 matching. This also applied to assertions used as conditions.
758
759 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
760 assertion subpattern, including such a pattern used as a condition,
761 unpredictable results occurred, instead of the error return
762 PCRE_ERROR_DFA_UITEM.
763
764 5. The C++ GlobalReplace function was not working like Perl for the special
765 situation when an empty string is matched. It now does the fancy magic
766 stuff that is necessary.
767
768 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
769 removed. (These were left over from very, very early versions of PCRE.)
770
771 7. Some cosmetic changes to the code to make life easier when compiling it
772 as part of something else:
773
774 (a) Change DEBUG to PCRE_DEBUG.
775
776 (b) In pcre_compile(), rename the member of the "branch_chain" structure
777 called "current" as "current_branch", to prevent a collision with the
778 Linux macro when compiled as a kernel module.
779
780 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
781 prevent a collision with the Linux macro when compiled as a kernel
782 module.
783
784 8. In pcre_compile() there are some checks for integer overflows that used to
785 cast potentially large values to (double). This has been changed to that
786 when building, a check for int64_t is made, and if it is found, it is used
787 instead, thus avoiding the use of floating point arithmetic. (There is no
788 other use of FP in PCRE.) If int64_t is not found, the fallback is to
789 double.
790
791 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
792 2005 (difference between two addresses compared to an unsigned value).
793
794 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
795 custom one, because of the following reported problem in Windows:
796
797 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
798 under Win32.
799 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
800 therefore missing the function definition.
801 - The compiler thus generates a "C" signature for the test function.
802 - The linker fails to find the "C" function.
803 - PCRE fails to configure if asked to do so against libbz2.
804
805 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
806 messages were output:
807
808 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
809 rerunning libtoolize, to keep the correct libtool macros in-tree.
810 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
811
812 I have done both of these things.
813
814 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
815 most of the time, it *can* run out if it is given a pattern that contains a
816 runaway infinite recursion. I updated the discussion in the pcrestack man
817 page.
818
819 13. Now that we have gone to the x.xx style of version numbers, the minor
820 version may start with zero. Using 08 or 09 is a bad idea because users
821 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
822 interpreted as invalid octal numbers. I've updated the previous comment in
823 configure.ac, and also added a check that gives an error if 08 or 09 are
824 used.
825
826 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
827 causing partial matching to fail when the end of the subject matched \W
828 in a UTF-8 pattern where \W was quantified with a minimum of 3.
829
830 15. There were some discrepancies between the declarations in pcre_internal.h
831 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
832 their definitions. The declarations used "const uschar *" and the
833 definitions used USPTR. Even though USPTR is normally defined as "const
834 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
835 reported that: "This difference in casting confuses some C++ compilers, for
836 example, SunCC recognizes above declarations as different functions and
837 generates broken code for hbpcre." I have changed the declarations to use
838 USPTR.
839
840 16. GNU libtool is named differently on some systems. The autogen.sh script now
841 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
842 (FreeBSD).
843
844 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
845 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
846 comment: "Figure out how to create a longlong from a string: strtoll and
847 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
848 instance, but it only takes 2 args instead of 3!"
849
850 18. A subtle bug concerned with back references has been fixed by a change of
851 specification, with a corresponding code fix. A pattern such as
852 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
853 refers, was giving matches when it shouldn't. For example, xa=xaaa would
854 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
855 same bug. Such groups have to be quantified to be useful, or contained
856 inside another quantified group. (If there's no repetition, the reference
857 can never match.) The problem arises because, having left the group and
858 moved on to the rest of the pattern, a later failure that backtracks into
859 the group uses the captured value from the final iteration of the group
860 rather than the correct earlier one. I have fixed this in PCRE by forcing
861 any group that contains a reference to itself to be an atomic group; that
862 is, there cannot be any backtracking into it once it has completed. This is
863 similar to recursive and subroutine calls.
864
865
866 Version 8.00 19-Oct-09
867 ----------------------
868
869 1. The table for translating pcre_compile() error codes into POSIX error codes
870 was out-of-date, and there was no check on the pcre_compile() error code
871 being within the table. This could lead to an OK return being given in
872 error.
873
874 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
875 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
876 in a Windows environment.
877
878 3. The pcregrep --count option prints the count for each file even when it is
879 zero, as does GNU grep. However, pcregrep was also printing all files when
880 --files-with-matches was added. Now, when both options are given, it prints
881 counts only for those files that have at least one match. (GNU grep just
882 prints the file name in this circumstance, but including the count seems
883 more useful - otherwise, why use --count?) Also ensured that the
884 combination -clh just lists non-zero counts, with no names.
885
886 4. The long form of the pcregrep -F option was incorrectly implemented as
887 --fixed_strings instead of --fixed-strings. This is an incompatible change,
888 but it seems right to fix it, and I didn't think it was worth preserving
889 the old behaviour.
890
891 5. The command line items --regex=pattern and --regexp=pattern were not
892 recognized by pcregrep, which required --regex pattern or --regexp pattern
893 (with a space rather than an '='). The man page documented the '=' forms,
894 which are compatible with GNU grep; these now work.
895
896 6. No libpcreposix.pc file was created for pkg-config; there was just
897 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
898
899 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
900 when UCP support is not needed, by modifying the Python script that
901 generates it from Unicode data files. This should not matter if the module
902 is correctly used as a library, but I received one complaint about 50K of
903 unwanted data. My guess is that the person linked everything into his
904 program rather than using a library. Anyway, it does no harm.
905
906 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
907 was a minimum greater than 1 for a wide character in a possessive
908 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
909 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
910 character. Chaos in the form of incorrect output or a compiling loop could
911 result.
912
913 9. The restrictions on what a pattern can contain when partial matching is
914 requested for pcre_exec() have been removed. All patterns can now be
915 partially matched by this function. In addition, if there are at least two
916 slots in the offset vector, the offset of the earliest inspected character
917 for the match and the offset of the end of the subject are set in them when
918 PCRE_ERROR_PARTIAL is returned.
919
920 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
921 synonymous with PCRE_PARTIAL, for backwards compatibility, and
922 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
923 and may be more useful for multi-segment matching.
924
925 11. Partial matching with pcre_exec() is now more intuitive. A partial match
926 used to be given if ever the end of the subject was reached; now it is
927 given only if matching could not proceed because another character was
928 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
929 string "Z", which now yields "no match" instead of "partial match". In the
930 case of pcre_dfa_exec(), "no match" is given if every matching path for the
931 final character ended with (*FAIL).
932
933 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
934 if the pattern had a "must contain" character that was already found in the
935 earlier partial match, unless partial matching was again requested. For
936 example, with the pattern /dog.(body)?/, the "must contain" character is
937 "g". If the first part-match was for the string "dog", restarting with
938 "sbody" failed. This bug has been fixed.
939
940 13. The string returned by pcre_dfa_exec() after a partial match has been
941 changed so that it starts at the first inspected character rather than the
942 first character of the match. This makes a difference only if the pattern
943 starts with a lookbehind assertion or \b or \B (\K is not supported by
944 pcre_dfa_exec()). It's an incompatible change, but it makes the two
945 matching functions compatible, and I think it's the right thing to do.
946
947 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
948 so that the demonstration program is easily available in environments where
949 PCRE has not been installed from source.
950
951 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
952 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
953 library.
954
955 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
956 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
957 is not the first non-POSIX option to be added. Clearly some people find
958 these options useful.
959
960 17. If a caller to the POSIX matching function regexec() passes a non-zero
961 value for nmatch with a NULL value for pmatch, the value of
962 nmatch is forced to zero.
963
964 18. RunGrepTest did not have a test for the availability of the -u option of
965 the diff command, as RunTest does. It now checks in the same way as
966 RunTest, and also checks for the -b option.
967
968 19. If an odd number of negated classes containing just a single character
969 interposed, within parentheses, between a forward reference to a named
970 subpattern and the definition of the subpattern, compilation crashed with
971 an internal error, complaining that it could not find the referenced
972 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
973 [The bug was that it was starting one character too far in when skipping
974 over the character class, thus treating the ] as data rather than
975 terminating the class. This meant it could skip too much.]
976
977 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
978 /g option in pcretest when the pattern contains \K, which makes it possible
979 to have an empty string match not at the start, even when the pattern is
980 anchored. Updated pcretest and pcredemo to use this option.
981
982 21. If the maximum number of capturing subpatterns in a recursion was greater
983 than the maximum at the outer level, the higher number was returned, but
984 with unset values at the outer level. The correct (outer level) value is
985 now given.
986
987 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
988 PCRE did not set those parentheses (unlike Perl). I have now found a way to
989 make it do so. The string so far is captured, making this feature
990 compatible with Perl.
991
992 23. The tests have been re-organized, adding tests 11 and 12, to make it
993 possible to check the Perl 5.10 features against Perl 5.10.
994
995 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
996 pattern matches a fixed length string. PCRE did not allow this; now it
997 does. Neither allows recursion.
998
999 25. I finally figured out how to implement a request to provide the minimum
1000 length of subject string that was needed in order to match a given pattern.
1001 (It was back references and recursion that I had previously got hung up
1002 on.) This code has now been added to pcre_study(); it finds a lower bound
1003 to the length of subject needed. It is not necessarily the greatest lower
1004 bound, but using it to avoid searching strings that are too short does give
1005 some useful speed-ups. The value is available to calling programs via
1006 pcre_fullinfo().
1007
1008 26. While implementing 25, I discovered to my embarrassment that pcretest had
1009 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1010 study optimizations had never been tested with that matching function.
1011 Oops. What is worse, even when it was passed study data, there was a bug in
1012 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1013 were also very few tests of studied patterns with pcre_dfa_exec().
1014
1015 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1016 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1017 on the other side of the coin, they are no longer allowed to have different
1018 names, because these cannot be distinguished in PCRE, and this has caused
1019 confusion. (This is a difference from Perl.)
1020
1021 28. When duplicate subpattern names are present (necessarily with different
1022 numbers, as required by 27 above), and a test is made by name in a
1023 conditional pattern, either for a subpattern having been matched, or for
1024 recursion in such a pattern, all the associated numbered subpatterns are
1025 tested, and the overall condition is true if the condition is true for any
1026 one of them. This is the way Perl works, and is also more like the way
1027 testing by number works.
1028
1029
1030 Version 7.9 11-Apr-09
1031 ---------------------
1032
1033 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1034 (pcretest), all targets were linked against these libraries. This included
1035 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1036 libraries. This caused unwanted dependencies to be created. This problem
1037 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1038 pcretest is linked with readline.
1039
1040 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1041 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1042 moved outside it again, because FALSE and TRUE are already defined in AIX,
1043 but BOOL is not.
1044
1045 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1046 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1047
1048 4. The pcregrep documentation said spaces were inserted as well as colons (or
1049 hyphens) following file names and line numbers when outputting matching
1050 lines. This is not true; no spaces are inserted. I have also clarified the
1051 wording for the --colour (or --color) option.
1052
1053 5. In pcregrep, when --colour was used with -o, the list of matching strings
1054 was not coloured; this is different to GNU grep, so I have changed it to be
1055 the same.
1056
1057 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1058 each matching line was coloured. Now it goes on to look for further matches
1059 of any of the test patterns, which is the same behaviour as GNU grep.
1060
1061 7. A pattern that could match an empty string could cause pcregrep to loop; it
1062 doesn't make sense to accept an empty string match in pcregrep, so I have
1063 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1064 seems to be how GNU grep behaves.
1065
1066 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1067 start or after a newline", because the conditional assertion was not being
1068 correctly handled. The rule now is that both the assertion and what follows
1069 in the first alternative must satisfy the test.
1070
1071 9. If auto-callout was enabled in a pattern with a conditional group whose
1072 condition was an assertion, PCRE could crash during matching, both with
1073 pcre_exec() and pcre_dfa_exec().
1074
1075 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1076 used for matching.
1077
1078 11. Unicode property support in character classes was not working for
1079 characters (bytes) greater than 127 when not in UTF-8 mode.
1080
1081 12. Added the -M command line option to pcretest.
1082
1083 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1084
1085 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1086
1087 16. Added comments and documentation about mis-use of no_arg in the C++
1088 wrapper.
1089
1090 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1091 from Martin Jerabek that uses macro names for all relevant character and
1092 string constants.
1093
1094 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1095 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1096 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1097 these, but not everybody uses configure.
1098
1099 19. A conditional group that had only one branch was not being correctly
1100 recognized as an item that could match an empty string. This meant that an
1101 enclosing group might also not be so recognized, causing infinite looping
1102 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1103 with the subject "ab", where knowledge that the repeated group can match
1104 nothing is needed in order to break the loop.
1105
1106 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1107 exec(), but without supplying a callout function, matching went wrong.
1108
1109 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1110 leak if the size of the offset vector was greater than 30. When the vector
1111 is smaller, the saved offsets during recursion go onto a local stack
1112 vector, but for larger vectors malloc() is used. It was failing to free
1113 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1114 error, in fact).
1115
1116 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1117 heapframe that is used only when UTF-8 support is enabled. This caused no
1118 problem, but was untidy.
1119
1120 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1121 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1122 included within another project.
1123
1124 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1125 slightly modified by me:
1126
1127 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1128 not building pcregrep.
1129
1130 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1131 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1132
1133 25. Forward references, both numeric and by name, in patterns that made use of
1134 duplicate group numbers, could behave incorrectly or give incorrect errors,
1135 because when scanning forward to find the reference group, PCRE was not
1136 taking into account the duplicate group numbers. A pattern such as
1137 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1138
1139 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1140 the feature of a custom pointer more persuasive (as requested by a user).
1141
1142 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1143 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1144 pre-defined.
1145
1146 28. Added support for (*UTF8) at the start of a pattern.
1147
1148 29. Arrange for flags added by the "release type" setting in CMake to be shown
1149 in the configuration summary.
1150
1151
1152 Version 7.8 05-Sep-08
1153 ---------------------
1154
1155 1. Replaced UCP searching code with optimized version as implemented for Ad
1156 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1157 stage table and inline lookup instead of a function, giving speed ups of 2
1158 to 5 times on some simple patterns that I tested. Permission was given to
1159 distribute the MultiStage2.py script that generates the tables (it's not in
1160 the tarball, but is in the Subversion repository).
1161
1162 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1163 scripts.
1164
1165 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1166 a group with a zero qualifier. The result of the study could be incorrect,
1167 or the function might crash, depending on the pattern.
1168
1169 4. Caseless matching was not working for non-ASCII characters in back
1170 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1171 It now works when Unicode Property Support is available.
1172
1173 5. In pcretest, an escape such as \x{de} in the data was always generating
1174 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1175 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1176 truncation.
1177
1178 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1179
1180 7. Added two (int) casts to pcregrep when printing the difference of two
1181 pointers, in case they are 64-bit values.
1182
1183 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1184 test 2 if it fails.
1185
1186 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1187 and a #define of that name to empty if it is not externally set. This is to
1188 allow users of MSVC to set it if necessary.
1189
1190 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1191 the convenience functions in the pcre_get.c source file.
1192
1193 11. An option change at the start of a pattern that had top-level alternatives
1194 could cause overwriting and/or a crash. This command provoked a crash in
1195 some environments:
1196
1197 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1198
1199 This potential security problem was recorded as CVE-2008-2371.
1200
1201 12. For a pattern where the match had to start at the beginning or immediately
1202 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1203 pcre_dfa_exec() could read past the end of the passed subject if there was
1204 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1205 pcretest so that it places the subject at the end of its malloc-ed buffer.
1206
1207 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1208 exec() might read past the end of the data buffer in UTF-8 mode.
1209
1210 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1211 the data contained the byte 0x85 as part of a UTF-8 character within its
1212 first line. This applied both to normal and DFA matching.
1213
1214 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1215 /^[^d]*?$/8 failed to match "abc".
1216
1217 16. Added a missing copyright notice to pcrecpp_internal.h.
1218
1219 17. Make it more clear in the documentation that values returned from
1220 pcre_exec() in ovector are byte offsets, not character counts.
1221
1222 18. Tidied a few places to stop certain compilers from issuing warnings.
1223
1224 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1225 supplied by Stefan Weber. I made a further small update for 7.8 because
1226 there is a change of source arrangements: the pcre_searchfuncs.c module is
1227 replaced by pcre_ucd.c.
1228
1229
1230 Version 7.7 07-May-08
1231 ---------------------
1232
1233 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1234 a string to a long long, pretend we don't even have a long long." This is
1235 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1236
1237 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1238 pre-7.6 versions, which defined a global no_arg variable instead of putting
1239 it in the RE class. (See also #8 below.)
1240
1241 3. Remove a line of dead code, identified by coverity and reported by Nuno
1242 Lopes.
1243
1244 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1245
1246 (1) The include/exclude patterns were being applied to the whole pathnames
1247 of files, instead of just to the final components.
1248
1249 (2) If there was more than one level of directory, the subdirectories were
1250 skipped unless they satisfied the include/exclude conditions. This is
1251 inconsistent with GNU grep (and could even be seen as contrary to the
1252 pcregrep specification - which I improved to make it absolutely clear).
1253 The action now is always to scan all levels of directory, and just
1254 apply the include/exclude patterns to regular files.
1255
1256 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1257 --exclude_dir in the tests to avoid scanning .svn directories.
1258
1259 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1260 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1261 doesn't support NULs in patterns.
1262
1263 7. Added some missing "const"s to declarations of static tables in
1264 pcre_compile.c and pcre_dfa_exec.c.
1265
1266 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1267 caused by fix #2 above. (Subsequently also a second patch to fix the
1268 first patch. And a third patch - this was a messy problem.)
1269
1270 9. Applied Craig's patch to remove the use of push_back().
1271
1272 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1273 matching function regexec().
1274
1275 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1276 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1277 references. PCRE supports relative numbers with this syntax (I don't think
1278 Oniguruma does).
1279
1280 12. Previously, a group with a zero repeat such as (...){0} was completely
1281 omitted from the compiled regex. However, this means that if the group
1282 was called as a subroutine from elsewhere in the pattern, things went wrong
1283 (an internal error was given). Such groups are now left in the compiled
1284 pattern, with a new opcode that causes them to be skipped at execution
1285 time.
1286
1287 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1288 to the way PCRE behaves:
1289
1290 (a) A lone ] character is dis-allowed (Perl treats it as data).
1291
1292 (b) A back reference to an unmatched subpattern matches an empty string
1293 (Perl fails the current match path).
1294
1295 (c) A data ] in a character class must be notated as \] because if the
1296 first data character in a class is ], it defines an empty class. (In
1297 Perl it is not possible to have an empty class.) The empty class []
1298 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1299 The negative empty class [^] matches any one character, independently
1300 of the DOTALL setting.
1301
1302 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1303 non-existent subpattern following a character class starting with ']' and
1304 containing () gave an internal compiling error instead of "reference to
1305 non-existent subpattern". Fortunately, when the pattern did exist, the
1306 compiled code was correct. (When scanning forwards to check for the
1307 existencd of the subpattern, it was treating the data ']' as terminating
1308 the class, so got the count wrong. When actually compiling, the reference
1309 was subsequently set up correctly.)
1310
1311 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1312 it was being rejected as not supported by pcre_dfa_exec(), even though
1313 other assertions are supported. I have made pcre_dfa_exec() support
1314 (*FAIL).
1315
1316 16. The implementation of 13c above involved the invention of a new opcode,
1317 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1318 cannot be changed at match time, I realized I could make a small
1319 improvement to matching performance by compiling OP_ALLANY instead of
1320 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1321 on the OP_ANY path.
1322
1323 17. Compiling pcretest on Windows with readline support failed without the
1324 following two fixes: (1) Make the unistd.h include conditional on
1325 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1326
1327 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1328 ncurses library to be included for pcretest when ReadLine support is
1329 requested, but also to allow for it to be overridden. This patch came from
1330 Daniel Bergström.
1331
1332 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1333 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1334 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1335 spotting this.
1336
1337
1338 Version 7.6 28-Jan-08
1339 ---------------------
1340
1341 1. A character class containing a very large number of characters with
1342 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1343 overflow.
1344
1345 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1346 HAVE_LONG_LONG is not defined.
1347
1348 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1349 bring them up to date and include new features. This patch includes:
1350
1351 - Fixed PH's badly added libz and libbz2 support.
1352 - Fixed a problem with static linking.
1353 - Added pcredemo. [But later removed - see 7 below.]
1354 - Fixed dftables problem and added an option.
1355 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1356 HAVE_LONG_LONG.
1357 - Added readline support for pcretest.
1358 - Added an listing of the option settings after cmake has run.
1359
1360 4. A user submitted a patch to Makefile that makes it easy to create
1361 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1362 Makefile.am that cause it to include this special target, without
1363 affecting anything else. Note that the same mingw target plus all
1364 the other distribution libraries and programs are now supported
1365 when configuring with CMake (see 6 below) instead of with
1366 Configure/Make.
1367
1368 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1369 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1370 exported in the Windows port". It has not yet been confirmed that the patch
1371 solves the problem, but it does no harm.
1372
1373 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1374 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1375 with CMake, and also correct the comment about stack recursion.
1376
1377 7. Remove the automatic building of pcredemo from the ./configure system and
1378 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1379 of a program that users should build themselves after PCRE is installed, so
1380 building it automatically is not really right. What is more, it gave
1381 trouble in some build environments.
1382
1383 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1384
1385
1386 Version 7.5 10-Jan-08
1387 ---------------------
1388
1389 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1390 values in parens when parsing an RE using the C++ wrapper."
1391
1392 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1393 Characters greater than 255 were excluded from the class instead of being
1394 included.
1395
1396 3. The same bug as (2) above applied to negated POSIX classes such as
1397 [:^space:].
1398
1399 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1400 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1401 I have changed it.
1402
1403 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1404 first named subpattern) and a construct such as (?&a) would reference the
1405 first named subpattern whose name started with "a" (in other words, the
1406 length check was missing). Both these problems are fixed. "Subpattern name
1407 expected" is now given for (?&) (a zero-length name), and this patch also
1408 makes it give the same error for \k'' (previously it complained that that
1409 was a reference to a non-existent subpattern).
1410
1411 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1412 this is right because (?- can be followed by option settings as well as by
1413 digits. I have, however, made the messages clearer.
1414
1415 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1416 than the number used in the conditional) now cause a compile-time error.
1417 This is actually not compatible with Perl, which accepts such patterns, but
1418 treats the conditional as always being FALSE (as PCRE used to), but it
1419 seems to me that giving a diagnostic is better.
1420
1421 8. Change "alphameric" to the more common word "alphanumeric" in comments
1422 and messages.
1423
1424 9. Fix two occurrences of "backslash" in comments that should have been
1425 "backspace".
1426
1427 10. Remove two redundant lines of code that can never be obeyed (their function
1428 was moved elsewhere).
1429
1430 11. The program that makes PCRE's Unicode character property table had a bug
1431 which caused it to generate incorrect table entries for sequences of
1432 characters that have the same character type, but are in different scripts.
1433 It amalgamated them into a single range, with the script of the first of
1434 them. In other words, some characters were in the wrong script. There were
1435 thirteen such cases, affecting characters in the following ranges:
1436
1437 U+002b0 - U+002c1
1438 U+0060c - U+0060d
1439 U+0061e - U+00612
1440 U+0064b - U+0065e
1441 U+0074d - U+0076d
1442 U+01800 - U+01805
1443 U+01d00 - U+01d77
1444 U+01d9b - U+01dbf
1445 U+0200b - U+0200f
1446 U+030fc - U+030fe
1447 U+03260 - U+0327f
1448 U+0fb46 - U+0fbb1
1449 U+10450 - U+1049d
1450
1451 12. The -o option (show only the matching part of a line) for pcregrep was not
1452 compatible with GNU grep in that, if there was more than one match in a
1453 line, it showed only the first of them. It now behaves in the same way as
1454 GNU grep.
1455
1456 13. If the -o and -v options were combined for pcregrep, it printed a blank
1457 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1458 does the same. The return code can be used to tell if there were any
1459 non-matching lines.
1460
1461 14. Added --file-offsets and --line-offsets to pcregrep.
1462
1463 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1464 infinitely looping recursion. The bug was that positive lookaheads were not
1465 being skipped when checking for a possible empty match (negative lookaheads
1466 and both kinds of lookbehind were skipped).
1467
1468 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1469 inclusion of <windows.h> to before rather than after the definition of
1470 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1471
1472 17. Specifying a possessive quantifier with a specific limit for a Unicode
1473 character property caused pcre_compile() to compile bad code, which led at
1474 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1475 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1476 caused the error; without that there was no problem.
1477
1478 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1479
1480 19. Added --enable-pcretest-libreadline.
1481
1482 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1483 RE::GlobalReplace(). As a result, the number of replacements returned was
1484 double what it should be. I removed one of the increments, but Craig sent a
1485 later patch that removed the other one (the right fix) and added unit tests
1486 that check the return values (which was not done before).
1487
1488 21. Several CMake things:
1489
1490 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1491 the names libpcre and libpcreposix, not just pcre and pcreposix.
1492
1493 (2) The above change means that pcretest and pcregrep are now correctly
1494 linked with the newly-built libraries, not previously installed ones.
1495
1496 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1497
1498 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1499 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1500 UTF-8 newline character). The key issue is that the pattern starts .*;
1501 this means that the match must be either at the beginning, or after a
1502 newline. The bug was in the code for advancing after a failed match and
1503 checking that the new position followed a newline. It was not taking
1504 account of UTF-8 characters correctly.
1505
1506 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1507 character classes. PCRE was not treating the sequence [:...:] as a
1508 character class unless the ... were all letters. Perl, however, seems to
1509 allow any characters between [: and :], though of course it rejects as
1510 unknown any "names" that contain non-letters, because all the known class
1511 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1512 for example, whereas PCRE did not - it did not recognize a POSIX character
1513 class. This seemed a bit dangerous, so the code has been changed to be
1514 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1515 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1516 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1517 Perl does, and where it didn't before.
1518
1519 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1520 Windows environments %n is disabled by default.
1521
1522
1523 Version 7.4 21-Sep-07
1524 ---------------------
1525
1526 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1527 means that a class such as [\s] counted as "explicit reference to CR or
1528 LF". That isn't really right - the whole point of the change was to try to
1529 help when there was an actual mention of one of the two characters. So now
1530 the change happens only if \r or \n (or a literal CR or LF) character is
1531 encountered.
1532
1533 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1534 of both had grown to the point where there were only 3 bits left.
1535 Fortunately, there was spare space in the data structure, and so I have
1536 moved the internal flags into a new 16-bit field to free up more option
1537 bits.
1538
1539 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1540 but did not set the internal JCHANGED flag - either of these is enough to
1541 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1542 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1543 start sets both bits.
1544
1545 4. Added options (at build time, compile time, exec time) to change \R from
1546 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1547
1548 5. doc/pcresyntax.html was missing from the distribution.
1549
1550 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1551 compatibility, even though it is no longer used.
1552
1553 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1554 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1555 windows.h file is present (where different names are used). [This was
1556 reversed later after testing - see 16 below.]
1557
1558 8. Changed all #include <config.h> to #include "config.h". There were also
1559 some further <pcre.h> cases that I changed to "pcre.h".
1560
1561 9. When pcregrep was used with the --colour option, it missed the line ending
1562 sequence off the lines that it output.
1563
1564 10. It was pointed out to me that arrays of string pointers cause lots of
1565 relocations when a shared library is dynamically loaded. A technique of
1566 using a single long string with a table of offsets can drastically reduce
1567 these. I have refactored PCRE in four places to do this. The result is
1568 dramatic:
1569
1570 Originally: 290
1571 After changing UCP table: 187
1572 After changing error message table: 43
1573 After changing table of "verbs" 36
1574 After changing table of Posix names 22
1575
1576 Thanks to the folks working on Gregex for glib for this insight.
1577
1578 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1579 unicode-properties was also set.
1580
1581 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1582
1583 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1584 checked only for CRLF.
1585
1586 14. Added casts to pcretest.c to avoid compiler warnings.
1587
1588 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1589
1590 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1591 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1592 entirely. This removes changes made in 7 above.
1593
1594 17. The CMake files have been updated, and there is now more information about
1595 building with CMake in the NON-UNIX-USE document.
1596
1597
1598 Version 7.3 28-Aug-07
1599 ---------------------
1600
1601 1. In the rejigging of the build system that eventually resulted in 7.1, the
1602 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1603 brackets there is not right, since it causes compilers to look for an
1604 installed pcre.h, not the version that is in the source that is being
1605 compiled (which of course may be different). I have changed it back to:
1606
1607 #include "pcre.h"
1608
1609 I have a vague recollection that the change was concerned with compiling in
1610 different directories, but in the new build system, that is taken care of
1611 by the VPATH setting the Makefile.
1612
1613 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1614 when the subject happened to end in the byte 0x85 (e.g. if the last
1615 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1616 characters but of course it shouldn't be taken as a newline when it is part
1617 of another character. The bug was that, for an unlimited repeat of . in
1618 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1619 characters when looking for a newline.
1620
1621 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1622
1623 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1624 in debug output.
1625
1626 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1627 long printing in the pcrecpp unittest when running under MinGW.
1628
1629 6. ESC_K was left out of the EBCDIC table.
1630
1631 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1632 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1633 limit also applies to "virtual nesting" when a pattern is recursive, and in
1634 this case 1000 isn't so big. I have been able to remove this limit at the
1635 expense of backing off one optimization in certain circumstances. Normally,
1636 when pcre_exec() would call its internal match() function recursively and
1637 immediately return the result unconditionally, it uses a "tail recursion"
1638 feature to save stack. However, when a subpattern that can match an empty
1639 string has an unlimited repetition quantifier, it no longer makes this
1640 optimization. That gives it a stack frame in which to save the data for
1641 checking that an empty string has been matched. Previously this was taken
1642 from the 1000-entry workspace that had been reserved. So now there is no
1643 explicit limit, but more stack is used.
1644
1645 8. Applied Daniel's patches to solve problems with the import/export magic
1646 syntax that is required for Windows, and which was going wrong for the
1647 pcreposix and pcrecpp parts of the library. These were overlooked when this
1648 problem was solved for the main library.
1649
1650 9. There were some crude static tests to avoid integer overflow when computing
1651 the size of patterns that contain repeated groups with explicit upper
1652 limits. As the maximum quantifier is 65535, the maximum group length was
1653 set at 30,000 so that the product of these two numbers did not overflow a
1654 32-bit integer. However, it turns out that people want to use groups that
1655 are longer than 30,000 bytes (though not repeat them that many times).
1656 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1657 made it possible to implement the integer overflow checks in a much more
1658 dynamic way, which I have now done. The artificial limitation on group
1659 length has been removed - we now have only the limit on the total length of
1660 the compiled pattern, which depends on the LINK_SIZE setting.
1661
1662 10. Fixed a bug in the documentation for get/copy named substring when
1663 duplicate names are permitted. If none of the named substrings are set, the
1664 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1665 empty string.
1666
1667 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1668 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1669 because the ] is interpreted as the first data character and the
1670 terminating ] is not found. PCRE has been made compatible with Perl in this
1671 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1672 cause memory overwriting.
1673
1674 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1675 string has been matched (to stop an infinite loop). It was not recognizing
1676 a conditional subpattern that could match an empty string if that
1677 subpattern was within another subpattern. For example, it looped when
1678 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1679 condition was not nested. This bug has been fixed.
1680
1681 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1682 past the start of the subject in the presence of bytes with the top bit
1683 set, for example "\x8aBCD".
1684
1685 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1686 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1687
1688 14. Optimized (?!) to (*FAIL).
1689
1690 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1691 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1692 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1693 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1694 does: it's just the validity check that is more restrictive.
1695
1696 16. Inserted checks for integer overflows during escape sequence (backslash)
1697 processing, and also fixed erroneous offset values for syntax errors during
1698 backslash processing.
1699
1700 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1701 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1702
1703 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1704 caused an overrun.
1705
1706 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1707 something other than just ASCII characters) inside a group that had an
1708 unlimited repeat caused a loop at compile time (while checking to see
1709 whether the group could match an empty string).
1710
1711 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1712 [\P{Any}] did so. (Error in the code for printing property names.)
1713
1714 21. An orphan \E inside a character class could cause a crash.
1715
1716 22. A repeated capturing bracket such as (A)? could cause a wild memory
1717 reference during compilation.
1718
1719 23. There are several functions in pcre_compile() that scan along a compiled
1720 expression for various reasons (e.g. to see if it's fixed length for look
1721 behind). There were bugs in these functions when a repeated \p or \P was
1722 present in the pattern. These operators have additional parameters compared
1723 with \d, etc, and these were not being taken into account when moving along
1724 the compiled data. Specifically:
1725
1726 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1727 length.
1728
1729 (b) An item such as \pL+ within a repeated group could cause crashes or
1730 loops.
1731
1732 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1733 "reference to non-existent subpattern" error.
1734
1735 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1736
1737 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1738 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1739
1740 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1741
1742 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1743 character were causing crashes (broken optimization).
1744
1745 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1746 \p or \P) caused a compile-time loop.
1747
1748 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1749 break. For example, the unstudied pattern [\r\n]A does not match the string
1750 "\r\nA" because change 7.0/46 below moves the current point on by two
1751 characters after failing to match at the start. However, the pattern \nA
1752 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1753 the same is true. There doesn't seem any very clean way out of this, but
1754 what I have chosen to do makes the common cases work: PCRE now takes note
1755 of whether there can be an explicit match for \r or \n anywhere in the
1756 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1757 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1758 pattern has explicit CR or LF references.
1759
1760 29. Added (*CR) etc for changing newline setting at start of pattern.
1761
1762
1763 Version 7.2 19-Jun-07
1764 ---------------------
1765
1766 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1767 which is apparently normally available under Windows.
1768
1769 2. Re-jig the pcregrep tests with different newline settings in an attempt
1770 to make them independent of the local environment's newline setting.
1771
1772 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1773
1774 4. Some of the "internals" tests were previously cut out when the link size
1775 was not 2, because the output contained actual offsets. The recent new
1776 "Z" feature of pcretest means that these can be cut out, making the tests
1777 usable with all link sizes.
1778
1779 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1780 stack recursion. This gives a massive performance boost under BSD, but just
1781 a small improvement under Linux. However, it saves one field in the frame
1782 in all cases.
1783
1784 6. Added more features from the forthcoming Perl 5.10:
1785
1786 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1787 recursion call. It refers to the nth most recently opened parentheses.
1788
1789 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1790 to be opened parentheses.
1791
1792 (c) Conditions that refer to capturing parentheses can be specified
1793 relatively, for example, (?(-2)... or (?(+3)...
1794
1795 (d) \K resets the start of the current match so that everything before
1796 is not part of it.
1797
1798 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1799
1800 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1801 reference syntax.
1802
1803 (g) (?| introduces a group in which the numbering of parentheses in each
1804 alternative starts with the same number.
1805
1806 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1807
1808 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1809 PCRE_INFO_JCHANGED.
1810
1811 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1812 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1813 for detecting groups that can match an empty string.
1814
1815 9. A pattern with a very large number of alternatives (more than several
1816 hundred) was running out of internal workspace during the pre-compile
1817 phase, where pcre_compile() figures out how much memory will be needed. A
1818 bit of new cunning has reduced the workspace needed for groups with
1819 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1820 workspace instead of running out of the 4096 that are available.
1821
1822 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1823
1824 11. Applied patch from Google to remove an optimization that didn't quite work.
1825 The report of the bug said:
1826
1827 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1828 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1829 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1830
1831 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1832 it matched the wrong number of bytes.
1833
1834
1835 Version 7.1 24-Apr-07
1836 ---------------------
1837
1838 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1839 that is more "standard", making use of automake and other Autotools. There
1840 is some re-arrangement of the files and adjustment of comments consequent
1841 on this.
1842
1843 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1844 for recursive directory scanning broke on some systems because the files
1845 are not scanned in any specific order and on different systems the order
1846 was different. A call to "sort" has been inserted into RunGrepTest for the
1847 approprate test as a short-term fix. In the longer term there may be an
1848 alternative.
1849
1850 3. I had an email from Eric Raymond about problems translating some of PCRE's
1851 man pages to HTML (despite the fact that I distribute HTML pages, some
1852 people do their own conversions for various reasons). The problems
1853 concerned the use of low-level troff macros .br and .in. I have therefore
1854 removed all such uses from the man pages (some were redundant, some could
1855 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1856 HTML has been updated to handle .nf/.fi and to complain if it encounters
1857 .br or .in.
1858
1859 4. Updated comments in configure.ac that get placed in config.h.in and also
1860 arranged for config.h to be included in the distribution, with the name
1861 config.h.generic, for the benefit of those who have to compile without
1862 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1863
1864 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1865 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1866 makevp.bat for the current PCRE, using the additional files
1867 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1868
1869 6. A Windows user reported a minor discrepancy with test 2, which turned out
1870 to be caused by a trailing space on an input line that had got lost in his
1871 copy. The trailing space was an accident, so I've just removed it.
1872
1873 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1874 that is needed.
1875
1876 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1877 as "const" (a) because they are and (b) because it helps the PHP
1878 maintainers who have recently made a script to detect big data structures
1879 in the php code that should be moved to the .rodata section. I remembered
1880 to update Builducptable as well, so it won't revert if ucptable.h is ever
1881 re-created.
1882
1883 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1884 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1885 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1886 support is not required. This saves 1.5-2K of code, which is important in
1887 some applications.
1888
1889 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1890 so as not to refer to the tables, even though these functions will never be
1891 called when UTF-8 support is disabled. Otherwise there are problems with a
1892 shared library.
1893
1894 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1895
1896 (a) It was defining its arguments as char * instead of void *.
1897
1898 (b) It was assuming that all moves were upwards in memory; this was true
1899 a long time ago when I wrote it, but is no longer the case.
1900
1901 The emulated memove() is provided for those environments that have neither
1902 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1903 is clearly not the case, as these two bugs were recently reported.
1904
1905 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1906 and Detrail to create the HTML documentation, the .txt form of the man
1907 pages, and it removes trailing spaces from listed files. It also creates
1908 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1909 case, it wraps all the #defines with #ifndefs. This script should be run
1910 before "make dist".
1911
1912 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1913 with Unicode property support.
1914
1915 (a) For a maximizing quantifier, if the two different cases of the
1916 character were of different lengths in their UTF-8 codings (there are
1917 some cases like this - I found 11), and the matching function had to
1918 back up over a mixture of the two cases, it incorrectly assumed they
1919 were both the same length.
1920
1921 (b) When PCRE was configured to use the heap rather than the stack for
1922 recursion during matching, it was not correctly preserving the data for
1923 the other case of a UTF-8 character when checking ahead for a match
1924 while processing a minimizing repeat. If the check also involved
1925 matching a wide character, but failed, corruption could cause an
1926 erroneous result when trying to check for a repeat of the original
1927 character.
1928
1929 13. Some tidying changes to the testing mechanism:
1930
1931 (a) The RunTest script now detects the internal link size and whether there
1932 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1933 values substituted by "configure". (The RunGrepTest script already did
1934 this for UTF-8.) The configure.ac script no longer substitutes the
1935 relevant variables.
1936
1937 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1938 with length and offset values. This means that the output is different
1939 for different internal link sizes. Test 2 is skipped for link sizes
1940 other than 2 because of this, bypassing the problem. Unfortunately,
1941 there was also a test in test 3 (the locale tests) that used /B and
1942 failed for link sizes other than 2. Rather than cut the whole test out,
1943 I have added a new /Z option to pcretest that replaces the length and
1944 offset values with spaces. This is now used to make test 3 independent
1945 of link size. (Test 2 will be tidied up later.)
1946
1947 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1948 segmentation fault instead of returning the appropriate error message.
1949
1950 15. In multiline mode when the newline sequence was set to "any", the pattern
1951 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1952 This doesn't seem right; it now treats the CRLF combination as the line
1953 ending, and so does not match in that case. It's only a pattern such as ^$
1954 that would hit this one: something like ^ABC$ would have failed after \r
1955 and then tried again after \r\n.
1956
1957 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1958 in an attempt to make files that differ only in their line terminators
1959 compare equal. This works on Linux.
1960
1961 17. Under certain error circumstances pcregrep might try to free random memory
1962 as it exited. This is now fixed, thanks to valgrind.
1963
1964 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1965 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1966 was because its rules for how to advance for /g after matching an empty
1967 string at the end of a line did not allow for this case. They now check for
1968 it specially.
1969
1970 20. pcretest is supposed to handle patterns and data of any length, by
1971 extending its buffers when necessary. It was getting this wrong when the
1972 buffer for a data line had to be extended.
1973
1974 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1975 CRLF as a newline sequence.
1976
1977 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1978 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1979 I have nevertheless tidied it up.
1980
1981 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1982
1983 24. Added a man page for pcre-config.
1984
1985
1986 Version 7.0 19-Dec-06
1987 ---------------------
1988
1989 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1990 moving to gcc 4.1.1.
1991
1992 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1993 sys/time.h, which is documented as needed for this function. It doesn't
1994 seem to matter on Linux, but it showed up on some releases of OS X.
1995
1996 3. It seems that there are systems where bytes whose values are greater than
1997 127 match isprint() in the "C" locale. The "C" locale should be the
1998 default when a C program starts up. In most systems, only ASCII printing
1999 characters match isprint(). This difference caused the output from pcretest
2000 to vary, making some of the tests fail. I have changed pcretest so that:
2001
2002 (a) When it is outputting text in the compiled version of a pattern, bytes
2003 other than 32-126 are always shown as hex escapes.
2004
2005 (b) When it is outputting text that is a matched part of a subject string,
2006 it does the same, unless a different locale has been set for the match
2007 (using the /L modifier). In this case, it uses isprint() to decide.
2008
2009 4. Fixed a major bug that caused incorrect computation of the amount of memory
2010 required for a compiled pattern when options that changed within the
2011 pattern affected the logic of the preliminary scan that determines the
2012 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2013 that the computed length was too small. The symptoms of this bug were
2014 either the PCRE error "internal error: code overflow" from pcre_compile(),
2015 or a glibc crash with a message such as "pcretest: free(): invalid next
2016 size (fast)". Examples of patterns that provoked this bug (shown in
2017 pcretest format) are:
2018
2019 /(?-x: )/x
2020 /(?x)(?-x: \s*#\s*)/
2021 /((?i)[\x{c0}])/8
2022 /(?i:[\x{c0}])/8
2023
2024 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2025 is now done differently.
2026
2027 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2028 wrapper classes; (b) implement a new function in the C++ scanner that is
2029 more efficient than the old way of doing things because it avoids levels of
2030 recursion in the regex matching; (c) add a paragraph to the documentation
2031 for the FullMatch() function.
2032
2033 6. The escape sequence \n was being treated as whatever was defined as
2034 "newline". Not only was this contrary to the documentation, which states
2035 that \n is character 10 (hex 0A), but it also went horribly wrong when
2036 "newline" was defined as CRLF. This has been fixed.
2037
2038 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2039 was being set to -1 for the "end of line" case (supposedly a value that no
2040 character can have). Though this value is never used (the check for end of
2041 line is "zero bytes in current character"), it caused compiler complaints.
2042 I've changed it to 0xffffffff.
2043
2044 8. In pcre_version.c, the version string was being built by a sequence of
2045 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2046 string (as it is for production releases) called a macro with an empty
2047 argument. The C standard says the result of this is undefined. The gcc
2048 compiler treats it as an empty string (which was what was wanted) but it is
2049 reported that Visual C gives an error. The source has been hacked around to
2050 avoid this problem.
2051
2052 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2053 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2054 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2055 of them did).
2056
2057 10. Originally, pcretest opened its input and output without "b"; then I was
2058 told that "b" was needed in some environments, so it was added for release
2059 5.0 to both the input and output. (It makes no difference on Unix-like
2060 systems.) Later I was told that it is wrong for the input on Windows. I've
2061 now abstracted the modes into two macros, to make it easier to fiddle with
2062 them, and removed "b" from the input mode under Windows.
2063
2064 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2065
2066 12. Added -help and --help to pcretest as an official way of being reminded
2067 of the options.
2068
2069 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2070 and pcrecpp.cc because they annoy compilers at high warning levels.
2071
2072 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2073
2074 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2075 scripts are not C programs :-) and which was not noticed because it works
2076 on Linux.
2077
2078 16. pcretest is supposed to handle any length of pattern and data line (as one
2079 line or as a continued sequence of lines) by extending its input buffer if
2080 necessary. This feature was broken for very long pattern lines, leading to
2081 a string of junk being passed to pcre_compile() if the pattern was longer
2082 than about 50K.
2083
2084 17. I have done a major re-factoring of the way pcre_compile() computes the
2085 amount of memory needed for a compiled pattern. Previously, there was code
2086 that made a preliminary scan of the pattern in order to do this. That was
2087 OK when PCRE was new, but as the facilities have expanded, it has become
2088 harder and harder to keep it in step with the real compile phase, and there
2089 have been a number of bugs (see for example, 4 above). I have now found a
2090 cunning way of running the real compile function in a "fake" mode that
2091 enables it to compute how much memory it would need, while actually only
2092 ever using a few hundred bytes of working memory and without too many
2093 tests of the mode. This should make future maintenance and development
2094 easier. A side effect of this work is that the limit of 200 on the nesting
2095 depth of parentheses has been removed (though this was never a serious
2096 limitation, I suspect). However, there is a downside: pcre_compile() now
2097 runs more slowly than before (30% or more, depending on the pattern). I
2098 hope this isn't a big issue. There is no effect on runtime performance.
2099
2100 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2101 newline (only possible for the last line of a file) and it was a
2102 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2103
2104 19. Added additional timing features to pcretest. (1) The -tm option now times
2105 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2106 separate command line item, by a number that specifies the number of
2107 repeats to use when timing. The default is 50000; this gives better
2108 precision, but takes uncomfortably long for very large patterns.
2109
2110 20. Extended pcre_study() to be more clever in cases where a branch of a
2111 subpattern has no definite first character. For example, (a*|b*)[cd] would
2112 previously give no result from pcre_study(). Now it recognizes that the
2113 first character must be a, b, c, or d.
2114
2115 21. There was an incorrect error "recursive call could loop indefinitely" if
2116 a subpattern (or the entire pattern) that was being tested for matching an
2117 empty string contained only one non-empty item after a nested subpattern.
2118 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2119 incorrectly, because the \d was being skipped in the check.
2120
2121 22. The pcretest program now has a new pattern option /B and a command line
2122 option -b, which is equivalent to adding /B to every pattern. This causes
2123 it to show the compiled bytecode, without the additional information that
2124 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2125 is the same as /B/I).
2126
2127 23. A new optimization is now able automatically to treat some sequences such
2128 as a*b as a*+b. More specifically, if something simple (such as a character
2129 or a simple class like \d) has an unlimited quantifier, and is followed by
2130 something that cannot possibly match the quantified thing, the quantifier
2131 is automatically "possessified".
2132
2133 24. A recursive reference to a subpattern whose number was greater than 39
2134 went wrong under certain circumstances in UTF-8 mode. This bug could also
2135 have affected the operation of pcre_study().
2136
2137 25. Realized that a little bit of performance could be had by replacing
2138 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2139
2140 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2141
2142 27. Possessive quantifiers such as a++ were previously implemented by turning
2143 them into atomic groups such as ($>a+). Now they have their own opcodes,
2144 which improves performance. This includes the automatically created ones
2145 from 23 above.
2146
2147 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2148 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2149 the first matched character to be a colon. This applied both to named and
2150 numbered groups.
2151
2152 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2153
2154 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2155 building PCRE on Windows easier, so I have included it in the distribution.
2156
2157 31. There is now a check in pcretest against a ridiculously large number being
2158 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2159 loop, the loop is abandoned.
2160
2161 32. Forward references to subpatterns in conditions such as (?(2)...) where
2162 subpattern 2 is defined later cause pcre_compile() to search forwards in
2163 the pattern for the relevant set of parentheses. This search went wrong
2164 when there were unescaped parentheses in a character class, parentheses
2165 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2166
2167 33. "Subroutine" calls and backreferences were previously restricted to
2168 referencing subpatterns earlier in the regex. This restriction has now
2169 been removed.
2170
2171 34. Added a number of extra features that are going to be in Perl 5.10. On the
2172 whole, these are just syntactic alternatives for features that PCRE had
2173 previously implemented using the Python syntax or my own invention. The
2174 other formats are all retained for compatibility.
2175
2176 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2177 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2178 also .NET compatible.
2179
2180 (b) A recursion or subroutine call to a named group can now be defined as
2181 (?&name) as well as (?P>name).
2182
2183 (c) A backreference to a named group can now be defined as \k<name> or
2184 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2185 5.10, are also .NET compatible.
2186
2187 (d) A conditional reference to a named group can now use the syntax
2188 (?(<name>) or (?('name') as well as (?(name).
2189
2190 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2191 groups (named and numbered) that are never evaluated inline, but can be
2192 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2193 is always false. There may be only one alternative in such a group.
2194
2195 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2196 as the simple (?(R). The condition is true only if the most recent
2197 recursion is that of the given number or name. It does not search out
2198 through the entire recursion stack.
2199
2200 (g) The escape \gN or \g{N} has been added, where N is a positive or
2201 negative number, specifying an absolute or relative reference.
2202
2203 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2204 some "unreachable code" warnings.
2205
2206 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2207 things, this adds five new scripts.
2208
2209 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2210 There were also incompatibilities regarding the handling of \Q..\E inside
2211 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2212 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2213
2214 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2215 matches an empty string, and forcibly breaks the loop. There were bugs in
2216 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2217 against aaaa the result was just "a" rather than "aaaa", for example. Two
2218 separate and independent bugs (that affected different cases) have been
2219 fixed.
2220
2221 39. Refactored the code to abolish the use of different opcodes for small
2222 capturing bracket numbers. This is a tidy that I avoided doing when I
2223 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2224 The new approach is not only tidier, it makes it possible to reduce the
2225 memory needed to fix the previous bug (38).
2226
2227 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2228 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2229 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2230 mode.
2231
2232 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2233 report.
2234
2235 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2236 copy construction and assignment in the C++ wrapper.
2237
2238 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2239 couple of bugs that could have given wrong results in the "--newline=crlf"
2240 case.
2241
2242 44. Added a number of casts and did some reorganization of signed/unsigned int
2243 variables following suggestions from Dair Grant. Also renamed the variable
2244 "this" as "item" because it is a C++ keyword.
2245
2246 45. Arranged for dftables to add
2247
2248 #include "pcre_internal.h"
2249
2250 to pcre_chartables.c because without it, gcc 4.x may remove the array
2251 definition from the final binary if PCRE is built into a static library and
2252 dead code stripping is activated.
2253
2254 46. For an unanchored pattern, if a match attempt fails at the start of a
2255 newline sequence, and the newline setting is CRLF or ANY, and the next two
2256 characters are CRLF, advance by two characters instead of one.
2257
2258
2259 Version 6.7 04-Jul-06
2260 ---------------------
2261
2262 1. In order to handle tests when input lines are enormously long, pcretest has
2263 been re-factored so that it automatically extends its buffers when
2264 necessary. The code is crude, but this _is_ just a test program. The
2265 default size has been increased from 32K to 50K.
2266
2267 2. The code in pcre_study() was using the value of the re argument before
2268 testing it for NULL. (Of course, in any sensible call of the function, it
2269 won't be NULL.)
2270
2271 3. The memmove() emulation function in pcre_internal.h, which is used on
2272 systems that lack both memmove() and bcopy() - that is, hardly ever -
2273 was missing a "static" storage class specifier.
2274
2275 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2276 containing an extended class (one that cannot be represented by a bitmap
2277 because it contains high-valued characters or Unicode property items, e.g.
2278 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2279 pattern, but PCRE should not loop if you do not (it no longer does).
2280 [Detail: two cases were found: (a) a repeated subpattern containing an
2281 extended class; (b) a recursive reference to a subpattern that followed a
2282 previous extended class. It wasn't skipping over the extended class
2283 correctly when UTF-8 mode was not set.]
2284
2285 5. A negated single-character class was not being recognized as fixed-length
2286 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2287 compile error "lookbehind assertion is not fixed length".
2288
2289 6. The RunPerlTest auxiliary script was showing an unexpected difference
2290 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2291 write a Perl script that can interpret lines of an input file either as
2292 byte characters or as UTF-8, which is what "perltest" was being required to
2293 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2294 can't do is switch easily at run time between having the "use utf8;" pragma
2295 or not. In the end, I fudged it by using the RunPerlTest script to insert
2296 "use utf8;" explicitly for the UTF-8 tests.
2297
2298 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2299 the end of the subject string, contrary to the documentation and to what
2300 Perl does. This was true of both matching functions. Now it matches only at
2301 the start of the subject and immediately after *internal* newlines.
2302
2303 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2304 a pointer to an int instead of a pointer to an unsigned long int. This
2305 caused problems on 64-bit systems.
2306
2307 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2308 instance of the 'standard' template library not being so standard".
2309
2310 10. There was no check on the number of named subpatterns nor the maximum
2311 length of a subpattern name. The product of these values is used to compute
2312 the size of the memory block for a compiled pattern. By supplying a very
2313 long subpattern name and a large number of named subpatterns, the size
2314 computation could be caused to overflow. This is now prevented by limiting
2315 the length of names to 32 characters, and the number of named subpatterns
2316 to 10,000.
2317
2318 11. Subpatterns that are repeated with specific counts have to be replicated in
2319 the compiled pattern. The size of memory for this was computed from the
2320 length of the subpattern and the repeat count. The latter is limited to
2321 65535, but there was no limit on the former, meaning that integer overflow
2322 could in principle occur. The compiled length of a repeated subpattern is
2323 now limited to 30,000 bytes in order to prevent this.
2324
2325 12. Added the optional facility to have named substrings with the same name.
2326
2327 13. Added the ability to use a named substring as a condition, using the
2328 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2329 are numbers (not recommended). Forward references are permitted.
2330
2331 14. Added forward references in named backreferences (if you see what I mean).
2332
2333 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2334 pattern could run off the end of the subject. For example, the pattern
2335 "(?s)(.{1,5})"8 did this with the subject "ab".
2336
2337 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2338 PCRE_CASELESS was set when matching characters that were quantified with ?
2339 or *.
2340
2341 17. A character class other than a single negated character that had a minimum
2342 but no maximum quantifier - for example [ab]{6,} - was not handled
2343 correctly by pce_dfa_exec(). It would match only one character.
2344
2345 18. A valid (though odd) pattern that looked like a POSIX character
2346 class but used an invalid character after [ (for example [[,abc,]]) caused
2347 pcre_compile() to give the error "Failed: internal error: code overflow" or
2348 in some cases to crash with a glibc free() error. This could even happen if
2349 the pattern terminated after [[ but there just happened to be a sequence of
2350 letters, a binary zero, and a closing ] in the memory that followed.
2351
2352 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2353 over the years. Originally (before any Unicode support), just the bottom 8
2354 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2355 output from "man perlunicode" includes this:
2356
2357 The regular expression compiler produces polymorphic opcodes. That
2358 is, the pattern adapts to the data and automatically switches to
2359 the Unicode character scheme when presented with Unicode data--or
2360 instead uses a traditional byte scheme when presented with byte
2361 data.
2362
2363 Sadly, a wide octal escape does not cause a switch, and in a string with
2364 no other multibyte characters, these octal escapes are treated as before.
2365 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2366 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2367 Unicode string.
2368
2369 I have not perpetrated such confusion in PCRE. Up till now, it took just
2370 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2371 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2372 translate to the appropriate multibyte character.
2373
2374 29. Applied some refactoring to reduce the number of warnings from Microsoft
2375 and Borland compilers. This has included removing the fudge introduced
2376 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2377 a warning about an unused variable.
2378
2379 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2380 characters since release 4.0, because Perl (from release 5.004) does not.
2381 [Or at least, is documented not to: some releases seem to be in conflict
2382 with the documentation.] However, when a pattern was studied with
2383 pcre_study() and all its branches started with \s, PCRE still included VT
2384 as a possible starting character. Of course, this did no harm; it just
2385 caused an unnecessary match attempt.
2386
2387 22. Removed a now-redundant internal flag bit that recorded the fact that case
2388 dependency changed within the pattern. This was once needed for "required
2389 byte" processing, but is no longer used. This recovers a now-scarce options
2390 bit. Also moved the least significant internal flag bit to the most-
2391 significant bit of the word, which was not previously used (hangover from
2392 the days when it was an int rather than a uint) to free up another bit for
2393 the future.
2394
2395 23. Added support for CRLF line endings as well as CR and LF. As well as the
2396 default being selectable at build time, it can now be changed at runtime
2397 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2398 specify that it is scanning data with non-default line endings.
2399
2400 24. Changed the definition of CXXLINK to make it agree with the definition of
2401 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2402
2403 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2404 recursions. This makes a big different to stack usage for some patterns.
2405
2406 26. If a subpattern containing a named recursion or subroutine reference such
2407 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2408 the space required for the compiled pattern went wrong and gave too small a
2409 value. Depending on the environment, this could lead to "Failed: internal
2410 error: code overflow at offset 49" or "glibc detected double free or
2411 corruption" errors.
2412
2413 27. Applied patches from Google (a) to support the new newline modes and (b) to
2414 advance over multibyte UTF-8 characters in GlobalReplace.
2415
2416 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2417 difference for some implementation of PCRE in some Windows version.
2418
2419 29. Added some extra testing facilities to pcretest:
2420
2421 \q<number> in a data line sets the "match limit" value
2422 \Q<number> in a data line sets the "match recursion limt" value
2423 -S <number> sets the stack size, where <number> is in megabytes
2424
2425 The -S option isn't available for Windows.
2426
2427
2428 Version 6.6 06-Feb-06
2429 ---------------------
2430
2431 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2432 in pcreposix.h. I have copied the definition from pcre.h.
2433
2434 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2435 because pcre.h is no longer a built file.
2436
2437 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2438 not normally included in the compiled code.
2439
2440
2441 Version 6.5 01-Feb-06
2442 ---------------------
2443
2444 1. When using the partial match feature with pcre_dfa_exec(), it was not
2445 anchoring the second and subsequent partial matches at the new starting
2446 point. This could lead to incorrect results. For example, with the pattern
2447 /1234/, partially matching against "123" and then "a4" gave a match.
2448
2449 2. Changes to pcregrep:
2450
2451 (a) All non-match returns from pcre_exec() were being treated as failures
2452 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2453 error message is output. Some extra information is given for the
2454 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2455 probably the only errors that are likely to be caused by users (by
2456 specifying a regex that has nested indefinite repeats, for instance).
2457 If there are more than 20 of these errors, pcregrep is abandoned.
2458
2459 (b) A binary zero was treated as data while matching, but terminated the
2460 output line if it was written out. This has been fixed: binary zeroes
2461 are now no different to any other data bytes.
2462
2463 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2464 used to set a locale for matching. The --locale=xxxx long option has
2465 been added (no short equivalent) to specify a locale explicitly on the
2466 pcregrep command, overriding the environment variables.
2467
2468 (d) When -B was used with -n, some line numbers in the output were one less
2469 than they should have been.
2470
2471 (e) Added the -o (--only-matching) option.
2472
2473 (f) If -A or -C was used with -c (count only), some lines of context were
2474 accidentally printed for the final match.
2475
2476 (g) Added the -H (--with-filename) option.
2477
2478 (h) The combination of options -rh failed to suppress file names for files
2479 that were found from directory arguments.
2480
2481 (i) Added the -D (--devices) and -d (--directories) options.
2482
2483 (j) Added the -F (--fixed-strings) option.
2484
2485 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2486
2487 (l) Added the --colo(u)r option.
2488
2489 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2490 is not present by default.
2491
2492 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2493 items such as (?R) or (?1), when the recursion could match a number of
2494 alternatives. If it matched one of the alternatives, but subsequently,
2495 outside the recursion, there was a failure, the code tried to back up into
2496 the recursion. However, because of the way PCRE is implemented, this is not
2497 possible, and the result was an incorrect result from the match.
2498
2499 In order to prevent this happening, the specification of recursion has
2500 been changed so that all such subpatterns are automatically treated as
2501 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2502
2503 4. I had overlooked the fact that, in some locales, there are characters for
2504 which isalpha() is true but neither isupper() nor islower() are true. In
2505 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2506 and ordfeminine) are like this. This affected the treatment of \w and \W
2507 when they appeared in character classes, but not when they appeared outside
2508 a character class. The bit map for "word" characters is now created
2509 separately from the results of isalnum() instead of just taking it from the
2510 upper, lower, and digit maps. (Plus the underscore character, of course.)
2511
2512 5. The above bug also affected the handling of POSIX character classes such as
2513 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2514 permanent tables. Instead, the bit maps for such a class were previously
2515 created as the appropriate unions of the upper, lower, and digit bitmaps.
2516 Now they are created by subtraction from the [[:word:]] class, which has
2517 its own bitmap.
2518
2519 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2520 It is created by subtracting the vertical space characters (\x09, \x0a,
2521 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2522 subtraction was done in the overall bitmap for a character class, meaning
2523 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2524 be recognized. This bug has been fixed.
2525
2526 7. Patches from the folks at Google:
2527
2528 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2529 real life, but is still worth protecting against".
2530
2531 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2532 regular expressions".
2533
2534 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2535 have it.
2536
2537 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2538 "configure" and the latter not, in order to fix a problem somebody had
2539 with compiling the Arg class on HP-UX.
2540
2541 (e) Improve the error-handling of the C++ wrapper a little bit.
2542
2543 (f) New tests for checking recursion limiting.
2544
2545 8. The pcre_memmove() function, which is used only if the environment does not
2546 have a standard memmove() function (and is therefore rarely compiled),
2547 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2548 returning a result (though PCRE never actually uses the result).
2549
2550 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2551 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2552 returned instead of calling malloc() with an overflowing number that would
2553 most likely cause subsequent chaos.
2554
2555 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2556
2557 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2558 with this option is matched, the nmatch and pmatch options of regexec() are
2559 ignored.
2560
2561 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2562 provided in case anyone wants to the the POSIX interface with UTF-8
2563 strings.
2564
2565 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2566 C++ linking (needed for some HP-UX environments).
2567
2568 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2569 (unused parameter) and in the pcre_printint() function (omitted "default"
2570 switch label when the default is to do nothing).
2571
2572 15. Added some code to make it possible, when PCRE is compiled as a C++
2573 library, to replace subject pointers for pcre_exec() with a smart pointer
2574 class, thus making it possible to process discontinuous strings.
2575
2576 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2577 much the same function. They were added by different people who were trying
2578 to make PCRE easy to compile on non-Unix systems. It has been suggested
2579 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2580 for compiling on Windows systems. I have therefore replaced it with
2581 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2582 defaults to "extern" for C or "extern C" for C++, which works fine on
2583 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2584 SCOPE with something explicit in config.h. In addition:
2585
2586 (a) pcreposix.h still had just "extern" instead of either of these macros;
2587 I have replaced it with PCRE_DATA_SCOPE.
2588
2589 (b) Functions such as _pcre_xclass(), which are internal to the library,
2590 but external in the C sense, all had PCRE_EXPORT in their definitions.
2591 This is apparently wrong for the Windows case, so I have removed it.
2592 (It makes no difference on Unix-like systems.)
2593
2594 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2595 of recursive calls to match(). This is different to MATCH_LIMIT because
2596 that limits the total number of calls to match(), not all of which increase
2597 the depth of recursion. Limiting the recursion depth limits the amount of
2598 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2599 when PCRE is compiled, and changed at run time. A patch from Google adds
2600 this functionality to the C++ interface.
2601
2602 18. Changes to the handling of Unicode character properties:
2603
2604 (a) Updated the table to Unicode 4.1.0.
2605
2606 (b) Recognize characters that are not in the table as "Cn" (undefined).
2607
2608 (c) I revised the way the table is implemented to a much improved format
2609 which includes recognition of ranges. It now supports the ranges that
2610 are defined in UnicodeData.txt, and it also amalgamates other
2611 characters into ranges. This has reduced the number of entries in the
2612 table from around 16,000 to around 3,000, thus reducing its size
2613 considerably. I realized I did not need to use a tree structure after
2614 all - a binary chop search is just as efficient. Having reduced the
2615 number of entries, I extended their size from 6 bytes to 8 bytes to
2616 allow for more data.
2617
2618 (d) Added support for Unicode script names via properties such as \p{Han}.
2619
2620 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2621 matching that character.
2622
2623 20. When matching a repeated Unicode property with a minimum greater than zero,
2624 (for example \pL{2,}), PCRE could look past the end of the subject if it
2625 reached it while seeking the minimum number of characters. This could
2626 happen only if some of the characters were more than one byte long, because
2627 there is a check for at least the minimum number of bytes.
2628
2629 21. Refactored the implementation of \p and \P so as to be more general, to
2630 allow for more different types of property in future. This has changed the
2631 compiled form incompatibly. Anybody with saved compiled patterns that use
2632 \p or \P will have to recompile them.
2633
2634 22. Added "Any" and "L&" to the supported property types.
2635
2636 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2637 but give a compile time error if the value is greater than 0xff.
2638
2639 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2640 accidentally not being installed or uninstalled.
2641
2642 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2643 made were to insert the current release number. This seemed silly, because
2644 it made things harder for people building PCRE on systems that don't run
2645 "configure". I have turned pcre.h into a distributed file, no longer built
2646 by "configure", with the version identification directly included. There is
2647 no longer a pcre.h.in file.
2648
2649 However, this change necessitated a change to the pcre-config script as
2650 well. It is built from pcre-config.in, and one of the substitutions was the
2651 release number. I have updated configure.ac so that ./configure now finds
2652 the release number by grepping pcre.h.
2653
2654 26. Added the ability to run the tests under valgrind.
2655
2656
2657 Version 6.4 05-Sep-05
2658 ---------------------
2659
2660 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2661 "--" to be printed when multiple files were scanned, even when none of the
2662 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2663 consider it to be a bug, and have restored the previous behaviour.
2664
2665 2. A couple of code tidies to get rid of compiler warnings.
2666
2667 3. The pcretest program used to cheat by referring to symbols in the library
2668 whose names begin with _pcre_. These are internal symbols that are not
2669 really supposed to be visible externally, and in some environments it is
2670 possible to suppress them. The cheating is now confined to including
2671 certain files from the library's source, which is a bit cleaner.
2672
2673 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2674 file's purpose clearer.
2675
2676 5. Reorganized pcre_ucp_findchar().
2677
2678
2679 Version 6.3 15-Aug-05
2680 ---------------------
2681
2682 1. The file libpcre.pc.in did not have general read permission in the tarball.
2683
2684 2. There were some problems when building without C++ support:
2685
2686 (a) If C++ support was not built, "make install" and "make test" still
2687 tried to test it.
2688
2689 (b) There were problems when the value of CXX was explicitly set. Some
2690 changes have been made to try to fix these, and ...
2691
2692 (c) --disable-cpp can now be used to explicitly disable C++ support.
2693
2694 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2695 backslash in a target when C++ was disabled. This confuses some
2696 versions of "make", apparently. Using an intermediate variable solves
2697 this. (Same for CPP_LOBJ.)
2698
2699 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2700 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2701 necessary on certain architectures.
2702
2703 4. Added a setting of -export-symbols-regex to the link command to remove
2704 those symbols that are exported in the C sense, but actually are local
2705 within the library, and not documented. Their names all begin with
2706 "_pcre_". This is not a perfect job, because (a) we have to except some
2707 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2708 available (and never for static libraries). I have made a note to try to
2709 find a way round (a) in the future.
2710
2711
2712 Version 6.2 01-Aug-05
2713 ---------------------
2714
2715 1. There was no test for integer overflow of quantifier values. A construction
2716 such as {1111111111111111} would give undefined results. What is worse, if
2717 a minimum quantifier for a parenthesized subpattern overflowed and became
2718 negative, the calculation of the memory size went wrong. This could have
2719 led to memory overwriting.
2720
2721 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2722
2723 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2724 operating environments where this matters.
2725
2726 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2727 PCRE options from within the C++ wrapper.
2728
2729 5. Named capturing subpatterns were not being correctly counted when a pattern
2730 was compiled. This caused two problems: (a) If there were more than 100
2731 such subpatterns, the calculation of the memory needed for the whole
2732 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2733 back references of the form \12, where the number was greater than 9, were
2734 not recognized as back references, even though there were sufficient
2735 previous subpatterns.
2736
2737 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2738 versions of gcc, e.g. 2.95.4.
2739
2740
2741 Version 6.1 21-Jun-05
2742 ---------------------
2743
2744 1. There was one reference to the variable "posix" in pcretest.c that was not
2745 surrounded by "#if !defined NOPOSIX".
2746
2747 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2748 the cross-check on the old pcre_info() function, for the benefit of the
2749 cut-down version of PCRE that is currently imported into Exim.
2750
2751 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2752 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2753 patterns that start (?i) but I don't think that matters. The use of (?i) is
2754 just an example; this all applies to the other options as well.
2755
2756 4. Since libtool seems to echo the compile commands it is issuing, the output
2757 from "make" can be reduced a bit by putting "@" in front of each libtool
2758 compile command.
2759
2760 5. Patch from the folks at Google for configure.in to be a bit more thorough
2761 in checking for a suitable C++ installation before trying to compile the
2762 C++ stuff. This should fix a reported problem when a compiler was present,
2763 but no suitable headers.
2764
2765 6. The man pages all had just "PCRE" as their title. I have changed them to
2766 be the relevant file name. I have also arranged that these names are
2767 retained in the file doc/pcre.txt, which is a concatenation in text format
2768 of all the man pages except the little individual ones for each function.
2769
2770 7. The NON-UNIX-USE file had not been updated for the different set of source
2771 files that come with release 6. I also added a few comments about the C++
2772 wrapper.
2773
2774
2775 Version 6.0 07-Jun-05
2776 ---------------------
2777
2778 1. Some minor internal re-organization to help with my DFA experiments.
2779
2780 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2781 didn't matter for the library itself when fully configured, but did matter
2782 when compiling without UCP support, or within Exim, where the ucp files are
2783 not imported.
2784
2785 3. Refactoring of the library code to split up the various functions into
2786 different source modules. The addition of the new DFA matching code (see
2787 below) to a single monolithic source would have made it really too
2788 unwieldy, quite apart from causing all the code to be include in a
2789 statically linked application, when only some functions are used. This is
2790 relevant even without the DFA addition now that patterns can be compiled in
2791 one application and matched in another.
2792
2793 The downside of splitting up is that there have to be some external
2794 functions and data tables that are used internally in different modules of
2795 the library but which are not part of the API. These have all had their
2796 names changed to start with "_pcre_" so that they are unlikely to clash
2797 with other external names.
2798
2799 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2800 a different (DFA) algorithm. Although it is slower than the original
2801 function, it does have some advantages for certain types of matching
2802 problem.
2803
2804 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2805 including restarting after a partial match.
2806
2807 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2808 defined when compiling for Windows was sent to me. I have put it into the
2809 code, though I have no means of testing or verifying it.
2810
2811 7. Added the pcre_refcount() auxiliary function.
2812
2813 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2814 match before or at the first newline in the subject string. In pcretest,
2815 the /f option on a pattern can be used to set this.
2816
2817 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2818 would behave wrongly. This has been present in PCRE since release 4.0.
2819
2820 10. A number of changes to the pcregrep command:
2821
2822 (a) Refactored how -x works; insert ^(...)$ instead of setting
2823 PCRE_ANCHORED and checking the length, in preparation for adding
2824 something similar for -w.
2825
2826 (b) Added the -w (match as a word) option.
2827
2828 (c) Refactored the way lines are read and buffered so as to have more
2829 than one at a time available.
2830
2831 (d) Implemented a pcregrep test script.
2832
2833 (e) Added the -M (multiline match) option. This allows patterns to match
2834 over several lines of the subject. The buffering ensures that at least
2835 8K, or the rest of the document (whichever is the shorter) is available
2836 for matching (and similarly the previous 8K for lookbehind assertions).
2837
2838 (f) Changed the --help output so that it now says
2839
2840 -w, --word-regex(p)
2841
2842 instead of two lines, one with "regex" and the other with "regexp"
2843 because that confused at least one person since the short forms are the
2844 same. (This required a bit of code, as the output is generated
2845 automatically from a table. It wasn't just a text change.)
2846
2847 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2848 option but starts with a hyphen. Could be a pattern or a path name
2849 starting with a hyphen, for instance.
2850
2851 (h) "-" can be given as a file name to represent stdin.
2852
2853 (i) When file names are being printed, "(standard input)" is used for
2854 the standard input, for compatibility with GNU grep. Previously
2855 "<stdin>" was used.
2856
2857 (j) The option --label=xxx can be used to supply a name to be used for
2858 stdin when file names are being printed. There is no short form.
2859
2860 (k) Re-factored the options decoding logic because we are going to add
2861 two more options that take data. Such options can now be given in four
2862 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2863
2864 (l) Added the -A, -B, and -C options for requesting that lines of context
2865 around matches be printed.
2866
2867 (m) Added the -L option to print the names of files that do not contain
2868 any matching lines, that is, the complement of -l.
2869
2870 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2871 continue to scan other files.
2872
2873 (o) The -s option was incorrectly implemented. For compatibility with other
2874 greps, it now suppresses the error message for a non-existent or non-
2875 accessible file (but not the return code). There is a new option called
2876 -q that suppresses the output of matching lines, which was what -s was
2877 previously doing.
2878
2879 (p) Added --include and --exclude options to specify files for inclusion
2880 and exclusion when recursing.
2881
2882 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2883 Hopefully, it now does.
2884
2885 12. Missing cast in pcre_study().
2886
2887 13. Added an "uninstall" target to the makefile.
2888
2889 14. Replaced "extern" in the function prototypes in Makefile.in with
2890 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2891 world, but is set differently for Windows.
2892
2893 15. Added a second compiling function called pcre_compile2(). The only
2894 difference is that it has an extra argument, which is a pointer to an
2895 integer error code. When there is a compile-time failure, this is set
2896 non-zero, in addition to the error test pointer being set to point to an
2897 error message. The new argument may be NULL if no error number is required
2898 (but then you may as well call pcre_compile(), which is now just a
2899 wrapper). This facility is provided because some applications need a
2900 numeric error indication, but it has also enabled me to tidy up the way
2901 compile-time errors are handled in the POSIX wrapper.
2902
2903 16. Added VPATH=.libs to the makefile; this should help when building with one
2904 prefix path and installing with another. (Or so I'm told by someone who
2905 knows more about this stuff than I do.)
2906
2907 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2908 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2909 match everything, including newlines. This is not POSIX-compatible, but
2910 somebody wanted the feature. From pcretest it can be activated by using
2911 both the P and the s flags.
2912
2913 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2914
2915 19. libpcre.pc was being incorrectly installed as executable.
2916
2917 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2918 it now also looks for '\r' so that it will work unmodified on Windows.
2919
2920 21. Added Google's contributed C++ wrapper to the distribution.
2921
2922 22. Added some untidy missing memory free() calls in pcretest, to keep
2923 Electric Fence happy when testing.
2924
2925
2926
2927 Version 5.0 13-Sep-04
2928 ---------------------
2929
2930 1. Internal change: literal characters are no longer packed up into items
2931 containing multiple characters in a single byte-string. Each character
2932 is now matched using a separate opcode. However, there may be more than one
2933 byte in the character in UTF-8 mode.
2934
2935 2. The pcre_callout_block structure has two new fields: pattern_position and
2936 next_item_length. These contain the offset in the pattern to the next match
2937 item, and its length, respectively.
2938
2939 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2940 insertion of callouts before each pattern item. Added the /C option to
2941 pcretest to make use of this.
2942
2943 4. On the advice of a Windows user, the lines
2944
2945 #if defined(_WIN32) || defined(WIN32)
2946 _setmode( _fileno( stdout ), 0x8000 );
2947 #endif /* defined(_WIN32) || defined(WIN32) */
2948
2949 have been added to the source of pcretest. This apparently does useful
2950 magic in relation to line terminators.
2951
2952 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2953 for the benefit of those environments where the "b" makes a difference.
2954
2955 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2956 to know about it. I have put a hack into configure.in that adds in code
2957 to set GCC=yes if CC=icc. This seems to end up at a point in the
2958 generated configure script that is early enough to affect the setting of
2959 compiler options, which is what is needed, but I have no means of testing
2960 whether it really works. (The user who reported this had patched the
2961 generated configure script, which of course I cannot do.)
2962
2963 LATER: After change 22 below (new libtool files), the configure script
2964 seems to know about icc (and also ecc). Therefore, I have commented out
2965 this hack in configure.in.
2966
2967 7. Added support for pkg-config (2 patches were sent in).
2968
2969 8. Negated POSIX character classes that used a combination of internal tables
2970 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2971 [[:^ascii]]. Typically, they would match almost any characters. The other
2972 POSIX classes were not broken in this way.
2973
2974 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2975 to find the match, as PCRE was deluded into thinking that the match had to
2976 start at the start point or following a newline. The same bug applied to
2977 patterns with negative forward assertions or any backward assertions
2978 preceding ".*" at the start, unless the pattern required a fixed first
2979 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2980
2981 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2982 starting at the last subject character, bytes beyond the end of the subject
2983 string were read.
2984
2985 11. Renamed the variable "class" as "classbits" to make life easier for C++
2986 users. (Previously there was a macro definition, but it apparently wasn't
2987 enough.)
2988
2989 12. Added the new field "tables" to the extra data so that tables can be passed
2990 in at exec time, or the internal tables can be re-selected. This allows
2991 a compiled regex to be saved and re-used at a later time by a different
2992 program that might have everything at different addresses.
2993
2994 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2995 -R library as well as a -L library.
2996
2997 14. The debugging options of pcretest (-d on the command line or D on a
2998 pattern) showed incorrect output for anything following an extended class
2999 that contained multibyte characters and which was followed by a quantifier.
3000
3001 15. Added optional support for general category Unicode character properties
3002 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3003 support. It adds about 90K to the size of the library. The meanings of the
3004 inbuilt class escapes such as \d and \s have NOT been changed.
3005
3006 16. Updated pcredemo.c to include calls to free() to release the memory for the
3007 compiled pattern.
3008
3009 17. The generated file chartables.c was being created in the source directory
3010 instead of in the building directory. This caused the build to fail if the
3011 source directory was different from the building directory, and was
3012 read-only.
3013
3014 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3015 file. No doubt somebody will tell me if they don't make sense... Also added
3016 Dan Mooney's comments about building on OpenVMS.
3017
3018 19. Added support for partial matching via the PCRE_PARTIAL option for
3019 pcre_exec() and the \P data escape in pcretest.
3020
3021 20. Extended pcretest with 3 new pattern features:
3022
3023 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3024 write the compiled pattern to the file whose name is "rest-of-line".
3025 This is a straight binary dump of the data, with the saved pointer to
3026 the character tables forced to be NULL. The study data, if any, is
3027 written too. After writing, pcretest reads a new pattern.
3028
3029 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3030 compiled pattern from the given file. There must not be any
3031 occurrences of "<" in the file name (pretty unlikely); if there are,
3032 pcretest will instead treat the initial "<" as a pattern delimiter.
3033 After reading in the pattern, pcretest goes on to read data lines as
3034 usual.
3035
3036 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3037 and 16-bit fields in a compiled pattern, to simulate a pattern that
3038 was compiled on a host of opposite endianness.
3039
3040 21. The pcre-exec() function can now cope with patterns that were compiled on
3041 hosts of opposite endianness, with this restriction:
3042
3043 As for any compiled expression that is saved and used later, the tables
3044 pointer field cannot be preserved; the extra_data field in the arguments
3045 to pcre_exec() should be used to pass in a tables address if a value
3046 other than the default internal tables were used at compile time.
3047
3048 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3049 now diagnosed as an error. Previously, most of the time, a negative number
3050 would have been treated as zero, but if in addition "ovector" was passed as
3051 NULL, a crash could occur.
3052
3053 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3054 new versions from the libtool 1.5 distribution (the last one is a copy of
3055 a file called libtool.m4). This seems to have fixed the need to patch
3056 "configure" to support Darwin 1.3 (which I used to do). However, I still
3057 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3058 workstation).
3059
3060 24. Changed the PCRE licence to be the more standard "BSD" licence.
3061
3062
3063 Version 4.5 01-Dec-03
3064 ---------------------
3065
3066 1. There has been some re-arrangement of the code for the match() function so
3067 that it can be compiled in a version that does not call itself recursively.
3068 Instead, it keeps those local variables that need separate instances for
3069 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3070 needs to "recurse". Keeping track of where control must go is done by means
3071 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3072 hide most of the details from the main code, and operates only if
3073 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3074 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3075 operating.
3076
3077 To make it easier for callers to provide specially tailored get/free
3078 functions for this usage, two new functions, pcre_stack_malloc, and
3079 pcre_stack_free, are used. They are always called in strict stacking order,
3080 and the size of block requested is always the same.
3081
3082 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3083 PCRE has been compiled to use the stack or the heap for recursion. The
3084 -C option of pcretest uses this to show which version is compiled.
3085
3086 A new data escape \S, is added to pcretest; it causes the amounts of store
3087 obtained and freed by both kinds of malloc/free at match time to be added
3088 to the output.
3089
3090 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3091 what's available on my current Linux desktop machine.
3092
3093 3. When matching a UTF-8 string, the test for a valid string at the start has
3094 been extended. If start_offset is not zero, PCRE now checks that it points
3095 to a byte that is the start of a UTF-8 character. If not, it returns
3096 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3097 this is necessary because there may be backward assertions in the pattern.
3098 When matching the same subject several times, it may save resources to use
3099 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3100
3101 4. The code for checking the validity of UTF-8 strings has been tightened so
3102 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3103 containing "overlong sequences".
3104
3105 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3106 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3107 should have been "&", but it just so happened that all the cases this let
3108 through by mistake were picked up later in the function.
3109
3110 6. I had used a variable called "isblank" - this is a C99 function, causing
3111 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3112
3113 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3114 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3115 so the version is shown; (c) stop "make test" echoing "./RunTest".
3116
3117 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3118
3119 9. The prototype for memmove() for systems that don't have it was using
3120 size_t, but the inclusion of the header that defines size_t was later. I've
3121 moved the #includes for the C headers earlier to avoid this.
3122
3123 10. Added some adjustments to the code to make it easier to compiler on certain
3124 special systems:
3125
3126 (a) Some "const" qualifiers were missing.
3127 (b) Added the macro EXPORT before all exported functions; by default this
3128 is defined to be empty.
3129 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3130 that it reads its output file name as an argument instead of writing
3131 to the standard output and assuming this can be redirected.
3132
3133 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3134 class containing characters with values greater than 255, PCRE compilation
3135 went into a loop.
3136
3137 12. A recursive reference to a subpattern that was within another subpattern
3138 that had a minimum quantifier of zero caused PCRE to crash. For example,
3139 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3140 recursion. If the recursively-called subpattern itself had a zero repeat,
3141 that was OK.
3142
3143 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3144 buffer into which it was copied (for escape processing) was still set at
3145 1024, so long lines caused crashes.
3146
3147 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3148 "internal error: code overflow...". This applied to any character class
3149 that was followed by a possessive quantifier.
3150
3151 15. Modified the Makefile to add libpcre.la as a prerequisite for
3152 libpcreposix.la because I was told this is needed for a parallel build to
3153 work.
3154
3155 16. If a pattern that contained .* following optional items at the start was
3156 studied, the wrong optimizing data was generated, leading to matching
3157 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3158 matching string must start with a or b or c. The correct conclusion for
3159 this pattern is that a match can start with any character.
3160
3161
3162 Version 4.4 13-Aug-03
3163 ---------------------
3164
3165 1. In UTF-8 mode, a character class containing characters with values between
3166 127 and 255 was not handled correctly if the compiled pattern was studied.
3167 In fixing this, I have also improved the studying algorithm for such
3168 classes (slightly).
3169
3170 2. Three internal functions had redundant arguments passed to them. Removal
3171 might give a very teeny performance improvement.
3172
3173 3. Documentation bug: the value of the capture_top field in a callout is *one
3174 more than* the number of the hightest numbered captured substring.
3175
3176 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3177 in incorrectly linking with a previously installed version. They now link
3178 explicitly with libpcre.la.
3179
3180 5. configure.in no longer needs to recognize Cygwin specially.
3181
3182 6. A problem in pcre.in for Windows platforms is fixed.
3183
3184 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3185 pcretest, it used to include the size of the study block as part of its
3186 output. Unfortunately, the structure contains a field that has a different
3187 size on different hardware architectures. This meant that the tests that
3188 showed this size failed. As the block is currently always of a fixed size,
3189 this information isn't actually particularly useful in pcretest output, so
3190 I have just removed it.
3191
3192 8. Three pre-processor statements accidentally did not start in column 1.
3193 Sadly, there are *still* compilers around that complain, even though
3194 standard C has not required this for well over a decade. Sigh.
3195
3196 9. In pcretest, the code for checking callouts passed small integers in the
3197 callout_data field, which is a void * field. However, some picky compilers
3198 complained about the casts involved for this on 64-bit systems. Now
3199 pcretest passes the address of the small integer instead, which should get
3200 rid of the warnings.
3201
3202 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3203 both compile and run time, and gives an error if an invalid UTF-8 sequence
3204 is found. There is a option for disabling this check in cases where the
3205 string is known to be correct and/or the maximum performance is wanted.
3206
3207 11. In response to a bug report, I changed one line in Makefile.in from
3208
3209 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3210 to
3211 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3212
3213 to look similar to other lines, but I have no way of telling whether this
3214 is the right thing to do, as I do not use Windows. No doubt I'll get told
3215 if it's wrong...
3216
3217
3218 Version 4.3 21-May-03
3219 ---------------------
3220
3221 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3222 Makefile.
3223
3224 2. Some refactoring to improve the quality of the code:
3225
3226 (i) The utf8_table... variables are now declared "const".
3227
3228 (ii) The code for \cx, which used the "case flipping" table to upper case
3229 lower case letters, now just substracts 32. This is ASCII-specific,
3230 but the whole concept of \cx is ASCII-specific, so it seems
3231 reasonable.
3232
3233 (iii) PCRE was using its character types table to recognize decimal and
3234 hexadecimal digits in the pattern. This is silly, because it handles
3235 only 0-9, a-f, and A-F, but the character types table is locale-
3236 specific, which means strange things might happen. A private
3237 table is now used for this - though it costs 256 bytes, a table is
3238 much faster than multiple explicit tests. Of course, the standard
3239 character types table is still used for matching digits in subject
3240 strings against \d.
3241
3242 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3243 ending in _t are). So I've renamed it as ESC_tee.
3244
3245 3. The first argument for regexec() in the POSIX wrapper should have been
3246 defined as "const".
3247
3248 4. Changed pcretest to use malloc() for its buffers so that they can be
3249 Electric Fenced for debugging.
3250
3251 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3252 to read one or more bytes before the start of the subject string. Often this
3253 had no effect on PCRE's behaviour, but in some circumstances it could
3254 provoke a segmentation fault.
3255
3256 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3257 to try to read one or more bytes before the start of the subject string.
3258
3259 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3260 UTF-8 support could misbehave in various ways if the subject string
3261 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3262 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3263 back over UTF-8 characters.)
3264
3265
3266 Version 4.2 14-Apr-03
3267 ---------------------
3268
3269 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3270
3271 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3272 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3273 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3274 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3275 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3276 and BUILD_EXEEXT
3277 Note: automatic setting of the BUILD variables is not yet working
3278 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3279 compile-time but not at link-time
3280 [LINK]: use for linking executables only
3281 make different versions for Windows and non-Windows
3282 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3283 libraries
3284 [LINK_FOR_BUILD]: new variable
3285 [OBJEXT]: use throughout
3286 [EXEEXT]: use throughout
3287 <winshared>: new target
3288 <wininstall>: new target
3289 <dftables.o>: use native compiler
3290 <dftables>: use native linker
3291 <install>: handle Windows platform correctly
3292 <clean>: ditto
3293 <check>: ditto
3294 copy DLL to top builddir before testing
3295
3296 As part of these changes, -no-undefined was removed again. This was reported
3297 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3298 in any case.
3299
3300 3. Some tidies to get rid of compiler warnings:
3301
3302 . In the match_data structure, match_limit was an unsigned long int, whereas
3303 match_call_count was an int. I've made them both unsigned long ints.
3304
3305 . In pcretest the fact that a const uschar * doesn't automatically cast to
3306 a void * provoked a warning.
3307
3308 . Turning on some more compiler warnings threw up some "shadow" variables
3309 and a few more missing casts.
3310
3311 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3312 option, a class that contained a single character with a value between 128
3313 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3314
3315 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3316 option, a class that contained several characters, but with at least one
3317 whose value was between 128 and 255 caused PCRE to crash.
3318
3319
3320 Version 4.1 12-Mar-03
3321 ---------------------
3322
3323 1. Compiling with gcc -pedantic found a couple of places where casts were
3324 needed, and a string in dftables.c that was longer than standard compilers are
3325 required to support.
3326
3327 2. Compiling with Sun's compiler found a few more places where the code could
3328 be tidied up in order to avoid warnings.
3329
3330 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3331 first of these names is deprecated in the latest Autoconf in favour of the name
3332 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3333 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3334 analogy I have changed it to CFLAGS_FOR_BUILD.
3335
3336 4. Added -no-undefined to the linking command in the Makefile, because this is
3337 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3338 linking step for the pcreposix library.
3339
3340 5. PCRE was failing to diagnose the case of two named groups with the same
3341 name.
3342
3343 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3344 literal character that is needed in the subject for a match, and scans along to
3345 ensure that it is present before embarking on the full matching process. This
3346 saves time in cases of nested unlimited repeats that are never going to match.
3347 Problem: the scan can take a lot of time if the subject is very long (e.g.
3348 megabytes), thus penalizing straightforward matches. It is now done only if the
3349 amount of subject to be scanned is less than 1000 bytes.
3350
3351 7. A lesser problem with the same optimization is that it was recording the
3352 first character of an anchored pattern as "needed", thus provoking a search
3353 right along the subject, even when the first match of the pattern was going to
3354 fail. The "needed" character is now not set for anchored patterns, unless it
3355 follows something in the pattern that is of non-fixed length. Thus, it still
3356 fulfils its original purpose of finding quick non-matches in cases of nested
3357 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3358
3359
3360 Version 4.0 17-Feb-03
3361 ---------------------
3362
3363 1. If a comment in an extended regex that started immediately after a meta-item
3364 extended to the end of string, PCRE compiled incorrect data. This could lead to
3365 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3366
3367 2. Moved to autoconf 2.53 and libtool 1.4.2.
3368
3369 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3370 the special perltest8 script is no longer needed - all the tests can be run
3371 from a single perltest script.
3372
3373 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3374 by \s. It has now been removed in PCRE. This means it isn't recognized as
3375 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3376 class [:space:] *does* include VT, thereby creating a mess.
3377
3378 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3379 space and tab.
3380
3381 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3382 its new features into the main test script, reducing the number of scripts.
3383
3384 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3385 were backward compatible, and made the (?i) apply to the whole pattern, as if
3386 /i were given. Now it behaves more logically, and applies the option setting
3387 only to what follows. PCRE has been changed to follow suit. However, if it
3388 finds options settings right at the start of the pattern, it extracts them into
3389 the global options, as before. Thus, they show up in the info data.
3390
3391 8. Added support for the \Q...\E escape sequence. Characters in between are
3392 treated as literals. This is slightly different from Perl in that $ and @ are
3393 also handled as literals inside the quotes. In Perl, they will cause variable
3394 interpolation. Note the following examples:
3395
3396 Pattern PCRE matches Perl matches
3397
3398 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3399 \Qabc\$xyz\E abc\$xyz abc\$xyz
3400 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3401
3402 For compatibility with Perl, \Q...\E sequences are recognized inside character
3403 classes as well as outside them.
3404
3405 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3406 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3407 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3408 signed/unsigned warnings.
3409
3410 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3411 option for pcretest, so I've replaced it by a simple function that does just
3412 that job.
3413
3414 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3415 "pcregrep -".
3416
3417 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3418 Java package. This provides some syntactic sugar for simple cases of what my
3419 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3420 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3421 item, you can use this simplified notation. Note that only makes sense with
3422 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3423 greediness, whatever the setting of the PCRE_UNGREEDY option.
3424
3425 13. A change of greediness default within a pattern was not taking effect at
3426 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3427 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3428 was abstracted outside.
3429
3430 14. PCRE now supports the \G assertion. It is true when the current matching
3431 position is at the start point of the match. This differs from \A when the
3432 starting offset is non-zero. Used with the /g option of pcretest (or similar
3433 code), it works in the same way as it does for Perl's /g option. If all
3434 alternatives of a regex begin with \G, the expression is anchored to the start
3435 match position, and the "anchored" flag is set in the compiled expression.
3436
3437 15. Some bugs concerning the handling of certain option changes within patterns
3438 have been fixed. These applied to options other than (?ims). For example,
3439 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3440 been the other way round. Some of this was related to change 7 above.
3441
3442 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3443 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3444 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3445 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3446
3447 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3448 mode. Unlike ".", it always matches newline, whatever the setting of
3449 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3450 assertions. Perl allows it, but it doesn't (in general) work because it can't
3451 calculate the length of the lookbehind. At least, that's the case for Perl
3452 5.8.0 - I've been told they are going to document that it doesn't work in
3453 future.
3454
3455 18. Added an error diagnosis for escapes that PCRE does not support: these are
3456 \L, \l, \N, \P, \p, \U, \u, and \X.
3457
3458 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3459 reading past the end of the pattern in cases such as /[abcd/.
3460
3461 20. PCRE was getting more memory than necessary for patterns with classes that
3462 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3463
3464 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3465 compiling PCRE for use with Virtual Pascal.
3466
3467 22. Small fix to the Makefile to make it work properly if the build is done
3468 outside the source tree.
3469
3470 23. Added a new extension: a condition to go with recursion. If a conditional
3471 subpattern starts with (?(R) the "true" branch is used if recursion has
3472 happened, whereas the "false" branch is used only at the top level.
3473
3474 24. When there was a very long string of literal characters (over 255 bytes
3475 without UTF support, over 250 bytes with UTF support), the computation of how
3476 much memory was required could be incorrect, leading to segfaults or other
3477 strange effects.
3478
3479 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3480 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3481 there was a subsequent back reference to those brackets. This meant that, for
3482 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3483 possible to check for precisely this case. All we can do is abandon the
3484 optimization if .* occurs inside capturing brackets when there are any back
3485 references whatsoever. (See below for a better fix that came later.)
3486
3487 26. The handling of the optimization for finding the first character of a
3488 non-anchored pattern, and for finding a character that is required later in the
3489 match were failing in some cases. This didn't break the matching; it just
3490 failed to optimize when it could. The way this is done has been re-implemented.
3491
3492 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3493
3494 28. Added a new feature that provides some of the functionality that Perl
3495 provides with (?{...}). The facility is termed a "callout". The way it is done
3496 in PCRE is for the caller to provide an optional function, by setting
3497 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3498 global variable. By default it is unset, which disables all calling out. To get
3499 the function called, the regex must include (?C) at appropriate points. This
3500 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3501 This provides a means of identifying different callout points. When PCRE
3502 reaches such a point in the regex, if pcre_callout has been set, the external
3503 function is called. It is provided with data in a structure called
3504 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3505 matching continues; if it returns a non-zero value, the match at the current
3506 point fails. However, backtracking will occur if possible. [This was changed
3507 later and other features added - see item 49 below.]
3508
3509 29. pcretest is upgraded to test the callout functionality. It provides a
3510 callout function that displays information. By default, it shows the start of
3511 the match and the current position in the text. There are some new data escapes
3512 to vary what happens:
3513
3514 \C+ in addition, show current contents of captured substrings
3515 \C- do not supply a callout function
3516 \C!n return 1 when callout number n is reached
3517 \C!n!m return 1 when callout number n is reached for the mth time
3518
3519 30. If pcregrep was called with the -l option and just a single file name, it
3520 output "<stdin>" if a match was found, instead of the file name.
3521
3522 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3523 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3524 pcre_exec(). This saves a malloc/free per call. The default value of
3525 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3526 when configuring.
3527
3528 32. The default maximum size of a compiled pattern is 64K. There have been a
3529 few cases of people hitting this limit. The code now uses macros to handle the
3530 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3531 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3532 configuring. Tests 2 and 5 work only with 2-byte links because they output
3533 debugging information about compiled patterns.
3534
3535 33. Internal code re-arrangements:
3536
3537 (a) Moved the debugging function for printing out a compiled regex into
3538 its own source file (printint.c) and used #include to pull it into
3539 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3540 separate copies.
3541
3542 (b) Defined the list of op-code names for debugging as a macro in
3543 internal.h so that it is next to the definition of the opcodes.
3544
3545 (c) Defined a table of op-code lengths for simpler skipping along compiled
3546 code. This is again a macro in internal.h so that it is next to the
3547 definition of the opcodes.
3548
3549 34. Added support for recursive calls to individual subpatterns, along the
3550 lines of Robin Houston's patch (but implemented somewhat differently).
3551
3552 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3553 allow it to read and process whole directories in Win32. This code was
3554 contributed by Lionel Fourquaux; it has not been tested by me.
3555
3556 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3557 used to name a group. Names consist of alphanumerics and underscores, and must
3558 be unique. Back references use the syntax (?P=name) and recursive calls use
3559 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3560 numbers. The function pcre_fullinfo() can be used after compilation to extract
3561 a name/number map. There are three relevant calls:
3562
3563 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3564 PCRE_INFO_NAMECOUNT yields the number of entries
3565 PCRE_INFO_NAMETABLE yields a pointer to the map.
3566
3567 The map is a vector of fixed-size entries. The size of each entry depends on
3568 the length of the longest name used. The first two bytes of each entry are the
3569 group number, most significant byte first. There follows the corresponding
3570 name, zero terminated. The names are in alphabetical order.
3571
3572 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3573 case instead of 255. Making it the same both with and without UTF-8 support
3574 means that the same test output works with both.
3575
3576 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3577 calling malloc() with a zero argument.
3578
3579 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3580 optimization. I've improved things by keeping a bitmap of backreferences with
3581 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3582 fact referenced, the optimization can be applied. It is unlikely that a
3583 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3584 the match to follow \n) will appear inside brackets with a number greater than
3585 31, but if it does, any back reference > 31 suppresses the optimization.
3586
3587 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3588 of disabling numbered capturing parentheses. Any opening parenthesis that is
3589 not followed by ? behaves as if it were followed by ?: but named parentheses
3590 can still be used for capturing (and they will acquire numbers in the usual
3591 way).
3592
3593 41. Redesigned the return codes from the match() function into yes/no/error so
3594 that errors can be passed back from deep inside the nested calls. A malloc
3595 failure while inside a recursive subpattern call now causes the
3596 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3597
3598 42. It is now possible to set a limit on the number of times the match()
3599 function is called in a call to pcre_exec(). This facility makes it possible to
3600 limit the amount of recursion and backtracking, though not in a directly
3601 obvious way, because the match() function is used in a number of different
3602 circumstances. The count starts from zero for each position in the subject
3603 string (for non-anchored patterns). The default limit is, for compatibility, a
3604 large number, namely 10 000 000. You can change this in two ways:
3605
3606 (a) When configuring PCRE before making, you can use --with-match-limit=n
3607 to set a default value for the compiled library.
3608
3609 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3610 a different value is set. See 45 below.
3611
3612 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3613
3614 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3615 of things that can be changed at compile time. The first argument specifies
3616 what is wanted and the second points to where the information is to be placed.
3617 The current list of available information is:
3618
3619 PCRE_CONFIG_UTF8
3620
3621 The output is an integer that is set to one if UTF-8 support is available;
3622 otherwise it is set to zero.
3623
3624 PCRE_CONFIG_NEWLINE
3625
3626 The output is an integer that it set to the value of the code that is used for
3627 newline. It is either LF (10) or CR (13).
3628
3629 PCRE_CONFIG_LINK_SIZE
3630
3631 The output is an integer that contains the number of bytes used for internal
3632 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3633
3634 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3635
3636 The output is an integer that contains the threshold above which the POSIX
3637 interface uses malloc() for output vectors. See item 31 above.
3638
3639 PCRE_CONFIG_MATCH_LIMIT
3640
3641 The output is an unsigned integer that contains the default limit of the number
3642 of match() calls in a pcre_exec() execution. See 42 above.
3643
3644 44. pcretest has been upgraded by the addition of the -C option. This causes it
3645 to extract all the available output from the new pcre_config() function, and to
3646 output it. The program then exits immediately.
3647
3648 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3649 order to support additional features. One way would have been to define
3650 pcre_exec2() (for example) with extra arguments, but this would not have been
3651 extensible, and would also have required all calls to the original function to
3652 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3653 is used for passing in "extra" data from pcre_study().
3654
3655 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3656 contains the following fields:
3657
3658 flags a bitmap indicating which of the following fields are set
3659 study_data opaque data from pcre_study()
3660 match_limit a way of specifying a limit on match() calls for a specific
3661 call to pcre_exec()
3662 callout_data data for callouts (see 49 below)
3663
3664 The flag bits are also defined in pcre.h, and are
3665
3666 PCRE_EXTRA_STUDY_DATA
3667 PCRE_EXTRA_MATCH_LIMIT
3668 PCRE_EXTRA_CALLOUT_DATA
3669
3670 The pcre_study() function now returns one of these new pcre_extra blocks, with
3671 the actual study data pointed to by the study_data field, and the
3672 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3673 before. That is, this change is entirely upwards-compatible and requires no
3674 change to existing code.
3675
3676 If you want to pass in additional data to pcre_exec(), you can either place it
3677 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3678 block.
3679
3680 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3681 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3682 times with different match limits, until it finds the minimum value needed for
3683 pcre_exec() to complete. The value is then output. This can be instructive; for
3684 most simple matches the number is quite small, but for pathological cases it
3685 gets very large very quickly.
3686
3687 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3688 returns the size of the data block pointed to by the study_data field in a
3689 pcre_extra block, that is, the value that was passed as the argument to
3690 pcre_malloc() when PCRE was getting memory in which to place the information
3691 created by pcre_study(). The fourth argument should point to a size_t variable.
3692 pcretest has been extended so that this information is shown after a successful
3693 pcre_study() call when information about the compiled regex is being displayed.
3694
3695 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3696 because what follows is always an absolute path. (Later: it turns out that this
3697 is more than cosmetic for MinGW, because it doesn't like empty path
3698 components.)
3699
3700 49. Some changes have been made to the callout feature (see 28 above):
3701
3702 (i) A callout function now has three choices for what it returns:
3703
3704 0 => success, carry on matching
3705 > 0 => failure at this point, but backtrack if possible
3706 < 0 => serious error, return this value from pcre_exec()
3707
3708 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3709 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3710 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3711 use by callout functions. It will never be used by PCRE itself.
3712
3713 (ii) The pcre_extra structure (see 45 above) has a void * field called
3714 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3715 pcre_callout_block structure has a field of the same name. The contents of
3716 the field passed in the pcre_extra structure are passed to the callout
3717 function in the corresponding field in the callout block. This makes it
3718 easier to use the same callout-containing regex from multiple threads. For
3719 testing, the pcretest program has a new data escape
3720
3721 \C*n pass the number n (may be negative) as callout_data
3722
3723 If the callout function in pcretest receives a non-zero value as
3724 callout_data, it returns that value.
3725
3726 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3727 there were some redundant $(CFLAGS) in commands that are now specified as
3728 $(LINK), which already includes $(CFLAGS).
3729
3730 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3731 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3732 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3733 one-byte characters throughout. Note that case-insensitive matching applies
3734 only to characters whose values are less than 256. PCRE doesn't support the
3735 notion of cases for higher-valued characters.
3736
3737 (i) A character class whose characters are all within 0-255 is handled as
3738 a bit map, and the map is inverted for negative classes. Previously, a
3739 character > 255 always failed to match such a class; however it should
3740 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3741
3742 (ii) A negated character class with a single character < 255 is coded as
3743 "not this character" (OP_NOT). This wasn't working properly when the test
3744 character was multibyte, either singly or repeated.
3745
3746 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3747 mode, for example: \x{100}{2,3}.
3748
3749 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3750 singly or repeated) now correctly test multibyte characters. However,
3751 PCRE doesn't recognize any characters with values greater than 255 as
3752 digits, spaces, or word characters. Such characters always match \D, \S,
3753 and \W, and never match \d, \s, or \w.
3754
3755 (v) Classes may now contain characters and character ranges with values
3756 greater than 255. For example: [ab\x{100}-\x{400}].
3757
3758 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3759 PCRE in UTF-8 mode.
3760
3761 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3762 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3763 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3764 value.)
3765
3766 53. The single man page has become too large. I have therefore split it up into
3767 a number of separate man pages. These also give rise to individual HTML pages;
3768 these are now put in a separate directory, and there is an index.html page that
3769 lists them all. Some hyperlinking between the pages has been installed.
3770
3771 54. Added convenience functions for handling named capturing parentheses.
3772
3773 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3774 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3775 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3776 are faulted.
3777
3778 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3779 calling configure. These values are used when compiling the dftables.c program
3780 which is run to generate the source of the default character tables. They
3781 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3782 you will need to set these values.
3783
3784 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3785
3786
3787 Version 3.9 02-Jan-02
3788 ---------------------
3789
3790 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3791
3792 2. If --disable-static was given, the building process failed when trying to
3793 build pcretest and pcregrep. (For some reason it was using libtool to compile
3794 them, which is not right, as they aren't part of the library.)
3795
3796
3797 Version 3.8 18-Dec-01
3798 ---------------------
3799
3800 1. The experimental UTF-8 code was completely screwed up. It was packing the
3801 bytes in the wrong order. How dumb can you get?
3802
3803
3804 Version 3.7 29-Oct-01
3805 ---------------------
3806
3807 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3808 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3809 this didn't happen under Solaris 8, where I normally test things.
3810
3811 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3812 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3813 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3814 this fix an hour or so after the initial 3.7 release.)
3815
3816
3817 Version 3.6 23-Oct-01
3818 ---------------------
3819
3820 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3821 offsets passed as NULL with zero offset count.
3822
3823 2. The config.guess and config.sub files had not been updated when I moved to
3824 the latest autoconf.
3825
3826
3827 Version 3.5 15-Aug-01
3828 ---------------------
3829
3830 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3831 had been forgotten.
3832
3833 2. By using declared but undefined structures, we can avoid using "void"
3834 definitions in pcre.h while keeping the internal definitions of the structures
3835 private.
3836
3837 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3838 user point of view, this means that both static and shared libraries are built
3839 by default, but this can be individually controlled. More of the work of
3840 handling this static/shared cases is now inside libtool instead of PCRE's make
3841 file.
3842
3843 4. The pcretest utility is now installed along with pcregrep because it is
3844 useful for users (to test regexs) and by doing this, it automatically gets
3845 relinked by libtool. The documentation has been turned into a man page, so
3846 there are now .1, .txt, and .html versions in /doc.
3847
3848 5. Upgrades to pcregrep:
3849 (i) Added long-form option names like gnu grep.
3850 (ii) Added --help to list all options with an explanatory phrase.
3851 (iii) Added -r, --recursive to recurse into sub-directories.
3852 (iv) Added -f, --file to read patterns from a file.
3853
3854 6. pcre_exec() was referring to its "code" argument before testing that
3855 argument for NULL (and giving an error if it was NULL).
3856
3857 7. Upgraded Makefile.in to allow for compiling in a different directory from
3858 the source directory.
3859
3860 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3861 options bits, the pointer it was passed was to an int instead of to an unsigned
3862 long int. This mattered only on 64-bit systems.
3863
3864 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3865 generated) instead of pcre.in, which it its source. Also made the same change
3866 in several of the .c files.
3867
3868 10. A new release of gcc defines printf() as a macro, which broke pcretest
3869 because it had an ifdef in the middle of a string argument for printf(). Fixed
3870 by using separate calls to printf().
3871
3872 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3873 script, to force use of CR or LF instead of \n in the source. On non-Unix
3874 systems, the value can be set in config.h.
3875
3876 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3877 absolute limit. Changed the text of the error message to make this clear, and
3878 likewise updated the man page.
3879
3880 13. The limit of 99 on the number of capturing subpatterns has been removed.
3881 The new limit is 65535, which I hope will not be a "real" limit.
3882
3883
3884 Version 3.4 22-Aug-00
3885 ---------------------
3886
3887 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3888
3889 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3890
3891
3892 Version 3.3 01-Aug-00
3893 ---------------------
3894
3895 1. If an octal character was given, but the value was greater than \377, it
3896 was not getting masked to the least significant bits, as documented. This could
3897 lead to crashes in some systems.
3898
3899 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3900 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3901
3902 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3903 These just pass their arguments on to (pcre_free)(), but they are provided
3904 because some uses of PCRE bind it to non-C systems that can call its functions,
3905 but cannot call free() or pcre_free() directly.
3906
3907 4. Add "make test" as a synonym for "make check". Corrected some comments in
3908 the Makefile.
3909
3910 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3911 Makefile.
3912
3913 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3914 command called pgrep for grepping around the active processes.
3915
3916 7. Added the beginnings of support for UTF-8 character strings.
3917
3918 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3919 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3920 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3921 out for the ar command.)
3922
3923
3924 Version 3.2 12-May-00
3925 ---------------------
3926
3927 This is purely a bug fixing release.
3928
3929 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3930 of ZA. This was just one example of several cases that could provoke this bug,
3931 which was introduced by change 9 of version 2.00. The code for breaking
3932 infinite loops after an iteration that matches an empty string was't working
3933 correctly.
3934
3935 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3936 when matched against abbab (for example). After matching an empty string, it
3937 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3938 caused it to match further down the string than it should.
3939
3940 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3941 was there because it doesn't seem to be needed, and it causes trouble on some
3942 systems, as it is not a Standard C header. It has been removed.
3943
3944 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3945 were reported on the Macintosh. The changes were from
3946
3947 while ((c = *(++ptr)) != 0 && c != '\n');
3948 to
3949 while ((c = *(++ptr)) != 0 && c != '\n') ;
3950
3951 Totally extraordinary, but if that's what it takes...
3952
3953 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3954 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3955 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3956 assumes the way PCRE uses memmove() (always moving upwards).
3957
3958 6. PCRE is being used in one environment where strchr() is not available. There
3959 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3960 faster code anyway.
3961
3962
3963 Version 3.1 09-Feb-00
3964 ---------------------
3965
3966 The only change in this release is the fixing of some bugs in Makefile.in for
3967 the "install" target:
3968
3969 (1) It was failing to install pcreposix.h.
3970
3971 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3972
3973
3974 Version 3.0 01-Feb-00
3975 ---------------------
3976
3977 1. Add support for the /+ modifier to perltest (to output $` like it does in
3978 pcretest).
3979
3980 2. Add support for the /g modifier to perltest.
3981
3982 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3983 matches null strings.
3984
3985 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3986 pattern. Perl treats empty patterns specially - it reuses the most recent
3987 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3988 effect.
3989
3990 5. The POSIX interface was broken in that it was just handing over the POSIX
3991 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3992 required a bigger vector, with some working space on the end. This means that
3993 the POSIX wrapper now has to get and free some memory, and copy the results.
3994
3995 6. Added some simple autoconf support, placing the test data and the
3996 documentation in separate directories, re-organizing some of the
3997 information files, and making it build pcre-config (a GNU standard). Also added
3998 libtool support for building PCRE as a shared library, which is now the
3999 default.
4000
4001 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4002 09 are not valid octal constants. Single digits will be used for minor values
4003 less than 10.
4004
4005 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4006 existing programs that set these in the POSIX interface can use PCRE without
4007 modification.
4008
4009 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4010 return all that pcre_info() returns, plus additional data. The pcre_info()
4011 function is retained for compatibility, but is considered to be obsolete.
4012
4013 10. Added experimental recursion feature (?R) to handle one common case that
4014 Perl 5.6 will be able to do with (?p{...}).
4015
4016 11. Added support for POSIX character classes like [:alpha:], which Perl is
4017 adopting.
4018
4019
4020 Version 2.08 31-Aug-99
4021 ----------------------
4022
4023 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4024 trying to match at the startoffset position, but instead was moving forward to
4025 the next newline as if a previous match had failed.
4026
4027 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4028 and could get into a loop if a null string was matched other than at the start
4029 of the subject.
4030
4031 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4032 be distinguished at compile time, and for completeness also added PCRE_DATE.
4033
4034 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4035 in GnuWin32 environments.
4036
4037
4038 Version 2.07 29-Jul-99
4039 ----------------------
4040
4041 1. The documentation is now supplied in plain text form and HTML as well as in
4042 the form of man page sources.
4043
4044 2. C++ compilers don't like assigning (void *) values to other pointer types.
4045 In particular this affects malloc(). Although there is no problem in Standard
4046 C, I've put in casts to keep C++ compilers happy.
4047
4048 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4049 should be (const char *).
4050
4051 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4052 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4053 However, I haven't made this a standard facility. The documentation doesn't
4054 mention it, and the Makefile doesn't support it.
4055
4056 5. The Makefile now contains an "install" target, with editable destinations at
4057 the top of the file. The pcretest program is not installed.
4058
4059 6. pgrep -V now gives the PCRE version number and date.
4060
4061 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4062 causing the entire string to be ignored, instead of just the last character.
4063
4064 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4065 non-matching string, it can take a very, very long time, even for strings of
4066 quite modest length, because of the nested recursion. PCRE now does better in
4067 some of these cases. It does this by remembering the last required literal
4068 character in the pattern, and pre-searching the subject to ensure it is present
4069 before running the real match. In other words, it applies a heuristic to detect
4070 some types of certain failure quickly, and in the above example, if presented
4071 with a string that has no trailing " it gives "no match" very quickly.
4072
4073 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4074 other alternatives are tried instead.
4075
4076
4077 Version 2.06 09-Jun-99
4078 ----------------------
4079
4080 1. Change pcretest's output for amount of store used to show just the code
4081 space, because the remainder (the data block) varies in size between 32-bit and
4082 64-bit systems.
4083
4084 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4085 start matching at. This allows lookbehinds to work when searching for multiple
4086 occurrences in a string.
4087
4088 3. Added additional options to pcretest for testing multiple occurrences:
4089
4090 /+ outputs the rest of the string that follows a match
4091 /g loops for multiple occurrences, using the new startoffset argument
4092 /G loops for multiple occurrences by passing an incremented pointer
4093
4094 4. PCRE wasn't doing the "first character" optimization for patterns starting
4095 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4096 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4097 the letter 'x'. On long subject strings, this gives a significant speed-up.
4098
4099
4100 Version 2.05 21-Apr-99
4101 ----------------------
4102
4103 1. Changed the type of magic_number from int to long int so that it works
4104 properly on 16-bit systems.
4105
4106 2. Fixed a bug which caused patterns starting with .* not to work correctly
4107 when the subject string contained newline characters. PCRE was assuming
4108 anchoring for such patterns in all cases, which is not correct because .* will
4109 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4110 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4111 must be retried after every newline in the subject.
4112
4113
4114 Version 2.04 18-Feb-99
4115 ----------------------
4116
4117 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4118 computation of the store needed to hold the pattern was incorrect (too large).
4119 If such patterns were nested a few deep, this could multiply and become a real
4120 problem.
4121
4122 2. Added /M option to pcretest to show the memory requirement of a specific
4123 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4124
4125 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4126 compiled in such a way that the backtracking after subsequent failure was
4127 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4128 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4129
4130
4131 Version 2.03 02-Feb-99
4132 ----------------------
4133
4134 1. Fixed typo and small mistake in man page.
4135
4136 2. Added 4th condition (GPL supersedes if conflict) and created separate
4137 LICENCE file containing the conditions.
4138
4139 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4140 Perl, that is the internal \ allows the delimiter to be included in the
4141 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4142 the final delimiter, add \ to the end of the pattern (to test the error).
4143
4144 4. Added the convenience functions for extracting substrings after a successful
4145 match. Updated pcretest to make it able to test these functions.
4146
4147
4148 Version 2.02 14-Jan-99
4149 ----------------------
4150
4151 1. Initialized the working variables associated with each extraction so that
4152 their saving and restoring doesn't refer to uninitialized store.
4153
4154 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4155 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4156 fix the problem.
4157
4158 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4159 calls, and wasn't printing the correct value for compiling calls. Increased the
4160 default value of LOOPREPEAT, and the number of significant figures in the
4161 times.
4162
4163 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4164
4165 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4166 a building problem on Windows NT with a FAT file system.
4167
4168
4169 Version 2.01 21-Oct-98
4170 ----------------------
4171
4172 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4173 to character tables built by pcre_maketables() in the current locale. If NULL
4174 is passed, the default tables are used.
4175
4176
4177 Version 2.00 24-Sep-98
4178 ----------------------
4179
4180 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4181 it any more.
4182
4183 2. Allow quantification of (?>) groups, and make it work correctly.
4184
4185 3. The first character computation wasn't working for (?>) groups.
4186
4187 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4188 end of the subject) and add 5.005's \z, which really does match only at the
4189 very end of the subject.
4190
4191 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4192
4193 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4194 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4195 localized options. All options to pcre_study() were also removed.
4196
4197 7. Add other new features from 5.005:
4198
4199 $(?<= positive lookbehind
4200 $(?<! negative lookbehind
4201 (?imsx-imsx) added the unsetting capability
4202 such a setting is global if at outer level; local otherwise
4203 (?imsx-imsx:) non-capturing groups with option setting
4204 (?(cond)re|re) conditional pattern matching
4205
4206 A backreference to itself in a repeated group matches the previous
4207 captured string.
4208
4209 8. General tidying up of studying (both automatic and via "study")
4210 consequential on the addition of new assertions.
4211
4212 9. As in 5.005, unlimited repeated groups that could match an empty substring
4213 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4214 runtime if any iteration does actually match an empty substring.
4215
4216 10. Include the RunTest script in the distribution.
4217
4218 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4219 discrepancies, some of which were old and were also with respect to 5.004. They
4220 have now been fixed.
4221
4222
4223 Version 1.09 28-Apr-98
4224 ----------------------
4225
4226 1. A negated single character class followed by a quantifier with a minimum
4227 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4228 program crashes, or just wrong answers. This did not apply to negated classes
4229 containing more than one character, or to minima other than one.
4230
4231
4232 Version 1.08 27-Mar-98
4233 ----------------------
4234
4235 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4236
4237 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4238 latter must appear before anything that relies on it in the pattern.
4239
4240
4241 Version 1.07 16-Feb-98
4242 ----------------------
4243
4244 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4245 repeat of a potentially empty string).
4246
4247
4248 Version 1.06 23-Jan-98
4249 ----------------------
4250
4251 1. Added Markus Oberhumer's little patches for C++.
4252
4253 2. Literal strings longer than 255 characters were broken.
4254
4255
4256 Version 1.05 23-Dec-97
4257 ----------------------
4258
4259 1. Negated character classes containing more than one character were failing if
4260 PCRE_CASELESS was set at run time.
4261
4262
4263 Version 1.04 19-Dec-97
4264 ----------------------
4265
4266 1. Corrected the man page, where some "const" qualifiers had been omitted.
4267
4268 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4269 input syntax.
4270
4271 3. Fixed memory leak which occurred when a regex with back references was
4272 matched with an offsets vector that wasn't big enough. The temporary memory
4273 that is used in this case wasn't being freed if the match failed.
4274
4275 4. Tidied pcretest to ensure it frees memory that it gets.
4276
4277 5. Temporary memory was being obtained in the case where the passed offsets
4278 vector was exactly big enough.
4279
4280 6. Corrected definition of offsetof() from change 5 below.
4281
4282 7. I had screwed up change 6 below and broken the rules for the use of
4283 setjmp(). Now fixed.
4284
4285
4286 Version 1.03 18-Dec-97
4287 ----------------------
4288
4289 1. A erroneous regex with a missing opening parenthesis was correctly
4290 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4291 on some systems.
4292
4293 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4294 it was reported that one broken compiler failed on the former because "code" is
4295 also an independent variable.
4296
4297 3. The erroneous regex a[]b caused an array overrun reference.
4298
4299 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4300 fail on data ending with that character. (It was going on too far, and checking
4301 the next character, typically a binary zero.) This was specific to the
4302 optimized code for single-character negative classes.
4303
4304 5. Added a contributed patch from the TIN world which does the following:
4305
4306 + Add an undef for memmove, in case the the system defines a macro for it.
4307
4308 + Add a definition of offsetof(), in case there isn't one. (I don't know
4309 the reason behind this - offsetof() is part of the ANSI standard - but
4310 it does no harm).
4311
4312 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4313 most of the places where whitespace preceded '#'. I have given up and
4314 allowed the remaining 2 cases to be at the margin.
4315
4316 + Rename some variables in pcre to eliminate shadowing. This seems very
4317 pedantic, but does no harm, of course.
4318
4319 6. Moved the call to setjmp() into its own function, to get rid of warnings
4320 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4321
4322 7. Constructs such as \d{8,} were compiling into the equivalent of
4323 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4324 outcome, but in this particular case used more store than had been allocated,
4325 which caused the bug to be discovered because it threw up an internal error.
4326
4327 8. The debugging code in both pcre and pcretest for outputting the compiled
4328 form of a regex was going wrong in the case of back references followed by
4329 curly-bracketed repeats.
4330
4331
4332 Version 1.02 12-Dec-97
4333 ----------------------
4334
4335 1. Typos in pcre.3 and comments in the source fixed.
4336
4337 2. Applied a contributed patch to get rid of places where it used to remove
4338 'const' from variables, and fixed some signed/unsigned and uninitialized
4339 variable warnings.
4340
4341 3. Added the "runtest" target to Makefile.
4342
4343 4. Set default compiler flag to -O2 rather than just -O.
4344
4345
4346 Version 1.01 19-Nov-97
4347 ----------------------
4348
4349 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4350 like /([ab]*)*/, that is, for classes with more than one character in them.
4351
4352 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4353 as /((?>a*))*/ (a PCRE_EXTRA facility).
4354
4355
4356 Version 1.00 18-Nov-97
4357 ----------------------
4358
4359 1. Added compile-time macros to support systems such as SunOS4 which don't have
4360 memmove() or strerror() but have other things that can be used instead.
4361
4362 2. Arranged that "make clean" removes the executables.
4363
4364
4365 Version 0.99 27-Oct-97
4366 ----------------------
4367
4368 1. Fixed bug in code for optimizing classes with only one character. It was
4369 initializing a 32-byte map regardless, which could cause it to run off the end
4370 of the memory it had got.
4371
4372 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4373
4374
4375 Version 0.98 22-Oct-97
4376 ----------------------
4377
4378 1. Fixed bug in code for handling temporary memory usage when there are more
4379 back references than supplied space in the ovector. This could cause segfaults.
4380
4381
4382 Version 0.97 21-Oct-97
4383 ----------------------
4384
4385 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4386
4387 2. Optimized negated single characters not to use a bit map.
4388
4389 3. Brought error texts together as macro definitions; clarified some of them;
4390 fixed one that was wrong - it said "range out of order" when it meant "invalid
4391 escape sequence".
4392
4393 4. Changed some char * arguments to const char *.
4394
4395 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4396
4397 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4398 pcretest.
4399
4400
4401 Version 0.96 16-Oct-97
4402 ----------------------
4403
4404 1. Added a simple "pgrep" utility to the distribution.
4405
4406 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4407 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4408 where "ddd" means "one or more decimal digits".
4409
4410 3. Fixed serious bug. If a pattern had a back reference, but the call to
4411 pcre_exec() didn't supply a large enough ovector to record the related
4412 identifying subpattern, the match always failed. PCRE now remembers the number
4413 of the largest back reference, and gets some temporary memory in which to save
4414 the offsets during matching if necessary, in order to ensure that
4415 backreferences always work.
4416
4417 4. Increased the compatibility with Perl in a number of ways:
4418
4419 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4420 to request this handling. The option can be set at compile or exec time.
4421
4422 (b) $ matches before a terminating newline by default; an option
4423 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4424 mode). The option can be set at compile or exec time.
4425
4426 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4427 the same as Perl's. If the decimal number it represents is less than 10
4428 or there aren't that many previous left capturing parentheses, an octal
4429 escape is read. Inside a character class, it's always an octal escape,
4430 even if it is a single digit.
4431
4432 (d) An escaped but undefined alphabetic character is taken as a literal,
4433 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4434 escapes.
4435
4436 (e) {0} is now permitted. (The previous item is removed from the compiled
4437 pattern).
4438
4439 5. Changed all the names of code files so that the basic parts are no longer
4440 than 10 characters, and abolished the teeny "globals.c" file.
4441
4442 6. Changed the handling of character classes; they are now done with a 32-byte
4443 bit map always.
4444
4445 7. Added the -d and /D options to pcretest to make it possible to look at the
4446 internals of compilation without having to recompile pcre.
4447
4448
4449 Version 0.95 23-Sep-97
4450 ----------------------
4451
4452 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4453 \x20 at the start of a run of normal characters. These were being treated as
4454 real characters, instead of the source characters being re-checked.
4455
4456
4457 Version 0.94 18-Sep-97
4458 ----------------------
4459
4460 1. The functions are now thread-safe, with the caveat that the global variables
4461 containing pointers to malloc() and free() or alternative functions are the
4462 same for all threads.
4463
4464 2. Get pcre_study() to generate a bitmap of initial characters for non-
4465 anchored patterns when this is possible, and use it if passed to pcre_exec().
4466
4467
4468 Version 0.93 15-Sep-97
4469 ----------------------
4470
4471 1. /(b)|(:+)/ was computing an incorrect first character.
4472
4473 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4474 but not actually doing anything yet.
4475
4476 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4477 as Perl does (e.g. [-az] or [az-]).
4478
4479 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4480 all possible positions.
4481
4482 5. Split up into different modules to avoid including unneeded functions in a
4483 compiled binary. However, compile and exec are still in one module. The "study"
4484 function is split off.
4485
4486 6. The character tables are now in a separate module whose source is generated
4487 by an auxiliary program - but can then be edited by hand if required. There are
4488 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4489 toupper() in the code.
4490
4491 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4492 make them global. Abolish the function for setting them, as the caller can now
4493 set them directly.
4494
4495
4496 Version 0.92 11-Sep-97
4497 ----------------------
4498
4499 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4500 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4501
4502 2. Caseless matching was not working in character classes if the characters in
4503 the pattern were in upper case.
4504
4505 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4506
4507 4. Make PCRE_ANCHORED public and accept as a compile option.
4508
4509 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4510 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4511 pass them.
4512
4513 6. Give an error if bad option bits passed at compile or run time.
4514
4515 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4516 pcretest to cause it to pass that flag.
4517
4518 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4519 options, and the first character, if set.
4520
4521 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4522
4523
4524 Version 0.91 10-Sep-97
4525 ----------------------
4526
4527 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4528 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4529
4530 2. PCRE was looping on encountering an indefinitely repeated back reference to
4531 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4532 Perl does - treats the match as successful.
4533
4534 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5