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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 758 - (show annotations) (download)
Mon Nov 21 12:05:36 2011 UTC (2 years, 4 months ago) by ph10
File size: 220154 byte(s)
Disallow \N in character classes, for Perl compatibility.

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 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
50 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
51 is anything other than /usr, -L is still shown).
52
53 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
54
55 14. Perl does not support \N without a following name in a [] class; PCRE now
56 also gives an error.
57
58
59 Version 8.20 21-Oct-2011
60 ------------------------
61
62 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
63 a POSIX class. After further experiments with Perl, which convinced me that
64 Perl has bugs and confusions, a closing square bracket is no longer allowed
65 in a POSIX name. This bug also affected patterns with classes that started
66 with full stops.
67
68 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
69 captured substring, but while checking the failing first alternative,
70 substring 1 is temporarily captured. If the output vector supplied to
71 pcre_exec() was not big enough for this capture, the yield of the function
72 was still zero ("insufficient space for captured substrings"). This cannot
73 be totally fixed without adding another stack variable, which seems a lot
74 of expense for a edge case. However, I have improved the situation in cases
75 such as /(a)(b)x|abc/ matched against "abc", where the return code
76 indicates that fewer than the maximum number of slots in the ovector have
77 been set.
78
79 3. Related to (2) above: when there are more back references in a pattern than
80 slots in the output vector, pcre_exec() uses temporary memory during
81 matching, and copies in the captures as far as possible afterwards. It was
82 using the entire output vector, but this conflicts with the specification
83 that only 2/3 is used for passing back captured substrings. Now it uses
84 only the first 2/3, for compatibility. This is, of course, another edge
85 case.
86
87 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
88 main code base, and can be used by building with --enable-jit. When this is
89 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
90 runtime --no-jit option is given.
91
92 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
93 ovector, the return from the function was zero, implying that there were
94 other matches that did not fit. The correct "exactly full" value is now
95 returned.
96
97 6. If a subpattern that was called recursively or as a subroutine contained
98 (*PRUNE) or any other control that caused it to give a non-standard return,
99 invalid errors such as "Error -26 (nested recursion at the same subject
100 position)" or even infinite loops could occur.
101
102 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
103 computing the minimum length on reaching *ACCEPT, and so ended up with the
104 wrong value of 1 rather than 0. Further investigation indicates that
105 computing a minimum subject length in the presence of *ACCEPT is difficult
106 (think back references, subroutine calls), and so I have changed the code
107 so that no minimum is registered for a pattern that contains *ACCEPT.
108
109 8. If (*THEN) was present in the first (true) branch of a conditional group,
110 it was not handled as intended. [But see 16 below.]
111
112 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
113 Sheri Pierce.
114
115 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
116 the first byte in a match must be "a".
117
118 11. Change 17 for 8.13 increased the recursion depth for patterns like
119 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
120 pattern contains any instances of (*THEN). If it does not, the old
121 optimizations are restored. It would be nice to do this on a per-group
122 basis, but at the moment that is not feasible.
123
124 12. In some environments, the output of pcretest -C is CRLF terminated. This
125 broke RunTest's code that checks for the link size. A single white space
126 character after the value is now allowed for.
127
128 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
129 For "fr", it uses the Windows-specific input and output files.
130
131 14. If (*THEN) appeared in a group that was called recursively or as a
132 subroutine, it did not work as intended. [But see next item.]
133
134 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
135 pattern fragments (but not containing any | characters). If A and B are
136 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
137 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
138 D. In other words, Perl considers parentheses that do not contain any |
139 characters to be part of a surrounding alternative, whereas PCRE was
140 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
141 differently. PCRE now behaves in the same way as Perl, except in the case
142 of subroutine/recursion calls such as (?1) which have in any case always
143 been different (but PCRE had them first :-).
144
145 16. Related to 15 above: Perl does not treat the | in a conditional group as
146 creating alternatives. Such a group is treated in the same way as an
147 ordinary group without any | characters when processing (*THEN). PCRE has
148 been changed to match Perl's behaviour.
149
150 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
151 RunGrepTest script failed.
152
153 18. Change 22 for version 13 caused atomic groups to use more stack. This is
154 inevitable for groups that contain captures, but it can lead to a lot of
155 stack use in large patterns. The old behaviour has been restored for atomic
156 groups that do not contain any capturing parentheses.
157
158 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
159 suppress the check for a minimum subject length at run time. (If it was
160 given to pcre_exec() or pcre_dfa_exec() it did work.)
161
162 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
163 fail to work when decoding hex characters in data strings in EBCDIC
164 environments.
165
166 21. It appears that in at least one Mac OS environment, the isxdigit() function
167 is implemented as a macro that evaluates to its argument more than once,
168 contravening the C 90 Standard (I haven't checked a later standard). There
169 was an instance in pcretest which caused it to go wrong when processing
170 \x{...} escapes in subject strings. The has been rewritten to avoid using
171 things like p++ in the argument of isxdigit().
172
173
174 Version 8.13 16-Aug-2011
175 ------------------------
176
177 1. The Unicode data tables have been updated to Unicode 6.0.0.
178
179 2. Two minor typos in pcre_internal.h have been fixed.
180
181 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
182 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
183 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
184
185 4. There were a number of related bugs in the code for matching backrefences
186 caselessly in UTF-8 mode when codes for the characters concerned were
187 different numbers of bytes. For example, U+023A and U+2C65 are an upper
188 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
189 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
190 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
191 2-byte code at the end of the subject (it thought there wasn't enough data
192 left).
193
194 5. Comprehensive information about what went wrong is now returned by
195 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
196 as the output vector has at least 2 elements. The offset of the start of
197 the failing character and a reason code are placed in the vector.
198
199 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
200 now returned is for the first byte of the failing character, instead of the
201 last byte inspected. This is an incompatible change, but I hope it is small
202 enough not to be a problem. It makes the returned offset consistent with
203 pcre_exec() and pcre_dfa_exec().
204
205 7. pcretest now gives a text phrase as well as the error number when
206 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
207 failure, the offset and reason code are output.
208
209 8. When \R was used with a maximizing quantifier it failed to skip backwards
210 over a \r\n pair if the subsequent match failed. Instead, it just skipped
211 back over a single character (\n). This seems wrong (because it treated the
212 two characters as a single entity when going forwards), conflicts with the
213 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
214 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
215 has been changed.
216
217 9. Some internal refactoring has changed the processing so that the handling
218 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
219 time (the PCRE_DOTALL option was changed this way some time ago: version
220 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
221 which was always a bit of a fudge. It also means that there is one less
222 argument for the match() function, which reduces its stack requirements
223 slightly. This change also fixes an incompatibility with Perl: the pattern
224 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
225
226 10. More internal refactoring has drastically reduced the number of recursive
227 calls to match() for possessively repeated groups such as (abc)++ when
228 using pcre_exec().
229
230 11. While implementing 10, a number of bugs in the handling of groups were
231 discovered and fixed:
232
233 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
234 (a|)*(?1) gave a compile-time internal error.
235 ((a|)+)+ did not notice that the outer group could match an empty string.
236 (^a|^)+ was not marked as anchored.
237 (.*a|.*)+ was not marked as matching at start or after a newline.
238
239 12. Yet more internal refactoring has removed another argument from the match()
240 function. Special calls to this function are now indicated by setting a
241 value in a variable in the "match data" data block.
242
243 13. Be more explicit in pcre_study() instead of relying on "default" for
244 opcodes that mean there is no starting character; this means that when new
245 ones are added and accidentally left out of pcre_study(), testing should
246 pick them up.
247
248 14. The -s option of pcretest has been documented for ages as being an old
249 synonym of -m (show memory usage). I have changed it to mean "force study
250 for every regex", that is, assume /S for every regex. This is similar to -i
251 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
252 using it. It makes it easier to run collections of tests with and without
253 study enabled, and thereby test pcre_study() more easily. All the standard
254 tests are now run with and without -s (but some patterns can be marked as
255 "never study" - see 20 below).
256
257 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
258 restoration of the capturing data to the outer values was not happening
259 correctly.
260
261 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
262 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
263 pattern had matched an empty string, and so incorrectly returned a no
264 match.
265
266 17. There was optimizing code for the last branch of non-capturing parentheses,
267 and also for the obeyed branch of a conditional subexpression, which used
268 tail recursion to cut down on stack usage. Unfortunately, now that there is
269 the possibility of (*THEN) occurring in these branches, tail recursion is
270 no longer possible because the return has to be checked for (*THEN). These
271 two optimizations have therefore been removed. [But see 8.20/11 above.]
272
273 18. If a pattern containing \R was studied, it was assumed that \R always
274 matched two bytes, thus causing the minimum subject length to be
275 incorrectly computed because \R can also match just one byte.
276
277 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
278 was incorrectly computed.
279
280 20. If /S is present twice on a test pattern in pcretest input, it now
281 *disables* studying, thereby overriding the use of -s on the command line
282 (see 14 above). This is necessary for one or two tests to keep the output
283 identical in both cases.
284
285 21. When (*ACCEPT) was used in an assertion that matched an empty string and
286 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
287
288 22. When an atomic group that contained a capturing parenthesis was
289 successfully matched, but the branch in which it appeared failed, the
290 capturing was not being forgotten if a higher numbered group was later
291 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
292 group 1 to "a", when in fact it should be unset. This applied to multi-
293 branched capturing and non-capturing groups, repeated or not, and also to
294 positive assertions (capturing in negative assertions does not happen
295 in PCRE) and also to nested atomic groups.
296
297 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
298 subject after a captured substring, to make it easier to tell which of a
299 number of identical substrings has been captured.
300
301 24. The way atomic groups are processed by pcre_exec() has been changed so that
302 if they are repeated, backtracking one repetition now resets captured
303 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
304 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
305 "aaa". Previously, it would have been "a". As part of this code
306 refactoring, the way recursive calls are handled has also been changed.
307
308 25. If an assertion condition captured any substrings, they were not passed
309 back unless some other capturing happened later. For example, if
310 (?(?=(a))a) was matched against "a", no capturing was returned.
311
312 26. When studying a pattern that contained subroutine calls or assertions,
313 the code for finding the minimum length of a possible match was handling
314 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
315 group 1 called group 2 while simultaneously a separate group 2 called group
316 1). A stack overflow occurred in this case. I have fixed this by limiting
317 the recursion depth to 10.
318
319 27. Updated RunTest.bat in the distribution to the version supplied by Tom
320 Fortmann. This supports explicit test numbers on the command line, and has
321 argument validation and error reporting.
322
323 28. An instance of \X with an unlimited repeat could fail if at any point the
324 first character it looked at was a mark character.
325
326 29. Some minor code refactoring concerning Unicode properties and scripts
327 should reduce the stack requirement of match() slightly.
328
329 30. Added the '=' option to pcretest to check the setting of unused capturing
330 slots at the end of the pattern, which are documented as being -1, but are
331 not included in the return count.
332
333 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
334 compiled something random. Now it gives a compile-time error (as does
335 Perl).
336
337 32. A *MARK encountered during the processing of a positive assertion is now
338 recorded and passed back (compatible with Perl).
339
340 33. If --only-matching or --colour was set on a pcregrep call whose pattern
341 had alternative anchored branches, the search for a second match in a line
342 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
343 matched the line "0102" twice. The same bug affected patterns that started
344 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
345 twice.
346
347 34. Previously, PCRE did not allow quantification of assertions. However, Perl
348 does, and because of capturing effects, quantifying parenthesized
349 assertions may at times be useful. Quantifiers are now allowed for
350 parenthesized assertions.
351
352 35. A minor code tidy in pcre_compile() when checking options for \R usage.
353
354 36. \g was being checked for fancy things in a character class, when it should
355 just be a literal "g".
356
357 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
358 appearance of a nested POSIX class supersedes an apparent external class.
359 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
360 unescaped square brackets may also appear as part of class names. For
361 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
362 more like Perl. (But see 8.20/1 above.)
363
364 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
365 was because it thought it was \N{name}, which is not supported).
366
367 39. Add minix to OS list not supporting the -S option in pcretest.
368
369 40. PCRE tries to detect cases of infinite recursion at compile time, but it
370 cannot analyze patterns in sufficient detail to catch mutual recursions
371 such as ((?1))((?2)). There is now a runtime test that gives an error if a
372 subgroup is called recursively as a subpattern for a second time at the
373 same position in the subject string. In previous releases this might have
374 been caught by the recursion limit, or it might have run out of stack.
375
376 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
377 happen only once. PCRE was, however incorrectly giving a compile time error
378 "recursive call could loop indefinitely" because it cannot analyze the
379 pattern in sufficient detail. The compile time test no longer happens when
380 PCRE is compiling a conditional subpattern, but actual runaway loops are
381 now caught at runtime (see 40 above).
382
383 42. It seems that Perl allows any characters other than a closing parenthesis
384 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
385 has been changed to be the same.
386
387 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
388 as not to get warnings when autogen.sh is called. Also changed
389 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
390
391 44. To help people who use pcregrep to scan files containing exceedingly long
392 lines, the following changes have been made:
393
394 (a) The default value of the buffer size parameter has been increased from
395 8K to 20K. (The actual buffer used is three times this size.)
396
397 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
398 PCRE is built.
399
400 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
401 to be set at run time.
402
403 (d) Numerical values in pcregrep options can be followed by K or M, for
404 example --buffer-size=50K.
405
406 (e) If a line being scanned overflows pcregrep's buffer, an error is now
407 given and the return code is set to 2.
408
409 45. Add a pointer to the latest mark to the callout data block.
410
411 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
412 partial match of an empty string instead of no match. This was specific to
413 the use of ".".
414
415 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
416 complete match instead of a partial match. This bug was dependent on both
417 the PCRE_UTF8 and PCRE_DOTALL options being set.
418
419 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
420 starting byte set, because \b was not being ignored.
421
422
423 Version 8.12 15-Jan-2011
424 ------------------------
425
426 1. Fixed some typos in the markup of the man pages, and wrote a script that
427 checks for such things as part of the documentation building process.
428
429 2. On a big-endian 64-bit system, pcregrep did not correctly process the
430 --match-limit and --recursion-limit options (added for 8.11). In
431 particular, this made one of the standard tests fail. (The integer value
432 went into the wrong half of a long int.)
433
434 3. If the --colour option was given to pcregrep with -v (invert match), it
435 did strange things, either producing crazy output, or crashing. It should,
436 of course, ignore a request for colour when reporting lines that do not
437 match.
438
439 4. Another pcregrep bug caused similar problems if --colour was specified with
440 -M (multiline) and the pattern match finished with a line ending.
441
442 5. In pcregrep, when a pattern that ended with a literal newline sequence was
443 matched in multiline mode, the following line was shown as part of the
444 match. This seems wrong, so I have changed it.
445
446 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
447 the check for further matches in the same line (so they could be coloured)
448 to overrun the end of the current line. If another match was found, it was
449 incorrectly shown (and then shown again when found in the next line).
450
451 7. If pcregrep was compiled under Windows, there was a reference to the
452 function pcregrep_exit() before it was defined. I am assuming this was
453 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
454 reported by a user. I've moved the definition above the reference.
455
456
457 Version 8.11 10-Dec-2010
458 ------------------------
459
460 1. (*THEN) was not working properly if there were untried alternatives prior
461 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
462 backtracked to try for "b" instead of moving to the next alternative branch
463 at the same level (in this case, to look for "c"). The Perl documentation
464 is clear that when (*THEN) is backtracked onto, it goes to the "next
465 alternative in the innermost enclosing group".
466
467 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
468 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
469 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
470 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
471 (*THEN).
472
473 3. If \s appeared in a character class, it removed the VT character from
474 the class, even if it had been included by some previous item, for example
475 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
476 of \s, but is part of the POSIX "space" class.)
477
478 4. A partial match never returns an empty string (because you can always
479 match an empty string at the end of the subject); however the checking for
480 an empty string was starting at the "start of match" point. This has been
481 changed to the "earliest inspected character" point, because the returned
482 data for a partial match starts at this character. This means that, for
483 example, /(?<=abc)def/ gives a partial match for the subject "abc"
484 (previously it gave "no match").
485
486 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
487 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
488 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
489 has an implication that the given string is incomplete (because a partial
490 match is preferred over a full match). For this reason, these items now
491 give a partial match in this situation. [Aside: previously, the one case
492 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
493 match rather than a full match, which was wrong by the old rules, but is
494 now correct.]
495
496 6. There was a bug in the handling of #-introduced comments, recognized when
497 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
498 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
499 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
500 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
501 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
502 places in pcre_compile().
503
504 7. Related to (6) above, when pcre_compile() was skipping #-introduced
505 comments when looking ahead for named forward references to subpatterns,
506 the only newline sequence it recognized was NL. It now handles newlines
507 according to the set newline convention.
508
509 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
510 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
511 cater for a lack of strerror(). These oversights have been fixed.
512
513 9. Added --match-limit and --recursion-limit to pcregrep.
514
515 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
516
517 11. When the -o option was used, pcregrep was setting a return code of 1, even
518 when matches were found, and --line-buffered was not being honoured.
519
520 12. Added an optional parentheses number to the -o and --only-matching options
521 of pcregrep.
522
523 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
524 can match an empty string. The code to do it in pcretest and pcredemo
525 needed fixing:
526
527 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
528 only one byte after an empty string match just before CRLF (this case
529 just got forgotten; "any" and "anycrlf" were OK).
530
531 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
532 mode when an empty string match preceded an ASCII character followed by
533 a non-ASCII character. (The code for advancing by one character rather
534 than one byte was nonsense.)
535
536 (c) The pcredemo.c sample program did not have any code at all to handle
537 the cases when CRLF is a valid newline sequence.
538
539 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
540 as a starting offset was within the subject string. There is now a new
541 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
542 negative or greater than the length of the string. In order to test this,
543 pcretest is extended to allow the setting of negative starting offsets.
544
545 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
546 starting offset points to the beginning of a UTF-8 character was
547 unnecessarily clumsy. I tidied it up.
548
549 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
550 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
551
552 17. Nobody had reported that the --include_dir option, which was added in
553 release 7.7 should have been called --include-dir (hyphen, not underscore)
554 for compatibility with GNU grep. I have changed it to --include-dir, but
555 left --include_dir as an undocumented synonym, and the same for
556 --exclude-dir, though that is not available in GNU grep, at least as of
557 release 2.5.4.
558
559 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
560 characters from a string of bytes) have been redefined so as not to use
561 loops, in order to improve performance in some environments. At the same
562 time, I abstracted some of the common code into auxiliary macros to save
563 repetition (this should not affect the compiled code).
564
565 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
566 compile-time error is now given if \c is not followed by an ASCII
567 character, that is, a byte less than 128. (In EBCDIC mode, the code is
568 different, and any byte value is allowed.)
569
570 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
571 START_OPTIMIZE option, which is now allowed at compile time - but just
572 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
573 to pcregrep and other applications that have no direct access to PCRE
574 options. The new /Y option in pcretest sets this option when calling
575 pcre_compile().
576
577 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
578 back references. Groups containing recursive back references were forced to
579 be atomic by that change, but in the case of named groups, the amount of
580 memory required was incorrectly computed, leading to "Failed: internal
581 error: code overflow". This has been fixed.
582
583 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
584 pcretest.c, to avoid build problems in some Borland environments.
585
586
587 Version 8.10 25-Jun-2010
588 ------------------------
589
590 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
591 THEN.
592
593 2. (*ACCEPT) was not working when inside an atomic group.
594
595 3. Inside a character class, \B is treated as a literal by default, but
596 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
597 causes the error). The code is unchanged, but I tidied the documentation.
598
599 4. Inside a character class, PCRE always treated \R and \X as literals,
600 whereas Perl faults them if its -w option is set. I have changed PCRE so
601 that it faults them when PCRE_EXTRA is set.
602
603 5. Added support for \N, which always matches any character other than
604 newline. (It is the same as "." when PCRE_DOTALL is not set.)
605
606 6. When compiling pcregrep with newer versions of gcc which may have
607 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
608 declared with attribute warn_unused_result" were given. Just casting the
609 result to (void) does not stop the warnings; a more elaborate fudge is
610 needed. I've used a macro to implement this.
611
612 7. Minor change to pcretest.c to avoid a compiler warning.
613
614 8. Added four artifical Unicode properties to help with an option to make
615 \s etc use properties (see next item). The new properties are: Xan
616 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
617
618 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
619 use Unicode properties. (*UCP) at the start of a pattern can be used to set
620 this option. Modified pcretest to add /W to test this facility. Added
621 REG_UCP to make it available via the POSIX interface.
622
623 10. Added --line-buffered to pcregrep.
624
625 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
626 studied, and the match started with a letter with a code point greater than
627 127 whose first byte was different to the first byte of the other case of
628 the letter, the other case of this starting letter was not recognized
629 (#976).
630
631 12. If a pattern that was studied started with a repeated Unicode property
632 test, for example, \p{Nd}+, there was the theoretical possibility of
633 setting up an incorrect bitmap of starting bytes, but fortunately it could
634 not have actually happened in practice until change 8 above was made (it
635 added property types that matched character-matching opcodes).
636
637 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
638 possible starting bytes for non-anchored patterns.
639
640 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
641 \R, and also a number of cases that involve Unicode properties, both
642 explicit and implicit when PCRE_UCP is set.
643
644 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
645 input, it could crash or give wrong results if characters with values
646 greater than 0xc0 were present in the subject string. (Detail: it assumed
647 UTF-8 input when processing these items.)
648
649 16. Added a lot of (int) casts to avoid compiler warnings in systems where
650 size_t is 64-bit (#991).
651
652 17. Added a check for running out of memory when PCRE is compiled with
653 --disable-stack-for-recursion (#990).
654
655 18. If the last data line in a file for pcretest does not have a newline on
656 the end, a newline was missing in the output.
657
658 19. The default pcre_chartables.c file recognizes only ASCII characters (values
659 less than 128) in its various bitmaps. However, there is a facility for
660 generating tables according to the current locale when PCRE is compiled. It
661 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
662 characters, are recognized by isspace() and therefore were getting set in
663 these tables, and indeed these tables seem to approximate to ISO 8859. This
664 caused a problem in UTF-8 mode when pcre_study() was used to create a list
665 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
666 which of course cannot start UTF-8 characters. I have changed the code so
667 that only real ASCII characters (less than 128) and the correct starting
668 bytes for UTF-8 encodings are set for characters greater than 127 when in
669 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
670 altogether.)
671
672 20. Added the /T option to pcretest so as to be able to run tests with non-
673 standard character tables, thus making it possible to include the tests
674 used for 19 above in the standard set of tests.
675
676 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
677 reference to a subpattern the other side of a comment that contains an
678 opening parenthesis caused either an internal compiling error, or a
679 reference to the wrong subpattern.
680
681
682 Version 8.02 19-Mar-2010
683 ------------------------
684
685 1. The Unicode data tables have been updated to Unicode 5.2.0.
686
687 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
688 configured.
689
690 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
691 original author of that file, following a query about its status.
692
693 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
694 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
695
696 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
697 quantifier applied to a forward-referencing subroutine call, could compile
698 incorrect code or give the error "internal error: previously-checked
699 referenced subpattern not found".
700
701 6. Both MS Visual Studio and Symbian OS have problems with initializing
702 variables to point to external functions. For these systems, therefore,
703 pcre_malloc etc. are now initialized to local functions that call the
704 relevant global functions.
705
706 7. There were two entries missing in the vectors called coptable and poptable
707 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
708 I've fixed the data, and added a kludgy way of testing at compile time that
709 the lengths are correct (equal to the number of opcodes).
710
711 8. Following on from 7, I added a similar kludge to check the length of the
712 eint vector in pcreposix.c.
713
714 9. Error texts for pcre_compile() are held as one long string to avoid too
715 much relocation at load time. To find a text, the string is searched,
716 counting zeros. There was no check for running off the end of the string,
717 which could happen if a new error number was added without updating the
718 string.
719
720 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
721
722 11. \K was not working if it appeared in an atomic group or in a group that
723 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
724 \K is "not well defined" if used in an assertion. PCRE now accepts it if
725 the assertion is positive, but not if it is negative.
726
727 12. Change 11 fortuitously reduced the size of the stack frame used in the
728 "match()" function of pcre_exec.c by one pointer. Forthcoming
729 implementation of support for (*MARK) will need an extra pointer on the
730 stack; I have reserved it now, so that the stack frame size does not
731 decrease.
732
733 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
734 item in branch that calls a recursion is a subroutine call - as in the
735 second branch in the above example - was incorrectly given the compile-
736 time error "recursive call could loop indefinitely" because pcre_compile()
737 was not correctly checking the subroutine for matching a non-empty string.
738
739 14. The checks for overrunning compiling workspace could trigger after an
740 overrun had occurred. This is a "should never occur" error, but it can be
741 triggered by pathological patterns such as hundreds of nested parentheses.
742 The checks now trigger 100 bytes before the end of the workspace.
743
744 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
745
746
747 Version 8.01 19-Jan-2010
748 ------------------------
749
750 1. If a pattern contained a conditional subpattern with only one branch (in
751 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
752 computed the wrong minimum data length (which is of course zero for such
753 subpatterns). This could cause incorrect "no match" results.
754
755 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
756 the pattern is reset in the first branch, pcre_compile() failed with
757 "internal error: code overflow at offset...". This happened only when
758 the reset was to the original external option setting. (An optimization
759 abstracts leading options settings into an external setting, which was the
760 cause of this.)
761
762 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
763 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
764 assertion pattern did not match (meaning that the assertion was true), it
765 was incorrectly treated as false if the SKIP had been reached during the
766 matching. This also applied to assertions used as conditions.
767
768 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
769 assertion subpattern, including such a pattern used as a condition,
770 unpredictable results occurred, instead of the error return
771 PCRE_ERROR_DFA_UITEM.
772
773 5. The C++ GlobalReplace function was not working like Perl for the special
774 situation when an empty string is matched. It now does the fancy magic
775 stuff that is necessary.
776
777 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
778 removed. (These were left over from very, very early versions of PCRE.)
779
780 7. Some cosmetic changes to the code to make life easier when compiling it
781 as part of something else:
782
783 (a) Change DEBUG to PCRE_DEBUG.
784
785 (b) In pcre_compile(), rename the member of the "branch_chain" structure
786 called "current" as "current_branch", to prevent a collision with the
787 Linux macro when compiled as a kernel module.
788
789 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
790 prevent a collision with the Linux macro when compiled as a kernel
791 module.
792
793 8. In pcre_compile() there are some checks for integer overflows that used to
794 cast potentially large values to (double). This has been changed to that
795 when building, a check for int64_t is made, and if it is found, it is used
796 instead, thus avoiding the use of floating point arithmetic. (There is no
797 other use of FP in PCRE.) If int64_t is not found, the fallback is to
798 double.
799
800 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
801 2005 (difference between two addresses compared to an unsigned value).
802
803 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
804 custom one, because of the following reported problem in Windows:
805
806 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
807 under Win32.
808 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
809 therefore missing the function definition.
810 - The compiler thus generates a "C" signature for the test function.
811 - The linker fails to find the "C" function.
812 - PCRE fails to configure if asked to do so against libbz2.
813
814 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
815 messages were output:
816
817 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
818 rerunning libtoolize, to keep the correct libtool macros in-tree.
819 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
820
821 I have done both of these things.
822
823 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
824 most of the time, it *can* run out if it is given a pattern that contains a
825 runaway infinite recursion. I updated the discussion in the pcrestack man
826 page.
827
828 13. Now that we have gone to the x.xx style of version numbers, the minor
829 version may start with zero. Using 08 or 09 is a bad idea because users
830 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
831 interpreted as invalid octal numbers. I've updated the previous comment in
832 configure.ac, and also added a check that gives an error if 08 or 09 are
833 used.
834
835 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
836 causing partial matching to fail when the end of the subject matched \W
837 in a UTF-8 pattern where \W was quantified with a minimum of 3.
838
839 15. There were some discrepancies between the declarations in pcre_internal.h
840 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
841 their definitions. The declarations used "const uschar *" and the
842 definitions used USPTR. Even though USPTR is normally defined as "const
843 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
844 reported that: "This difference in casting confuses some C++ compilers, for
845 example, SunCC recognizes above declarations as different functions and
846 generates broken code for hbpcre." I have changed the declarations to use
847 USPTR.
848
849 16. GNU libtool is named differently on some systems. The autogen.sh script now
850 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
851 (FreeBSD).
852
853 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
854 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
855 comment: "Figure out how to create a longlong from a string: strtoll and
856 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
857 instance, but it only takes 2 args instead of 3!"
858
859 18. A subtle bug concerned with back references has been fixed by a change of
860 specification, with a corresponding code fix. A pattern such as
861 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
862 refers, was giving matches when it shouldn't. For example, xa=xaaa would
863 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
864 same bug. Such groups have to be quantified to be useful, or contained
865 inside another quantified group. (If there's no repetition, the reference
866 can never match.) The problem arises because, having left the group and
867 moved on to the rest of the pattern, a later failure that backtracks into
868 the group uses the captured value from the final iteration of the group
869 rather than the correct earlier one. I have fixed this in PCRE by forcing
870 any group that contains a reference to itself to be an atomic group; that
871 is, there cannot be any backtracking into it once it has completed. This is
872 similar to recursive and subroutine calls.
873
874
875 Version 8.00 19-Oct-09
876 ----------------------
877
878 1. The table for translating pcre_compile() error codes into POSIX error codes
879 was out-of-date, and there was no check on the pcre_compile() error code
880 being within the table. This could lead to an OK return being given in
881 error.
882
883 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
884 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
885 in a Windows environment.
886
887 3. The pcregrep --count option prints the count for each file even when it is
888 zero, as does GNU grep. However, pcregrep was also printing all files when
889 --files-with-matches was added. Now, when both options are given, it prints
890 counts only for those files that have at least one match. (GNU grep just
891 prints the file name in this circumstance, but including the count seems
892 more useful - otherwise, why use --count?) Also ensured that the
893 combination -clh just lists non-zero counts, with no names.
894
895 4. The long form of the pcregrep -F option was incorrectly implemented as
896 --fixed_strings instead of --fixed-strings. This is an incompatible change,
897 but it seems right to fix it, and I didn't think it was worth preserving
898 the old behaviour.
899
900 5. The command line items --regex=pattern and --regexp=pattern were not
901 recognized by pcregrep, which required --regex pattern or --regexp pattern
902 (with a space rather than an '='). The man page documented the '=' forms,
903 which are compatible with GNU grep; these now work.
904
905 6. No libpcreposix.pc file was created for pkg-config; there was just
906 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
907
908 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
909 when UCP support is not needed, by modifying the Python script that
910 generates it from Unicode data files. This should not matter if the module
911 is correctly used as a library, but I received one complaint about 50K of
912 unwanted data. My guess is that the person linked everything into his
913 program rather than using a library. Anyway, it does no harm.
914
915 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
916 was a minimum greater than 1 for a wide character in a possessive
917 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
918 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
919 character. Chaos in the form of incorrect output or a compiling loop could
920 result.
921
922 9. The restrictions on what a pattern can contain when partial matching is
923 requested for pcre_exec() have been removed. All patterns can now be
924 partially matched by this function. In addition, if there are at least two
925 slots in the offset vector, the offset of the earliest inspected character
926 for the match and the offset of the end of the subject are set in them when
927 PCRE_ERROR_PARTIAL is returned.
928
929 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
930 synonymous with PCRE_PARTIAL, for backwards compatibility, and
931 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
932 and may be more useful for multi-segment matching.
933
934 11. Partial matching with pcre_exec() is now more intuitive. A partial match
935 used to be given if ever the end of the subject was reached; now it is
936 given only if matching could not proceed because another character was
937 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
938 string "Z", which now yields "no match" instead of "partial match". In the
939 case of pcre_dfa_exec(), "no match" is given if every matching path for the
940 final character ended with (*FAIL).
941
942 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
943 if the pattern had a "must contain" character that was already found in the
944 earlier partial match, unless partial matching was again requested. For
945 example, with the pattern /dog.(body)?/, the "must contain" character is
946 "g". If the first part-match was for the string "dog", restarting with
947 "sbody" failed. This bug has been fixed.
948
949 13. The string returned by pcre_dfa_exec() after a partial match has been
950 changed so that it starts at the first inspected character rather than the
951 first character of the match. This makes a difference only if the pattern
952 starts with a lookbehind assertion or \b or \B (\K is not supported by
953 pcre_dfa_exec()). It's an incompatible change, but it makes the two
954 matching functions compatible, and I think it's the right thing to do.
955
956 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
957 so that the demonstration program is easily available in environments where
958 PCRE has not been installed from source.
959
960 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
961 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
962 library.
963
964 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
965 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
966 is not the first non-POSIX option to be added. Clearly some people find
967 these options useful.
968
969 17. If a caller to the POSIX matching function regexec() passes a non-zero
970 value for nmatch with a NULL value for pmatch, the value of
971 nmatch is forced to zero.
972
973 18. RunGrepTest did not have a test for the availability of the -u option of
974 the diff command, as RunTest does. It now checks in the same way as
975 RunTest, and also checks for the -b option.
976
977 19. If an odd number of negated classes containing just a single character
978 interposed, within parentheses, between a forward reference to a named
979 subpattern and the definition of the subpattern, compilation crashed with
980 an internal error, complaining that it could not find the referenced
981 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
982 [The bug was that it was starting one character too far in when skipping
983 over the character class, thus treating the ] as data rather than
984 terminating the class. This meant it could skip too much.]
985
986 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
987 /g option in pcretest when the pattern contains \K, which makes it possible
988 to have an empty string match not at the start, even when the pattern is
989 anchored. Updated pcretest and pcredemo to use this option.
990
991 21. If the maximum number of capturing subpatterns in a recursion was greater
992 than the maximum at the outer level, the higher number was returned, but
993 with unset values at the outer level. The correct (outer level) value is
994 now given.
995
996 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
997 PCRE did not set those parentheses (unlike Perl). I have now found a way to
998 make it do so. The string so far is captured, making this feature
999 compatible with Perl.
1000
1001 23. The tests have been re-organized, adding tests 11 and 12, to make it
1002 possible to check the Perl 5.10 features against Perl 5.10.
1003
1004 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1005 pattern matches a fixed length string. PCRE did not allow this; now it
1006 does. Neither allows recursion.
1007
1008 25. I finally figured out how to implement a request to provide the minimum
1009 length of subject string that was needed in order to match a given pattern.
1010 (It was back references and recursion that I had previously got hung up
1011 on.) This code has now been added to pcre_study(); it finds a lower bound
1012 to the length of subject needed. It is not necessarily the greatest lower
1013 bound, but using it to avoid searching strings that are too short does give
1014 some useful speed-ups. The value is available to calling programs via
1015 pcre_fullinfo().
1016
1017 26. While implementing 25, I discovered to my embarrassment that pcretest had
1018 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1019 study optimizations had never been tested with that matching function.
1020 Oops. What is worse, even when it was passed study data, there was a bug in
1021 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1022 were also very few tests of studied patterns with pcre_dfa_exec().
1023
1024 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1025 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1026 on the other side of the coin, they are no longer allowed to have different
1027 names, because these cannot be distinguished in PCRE, and this has caused
1028 confusion. (This is a difference from Perl.)
1029
1030 28. When duplicate subpattern names are present (necessarily with different
1031 numbers, as required by 27 above), and a test is made by name in a
1032 conditional pattern, either for a subpattern having been matched, or for
1033 recursion in such a pattern, all the associated numbered subpatterns are
1034 tested, and the overall condition is true if the condition is true for any
1035 one of them. This is the way Perl works, and is also more like the way
1036 testing by number works.
1037
1038
1039 Version 7.9 11-Apr-09
1040 ---------------------
1041
1042 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1043 (pcretest), all targets were linked against these libraries. This included
1044 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1045 libraries. This caused unwanted dependencies to be created. This problem
1046 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1047 pcretest is linked with readline.
1048
1049 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1050 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1051 moved outside it again, because FALSE and TRUE are already defined in AIX,
1052 but BOOL is not.
1053
1054 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1055 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1056
1057 4. The pcregrep documentation said spaces were inserted as well as colons (or
1058 hyphens) following file names and line numbers when outputting matching
1059 lines. This is not true; no spaces are inserted. I have also clarified the
1060 wording for the --colour (or --color) option.
1061
1062 5. In pcregrep, when --colour was used with -o, the list of matching strings
1063 was not coloured; this is different to GNU grep, so I have changed it to be
1064 the same.
1065
1066 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1067 each matching line was coloured. Now it goes on to look for further matches
1068 of any of the test patterns, which is the same behaviour as GNU grep.
1069
1070 7. A pattern that could match an empty string could cause pcregrep to loop; it
1071 doesn't make sense to accept an empty string match in pcregrep, so I have
1072 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1073 seems to be how GNU grep behaves.
1074
1075 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1076 start or after a newline", because the conditional assertion was not being
1077 correctly handled. The rule now is that both the assertion and what follows
1078 in the first alternative must satisfy the test.
1079
1080 9. If auto-callout was enabled in a pattern with a conditional group whose
1081 condition was an assertion, PCRE could crash during matching, both with
1082 pcre_exec() and pcre_dfa_exec().
1083
1084 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1085 used for matching.
1086
1087 11. Unicode property support in character classes was not working for
1088 characters (bytes) greater than 127 when not in UTF-8 mode.
1089
1090 12. Added the -M command line option to pcretest.
1091
1092 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1093
1094 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1095
1096 16. Added comments and documentation about mis-use of no_arg in the C++
1097 wrapper.
1098
1099 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1100 from Martin Jerabek that uses macro names for all relevant character and
1101 string constants.
1102
1103 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1104 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1105 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1106 these, but not everybody uses configure.
1107
1108 19. A conditional group that had only one branch was not being correctly
1109 recognized as an item that could match an empty string. This meant that an
1110 enclosing group might also not be so recognized, causing infinite looping
1111 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1112 with the subject "ab", where knowledge that the repeated group can match
1113 nothing is needed in order to break the loop.
1114
1115 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1116 exec(), but without supplying a callout function, matching went wrong.
1117
1118 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1119 leak if the size of the offset vector was greater than 30. When the vector
1120 is smaller, the saved offsets during recursion go onto a local stack
1121 vector, but for larger vectors malloc() is used. It was failing to free
1122 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1123 error, in fact).
1124
1125 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1126 heapframe that is used only when UTF-8 support is enabled. This caused no
1127 problem, but was untidy.
1128
1129 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1130 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1131 included within another project.
1132
1133 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1134 slightly modified by me:
1135
1136 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1137 not building pcregrep.
1138
1139 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1140 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1141
1142 25. Forward references, both numeric and by name, in patterns that made use of
1143 duplicate group numbers, could behave incorrectly or give incorrect errors,
1144 because when scanning forward to find the reference group, PCRE was not
1145 taking into account the duplicate group numbers. A pattern such as
1146 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1147
1148 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1149 the feature of a custom pointer more persuasive (as requested by a user).
1150
1151 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1152 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1153 pre-defined.
1154
1155 28. Added support for (*UTF8) at the start of a pattern.
1156
1157 29. Arrange for flags added by the "release type" setting in CMake to be shown
1158 in the configuration summary.
1159
1160
1161 Version 7.8 05-Sep-08
1162 ---------------------
1163
1164 1. Replaced UCP searching code with optimized version as implemented for Ad
1165 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1166 stage table and inline lookup instead of a function, giving speed ups of 2
1167 to 5 times on some simple patterns that I tested. Permission was given to
1168 distribute the MultiStage2.py script that generates the tables (it's not in
1169 the tarball, but is in the Subversion repository).
1170
1171 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1172 scripts.
1173
1174 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1175 a group with a zero qualifier. The result of the study could be incorrect,
1176 or the function might crash, depending on the pattern.
1177
1178 4. Caseless matching was not working for non-ASCII characters in back
1179 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1180 It now works when Unicode Property Support is available.
1181
1182 5. In pcretest, an escape such as \x{de} in the data was always generating
1183 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1184 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1185 truncation.
1186
1187 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1188
1189 7. Added two (int) casts to pcregrep when printing the difference of two
1190 pointers, in case they are 64-bit values.
1191
1192 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1193 test 2 if it fails.
1194
1195 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1196 and a #define of that name to empty if it is not externally set. This is to
1197 allow users of MSVC to set it if necessary.
1198
1199 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1200 the convenience functions in the pcre_get.c source file.
1201
1202 11. An option change at the start of a pattern that had top-level alternatives
1203 could cause overwriting and/or a crash. This command provoked a crash in
1204 some environments:
1205
1206 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1207
1208 This potential security problem was recorded as CVE-2008-2371.
1209
1210 12. For a pattern where the match had to start at the beginning or immediately
1211 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1212 pcre_dfa_exec() could read past the end of the passed subject if there was
1213 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1214 pcretest so that it places the subject at the end of its malloc-ed buffer.
1215
1216 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1217 exec() might read past the end of the data buffer in UTF-8 mode.
1218
1219 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1220 the data contained the byte 0x85 as part of a UTF-8 character within its
1221 first line. This applied both to normal and DFA matching.
1222
1223 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1224 /^[^d]*?$/8 failed to match "abc".
1225
1226 16. Added a missing copyright notice to pcrecpp_internal.h.
1227
1228 17. Make it more clear in the documentation that values returned from
1229 pcre_exec() in ovector are byte offsets, not character counts.
1230
1231 18. Tidied a few places to stop certain compilers from issuing warnings.
1232
1233 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1234 supplied by Stefan Weber. I made a further small update for 7.8 because
1235 there is a change of source arrangements: the pcre_searchfuncs.c module is
1236 replaced by pcre_ucd.c.
1237
1238
1239 Version 7.7 07-May-08
1240 ---------------------
1241
1242 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1243 a string to a long long, pretend we don't even have a long long." This is
1244 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1245
1246 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1247 pre-7.6 versions, which defined a global no_arg variable instead of putting
1248 it in the RE class. (See also #8 below.)
1249
1250 3. Remove a line of dead code, identified by coverity and reported by Nuno
1251 Lopes.
1252
1253 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1254
1255 (1) The include/exclude patterns were being applied to the whole pathnames
1256 of files, instead of just to the final components.
1257
1258 (2) If there was more than one level of directory, the subdirectories were
1259 skipped unless they satisfied the include/exclude conditions. This is
1260 inconsistent with GNU grep (and could even be seen as contrary to the
1261 pcregrep specification - which I improved to make it absolutely clear).
1262 The action now is always to scan all levels of directory, and just
1263 apply the include/exclude patterns to regular files.
1264
1265 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1266 --exclude_dir in the tests to avoid scanning .svn directories.
1267
1268 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1269 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1270 doesn't support NULs in patterns.
1271
1272 7. Added some missing "const"s to declarations of static tables in
1273 pcre_compile.c and pcre_dfa_exec.c.
1274
1275 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1276 caused by fix #2 above. (Subsequently also a second patch to fix the
1277 first patch. And a third patch - this was a messy problem.)
1278
1279 9. Applied Craig's patch to remove the use of push_back().
1280
1281 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1282 matching function regexec().
1283
1284 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1285 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1286 references. PCRE supports relative numbers with this syntax (I don't think
1287 Oniguruma does).
1288
1289 12. Previously, a group with a zero repeat such as (...){0} was completely
1290 omitted from the compiled regex. However, this means that if the group
1291 was called as a subroutine from elsewhere in the pattern, things went wrong
1292 (an internal error was given). Such groups are now left in the compiled
1293 pattern, with a new opcode that causes them to be skipped at execution
1294 time.
1295
1296 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1297 to the way PCRE behaves:
1298
1299 (a) A lone ] character is dis-allowed (Perl treats it as data).
1300
1301 (b) A back reference to an unmatched subpattern matches an empty string
1302 (Perl fails the current match path).
1303
1304 (c) A data ] in a character class must be notated as \] because if the
1305 first data character in a class is ], it defines an empty class. (In
1306 Perl it is not possible to have an empty class.) The empty class []
1307 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1308 The negative empty class [^] matches any one character, independently
1309 of the DOTALL setting.
1310
1311 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1312 non-existent subpattern following a character class starting with ']' and
1313 containing () gave an internal compiling error instead of "reference to
1314 non-existent subpattern". Fortunately, when the pattern did exist, the
1315 compiled code was correct. (When scanning forwards to check for the
1316 existencd of the subpattern, it was treating the data ']' as terminating
1317 the class, so got the count wrong. When actually compiling, the reference
1318 was subsequently set up correctly.)
1319
1320 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1321 it was being rejected as not supported by pcre_dfa_exec(), even though
1322 other assertions are supported. I have made pcre_dfa_exec() support
1323 (*FAIL).
1324
1325 16. The implementation of 13c above involved the invention of a new opcode,
1326 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1327 cannot be changed at match time, I realized I could make a small
1328 improvement to matching performance by compiling OP_ALLANY instead of
1329 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1330 on the OP_ANY path.
1331
1332 17. Compiling pcretest on Windows with readline support failed without the
1333 following two fixes: (1) Make the unistd.h include conditional on
1334 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1335
1336 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1337 ncurses library to be included for pcretest when ReadLine support is
1338 requested, but also to allow for it to be overridden. This patch came from
1339 Daniel Bergström.
1340
1341 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1342 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1343 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1344 spotting this.
1345
1346
1347 Version 7.6 28-Jan-08
1348 ---------------------
1349
1350 1. A character class containing a very large number of characters with
1351 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1352 overflow.
1353
1354 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1355 HAVE_LONG_LONG is not defined.
1356
1357 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1358 bring them up to date and include new features. This patch includes:
1359
1360 - Fixed PH's badly added libz and libbz2 support.
1361 - Fixed a problem with static linking.
1362 - Added pcredemo. [But later removed - see 7 below.]
1363 - Fixed dftables problem and added an option.
1364 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1365 HAVE_LONG_LONG.
1366 - Added readline support for pcretest.
1367 - Added an listing of the option settings after cmake has run.
1368
1369 4. A user submitted a patch to Makefile that makes it easy to create
1370 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1371 Makefile.am that cause it to include this special target, without
1372 affecting anything else. Note that the same mingw target plus all
1373 the other distribution libraries and programs are now supported
1374 when configuring with CMake (see 6 below) instead of with
1375 Configure/Make.
1376
1377 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1378 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1379 exported in the Windows port". It has not yet been confirmed that the patch
1380 solves the problem, but it does no harm.
1381
1382 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1383 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1384 with CMake, and also correct the comment about stack recursion.
1385
1386 7. Remove the automatic building of pcredemo from the ./configure system and
1387 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1388 of a program that users should build themselves after PCRE is installed, so
1389 building it automatically is not really right. What is more, it gave
1390 trouble in some build environments.
1391
1392 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1393
1394
1395 Version 7.5 10-Jan-08
1396 ---------------------
1397
1398 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1399 values in parens when parsing an RE using the C++ wrapper."
1400
1401 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1402 Characters greater than 255 were excluded from the class instead of being
1403 included.
1404
1405 3. The same bug as (2) above applied to negated POSIX classes such as
1406 [:^space:].
1407
1408 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1409 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1410 I have changed it.
1411
1412 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1413 first named subpattern) and a construct such as (?&a) would reference the
1414 first named subpattern whose name started with "a" (in other words, the
1415 length check was missing). Both these problems are fixed. "Subpattern name
1416 expected" is now given for (?&) (a zero-length name), and this patch also
1417 makes it give the same error for \k'' (previously it complained that that
1418 was a reference to a non-existent subpattern).
1419
1420 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1421 this is right because (?- can be followed by option settings as well as by
1422 digits. I have, however, made the messages clearer.
1423
1424 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1425 than the number used in the conditional) now cause a compile-time error.
1426 This is actually not compatible with Perl, which accepts such patterns, but
1427 treats the conditional as always being FALSE (as PCRE used to), but it
1428 seems to me that giving a diagnostic is better.
1429
1430 8. Change "alphameric" to the more common word "alphanumeric" in comments
1431 and messages.
1432
1433 9. Fix two occurrences of "backslash" in comments that should have been
1434 "backspace".
1435
1436 10. Remove two redundant lines of code that can never be obeyed (their function
1437 was moved elsewhere).
1438
1439 11. The program that makes PCRE's Unicode character property table had a bug
1440 which caused it to generate incorrect table entries for sequences of
1441 characters that have the same character type, but are in different scripts.
1442 It amalgamated them into a single range, with the script of the first of
1443 them. In other words, some characters were in the wrong script. There were
1444 thirteen such cases, affecting characters in the following ranges:
1445
1446 U+002b0 - U+002c1
1447 U+0060c - U+0060d
1448 U+0061e - U+00612
1449 U+0064b - U+0065e
1450 U+0074d - U+0076d
1451 U+01800 - U+01805
1452 U+01d00 - U+01d77
1453 U+01d9b - U+01dbf
1454 U+0200b - U+0200f
1455 U+030fc - U+030fe
1456 U+03260 - U+0327f
1457 U+0fb46 - U+0fbb1
1458 U+10450 - U+1049d
1459
1460 12. The -o option (show only the matching part of a line) for pcregrep was not
1461 compatible with GNU grep in that, if there was more than one match in a
1462 line, it showed only the first of them. It now behaves in the same way as
1463 GNU grep.
1464
1465 13. If the -o and -v options were combined for pcregrep, it printed a blank
1466 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1467 does the same. The return code can be used to tell if there were any
1468 non-matching lines.
1469
1470 14. Added --file-offsets and --line-offsets to pcregrep.
1471
1472 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1473 infinitely looping recursion. The bug was that positive lookaheads were not
1474 being skipped when checking for a possible empty match (negative lookaheads
1475 and both kinds of lookbehind were skipped).
1476
1477 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1478 inclusion of <windows.h> to before rather than after the definition of
1479 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1480
1481 17. Specifying a possessive quantifier with a specific limit for a Unicode
1482 character property caused pcre_compile() to compile bad code, which led at
1483 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1484 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1485 caused the error; without that there was no problem.
1486
1487 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1488
1489 19. Added --enable-pcretest-libreadline.
1490
1491 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1492 RE::GlobalReplace(). As a result, the number of replacements returned was
1493 double what it should be. I removed one of the increments, but Craig sent a
1494 later patch that removed the other one (the right fix) and added unit tests
1495 that check the return values (which was not done before).
1496
1497 21. Several CMake things:
1498
1499 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1500 the names libpcre and libpcreposix, not just pcre and pcreposix.
1501
1502 (2) The above change means that pcretest and pcregrep are now correctly
1503 linked with the newly-built libraries, not previously installed ones.
1504
1505 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1506
1507 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1508 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1509 UTF-8 newline character). The key issue is that the pattern starts .*;
1510 this means that the match must be either at the beginning, or after a
1511 newline. The bug was in the code for advancing after a failed match and
1512 checking that the new position followed a newline. It was not taking
1513 account of UTF-8 characters correctly.
1514
1515 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1516 character classes. PCRE was not treating the sequence [:...:] as a
1517 character class unless the ... were all letters. Perl, however, seems to
1518 allow any characters between [: and :], though of course it rejects as
1519 unknown any "names" that contain non-letters, because all the known class
1520 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1521 for example, whereas PCRE did not - it did not recognize a POSIX character
1522 class. This seemed a bit dangerous, so the code has been changed to be
1523 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1524 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1525 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1526 Perl does, and where it didn't before.
1527
1528 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1529 Windows environments %n is disabled by default.
1530
1531
1532 Version 7.4 21-Sep-07
1533 ---------------------
1534
1535 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1536 means that a class such as [\s] counted as "explicit reference to CR or
1537 LF". That isn't really right - the whole point of the change was to try to
1538 help when there was an actual mention of one of the two characters. So now
1539 the change happens only if \r or \n (or a literal CR or LF) character is
1540 encountered.
1541
1542 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1543 of both had grown to the point where there were only 3 bits left.
1544 Fortunately, there was spare space in the data structure, and so I have
1545 moved the internal flags into a new 16-bit field to free up more option
1546 bits.
1547
1548 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1549 but did not set the internal JCHANGED flag - either of these is enough to
1550 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1551 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1552 start sets both bits.
1553
1554 4. Added options (at build time, compile time, exec time) to change \R from
1555 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1556
1557 5. doc/pcresyntax.html was missing from the distribution.
1558
1559 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1560 compatibility, even though it is no longer used.
1561
1562 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1563 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1564 windows.h file is present (where different names are used). [This was
1565 reversed later after testing - see 16 below.]
1566
1567 8. Changed all #include <config.h> to #include "config.h". There were also
1568 some further <pcre.h> cases that I changed to "pcre.h".
1569
1570 9. When pcregrep was used with the --colour option, it missed the line ending
1571 sequence off the lines that it output.
1572
1573 10. It was pointed out to me that arrays of string pointers cause lots of
1574 relocations when a shared library is dynamically loaded. A technique of
1575 using a single long string with a table of offsets can drastically reduce
1576 these. I have refactored PCRE in four places to do this. The result is
1577 dramatic:
1578
1579 Originally: 290
1580 After changing UCP table: 187
1581 After changing error message table: 43
1582 After changing table of "verbs" 36
1583 After changing table of Posix names 22
1584
1585 Thanks to the folks working on Gregex for glib for this insight.
1586
1587 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1588 unicode-properties was also set.
1589
1590 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1591
1592 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1593 checked only for CRLF.
1594
1595 14. Added casts to pcretest.c to avoid compiler warnings.
1596
1597 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1598
1599 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1600 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1601 entirely. This removes changes made in 7 above.
1602
1603 17. The CMake files have been updated, and there is now more information about
1604 building with CMake in the NON-UNIX-USE document.
1605
1606
1607 Version 7.3 28-Aug-07
1608 ---------------------
1609
1610 1. In the rejigging of the build system that eventually resulted in 7.1, the
1611 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1612 brackets there is not right, since it causes compilers to look for an
1613 installed pcre.h, not the version that is in the source that is being
1614 compiled (which of course may be different). I have changed it back to:
1615
1616 #include "pcre.h"
1617
1618 I have a vague recollection that the change was concerned with compiling in
1619 different directories, but in the new build system, that is taken care of
1620 by the VPATH setting the Makefile.
1621
1622 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1623 when the subject happened to end in the byte 0x85 (e.g. if the last
1624 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1625 characters but of course it shouldn't be taken as a newline when it is part
1626 of another character. The bug was that, for an unlimited repeat of . in
1627 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1628 characters when looking for a newline.
1629
1630 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1631
1632 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1633 in debug output.
1634
1635 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1636 long printing in the pcrecpp unittest when running under MinGW.
1637
1638 6. ESC_K was left out of the EBCDIC table.
1639
1640 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1641 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1642 limit also applies to "virtual nesting" when a pattern is recursive, and in
1643 this case 1000 isn't so big. I have been able to remove this limit at the
1644 expense of backing off one optimization in certain circumstances. Normally,
1645 when pcre_exec() would call its internal match() function recursively and
1646 immediately return the result unconditionally, it uses a "tail recursion"
1647 feature to save stack. However, when a subpattern that can match an empty
1648 string has an unlimited repetition quantifier, it no longer makes this
1649 optimization. That gives it a stack frame in which to save the data for
1650 checking that an empty string has been matched. Previously this was taken
1651 from the 1000-entry workspace that had been reserved. So now there is no
1652 explicit limit, but more stack is used.
1653
1654 8. Applied Daniel's patches to solve problems with the import/export magic
1655 syntax that is required for Windows, and which was going wrong for the
1656 pcreposix and pcrecpp parts of the library. These were overlooked when this
1657 problem was solved for the main library.
1658
1659 9. There were some crude static tests to avoid integer overflow when computing
1660 the size of patterns that contain repeated groups with explicit upper
1661 limits. As the maximum quantifier is 65535, the maximum group length was
1662 set at 30,000 so that the product of these two numbers did not overflow a
1663 32-bit integer. However, it turns out that people want to use groups that
1664 are longer than 30,000 bytes (though not repeat them that many times).
1665 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1666 made it possible to implement the integer overflow checks in a much more
1667 dynamic way, which I have now done. The artificial limitation on group
1668 length has been removed - we now have only the limit on the total length of
1669 the compiled pattern, which depends on the LINK_SIZE setting.
1670
1671 10. Fixed a bug in the documentation for get/copy named substring when
1672 duplicate names are permitted. If none of the named substrings are set, the
1673 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1674 empty string.
1675
1676 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1677 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1678 because the ] is interpreted as the first data character and the
1679 terminating ] is not found. PCRE has been made compatible with Perl in this
1680 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1681 cause memory overwriting.
1682
1683 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1684 string has been matched (to stop an infinite loop). It was not recognizing
1685 a conditional subpattern that could match an empty string if that
1686 subpattern was within another subpattern. For example, it looped when
1687 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1688 condition was not nested. This bug has been fixed.
1689
1690 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1691 past the start of the subject in the presence of bytes with the top bit
1692 set, for example "\x8aBCD".
1693
1694 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1695 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1696
1697 14. Optimized (?!) to (*FAIL).
1698
1699 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1700 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1701 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1702 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1703 does: it's just the validity check that is more restrictive.
1704
1705 16. Inserted checks for integer overflows during escape sequence (backslash)
1706 processing, and also fixed erroneous offset values for syntax errors during
1707 backslash processing.
1708
1709 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1710 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1711
1712 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1713 caused an overrun.
1714
1715 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1716 something other than just ASCII characters) inside a group that had an
1717 unlimited repeat caused a loop at compile time (while checking to see
1718 whether the group could match an empty string).
1719
1720 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1721 [\P{Any}] did so. (Error in the code for printing property names.)
1722
1723 21. An orphan \E inside a character class could cause a crash.
1724
1725 22. A repeated capturing bracket such as (A)? could cause a wild memory
1726 reference during compilation.
1727
1728 23. There are several functions in pcre_compile() that scan along a compiled
1729 expression for various reasons (e.g. to see if it's fixed length for look
1730 behind). There were bugs in these functions when a repeated \p or \P was
1731 present in the pattern. These operators have additional parameters compared
1732 with \d, etc, and these were not being taken into account when moving along
1733 the compiled data. Specifically:
1734
1735 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1736 length.
1737
1738 (b) An item such as \pL+ within a repeated group could cause crashes or
1739 loops.
1740
1741 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1742 "reference to non-existent subpattern" error.
1743
1744 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1745
1746 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1747 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1748
1749 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1750
1751 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1752 character were causing crashes (broken optimization).
1753
1754 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1755 \p or \P) caused a compile-time loop.
1756
1757 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1758 break. For example, the unstudied pattern [\r\n]A does not match the string
1759 "\r\nA" because change 7.0/46 below moves the current point on by two
1760 characters after failing to match at the start. However, the pattern \nA
1761 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1762 the same is true. There doesn't seem any very clean way out of this, but
1763 what I have chosen to do makes the common cases work: PCRE now takes note
1764 of whether there can be an explicit match for \r or \n anywhere in the
1765 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1766 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1767 pattern has explicit CR or LF references.
1768
1769 29. Added (*CR) etc for changing newline setting at start of pattern.
1770
1771
1772 Version 7.2 19-Jun-07
1773 ---------------------
1774
1775 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1776 which is apparently normally available under Windows.
1777
1778 2. Re-jig the pcregrep tests with different newline settings in an attempt
1779 to make them independent of the local environment's newline setting.
1780
1781 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1782
1783 4. Some of the "internals" tests were previously cut out when the link size
1784 was not 2, because the output contained actual offsets. The recent new
1785 "Z" feature of pcretest means that these can be cut out, making the tests
1786 usable with all link sizes.
1787
1788 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1789 stack recursion. This gives a massive performance boost under BSD, but just
1790 a small improvement under Linux. However, it saves one field in the frame
1791 in all cases.
1792
1793 6. Added more features from the forthcoming Perl 5.10:
1794
1795 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1796 recursion call. It refers to the nth most recently opened parentheses.
1797
1798 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1799 to be opened parentheses.
1800
1801 (c) Conditions that refer to capturing parentheses can be specified
1802 relatively, for example, (?(-2)... or (?(+3)...
1803
1804 (d) \K resets the start of the current match so that everything before
1805 is not part of it.
1806
1807 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1808
1809 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1810 reference syntax.
1811
1812 (g) (?| introduces a group in which the numbering of parentheses in each
1813 alternative starts with the same number.
1814
1815 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1816
1817 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1818 PCRE_INFO_JCHANGED.
1819
1820 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1821 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1822 for detecting groups that can match an empty string.
1823
1824 9. A pattern with a very large number of alternatives (more than several
1825 hundred) was running out of internal workspace during the pre-compile
1826 phase, where pcre_compile() figures out how much memory will be needed. A
1827 bit of new cunning has reduced the workspace needed for groups with
1828 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1829 workspace instead of running out of the 4096 that are available.
1830
1831 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1832
1833 11. Applied patch from Google to remove an optimization that didn't quite work.
1834 The report of the bug said:
1835
1836 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1837 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1838 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1839
1840 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1841 it matched the wrong number of bytes.
1842
1843
1844 Version 7.1 24-Apr-07
1845 ---------------------
1846
1847 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1848 that is more "standard", making use of automake and other Autotools. There
1849 is some re-arrangement of the files and adjustment of comments consequent
1850 on this.
1851
1852 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1853 for recursive directory scanning broke on some systems because the files
1854 are not scanned in any specific order and on different systems the order
1855 was different. A call to "sort" has been inserted into RunGrepTest for the
1856 approprate test as a short-term fix. In the longer term there may be an
1857 alternative.
1858
1859 3. I had an email from Eric Raymond about problems translating some of PCRE's
1860 man pages to HTML (despite the fact that I distribute HTML pages, some
1861 people do their own conversions for various reasons). The problems
1862 concerned the use of low-level troff macros .br and .in. I have therefore
1863 removed all such uses from the man pages (some were redundant, some could
1864 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1865 HTML has been updated to handle .nf/.fi and to complain if it encounters
1866 .br or .in.
1867
1868 4. Updated comments in configure.ac that get placed in config.h.in and also
1869 arranged for config.h to be included in the distribution, with the name
1870 config.h.generic, for the benefit of those who have to compile without
1871 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1872
1873 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1874 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1875 makevp.bat for the current PCRE, using the additional files
1876 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1877
1878 6. A Windows user reported a minor discrepancy with test 2, which turned out
1879 to be caused by a trailing space on an input line that had got lost in his
1880 copy. The trailing space was an accident, so I've just removed it.
1881
1882 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1883 that is needed.
1884
1885 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1886 as "const" (a) because they are and (b) because it helps the PHP
1887 maintainers who have recently made a script to detect big data structures
1888 in the php code that should be moved to the .rodata section. I remembered
1889 to update Builducptable as well, so it won't revert if ucptable.h is ever
1890 re-created.
1891
1892 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1893 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1894 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1895 support is not required. This saves 1.5-2K of code, which is important in
1896 some applications.
1897
1898 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1899 so as not to refer to the tables, even though these functions will never be
1900 called when UTF-8 support is disabled. Otherwise there are problems with a
1901 shared library.
1902
1903 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1904
1905 (a) It was defining its arguments as char * instead of void *.
1906
1907 (b) It was assuming that all moves were upwards in memory; this was true
1908 a long time ago when I wrote it, but is no longer the case.
1909
1910 The emulated memove() is provided for those environments that have neither
1911 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1912 is clearly not the case, as these two bugs were recently reported.
1913
1914 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1915 and Detrail to create the HTML documentation, the .txt form of the man
1916 pages, and it removes trailing spaces from listed files. It also creates
1917 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1918 case, it wraps all the #defines with #ifndefs. This script should be run
1919 before "make dist".
1920
1921 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1922 with Unicode property support.
1923
1924 (a) For a maximizing quantifier, if the two different cases of the
1925 character were of different lengths in their UTF-8 codings (there are
1926 some cases like this - I found 11), and the matching function had to
1927 back up over a mixture of the two cases, it incorrectly assumed they
1928 were both the same length.
1929
1930 (b) When PCRE was configured to use the heap rather than the stack for
1931 recursion during matching, it was not correctly preserving the data for
1932 the other case of a UTF-8 character when checking ahead for a match
1933 while processing a minimizing repeat. If the check also involved
1934 matching a wide character, but failed, corruption could cause an
1935 erroneous result when trying to check for a repeat of the original
1936 character.
1937
1938 13. Some tidying changes to the testing mechanism:
1939
1940 (a) The RunTest script now detects the internal link size and whether there
1941 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1942 values substituted by "configure". (The RunGrepTest script already did
1943 this for UTF-8.) The configure.ac script no longer substitutes the
1944 relevant variables.
1945
1946 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1947 with length and offset values. This means that the output is different
1948 for different internal link sizes. Test 2 is skipped for link sizes
1949 other than 2 because of this, bypassing the problem. Unfortunately,
1950 there was also a test in test 3 (the locale tests) that used /B and
1951 failed for link sizes other than 2. Rather than cut the whole test out,
1952 I have added a new /Z option to pcretest that replaces the length and
1953 offset values with spaces. This is now used to make test 3 independent
1954 of link size. (Test 2 will be tidied up later.)
1955
1956 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1957 segmentation fault instead of returning the appropriate error message.
1958
1959 15. In multiline mode when the newline sequence was set to "any", the pattern
1960 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1961 This doesn't seem right; it now treats the CRLF combination as the line
1962 ending, and so does not match in that case. It's only a pattern such as ^$
1963 that would hit this one: something like ^ABC$ would have failed after \r
1964 and then tried again after \r\n.
1965
1966 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1967 in an attempt to make files that differ only in their line terminators
1968 compare equal. This works on Linux.
1969
1970 17. Under certain error circumstances pcregrep might try to free random memory
1971 as it exited. This is now fixed, thanks to valgrind.
1972
1973 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1974 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1975 was because its rules for how to advance for /g after matching an empty
1976 string at the end of a line did not allow for this case. They now check for
1977 it specially.
1978
1979 20. pcretest is supposed to handle patterns and data of any length, by
1980 extending its buffers when necessary. It was getting this wrong when the
1981 buffer for a data line had to be extended.
1982
1983 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1984 CRLF as a newline sequence.
1985
1986 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1987 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1988 I have nevertheless tidied it up.
1989
1990 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1991
1992 24. Added a man page for pcre-config.
1993
1994
1995 Version 7.0 19-Dec-06
1996 ---------------------
1997
1998 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1999 moving to gcc 4.1.1.
2000
2001 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2002 sys/time.h, which is documented as needed for this function. It doesn't
2003 seem to matter on Linux, but it showed up on some releases of OS X.
2004
2005 3. It seems that there are systems where bytes whose values are greater than
2006 127 match isprint() in the "C" locale. The "C" locale should be the
2007 default when a C program starts up. In most systems, only ASCII printing
2008 characters match isprint(). This difference caused the output from pcretest
2009 to vary, making some of the tests fail. I have changed pcretest so that:
2010
2011 (a) When it is outputting text in the compiled version of a pattern, bytes
2012 other than 32-126 are always shown as hex escapes.
2013
2014 (b) When it is outputting text that is a matched part of a subject string,
2015 it does the same, unless a different locale has been set for the match
2016 (using the /L modifier). In this case, it uses isprint() to decide.
2017
2018 4. Fixed a major bug that caused incorrect computation of the amount of memory
2019 required for a compiled pattern when options that changed within the
2020 pattern affected the logic of the preliminary scan that determines the
2021 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2022 that the computed length was too small. The symptoms of this bug were
2023 either the PCRE error "internal error: code overflow" from pcre_compile(),
2024 or a glibc crash with a message such as "pcretest: free(): invalid next
2025 size (fast)". Examples of patterns that provoked this bug (shown in
2026 pcretest format) are:
2027
2028 /(?-x: )/x
2029 /(?x)(?-x: \s*#\s*)/
2030 /((?i)[\x{c0}])/8
2031 /(?i:[\x{c0}])/8
2032
2033 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2034 is now done differently.
2035
2036 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2037 wrapper classes; (b) implement a new function in the C++ scanner that is
2038 more efficient than the old way of doing things because it avoids levels of
2039 recursion in the regex matching; (c) add a paragraph to the documentation
2040 for the FullMatch() function.
2041
2042 6. The escape sequence \n was being treated as whatever was defined as
2043 "newline". Not only was this contrary to the documentation, which states
2044 that \n is character 10 (hex 0A), but it also went horribly wrong when
2045 "newline" was defined as CRLF. This has been fixed.
2046
2047 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2048 was being set to -1 for the "end of line" case (supposedly a value that no
2049 character can have). Though this value is never used (the check for end of
2050 line is "zero bytes in current character"), it caused compiler complaints.
2051 I've changed it to 0xffffffff.
2052
2053 8. In pcre_version.c, the version string was being built by a sequence of
2054 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2055 string (as it is for production releases) called a macro with an empty
2056 argument. The C standard says the result of this is undefined. The gcc
2057 compiler treats it as an empty string (which was what was wanted) but it is
2058 reported that Visual C gives an error. The source has been hacked around to
2059 avoid this problem.
2060
2061 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2062 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2063 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2064 of them did).
2065
2066 10. Originally, pcretest opened its input and output without "b"; then I was
2067 told that "b" was needed in some environments, so it was added for release
2068 5.0 to both the input and output. (It makes no difference on Unix-like
2069 systems.) Later I was told that it is wrong for the input on Windows. I've
2070 now abstracted the modes into two macros, to make it easier to fiddle with
2071 them, and removed "b" from the input mode under Windows.
2072
2073 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2074
2075 12. Added -help and --help to pcretest as an official way of being reminded
2076 of the options.
2077
2078 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2079 and pcrecpp.cc because they annoy compilers at high warning levels.
2080
2081 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2082
2083 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2084 scripts are not C programs :-) and which was not noticed because it works
2085 on Linux.
2086
2087 16. pcretest is supposed to handle any length of pattern and data line (as one
2088 line or as a continued sequence of lines) by extending its input buffer if
2089 necessary. This feature was broken for very long pattern lines, leading to
2090 a string of junk being passed to pcre_compile() if the pattern was longer
2091 than about 50K.
2092
2093 17. I have done a major re-factoring of the way pcre_compile() computes the
2094 amount of memory needed for a compiled pattern. Previously, there was code
2095 that made a preliminary scan of the pattern in order to do this. That was
2096 OK when PCRE was new, but as the facilities have expanded, it has become
2097 harder and harder to keep it in step with the real compile phase, and there
2098 have been a number of bugs (see for example, 4 above). I have now found a
2099 cunning way of running the real compile function in a "fake" mode that
2100 enables it to compute how much memory it would need, while actually only
2101 ever using a few hundred bytes of working memory and without too many
2102 tests of the mode. This should make future maintenance and development
2103 easier. A side effect of this work is that the limit of 200 on the nesting
2104 depth of parentheses has been removed (though this was never a serious
2105 limitation, I suspect). However, there is a downside: pcre_compile() now
2106 runs more slowly than before (30% or more, depending on the pattern). I
2107 hope this isn't a big issue. There is no effect on runtime performance.
2108
2109 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2110 newline (only possible for the last line of a file) and it was a
2111 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2112
2113 19. Added additional timing features to pcretest. (1) The -tm option now times
2114 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2115 separate command line item, by a number that specifies the number of
2116 repeats to use when timing. The default is 50000; this gives better
2117 precision, but takes uncomfortably long for very large patterns.
2118
2119 20. Extended pcre_study() to be more clever in cases where a branch of a
2120 subpattern has no definite first character. For example, (a*|b*)[cd] would
2121 previously give no result from pcre_study(). Now it recognizes that the
2122 first character must be a, b, c, or d.
2123
2124 21. There was an incorrect error "recursive call could loop indefinitely" if
2125 a subpattern (or the entire pattern) that was being tested for matching an
2126 empty string contained only one non-empty item after a nested subpattern.
2127 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2128 incorrectly, because the \d was being skipped in the check.
2129
2130 22. The pcretest program now has a new pattern option /B and a command line
2131 option -b, which is equivalent to adding /B to every pattern. This causes
2132 it to show the compiled bytecode, without the additional information that
2133 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2134 is the same as /B/I).
2135
2136 23. A new optimization is now able automatically to treat some sequences such
2137 as a*b as a*+b. More specifically, if something simple (such as a character
2138 or a simple class like \d) has an unlimited quantifier, and is followed by
2139 something that cannot possibly match the quantified thing, the quantifier
2140 is automatically "possessified".
2141
2142 24. A recursive reference to a subpattern whose number was greater than 39
2143 went wrong under certain circumstances in UTF-8 mode. This bug could also
2144 have affected the operation of pcre_study().
2145
2146 25. Realized that a little bit of performance could be had by replacing
2147 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2148
2149 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2150
2151 27. Possessive quantifiers such as a++ were previously implemented by turning
2152 them into atomic groups such as ($>a+). Now they have their own opcodes,
2153 which improves performance. This includes the automatically created ones
2154 from 23 above.
2155
2156 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2157 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2158 the first matched character to be a colon. This applied both to named and
2159 numbered groups.
2160
2161 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2162
2163 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2164 building PCRE on Windows easier, so I have included it in the distribution.
2165
2166 31. There is now a check in pcretest against a ridiculously large number being
2167 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2168 loop, the loop is abandoned.
2169
2170 32. Forward references to subpatterns in conditions such as (?(2)...) where
2171 subpattern 2 is defined later cause pcre_compile() to search forwards in
2172 the pattern for the relevant set of parentheses. This search went wrong
2173 when there were unescaped parentheses in a character class, parentheses
2174 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2175
2176 33. "Subroutine" calls and backreferences were previously restricted to
2177 referencing subpatterns earlier in the regex. This restriction has now
2178 been removed.
2179
2180 34. Added a number of extra features that are going to be in Perl 5.10. On the
2181 whole, these are just syntactic alternatives for features that PCRE had
2182 previously implemented using the Python syntax or my own invention. The
2183 other formats are all retained for compatibility.
2184
2185 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2186 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2187 also .NET compatible.
2188
2189 (b) A recursion or subroutine call to a named group can now be defined as
2190 (?&name) as well as (?P>name).
2191
2192 (c) A backreference to a named group can now be defined as \k<name> or
2193 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2194 5.10, are also .NET compatible.
2195
2196 (d) A conditional reference to a named group can now use the syntax
2197 (?(<name>) or (?('name') as well as (?(name).
2198
2199 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2200 groups (named and numbered) that are never evaluated inline, but can be
2201 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2202 is always false. There may be only one alternative in such a group.
2203
2204 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2205 as the simple (?(R). The condition is true only if the most recent
2206 recursion is that of the given number or name. It does not search out
2207 through the entire recursion stack.
2208
2209 (g) The escape \gN or \g{N} has been added, where N is a positive or
2210 negative number, specifying an absolute or relative reference.
2211
2212 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2213 some "unreachable code" warnings.
2214
2215 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2216 things, this adds five new scripts.
2217
2218 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2219 There were also incompatibilities regarding the handling of \Q..\E inside
2220 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2221 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2222
2223 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2224 matches an empty string, and forcibly breaks the loop. There were bugs in
2225 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2226 against aaaa the result was just "a" rather than "aaaa", for example. Two
2227 separate and independent bugs (that affected different cases) have been
2228 fixed.
2229
2230 39. Refactored the code to abolish the use of different opcodes for small
2231 capturing bracket numbers. This is a tidy that I avoided doing when I
2232 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2233 The new approach is not only tidier, it makes it possible to reduce the
2234 memory needed to fix the previous bug (38).
2235
2236 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2237 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2238 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2239 mode.
2240
2241 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2242 report.
2243
2244 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2245 copy construction and assignment in the C++ wrapper.
2246
2247 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2248 couple of bugs that could have given wrong results in the "--newline=crlf"
2249 case.
2250
2251 44. Added a number of casts and did some reorganization of signed/unsigned int
2252 variables following suggestions from Dair Grant. Also renamed the variable
2253 "this" as "item" because it is a C++ keyword.
2254
2255 45. Arranged for dftables to add
2256
2257 #include "pcre_internal.h"
2258
2259 to pcre_chartables.c because without it, gcc 4.x may remove the array
2260 definition from the final binary if PCRE is built into a static library and
2261 dead code stripping is activated.
2262
2263 46. For an unanchored pattern, if a match attempt fails at the start of a
2264 newline sequence, and the newline setting is CRLF or ANY, and the next two
2265 characters are CRLF, advance by two characters instead of one.
2266
2267
2268 Version 6.7 04-Jul-06
2269 ---------------------
2270
2271 1. In order to handle tests when input lines are enormously long, pcretest has
2272 been re-factored so that it automatically extends its buffers when
2273 necessary. The code is crude, but this _is_ just a test program. The
2274 default size has been increased from 32K to 50K.
2275
2276 2. The code in pcre_study() was using the value of the re argument before
2277 testing it for NULL. (Of course, in any sensible call of the function, it
2278 won't be NULL.)
2279
2280 3. The memmove() emulation function in pcre_internal.h, which is used on
2281 systems that lack both memmove() and bcopy() - that is, hardly ever -
2282 was missing a "static" storage class specifier.
2283
2284 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2285 containing an extended class (one that cannot be represented by a bitmap
2286 because it contains high-valued characters or Unicode property items, e.g.
2287 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2288 pattern, but PCRE should not loop if you do not (it no longer does).
2289 [Detail: two cases were found: (a) a repeated subpattern containing an
2290 extended class; (b) a recursive reference to a subpattern that followed a
2291 previous extended class. It wasn't skipping over the extended class
2292 correctly when UTF-8 mode was not set.]
2293
2294 5. A negated single-character class was not being recognized as fixed-length
2295 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2296 compile error "lookbehind assertion is not fixed length".
2297
2298 6. The RunPerlTest auxiliary script was showing an unexpected difference
2299 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2300 write a Perl script that can interpret lines of an input file either as
2301 byte characters or as UTF-8, which is what "perltest" was being required to
2302 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2303 can't do is switch easily at run time between having the "use utf8;" pragma
2304 or not. In the end, I fudged it by using the RunPerlTest script to insert
2305 "use utf8;" explicitly for the UTF-8 tests.
2306
2307 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2308 the end of the subject string, contrary to the documentation and to what
2309 Perl does. This was true of both matching functions. Now it matches only at
2310 the start of the subject and immediately after *internal* newlines.
2311
2312 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2313 a pointer to an int instead of a pointer to an unsigned long int. This
2314 caused problems on 64-bit systems.
2315
2316 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2317 instance of the 'standard' template library not being so standard".
2318
2319 10. There was no check on the number of named subpatterns nor the maximum
2320 length of a subpattern name. The product of these values is used to compute
2321 the size of the memory block for a compiled pattern. By supplying a very
2322 long subpattern name and a large number of named subpatterns, the size
2323 computation could be caused to overflow. This is now prevented by limiting
2324 the length of names to 32 characters, and the number of named subpatterns
2325 to 10,000.
2326
2327 11. Subpatterns that are repeated with specific counts have to be replicated in
2328 the compiled pattern. The size of memory for this was computed from the
2329 length of the subpattern and the repeat count. The latter is limited to
2330 65535, but there was no limit on the former, meaning that integer overflow
2331 could in principle occur. The compiled length of a repeated subpattern is
2332 now limited to 30,000 bytes in order to prevent this.
2333
2334 12. Added the optional facility to have named substrings with the same name.
2335
2336 13. Added the ability to use a named substring as a condition, using the
2337 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2338 are numbers (not recommended). Forward references are permitted.
2339
2340 14. Added forward references in named backreferences (if you see what I mean).
2341
2342 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2343 pattern could run off the end of the subject. For example, the pattern
2344 "(?s)(.{1,5})"8 did this with the subject "ab".
2345
2346 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2347 PCRE_CASELESS was set when matching characters that were quantified with ?
2348 or *.
2349
2350 17. A character class other than a single negated character that had a minimum
2351 but no maximum quantifier - for example [ab]{6,} - was not handled
2352 correctly by pce_dfa_exec(). It would match only one character.
2353
2354 18. A valid (though odd) pattern that looked like a POSIX character
2355 class but used an invalid character after [ (for example [[,abc,]]) caused
2356 pcre_compile() to give the error "Failed: internal error: code overflow" or
2357 in some cases to crash with a glibc free() error. This could even happen if
2358 the pattern terminated after [[ but there just happened to be a sequence of
2359 letters, a binary zero, and a closing ] in the memory that followed.
2360
2361 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2362 over the years. Originally (before any Unicode support), just the bottom 8
2363 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2364 output from "man perlunicode" includes this:
2365
2366 The regular expression compiler produces polymorphic opcodes. That
2367 is, the pattern adapts to the data and automatically switches to
2368 the Unicode character scheme when presented with Unicode data--or
2369 instead uses a traditional byte scheme when presented with byte
2370 data.
2371
2372 Sadly, a wide octal escape does not cause a switch, and in a string with
2373 no other multibyte characters, these octal escapes are treated as before.
2374 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2375 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2376 Unicode string.
2377
2378 I have not perpetrated such confusion in PCRE. Up till now, it took just
2379 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2380 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2381 translate to the appropriate multibyte character.
2382
2383 29. Applied some refactoring to reduce the number of warnings from Microsoft
2384 and Borland compilers. This has included removing the fudge introduced
2385 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2386 a warning about an unused variable.
2387
2388 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2389 characters since release 4.0, because Perl (from release 5.004) does not.
2390 [Or at least, is documented not to: some releases seem to be in conflict
2391 with the documentation.] However, when a pattern was studied with
2392 pcre_study() and all its branches started with \s, PCRE still included VT
2393 as a possible starting character. Of course, this did no harm; it just
2394 caused an unnecessary match attempt.
2395
2396 22. Removed a now-redundant internal flag bit that recorded the fact that case
2397 dependency changed within the pattern. This was once needed for "required
2398 byte" processing, but is no longer used. This recovers a now-scarce options
2399 bit. Also moved the least significant internal flag bit to the most-
2400 significant bit of the word, which was not previously used (hangover from
2401 the days when it was an int rather than a uint) to free up another bit for
2402 the future.
2403
2404 23. Added support for CRLF line endings as well as CR and LF. As well as the
2405 default being selectable at build time, it can now be changed at runtime
2406 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2407 specify that it is scanning data with non-default line endings.
2408
2409 24. Changed the definition of CXXLINK to make it agree with the definition of
2410 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2411
2412 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2413 recursions. This makes a big different to stack usage for some patterns.
2414
2415 26. If a subpattern containing a named recursion or subroutine reference such
2416 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2417 the space required for the compiled pattern went wrong and gave too small a
2418 value. Depending on the environment, this could lead to "Failed: internal
2419 error: code overflow at offset 49" or "glibc detected double free or
2420 corruption" errors.
2421
2422 27. Applied patches from Google (a) to support the new newline modes and (b) to
2423 advance over multibyte UTF-8 characters in GlobalReplace.
2424
2425 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2426 difference for some implementation of PCRE in some Windows version.
2427
2428 29. Added some extra testing facilities to pcretest:
2429
2430 \q<number> in a data line sets the "match limit" value
2431 \Q<number> in a data line sets the "match recursion limt" value
2432 -S <number> sets the stack size, where <number> is in megabytes
2433
2434 The -S option isn't available for Windows.
2435
2436
2437 Version 6.6 06-Feb-06
2438 ---------------------
2439
2440 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2441 in pcreposix.h. I have copied the definition from pcre.h.
2442
2443 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2444 because pcre.h is no longer a built file.
2445
2446 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2447 not normally included in the compiled code.
2448
2449
2450 Version 6.5 01-Feb-06
2451 ---------------------
2452
2453 1. When using the partial match feature with pcre_dfa_exec(), it was not
2454 anchoring the second and subsequent partial matches at the new starting
2455 point. This could lead to incorrect results. For example, with the pattern
2456 /1234/, partially matching against "123" and then "a4" gave a match.
2457
2458 2. Changes to pcregrep:
2459
2460 (a) All non-match returns from pcre_exec() were being treated as failures
2461 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2462 error message is output. Some extra information is given for the
2463 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2464 probably the only errors that are likely to be caused by users (by
2465 specifying a regex that has nested indefinite repeats, for instance).
2466 If there are more than 20 of these errors, pcregrep is abandoned.
2467
2468 (b) A binary zero was treated as data while matching, but terminated the
2469 output line if it was written out. This has been fixed: binary zeroes
2470 are now no different to any other data bytes.
2471
2472 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2473 used to set a locale for matching. The --locale=xxxx long option has
2474 been added (no short equivalent) to specify a locale explicitly on the
2475 pcregrep command, overriding the environment variables.
2476
2477 (d) When -B was used with -n, some line numbers in the output were one less
2478 than they should have been.
2479
2480 (e) Added the -o (--only-matching) option.
2481
2482 (f) If -A or -C was used with -c (count only), some lines of context were
2483 accidentally printed for the final match.
2484
2485 (g) Added the -H (--with-filename) option.
2486
2487 (h) The combination of options -rh failed to suppress file names for files
2488 that were found from directory arguments.
2489
2490 (i) Added the -D (--devices) and -d (--directories) options.
2491
2492 (j) Added the -F (--fixed-strings) option.
2493
2494 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2495
2496 (l) Added the --colo(u)r option.
2497
2498 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2499 is not present by default.
2500
2501 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2502 items such as (?R) or (?1), when the recursion could match a number of
2503 alternatives. If it matched one of the alternatives, but subsequently,
2504 outside the recursion, there was a failure, the code tried to back up into
2505 the recursion. However, because of the way PCRE is implemented, this is not
2506 possible, and the result was an incorrect result from the match.
2507
2508 In order to prevent this happening, the specification of recursion has
2509 been changed so that all such subpatterns are automatically treated as
2510 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2511
2512 4. I had overlooked the fact that, in some locales, there are characters for
2513 which isalpha() is true but neither isupper() nor islower() are true. In
2514 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2515 and ordfeminine) are like this. This affected the treatment of \w and \W
2516 when they appeared in character classes, but not when they appeared outside
2517 a character class. The bit map for "word" characters is now created
2518 separately from the results of isalnum() instead of just taking it from the
2519 upper, lower, and digit maps. (Plus the underscore character, of course.)
2520
2521 5. The above bug also affected the handling of POSIX character classes such as
2522 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2523 permanent tables. Instead, the bit maps for such a class were previously
2524 created as the appropriate unions of the upper, lower, and digit bitmaps.
2525 Now they are created by subtraction from the [[:word:]] class, which has
2526 its own bitmap.
2527
2528 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2529 It is created by subtracting the vertical space characters (\x09, \x0a,
2530 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2531 subtraction was done in the overall bitmap for a character class, meaning
2532 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2533 be recognized. This bug has been fixed.
2534
2535 7. Patches from the folks at Google:
2536
2537 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2538 real life, but is still worth protecting against".
2539
2540 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2541 regular expressions".
2542
2543 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2544 have it.
2545
2546 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2547 "configure" and the latter not, in order to fix a problem somebody had
2548 with compiling the Arg class on HP-UX.
2549
2550 (e) Improve the error-handling of the C++ wrapper a little bit.
2551
2552 (f) New tests for checking recursion limiting.
2553
2554 8. The pcre_memmove() function, which is used only if the environment does not
2555 have a standard memmove() function (and is therefore rarely compiled),
2556 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2557 returning a result (though PCRE never actually uses the result).
2558
2559 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2560 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2561 returned instead of calling malloc() with an overflowing number that would
2562 most likely cause subsequent chaos.
2563
2564 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2565
2566 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2567 with this option is matched, the nmatch and pmatch options of regexec() are
2568 ignored.
2569
2570 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2571 provided in case anyone wants to the the POSIX interface with UTF-8
2572 strings.
2573
2574 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2575 C++ linking (needed for some HP-UX environments).
2576
2577 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2578 (unused parameter) and in the pcre_printint() function (omitted "default"
2579 switch label when the default is to do nothing).
2580
2581 15. Added some code to make it possible, when PCRE is compiled as a C++
2582 library, to replace subject pointers for pcre_exec() with a smart pointer
2583 class, thus making it possible to process discontinuous strings.
2584
2585 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2586 much the same function. They were added by different people who were trying
2587 to make PCRE easy to compile on non-Unix systems. It has been suggested
2588 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2589 for compiling on Windows systems. I have therefore replaced it with
2590 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2591 defaults to "extern" for C or "extern C" for C++, which works fine on
2592 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2593 SCOPE with something explicit in config.h. In addition:
2594
2595 (a) pcreposix.h still had just "extern" instead of either of these macros;
2596 I have replaced it with PCRE_DATA_SCOPE.
2597
2598 (b) Functions such as _pcre_xclass(), which are internal to the library,
2599 but external in the C sense, all had PCRE_EXPORT in their definitions.
2600 This is apparently wrong for the Windows case, so I have removed it.
2601 (It makes no difference on Unix-like systems.)
2602
2603 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2604 of recursive calls to match(). This is different to MATCH_LIMIT because
2605 that limits the total number of calls to match(), not all of which increase
2606 the depth of recursion. Limiting the recursion depth limits the amount of
2607 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2608 when PCRE is compiled, and changed at run time. A patch from Google adds
2609 this functionality to the C++ interface.
2610
2611 18. Changes to the handling of Unicode character properties:
2612
2613 (a) Updated the table to Unicode 4.1.0.
2614
2615 (b) Recognize characters that are not in the table as "Cn" (undefined).
2616
2617 (c) I revised the way the table is implemented to a much improved format
2618 which includes recognition of ranges. It now supports the ranges that
2619 are defined in UnicodeData.txt, and it also amalgamates other
2620 characters into ranges. This has reduced the number of entries in the
2621 table from around 16,000 to around 3,000, thus reducing its size
2622 considerably. I realized I did not need to use a tree structure after
2623 all - a binary chop search is just as efficient. Having reduced the
2624 number of entries, I extended their size from 6 bytes to 8 bytes to
2625 allow for more data.
2626
2627 (d) Added support for Unicode script names via properties such as \p{Han}.
2628
2629 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2630 matching that character.
2631
2632 20. When matching a repeated Unicode property with a minimum greater than zero,
2633 (for example \pL{2,}), PCRE could look past the end of the subject if it
2634 reached it while seeking the minimum number of characters. This could
2635 happen only if some of the characters were more than one byte long, because
2636 there is a check for at least the minimum number of bytes.
2637
2638 21. Refactored the implementation of \p and \P so as to be more general, to
2639 allow for more different types of property in future. This has changed the
2640 compiled form incompatibly. Anybody with saved compiled patterns that use
2641 \p or \P will have to recompile them.
2642
2643 22. Added "Any" and "L&" to the supported property types.
2644
2645 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2646 but give a compile time error if the value is greater than 0xff.
2647
2648 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2649 accidentally not being installed or uninstalled.
2650
2651 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2652 made were to insert the current release number. This seemed silly, because
2653 it made things harder for people building PCRE on systems that don't run
2654 "configure". I have turned pcre.h into a distributed file, no longer built
2655 by "configure", with the version identification directly included. There is
2656 no longer a pcre.h.in file.
2657
2658 However, this change necessitated a change to the pcre-config script as
2659 well. It is built from pcre-config.in, and one of the substitutions was the
2660 release number. I have updated configure.ac so that ./configure now finds
2661 the release number by grepping pcre.h.
2662
2663 26. Added the ability to run the tests under valgrind.
2664
2665
2666 Version 6.4 05-Sep-05
2667 ---------------------
2668
2669 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2670 "--" to be printed when multiple files were scanned, even when none of the
2671 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2672 consider it to be a bug, and have restored the previous behaviour.
2673
2674 2. A couple of code tidies to get rid of compiler warnings.
2675
2676 3. The pcretest program used to cheat by referring to symbols in the library
2677 whose names begin with _pcre_. These are internal symbols that are not
2678 really supposed to be visible externally, and in some environments it is
2679 possible to suppress them. The cheating is now confined to including
2680 certain files from the library's source, which is a bit cleaner.
2681
2682 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2683 file's purpose clearer.
2684
2685 5. Reorganized pcre_ucp_findchar().
2686
2687
2688 Version 6.3 15-Aug-05
2689 ---------------------
2690
2691 1. The file libpcre.pc.in did not have general read permission in the tarball.
2692
2693 2. There were some problems when building without C++ support:
2694
2695 (a) If C++ support was not built, "make install" and "make test" still
2696 tried to test it.
2697
2698 (b) There were problems when the value of CXX was explicitly set. Some
2699 changes have been made to try to fix these, and ...
2700
2701 (c) --disable-cpp can now be used to explicitly disable C++ support.
2702
2703 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2704 backslash in a target when C++ was disabled. This confuses some
2705 versions of "make", apparently. Using an intermediate variable solves
2706 this. (Same for CPP_LOBJ.)
2707
2708 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2709 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2710 necessary on certain architectures.
2711
2712 4. Added a setting of -export-symbols-regex to the link command to remove
2713 those symbols that are exported in the C sense, but actually are local
2714 within the library, and not documented. Their names all begin with
2715 "_pcre_". This is not a perfect job, because (a) we have to except some
2716 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2717 available (and never for static libraries). I have made a note to try to
2718 find a way round (a) in the future.
2719
2720
2721 Version 6.2 01-Aug-05
2722 ---------------------
2723
2724 1. There was no test for integer overflow of quantifier values. A construction
2725 such as {1111111111111111} would give undefined results. What is worse, if
2726 a minimum quantifier for a parenthesized subpattern overflowed and became
2727 negative, the calculation of the memory size went wrong. This could have
2728 led to memory overwriting.
2729
2730 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2731
2732 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2733 operating environments where this matters.
2734
2735 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2736 PCRE options from within the C++ wrapper.
2737
2738 5. Named capturing subpatterns were not being correctly counted when a pattern
2739 was compiled. This caused two problems: (a) If there were more than 100
2740 such subpatterns, the calculation of the memory needed for the whole
2741 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2742 back references of the form \12, where the number was greater than 9, were
2743 not recognized as back references, even though there were sufficient
2744 previous subpatterns.
2745
2746 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2747 versions of gcc, e.g. 2.95.4.
2748
2749
2750 Version 6.1 21-Jun-05
2751 ---------------------
2752
2753 1. There was one reference to the variable "posix" in pcretest.c that was not
2754 surrounded by "#if !defined NOPOSIX".
2755
2756 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2757 the cross-check on the old pcre_info() function, for the benefit of the
2758 cut-down version of PCRE that is currently imported into Exim.
2759
2760 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2761 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2762 patterns that start (?i) but I don't think that matters. The use of (?i) is
2763 just an example; this all applies to the other options as well.
2764
2765 4. Since libtool seems to echo the compile commands it is issuing, the output
2766 from "make" can be reduced a bit by putting "@" in front of each libtool
2767 compile command.
2768
2769 5. Patch from the folks at Google for configure.in to be a bit more thorough
2770 in checking for a suitable C++ installation before trying to compile the
2771 C++ stuff. This should fix a reported problem when a compiler was present,
2772 but no suitable headers.
2773
2774 6. The man pages all had just "PCRE" as their title. I have changed them to
2775 be the relevant file name. I have also arranged that these names are
2776 retained in the file doc/pcre.txt, which is a concatenation in text format
2777 of all the man pages except the little individual ones for each function.
2778
2779 7. The NON-UNIX-USE file had not been updated for the different set of source
2780 files that come with release 6. I also added a few comments about the C++
2781 wrapper.
2782
2783
2784 Version 6.0 07-Jun-05
2785 ---------------------
2786
2787 1. Some minor internal re-organization to help with my DFA experiments.
2788
2789 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2790 didn't matter for the library itself when fully configured, but did matter
2791 when compiling without UCP support, or within Exim, where the ucp files are
2792 not imported.
2793
2794 3. Refactoring of the library code to split up the various functions into
2795 different source modules. The addition of the new DFA matching code (see
2796 below) to a single monolithic source would have made it really too
2797 unwieldy, quite apart from causing all the code to be include in a
2798 statically linked application, when only some functions are used. This is
2799 relevant even without the DFA addition now that patterns can be compiled in
2800 one application and matched in another.
2801
2802 The downside of splitting up is that there have to be some external
2803 functions and data tables that are used internally in different modules of
2804 the library but which are not part of the API. These have all had their
2805 names changed to start with "_pcre_" so that they are unlikely to clash
2806 with other external names.
2807
2808 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2809 a different (DFA) algorithm. Although it is slower than the original
2810 function, it does have some advantages for certain types of matching
2811 problem.
2812
2813 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2814 including restarting after a partial match.
2815
2816 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2817 defined when compiling for Windows was sent to me. I have put it into the
2818 code, though I have no means of testing or verifying it.
2819
2820 7. Added the pcre_refcount() auxiliary function.
2821
2822 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2823 match before or at the first newline in the subject string. In pcretest,
2824 the /f option on a pattern can be used to set this.
2825
2826 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2827 would behave wrongly. This has been present in PCRE since release 4.0.
2828
2829 10. A number of changes to the pcregrep command:
2830
2831 (a) Refactored how -x works; insert ^(...)$ instead of setting
2832 PCRE_ANCHORED and checking the length, in preparation for adding
2833 something similar for -w.
2834
2835 (b) Added the -w (match as a word) option.
2836
2837 (c) Refactored the way lines are read and buffered so as to have more
2838 than one at a time available.
2839
2840 (d) Implemented a pcregrep test script.
2841
2842 (e) Added the -M (multiline match) option. This allows patterns to match
2843 over several lines of the subject. The buffering ensures that at least
2844 8K, or the rest of the document (whichever is the shorter) is available
2845 for matching (and similarly the previous 8K for lookbehind assertions).
2846
2847 (f) Changed the --help output so that it now says
2848
2849 -w, --word-regex(p)
2850
2851 instead of two lines, one with "regex" and the other with "regexp"
2852 because that confused at least one person since the short forms are the
2853 same. (This required a bit of code, as the output is generated
2854 automatically from a table. It wasn't just a text change.)
2855
2856 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2857 option but starts with a hyphen. Could be a pattern or a path name
2858 starting with a hyphen, for instance.
2859
2860 (h) "-" can be given as a file name to represent stdin.
2861
2862 (i) When file names are being printed, "(standard input)" is used for
2863 the standard input, for compatibility with GNU grep. Previously
2864 "<stdin>" was used.
2865
2866 (j) The option --label=xxx can be used to supply a name to be used for
2867 stdin when file names are being printed. There is no short form.
2868
2869 (k) Re-factored the options decoding logic because we are going to add
2870 two more options that take data. Such options can now be given in four
2871 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2872
2873 (l) Added the -A, -B, and -C options for requesting that lines of context
2874 around matches be printed.
2875
2876 (m) Added the -L option to print the names of files that do not contain
2877 any matching lines, that is, the complement of -l.
2878
2879 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2880 continue to scan other files.
2881
2882 (o) The -s option was incorrectly implemented. For compatibility with other
2883 greps, it now suppresses the error message for a non-existent or non-
2884 accessible file (but not the return code). There is a new option called
2885 -q that suppresses the output of matching lines, which was what -s was
2886 previously doing.
2887
2888 (p) Added --include and --exclude options to specify files for inclusion
2889 and exclusion when recursing.
2890
2891 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2892 Hopefully, it now does.
2893
2894 12. Missing cast in pcre_study().
2895
2896 13. Added an "uninstall" target to the makefile.
2897
2898 14. Replaced "extern" in the function prototypes in Makefile.in with
2899 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2900 world, but is set differently for Windows.
2901
2902 15. Added a second compiling function called pcre_compile2(). The only
2903 difference is that it has an extra argument, which is a pointer to an
2904 integer error code. When there is a compile-time failure, this is set
2905 non-zero, in addition to the error test pointer being set to point to an
2906 error message. The new argument may be NULL if no error number is required
2907 (but then you may as well call pcre_compile(), which is now just a
2908 wrapper). This facility is provided because some applications need a
2909 numeric error indication, but it has also enabled me to tidy up the way
2910 compile-time errors are handled in the POSIX wrapper.
2911
2912 16. Added VPATH=.libs to the makefile; this should help when building with one
2913 prefix path and installing with another. (Or so I'm told by someone who
2914 knows more about this stuff than I do.)
2915
2916 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2917 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2918 match everything, including newlines. This is not POSIX-compatible, but
2919 somebody wanted the feature. From pcretest it can be activated by using
2920 both the P and the s flags.
2921
2922 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2923
2924 19. libpcre.pc was being incorrectly installed as executable.
2925
2926 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2927 it now also looks for '\r' so that it will work unmodified on Windows.
2928
2929 21. Added Google's contributed C++ wrapper to the distribution.
2930
2931 22. Added some untidy missing memory free() calls in pcretest, to keep
2932 Electric Fence happy when testing.
2933
2934
2935
2936 Version 5.0 13-Sep-04
2937 ---------------------
2938
2939 1. Internal change: literal characters are no longer packed up into items
2940 containing multiple characters in a single byte-string. Each character
2941 is now matched using a separate opcode. However, there may be more than one
2942 byte in the character in UTF-8 mode.
2943
2944 2. The pcre_callout_block structure has two new fields: pattern_position and
2945 next_item_length. These contain the offset in the pattern to the next match
2946 item, and its length, respectively.
2947
2948 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2949 insertion of callouts before each pattern item. Added the /C option to
2950 pcretest to make use of this.
2951
2952 4. On the advice of a Windows user, the lines
2953
2954 #if defined(_WIN32) || defined(WIN32)
2955 _setmode( _fileno( stdout ), 0x8000 );
2956 #endif /* defined(_WIN32) || defined(WIN32) */
2957
2958 have been added to the source of pcretest. This apparently does useful
2959 magic in relation to line terminators.
2960
2961 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2962 for the benefit of those environments where the "b" makes a difference.
2963
2964 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2965 to know about it. I have put a hack into configure.in that adds in code
2966 to set GCC=yes if CC=icc. This seems to end up at a point in the
2967 generated configure script that is early enough to affect the setting of
2968 compiler options, which is what is needed, but I have no means of testing
2969 whether it really works. (The user who reported this had patched the
2970 generated configure script, which of course I cannot do.)
2971
2972 LATER: After change 22 below (new libtool files), the configure script
2973 seems to know about icc (and also ecc). Therefore, I have commented out
2974 this hack in configure.in.
2975
2976 7. Added support for pkg-config (2 patches were sent in).
2977
2978 8. Negated POSIX character classes that used a combination of internal tables
2979 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2980 [[:^ascii]]. Typically, they would match almost any characters. The other
2981 POSIX classes were not broken in this way.
2982
2983 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2984 to find the match, as PCRE was deluded into thinking that the match had to
2985 start at the start point or following a newline. The same bug applied to
2986 patterns with negative forward assertions or any backward assertions
2987 preceding ".*" at the start, unless the pattern required a fixed first
2988 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2989
2990 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2991 starting at the last subject character, bytes beyond the end of the subject
2992 string were read.
2993
2994 11. Renamed the variable "class" as "classbits" to make life easier for C++
2995 users. (Previously there was a macro definition, but it apparently wasn't
2996 enough.)
2997
2998 12. Added the new field "tables" to the extra data so that tables can be passed
2999 in at exec time, or the internal tables can be re-selected. This allows
3000 a compiled regex to be saved and re-used at a later time by a different
3001 program that might have everything at different addresses.
3002
3003 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3004 -R library as well as a -L library.
3005
3006 14. The debugging options of pcretest (-d on the command line or D on a
3007 pattern) showed incorrect output for anything following an extended class
3008 that contained multibyte characters and which was followed by a quantifier.
3009
3010 15. Added optional support for general category Unicode character properties
3011 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3012 support. It adds about 90K to the size of the library. The meanings of the
3013 inbuilt class escapes such as \d and \s have NOT been changed.
3014
3015 16. Updated pcredemo.c to include calls to free() to release the memory for the
3016 compiled pattern.
3017
3018 17. The generated file chartables.c was being created in the source directory
3019 instead of in the building directory. This caused the build to fail if the
3020 source directory was different from the building directory, and was
3021 read-only.
3022
3023 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3024 file. No doubt somebody will tell me if they don't make sense... Also added
3025 Dan Mooney's comments about building on OpenVMS.
3026
3027 19. Added support for partial matching via the PCRE_PARTIAL option for
3028 pcre_exec() and the \P data escape in pcretest.
3029
3030 20. Extended pcretest with 3 new pattern features:
3031
3032 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3033 write the compiled pattern to the file whose name is "rest-of-line".
3034 This is a straight binary dump of the data, with the saved pointer to
3035 the character tables forced to be NULL. The study data, if any, is
3036 written too. After writing, pcretest reads a new pattern.
3037
3038 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3039 compiled pattern from the given file. There must not be any
3040 occurrences of "<" in the file name (pretty unlikely); if there are,
3041 pcretest will instead treat the initial "<" as a pattern delimiter.
3042 After reading in the pattern, pcretest goes on to read data lines as
3043 usual.
3044
3045 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3046 and 16-bit fields in a compiled pattern, to simulate a pattern that
3047 was compiled on a host of opposite endianness.
3048
3049 21. The pcre-exec() function can now cope with patterns that were compiled on
3050 hosts of opposite endianness, with this restriction:
3051
3052 As for any compiled expression that is saved and used later, the tables
3053 pointer field cannot be preserved; the extra_data field in the arguments
3054 to pcre_exec() should be used to pass in a tables address if a value
3055 other than the default internal tables were used at compile time.
3056
3057 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3058 now diagnosed as an error. Previously, most of the time, a negative number
3059 would have been treated as zero, but if in addition "ovector" was passed as
3060 NULL, a crash could occur.
3061
3062 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3063 new versions from the libtool 1.5 distribution (the last one is a copy of
3064 a file called libtool.m4). This seems to have fixed the need to patch
3065 "configure" to support Darwin 1.3 (which I used to do). However, I still
3066 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3067 workstation).
3068
3069 24. Changed the PCRE licence to be the more standard "BSD" licence.
3070
3071
3072 Version 4.5 01-Dec-03
3073 ---------------------
3074
3075 1. There has been some re-arrangement of the code for the match() function so
3076 that it can be compiled in a version that does not call itself recursively.
3077 Instead, it keeps those local variables that need separate instances for
3078 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3079 needs to "recurse". Keeping track of where control must go is done by means
3080 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3081 hide most of the details from the main code, and operates only if
3082 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3083 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3084 operating.
3085
3086 To make it easier for callers to provide specially tailored get/free
3087 functions for this usage, two new functions, pcre_stack_malloc, and
3088 pcre_stack_free, are used. They are always called in strict stacking order,
3089 and the size of block requested is always the same.
3090
3091 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3092 PCRE has been compiled to use the stack or the heap for recursion. The
3093 -C option of pcretest uses this to show which version is compiled.
3094
3095 A new data escape \S, is added to pcretest; it causes the amounts of store
3096 obtained and freed by both kinds of malloc/free at match time to be added
3097 to the output.
3098
3099 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3100 what's available on my current Linux desktop machine.
3101
3102 3. When matching a UTF-8 string, the test for a valid string at the start has
3103 been extended. If start_offset is not zero, PCRE now checks that it points
3104 to a byte that is the start of a UTF-8 character. If not, it returns
3105 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3106 this is necessary because there may be backward assertions in the pattern.
3107 When matching the same subject several times, it may save resources to use
3108 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3109
3110 4. The code for checking the validity of UTF-8 strings has been tightened so
3111 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3112 containing "overlong sequences".
3113
3114 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3115 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3116 should have been "&", but it just so happened that all the cases this let
3117 through by mistake were picked up later in the function.
3118
3119 6. I had used a variable called "isblank" - this is a C99 function, causing
3120 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3121
3122 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3123 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3124 so the version is shown; (c) stop "make test" echoing "./RunTest".
3125
3126 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3127
3128 9. The prototype for memmove() for systems that don't have it was using
3129 size_t, but the inclusion of the header that defines size_t was later. I've
3130 moved the #includes for the C headers earlier to avoid this.
3131
3132 10. Added some adjustments to the code to make it easier to compiler on certain
3133 special systems:
3134
3135 (a) Some "const" qualifiers were missing.
3136 (b) Added the macro EXPORT before all exported functions; by default this
3137 is defined to be empty.
3138 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3139 that it reads its output file name as an argument instead of writing
3140 to the standard output and assuming this can be redirected.
3141
3142 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3143 class containing characters with values greater than 255, PCRE compilation
3144 went into a loop.
3145
3146 12. A recursive reference to a subpattern that was within another subpattern
3147 that had a minimum quantifier of zero caused PCRE to crash. For example,
3148 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3149 recursion. If the recursively-called subpattern itself had a zero repeat,
3150 that was OK.
3151
3152 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3153 buffer into which it was copied (for escape processing) was still set at
3154 1024, so long lines caused crashes.
3155
3156 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3157 "internal error: code overflow...". This applied to any character class
3158 that was followed by a possessive quantifier.
3159
3160 15. Modified the Makefile to add libpcre.la as a prerequisite for
3161 libpcreposix.la because I was told this is needed for a parallel build to
3162 work.
3163
3164 16. If a pattern that contained .* following optional items at the start was
3165 studied, the wrong optimizing data was generated, leading to matching
3166 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3167 matching string must start with a or b or c. The correct conclusion for
3168 this pattern is that a match can start with any character.
3169
3170
3171 Version 4.4 13-Aug-03
3172 ---------------------
3173
3174 1. In UTF-8 mode, a character class containing characters with values between
3175 127 and 255 was not handled correctly if the compiled pattern was studied.
3176 In fixing this, I have also improved the studying algorithm for such
3177 classes (slightly).
3178
3179 2. Three internal functions had redundant arguments passed to them. Removal
3180 might give a very teeny performance improvement.
3181
3182 3. Documentation bug: the value of the capture_top field in a callout is *one
3183 more than* the number of the hightest numbered captured substring.
3184
3185 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3186 in incorrectly linking with a previously installed version. They now link
3187 explicitly with libpcre.la.
3188
3189 5. configure.in no longer needs to recognize Cygwin specially.
3190
3191 6. A problem in pcre.in for Windows platforms is fixed.
3192
3193 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3194 pcretest, it used to include the size of the study block as part of its
3195 output. Unfortunately, the structure contains a field that has a different
3196 size on different hardware architectures. This meant that the tests that
3197 showed this size failed. As the block is currently always of a fixed size,
3198 this information isn't actually particularly useful in pcretest output, so
3199 I have just removed it.
3200
3201 8. Three pre-processor statements accidentally did not start in column 1.
3202 Sadly, there are *still* compilers around that complain, even though
3203 standard C has not required this for well over a decade. Sigh.
3204
3205 9. In pcretest, the code for checking callouts passed small integers in the
3206 callout_data field, which is a void * field. However, some picky compilers
3207 complained about the casts involved for this on 64-bit systems. Now
3208 pcretest passes the address of the small integer instead, which should get
3209 rid of the warnings.
3210
3211 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3212 both compile and run time, and gives an error if an invalid UTF-8 sequence
3213 is found. There is a option for disabling this check in cases where the
3214 string is known to be correct and/or the maximum performance is wanted.
3215
3216 11. In response to a bug report, I changed one line in Makefile.in from
3217
3218 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3219 to
3220 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3221
3222 to look similar to other lines, but I have no way of telling whether this
3223 is the right thing to do, as I do not use Windows. No doubt I'll get told
3224 if it's wrong...
3225
3226
3227 Version 4.3 21-May-03
3228 ---------------------
3229
3230 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3231 Makefile.
3232
3233 2. Some refactoring to improve the quality of the code:
3234
3235 (i) The utf8_table... variables are now declared "const".
3236
3237 (ii) The code for \cx, which used the "case flipping" table to upper case
3238 lower case letters, now just substracts 32. This is ASCII-specific,
3239 but the whole concept of \cx is ASCII-specific, so it seems
3240 reasonable.
3241
3242 (iii) PCRE was using its character types table to recognize decimal and
3243 hexadecimal digits in the pattern. This is silly, because it handles
3244 only 0-9, a-f, and A-F, but the character types table is locale-
3245 specific, which means strange things might happen. A private
3246 table is now used for this - though it costs 256 bytes, a table is
3247 much faster than multiple explicit tests. Of course, the standard
3248 character types table is still used for matching digits in subject
3249 strings against \d.
3250
3251 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3252 ending in _t are). So I've renamed it as ESC_tee.
3253
3254 3. The first argument for regexec() in the POSIX wrapper should have been
3255 defined as "const".
3256
3257 4. Changed pcretest to use malloc() for its buffers so that they can be
3258 Electric Fenced for debugging.
3259
3260 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3261 to read one or more bytes before the start of the subject string. Often this
3262 had no effect on PCRE's behaviour, but in some circumstances it could
3263 provoke a segmentation fault.
3264
3265 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3266 to try to read one or more bytes before the start of the subject string.
3267
3268 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3269 UTF-8 support could misbehave in various ways if the subject string
3270 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3271 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3272 back over UTF-8 characters.)
3273
3274
3275 Version 4.2 14-Apr-03
3276 ---------------------
3277
3278 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3279
3280 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3281 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3282 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3283 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3284 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3285 and BUILD_EXEEXT
3286 Note: automatic setting of the BUILD variables is not yet working
3287 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3288 compile-time but not at link-time
3289 [LINK]: use for linking executables only
3290 make different versions for Windows and non-Windows
3291 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3292 libraries
3293 [LINK_FOR_BUILD]: new variable
3294 [OBJEXT]: use throughout
3295 [EXEEXT]: use throughout
3296 <winshared>: new target
3297 <wininstall>: new target
3298 <dftables.o>: use native compiler
3299 <dftables>: use native linker
3300 <install>: handle Windows platform correctly
3301 <clean>: ditto
3302 <check>: ditto
3303 copy DLL to top builddir before testing
3304
3305 As part of these changes, -no-undefined was removed again. This was reported
3306 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3307 in any case.
3308
3309 3. Some tidies to get rid of compiler warnings:
3310
3311 . In the match_data structure, match_limit was an unsigned long int, whereas
3312 match_call_count was an int. I've made them both unsigned long ints.
3313
3314 . In pcretest the fact that a const uschar * doesn't automatically cast to
3315 a void * provoked a warning.
3316
3317 . Turning on some more compiler warnings threw up some "shadow" variables
3318 and a few more missing casts.
3319
3320 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3321 option, a class that contained a single character with a value between 128
3322 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3323
3324 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3325 option, a class that contained several characters, but with at least one
3326 whose value was between 128 and 255 caused PCRE to crash.
3327
3328
3329 Version 4.1 12-Mar-03
3330 ---------------------
3331
3332 1. Compiling with gcc -pedantic found a couple of places where casts were
3333 needed, and a string in dftables.c that was longer than standard compilers are
3334 required to support.
3335
3336 2. Compiling with Sun's compiler found a few more places where the code could
3337 be tidied up in order to avoid warnings.
3338
3339 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3340 first of these names is deprecated in the latest Autoconf in favour of the name
3341 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3342 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3343 analogy I have changed it to CFLAGS_FOR_BUILD.
3344
3345 4. Added -no-undefined to the linking command in the Makefile, because this is
3346 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3347 linking step for the pcreposix library.
3348
3349 5. PCRE was failing to diagnose the case of two named groups with the same
3350 name.
3351
3352 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3353 literal character that is needed in the subject for a match, and scans along to
3354 ensure that it is present before embarking on the full matching process. This
3355 saves time in cases of nested unlimited repeats that are never going to match.
3356 Problem: the scan can take a lot of time if the subject is very long (e.g.
3357 megabytes), thus penalizing straightforward matches. It is now done only if the
3358 amount of subject to be scanned is less than 1000 bytes.
3359
3360 7. A lesser problem with the same optimization is that it was recording the
3361 first character of an anchored pattern as "needed", thus provoking a search
3362 right along the subject, even when the first match of the pattern was going to
3363 fail. The "needed" character is now not set for anchored patterns, unless it
3364 follows something in the pattern that is of non-fixed length. Thus, it still
3365 fulfils its original purpose of finding quick non-matches in cases of nested
3366 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3367
3368
3369 Version 4.0 17-Feb-03
3370 ---------------------
3371
3372 1. If a comment in an extended regex that started immediately after a meta-item
3373 extended to the end of string, PCRE compiled incorrect data. This could lead to
3374 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3375
3376 2. Moved to autoconf 2.53 and libtool 1.4.2.
3377
3378 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3379 the special perltest8 script is no longer needed - all the tests can be run
3380 from a single perltest script.
3381
3382 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3383 by \s. It has now been removed in PCRE. This means it isn't recognized as
3384 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3385 class [:space:] *does* include VT, thereby creating a mess.
3386
3387 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3388 space and tab.
3389
3390 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3391 its new features into the main test script, reducing the number of scripts.
3392
3393 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3394 were backward compatible, and made the (?i) apply to the whole pattern, as if
3395 /i were given. Now it behaves more logically, and applies the option setting
3396 only to what follows. PCRE has been changed to follow suit. However, if it
3397 finds options settings right at the start of the pattern, it extracts them into
3398 the global options, as before. Thus, they show up in the info data.
3399
3400 8. Added support for the \Q...\E escape sequence. Characters in between are
3401 treated as literals. This is slightly different from Perl in that $ and @ are
3402 also handled as literals inside the quotes. In Perl, they will cause variable
3403 interpolation. Note the following examples:
3404
3405 Pattern PCRE matches Perl matches
3406
3407 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3408 \Qabc\$xyz\E abc\$xyz abc\$xyz
3409 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3410
3411 For compatibility with Perl, \Q...\E sequences are recognized inside character
3412 classes as well as outside them.
3413
3414 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3415 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3416 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3417 signed/unsigned warnings.
3418
3419 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3420 option for pcretest, so I've replaced it by a simple function that does just
3421 that job.
3422
3423 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3424 "pcregrep -".
3425
3426 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3427 Java package. This provides some syntactic sugar for simple cases of what my
3428 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3429 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3430 item, you can use this simplified notation. Note that only makes sense with
3431 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3432 greediness, whatever the setting of the PCRE_UNGREEDY option.
3433
3434 13. A change of greediness default within a pattern was not taking effect at
3435 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3436 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3437 was abstracted outside.
3438
3439 14. PCRE now supports the \G assertion. It is true when the current matching
3440 position is at the start point of the match. This differs from \A when the
3441 starting offset is non-zero. Used with the /g option of pcretest (or similar
3442 code), it works in the same way as it does for Perl's /g option. If all
3443 alternatives of a regex begin with \G, the expression is anchored to the start
3444 match position, and the "anchored" flag is set in the compiled expression.
3445
3446 15. Some bugs concerning the handling of certain option changes within patterns
3447 have been fixed. These applied to options other than (?ims). For example,
3448 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3449 been the other way round. Some of this was related to change 7 above.
3450
3451 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3452 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3453 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3454 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3455
3456 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3457 mode. Unlike ".", it always matches newline, whatever the setting of
3458 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3459 assertions. Perl allows it, but it doesn't (in general) work because it can't
3460 calculate the length of the lookbehind. At least, that's the case for Perl
3461 5.8.0 - I've been told they are going to document that it doesn't work in
3462 future.
3463
3464 18. Added an error diagnosis for escapes that PCRE does not support: these are
3465 \L, \l, \N, \P, \p, \U, \u, and \X.
3466
3467 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3468 reading past the end of the pattern in cases such as /[abcd/.
3469
3470 20. PCRE was getting more memory than necessary for patterns with classes that
3471 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3472
3473 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3474 compiling PCRE for use with Virtual Pascal.
3475
3476 22. Small fix to the Makefile to make it work properly if the build is done
3477 outside the source tree.
3478
3479 23. Added a new extension: a condition to go with recursion. If a conditional
3480 subpattern starts with (?(R) the "true" branch is used if recursion has
3481 happened, whereas the "false" branch is used only at the top level.
3482
3483 24. When there was a very long string of literal characters (over 255 bytes
3484 without UTF support, over 250 bytes with UTF support), the computation of how
3485 much memory was required could be incorrect, leading to segfaults or other
3486 strange effects.
3487
3488 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3489 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3490 there was a subsequent back reference to those brackets. This meant that, for
3491 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3492 possible to check for precisely this case. All we can do is abandon the
3493 optimization if .* occurs inside capturing brackets when there are any back
3494 references whatsoever. (See below for a better fix that came later.)
3495
3496 26. The handling of the optimization for finding the first character of a
3497 non-anchored pattern, and for finding a character that is required later in the
3498 match were failing in some cases. This didn't break the matching; it just
3499 failed to optimize when it could. The way this is done has been re-implemented.
3500
3501 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3502
3503 28. Added a new feature that provides some of the functionality that Perl
3504 provides with (?{...}). The facility is termed a "callout". The way it is done
3505 in PCRE is for the caller to provide an optional function, by setting
3506 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3507 global variable. By default it is unset, which disables all calling out. To get
3508 the function called, the regex must include (?C) at appropriate points. This
3509 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3510 This provides a means of identifying different callout points. When PCRE
3511 reaches such a point in the regex, if pcre_callout has been set, the external
3512 function is called. It is provided with data in a structure called
3513 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3514 matching continues; if it returns a non-zero value, the match at the current
3515 point fails. However, backtracking will occur if possible. [This was changed
3516 later and other features added - see item 49 below.]
3517
3518 29. pcretest is upgraded to test the callout functionality. It provides a
3519 callout function that displays information. By default, it shows the start of
3520 the match and the current position in the text. There are some new data escapes
3521 to vary what happens:
3522
3523 \C+ in addition, show current contents of captured substrings
3524 \C- do not supply a callout function
3525 \C!n return 1 when callout number n is reached
3526 \C!n!m return 1 when callout number n is reached for the mth time
3527
3528 30. If pcregrep was called with the -l option and just a single file name, it
3529 output "<stdin>" if a match was found, instead of the file name.
3530
3531 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3532 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3533 pcre_exec(). This saves a malloc/free per call. The default value of
3534 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3535 when configuring.
3536
3537 32. The default maximum size of a compiled pattern is 64K. There have been a
3538 few cases of people hitting this limit. The code now uses macros to handle the
3539 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3540 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3541 configuring. Tests 2 and 5 work only with 2-byte links because they output
3542 debugging information about compiled patterns.
3543
3544 33. Internal code re-arrangements:
3545
3546 (a) Moved the debugging function for printing out a compiled regex into
3547 its own source file (printint.c) and used #include to pull it into
3548 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3549 separate copies.
3550
3551 (b) Defined the list of op-code names for debugging as a macro in
3552 internal.h so that it is next to the definition of the opcodes.
3553
3554 (c) Defined a table of op-code lengths for simpler skipping along compiled
3555 code. This is again a macro in internal.h so that it is next to the
3556 definition of the opcodes.
3557
3558 34. Added support for recursive calls to individual subpatterns, along the
3559 lines of Robin Houston's patch (but implemented somewhat differently).
3560
3561 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3562 allow it to read and process whole directories in Win32. This code was
3563 contributed by Lionel Fourquaux; it has not been tested by me.
3564
3565 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3566 used to name a group. Names consist of alphanumerics and underscores, and must
3567 be unique. Back references use the syntax (?P=name) and recursive calls use
3568 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3569 numbers. The function pcre_fullinfo() can be used after compilation to extract
3570 a name/number map. There are three relevant calls:
3571
3572 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3573 PCRE_INFO_NAMECOUNT yields the number of entries
3574 PCRE_INFO_NAMETABLE yields a pointer to the map.
3575
3576 The map is a vector of fixed-size entries. The size of each entry depends on
3577 the length of the longest name used. The first two bytes of each entry are the
3578 group number, most significant byte first. There follows the corresponding
3579 name, zero terminated. The names are in alphabetical order.
3580
3581 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3582 case instead of 255. Making it the same both with and without UTF-8 support
3583 means that the same test output works with both.
3584
3585 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3586 calling malloc() with a zero argument.
3587
3588 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3589 optimization. I've improved things by keeping a bitmap of backreferences with
3590 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3591 fact referenced, the optimization can be applied. It is unlikely that a
3592 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3593 the match to follow \n) will appear inside brackets with a number greater than
3594 31, but if it does, any back reference > 31 suppresses the optimization.
3595
3596 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3597 of disabling numbered capturing parentheses. Any opening parenthesis that is
3598 not followed by ? behaves as if it were followed by ?: but named parentheses
3599 can still be used for capturing (and they will acquire numbers in the usual
3600 way).
3601
3602 41. Redesigned the return codes from the match() function into yes/no/error so
3603 that errors can be passed back from deep inside the nested calls. A malloc
3604 failure while inside a recursive subpattern call now causes the
3605 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3606
3607 42. It is now possible to set a limit on the number of times the match()
3608 function is called in a call to pcre_exec(). This facility makes it possible to
3609 limit the amount of recursion and backtracking, though not in a directly
3610 obvious way, because the match() function is used in a number of different
3611 circumstances. The count starts from zero for each position in the subject
3612 string (for non-anchored patterns). The default limit is, for compatibility, a
3613 large number, namely 10 000 000. You can change this in two ways:
3614
3615 (a) When configuring PCRE before making, you can use --with-match-limit=n
3616 to set a default value for the compiled library.
3617
3618 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3619 a different value is set. See 45 below.
3620
3621 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3622
3623 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3624 of things that can be changed at compile time. The first argument specifies
3625 what is wanted and the second points to where the information is to be placed.
3626 The current list of available information is:
3627
3628 PCRE_CONFIG_UTF8
3629
3630 The output is an integer that is set to one if UTF-8 support is available;
3631 otherwise it is set to zero.
3632
3633 PCRE_CONFIG_NEWLINE
3634
3635 The output is an integer that it set to the value of the code that is used for
3636 newline. It is either LF (10) or CR (13).
3637
3638 PCRE_CONFIG_LINK_SIZE
3639
3640 The output is an integer that contains the number of bytes used for internal
3641 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3642
3643 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3644
3645 The output is an integer that contains the threshold above which the POSIX
3646 interface uses malloc() for output vectors. See item 31 above.
3647
3648 PCRE_CONFIG_MATCH_LIMIT
3649
3650 The output is an unsigned integer that contains the default limit of the number
3651 of match() calls in a pcre_exec() execution. See 42 above.
3652
3653 44. pcretest has been upgraded by the addition of the -C option. This causes it
3654 to extract all the available output from the new pcre_config() function, and to
3655 output it. The program then exits immediately.
3656
3657 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3658 order to support additional features. One way would have been to define
3659 pcre_exec2() (for example) with extra arguments, but this would not have been
3660 extensible, and would also have required all calls to the original function to
3661 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3662 is used for passing in "extra" data from pcre_study().
3663
3664 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3665 contains the following fields:
3666
3667 flags a bitmap indicating which of the following fields are set
3668 study_data opaque data from pcre_study()
3669 match_limit a way of specifying a limit on match() calls for a specific
3670 call to pcre_exec()
3671 callout_data data for callouts (see 49 below)
3672
3673 The flag bits are also defined in pcre.h, and are
3674
3675 PCRE_EXTRA_STUDY_DATA
3676 PCRE_EXTRA_MATCH_LIMIT
3677 PCRE_EXTRA_CALLOUT_DATA
3678
3679 The pcre_study() function now returns one of these new pcre_extra blocks, with
3680 the actual study data pointed to by the study_data field, and the
3681 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3682 before. That is, this change is entirely upwards-compatible and requires no
3683 change to existing code.
3684
3685 If you want to pass in additional data to pcre_exec(), you can either place it
3686 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3687 block.
3688
3689 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3690 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3691 times with different match limits, until it finds the minimum value needed for
3692 pcre_exec() to complete. The value is then output. This can be instructive; for
3693 most simple matches the number is quite small, but for pathological cases it
3694 gets very large very quickly.
3695
3696 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3697 returns the size of the data block pointed to by the study_data field in a
3698 pcre_extra block, that is, the value that was passed as the argument to
3699 pcre_malloc() when PCRE was getting memory in which to place the information
3700 created by pcre_study(). The fourth argument should point to a size_t variable.
3701 pcretest has been extended so that this information is shown after a successful
3702 pcre_study() call when information about the compiled regex is being displayed.
3703
3704 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3705 because what follows is always an absolute path. (Later: it turns out that this
3706 is more than cosmetic for MinGW, because it doesn't like empty path
3707 components.)
3708
3709 49. Some changes have been made to the callout feature (see 28 above):
3710
3711 (i) A callout function now has three choices for what it returns:
3712
3713 0 => success, carry on matching
3714 > 0 => failure at this point, but backtrack if possible
3715 < 0 => serious error, return this value from pcre_exec()
3716
3717 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3718 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3719 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3720 use by callout functions. It will never be used by PCRE itself.
3721
3722 (ii) The pcre_extra structure (see 45 above) has a void * field called
3723 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3724 pcre_callout_block structure has a field of the same name. The contents of
3725 the field passed in the pcre_extra structure are passed to the callout
3726 function in the corresponding field in the callout block. This makes it
3727 easier to use the same callout-containing regex from multiple threads. For
3728 testing, the pcretest program has a new data escape
3729
3730 \C*n pass the number n (may be negative) as callout_data
3731
3732 If the callout function in pcretest receives a non-zero value as
3733 callout_data, it returns that value.
3734
3735 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3736 there were some redundant $(CFLAGS) in commands that are now specified as
3737 $(LINK), which already includes $(CFLAGS).
3738
3739 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3740 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3741 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3742 one-byte characters throughout. Note that case-insensitive matching applies
3743 only to characters whose values are less than 256. PCRE doesn't support the
3744 notion of cases for higher-valued characters.
3745
3746 (i) A character class whose characters are all within 0-255 is handled as
3747 a bit map, and the map is inverted for negative classes. Previously, a
3748 character > 255 always failed to match such a class; however it should
3749 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3750
3751 (ii) A negated character class with a single character < 255 is coded as
3752 "not this character" (OP_NOT). This wasn't working properly when the test
3753 character was multibyte, either singly or repeated.
3754
3755 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3756 mode, for example: \x{100}{2,3}.
3757
3758 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3759 singly or repeated) now correctly test multibyte characters. However,
3760 PCRE doesn't recognize any characters with values greater than 255 as
3761 digits, spaces, or word characters. Such characters always match \D, \S,
3762 and \W, and never match \d, \s, or \w.
3763
3764 (v) Classes may now contain characters and character ranges with values
3765 greater than 255. For example: [ab\x{100}-\x{400}].
3766
3767 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3768 PCRE in UTF-8 mode.
3769
3770 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3771 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3772 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3773 value.)
3774
3775 53. The single man page has become too large. I have therefore split it up into
3776 a number of separate man pages. These also give rise to individual HTML pages;
3777 these are now put in a separate directory, and there is an index.html page that
3778 lists them all. Some hyperlinking between the pages has been installed.
3779
3780 54. Added convenience functions for handling named capturing parentheses.
3781
3782 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3783 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3784 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3785 are faulted.
3786
3787 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3788 calling configure. These values are used when compiling the dftables.c program
3789 which is run to generate the source of the default character tables. They
3790 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3791 you will need to set these values.
3792
3793 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3794
3795
3796 Version 3.9 02-Jan-02
3797 ---------------------
3798
3799 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3800
3801 2. If --disable-static was given, the building process failed when trying to
3802 build pcretest and pcregrep. (For some reason it was using libtool to compile
3803 them, which is not right, as they aren't part of the library.)
3804
3805
3806 Version 3.8 18-Dec-01
3807 ---------------------
3808
3809 1. The experimental UTF-8 code was completely screwed up. It was packing the
3810 bytes in the wrong order. How dumb can you get?
3811
3812
3813 Version 3.7 29-Oct-01
3814 ---------------------
3815
3816 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3817 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3818 this didn't happen under Solaris 8, where I normally test things.
3819
3820 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3821 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3822 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3823 this fix an hour or so after the initial 3.7 release.)
3824
3825
3826 Version 3.6 23-Oct-01
3827 ---------------------
3828
3829 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3830 offsets passed as NULL with zero offset count.
3831
3832 2. The config.guess and config.sub files had not been updated when I moved to
3833 the latest autoconf.
3834
3835
3836 Version 3.5 15-Aug-01
3837 ---------------------
3838
3839 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3840 had been forgotten.
3841
3842 2. By using declared but undefined structures, we can avoid using "void"
3843 definitions in pcre.h while keeping the internal definitions of the structures
3844 private.
3845
3846 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3847 user point of view, this means that both static and shared libraries are built
3848 by default, but this can be individually controlled. More of the work of
3849 handling this static/shared cases is now inside libtool instead of PCRE's make
3850 file.
3851
3852 4. The pcretest utility is now installed along with pcregrep because it is
3853 useful for users (to test regexs) and by doing this, it automatically gets
3854 relinked by libtool. The documentation has been turned into a man page, so
3855 there are now .1, .txt, and .html versions in /doc.
3856
3857 5. Upgrades to pcregrep:
3858 (i) Added long-form option names like gnu grep.
3859 (ii) Added --help to list all options with an explanatory phrase.
3860 (iii) Added -r, --recursive to recurse into sub-directories.
3861 (iv) Added -f, --file to read patterns from a file.
3862
3863 6. pcre_exec() was referring to its "code" argument before testing that
3864 argument for NULL (and giving an error if it was NULL).
3865
3866 7. Upgraded Makefile.in to allow for compiling in a different directory from
3867 the source directory.
3868
3869 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3870 options bits, the pointer it was passed was to an int instead of to an unsigned
3871 long int. This mattered only on 64-bit systems.
3872
3873 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3874 generated) instead of pcre.in, which it its source. Also made the same change
3875 in several of the .c files.
3876
3877 10. A new release of gcc defines printf() as a macro, which broke pcretest
3878 because it had an ifdef in the middle of a string argument for printf(). Fixed
3879 by using separate calls to printf().
3880
3881 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3882 script, to force use of CR or LF instead of \n in the source. On non-Unix
3883 systems, the value can be set in config.h.
3884
3885 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3886 absolute limit. Changed the text of the error message to make this clear, and
3887 likewise updated the man page.
3888
3889 13. The limit of 99 on the number of capturing subpatterns has been removed.
3890 The new limit is 65535, which I hope will not be a "real" limit.
3891
3892
3893 Version 3.4 22-Aug-00
3894 ---------------------
3895
3896 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3897
3898 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3899
3900
3901 Version 3.3 01-Aug-00
3902 ---------------------
3903
3904 1. If an octal character was given, but the value was greater than \377, it
3905 was not getting masked to the least significant bits, as documented. This could
3906 lead to crashes in some systems.
3907
3908 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3909 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3910
3911 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3912 These just pass their arguments on to (pcre_free)(), but they are provided
3913 because some uses of PCRE bind it to non-C systems that can call its functions,
3914 but cannot call free() or pcre_free() directly.
3915
3916 4. Add "make test" as a synonym for "make check". Corrected some comments in
3917 the Makefile.
3918
3919 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3920 Makefile.
3921
3922 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3923 command called pgrep for grepping around the active processes.
3924
3925 7. Added the beginnings of support for UTF-8 character strings.
3926
3927 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3928 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3929 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3930 out for the ar command.)
3931
3932
3933 Version 3.2 12-May-00
3934 ---------------------
3935
3936 This is purely a bug fixing release.
3937
3938 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3939 of ZA. This was just one example of several cases that could provoke this bug,
3940 which was introduced by change 9 of version 2.00. The code for breaking
3941 infinite loops after an iteration that matches an empty string was't working
3942 correctly.
3943
3944 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3945 when matched against abbab (for example). After matching an empty string, it
3946 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3947 caused it to match further down the string than it should.
3948
3949 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3950 was there because it doesn't seem to be needed, and it causes trouble on some
3951 systems, as it is not a Standard C header. It has been removed.
3952
3953 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3954 were reported on the Macintosh. The changes were from
3955
3956 while ((c = *(++ptr)) != 0 && c != '\n');
3957 to
3958 while ((c = *(++ptr)) != 0 && c != '\n') ;
3959
3960 Totally extraordinary, but if that's what it takes...
3961
3962 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3963 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3964 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3965 assumes the way PCRE uses memmove() (always moving upwards).
3966
3967 6. PCRE is being used in one environment where strchr() is not available. There
3968 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3969 faster code anyway.
3970
3971
3972 Version 3.1 09-Feb-00
3973 ---------------------
3974
3975 The only change in this release is the fixing of some bugs in Makefile.in for
3976 the "install" target:
3977
3978 (1) It was failing to install pcreposix.h.
3979
3980 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3981
3982
3983 Version 3.0 01-Feb-00
3984 ---------------------
3985
3986 1. Add support for the /+ modifier to perltest (to output $` like it does in
3987 pcretest).
3988
3989 2. Add support for the /g modifier to perltest.
3990
3991 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3992 matches null strings.
3993
3994 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3995 pattern. Perl treats empty patterns specially - it reuses the most recent
3996 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3997 effect.
3998
3999 5. The POSIX interface was broken in that it was just handing over the POSIX
4000 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4001 required a bigger vector, with some working space on the end. This means that
4002 the POSIX wrapper now has to get and free some memory, and copy the results.
4003
4004 6. Added some simple autoconf support, placing the test data and the
4005 documentation in separate directories, re-organizing some of the
4006 information files, and making it build pcre-config (a GNU standard). Also added
4007 libtool support for building PCRE as a shared library, which is now the
4008 default.
4009
4010 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4011 09 are not valid octal constants. Single digits will be used for minor values
4012 less than 10.
4013
4014 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4015 existing programs that set these in the POSIX interface can use PCRE without
4016 modification.
4017
4018 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4019 return all that pcre_info() returns, plus additional data. The pcre_info()
4020 function is retained for compatibility, but is considered to be obsolete.
4021
4022 10. Added experimental recursion feature (?R) to handle one common case that
4023 Perl 5.6 will be able to do with (?p{...}).
4024
4025 11. Added support for POSIX character classes like [:alpha:], which Perl is
4026 adopting.
4027
4028
4029 Version 2.08 31-Aug-99
4030 ----------------------
4031
4032 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4033 trying to match at the startoffset position, but instead was moving forward to
4034 the next newline as if a previous match had failed.
4035
4036 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4037 and could get into a loop if a null string was matched other than at the start
4038 of the subject.
4039
4040 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4041 be distinguished at compile time, and for completeness also added PCRE_DATE.
4042
4043 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4044 in GnuWin32 environments.
4045
4046
4047 Version 2.07 29-Jul-99
4048 ----------------------
4049
4050 1. The documentation is now supplied in plain text form and HTML as well as in
4051 the form of man page sources.
4052
4053 2. C++ compilers don't like assigning (void *) values to other pointer types.
4054 In particular this affects malloc(). Although there is no problem in Standard
4055 C, I've put in casts to keep C++ compilers happy.
4056
4057 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4058 should be (const char *).
4059
4060 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4061 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4062 However, I haven't made this a standard facility. The documentation doesn't
4063 mention it, and the Makefile doesn't support it.
4064
4065 5. The Makefile now contains an "install" target, with editable destinations at
4066 the top of the file. The pcretest program is not installed.
4067
4068 6. pgrep -V now gives the PCRE version number and date.
4069
4070 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4071 causing the entire string to be ignored, instead of just the last character.
4072
4073 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4074 non-matching string, it can take a very, very long time, even for strings of
4075 quite modest length, because of the nested recursion. PCRE now does better in
4076 some of these cases. It does this by remembering the last required literal
4077 character in the pattern, and pre-searching the subject to ensure it is present
4078 before running the real match. In other words, it applies a heuristic to detect
4079 some types of certain failure quickly, and in the above example, if presented
4080 with a string that has no trailing " it gives "no match" very quickly.
4081
4082 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4083 other alternatives are tried instead.
4084
4085
4086 Version 2.06 09-Jun-99
4087 ----------------------
4088
4089 1. Change pcretest's output for amount of store used to show just the code
4090 space, because the remainder (the data block) varies in size between 32-bit and
4091 64-bit systems.
4092
4093 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4094 start matching at. This allows lookbehinds to work when searching for multiple
4095 occurrences in a string.
4096
4097 3. Added additional options to pcretest for testing multiple occurrences:
4098
4099 /+ outputs the rest of the string that follows a match
4100 /g loops for multiple occurrences, using the new startoffset argument
4101 /G loops for multiple occurrences by passing an incremented pointer
4102
4103 4. PCRE wasn't doing the "first character" optimization for patterns starting
4104 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4105 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4106 the letter 'x'. On long subject strings, this gives a significant speed-up.
4107
4108
4109 Version 2.05 21-Apr-99
4110 ----------------------
4111
4112 1. Changed the type of magic_number from int to long int so that it works
4113 properly on 16-bit systems.
4114
4115 2. Fixed a bug which caused patterns starting with .* not to work correctly
4116 when the subject string contained newline characters. PCRE was assuming
4117 anchoring for such patterns in all cases, which is not correct because .* will
4118 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4119 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4120 must be retried after every newline in the subject.
4121
4122
4123 Version 2.04 18-Feb-99
4124 ----------------------
4125
4126 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4127 computation of the store needed to hold the pattern was incorrect (too large).
4128 If such patterns were nested a few deep, this could multiply and become a real
4129 problem.
4130
4131 2. Added /M option to pcretest to show the memory requirement of a specific
4132 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4133
4134 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4135 compiled in such a way that the backtracking after subsequent failure was
4136 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4137 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4138
4139
4140 Version 2.03 02-Feb-99
4141 ----------------------
4142
4143 1. Fixed typo and small mistake in man page.
4144
4145 2. Added 4th condition (GPL supersedes if conflict) and created separate
4146 LICENCE file containing the conditions.
4147
4148 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4149 Perl, that is the internal \ allows the delimiter to be included in the
4150 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4151 the final delimiter, add \ to the end of the pattern (to test the error).
4152
4153 4. Added the convenience functions for extracting substrings after a successful
4154 match. Updated pcretest to make it able to test these functions.
4155
4156
4157 Version 2.02 14-Jan-99
4158 ----------------------
4159
4160 1. Initialized the working variables associated with each extraction so that
4161 their saving and restoring doesn't refer to uninitialized store.
4162
4163 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4164 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4165 fix the problem.
4166
4167 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4168 calls, and wasn't printing the correct value for compiling calls. Increased the
4169 default value of LOOPREPEAT, and the number of significant figures in the
4170 times.
4171
4172 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4173
4174 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4175 a building problem on Windows NT with a FAT file system.
4176
4177
4178 Version 2.01 21-Oct-98
4179 ----------------------
4180
4181 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4182 to character tables built by pcre_maketables() in the current locale. If NULL
4183 is passed, the default tables are used.
4184
4185
4186 Version 2.00 24-Sep-98
4187 ----------------------
4188
4189 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4190 it any more.
4191
4192 2. Allow quantification of (?>) groups, and make it work correctly.
4193
4194 3. The first character computation wasn't working for (?>) groups.
4195
4196 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4197 end of the subject) and add 5.005's \z, which really does match only at the
4198 very end of the subject.
4199
4200 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4201
4202 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4203 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4204 localized options. All options to pcre_study() were also removed.
4205
4206 7. Add other new features from 5.005:
4207
4208 $(?<= positive lookbehind
4209 $(?<! negative lookbehind
4210 (?imsx-imsx) added the unsetting capability
4211 such a setting is global if at outer level; local otherwise
4212 (?imsx-imsx:) non-capturing groups with option setting
4213 (?(cond)re|re) conditional pattern matching
4214
4215 A backreference to itself in a repeated group matches the previous
4216 captured string.
4217
4218 8. General tidying up of studying (both automatic and via "study")
4219 consequential on the addition of new assertions.
4220
4221 9. As in 5.005, unlimited repeated groups that could match an empty substring
4222 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4223 runtime if any iteration does actually match an empty substring.
4224
4225 10. Include the RunTest script in the distribution.
4226
4227 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4228 discrepancies, some of which were old and were also with respect to 5.004. They
4229 have now been fixed.
4230
4231
4232 Version 1.09 28-Apr-98
4233 ----------------------
4234
4235 1. A negated single character class followed by a quantifier with a minimum
4236 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4237 program crashes, or just wrong answers. This did not apply to negated classes
4238 containing more than one character, or to minima other than one.
4239
4240
4241 Version 1.08 27-Mar-98
4242 ----------------------
4243
4244 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4245
4246 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4247 latter must appear before anything that relies on it in the pattern.
4248
4249
4250 Version 1.07 16-Feb-98
4251 ----------------------
4252
4253 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4254 repeat of a potentially empty string).
4255
4256
4257 Version 1.06 23-Jan-98
4258 ----------------------
4259
4260 1. Added Markus Oberhumer's little patches for C++.
4261
4262 2. Literal strings longer than 255 characters were broken.
4263
4264
4265 Version 1.05 23-Dec-97
4266 ----------------------
4267
4268 1. Negated character classes containing more than one character were failing if
4269 PCRE_CASELESS was set at run time.
4270
4271
4272 Version 1.04 19-Dec-97
4273 ----------------------
4274
4275 1. Corrected the man page, where some "const" qualifiers had been omitted.
4276
4277 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4278 input syntax.
4279
4280 3. Fixed memory leak which occurred when a regex with back references was
4281 matched with an offsets vector that wasn't big enough. The temporary memory
4282 that is used in this case wasn't being freed if the match failed.
4283
4284 4. Tidied pcretest to ensure it frees memory that it gets.
4285
4286 5. Temporary memory was being obtained in the case where the passed offsets
4287 vector was exactly big enough.
4288
4289 6. Corrected definition of offsetof() from change 5 below.
4290
4291 7. I had screwed up change 6 below and broken the rules for the use of
4292 setjmp(). Now fixed.
4293
4294
4295 Version 1.03 18-Dec-97
4296 ----------------------
4297
4298 1. A erroneous regex with a missing opening parenthesis was correctly
4299 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4300 on some systems.
4301
4302 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4303 it was reported that one broken compiler failed on the former because "code" is
4304 also an independent variable.
4305
4306 3. The erroneous regex a[]b caused an array overrun reference.
4307
4308 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4309 fail on data ending with that character. (It was going on too far, and checking
4310 the next character, typically a binary zero.) This was specific to the
4311 optimized code for single-character negative classes.
4312
4313 5. Added a contributed patch from the TIN world which does the following:
4314
4315 + Add an undef for memmove, in case the the system defines a macro for it.
4316
4317 + Add a definition of offsetof(), in case there isn't one. (I don't know
4318 the reason behind this - offsetof() is part of the ANSI standard - but
4319 it does no harm).
4320
4321 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4322 most of the places where whitespace preceded '#'. I have given up and
4323 allowed the remaining 2 cases to be at the margin.
4324
4325 + Rename some variables in pcre to eliminate shadowing. This seems very
4326 pedantic, but does no harm, of course.
4327
4328 6. Moved the call to setjmp() into its own function, to get rid of warnings
4329 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4330
4331 7. Constructs such as \d{8,} were compiling into the equivalent of
4332 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4333 outcome, but in this particular case used more store than had been allocated,
4334 which caused the bug to be discovered because it threw up an internal error.
4335
4336 8. The debugging code in both pcre and pcretest for outputting the compiled
4337 form of a regex was going wrong in the case of back references followed by
4338 curly-bracketed repeats.
4339
4340
4341 Version 1.02 12-Dec-97
4342 ----------------------
4343
4344 1. Typos in pcre.3 and comments in the source fixed.
4345
4346 2. Applied a contributed patch to get rid of places where it used to remove
4347 'const' from variables, and fixed some signed/unsigned and uninitialized
4348 variable warnings.
4349
4350 3. Added the "runtest" target to Makefile.
4351
4352 4. Set default compiler flag to -O2 rather than just -O.
4353
4354
4355 Version 1.01 19-Nov-97
4356 ----------------------
4357
4358 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4359 like /([ab]*)*/, that is, for classes with more than one character in them.
4360
4361 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4362 as /((?>a*))*/ (a PCRE_EXTRA facility).
4363
4364
4365 Version 1.00 18-Nov-97
4366 ----------------------
4367
4368 1. Added compile-time macros to support systems such as SunOS4 which don't have
4369 memmove() or strerror() but have other things that can be used instead.
4370
4371 2. Arranged that "make clean" removes the executables.
4372
4373
4374 Version 0.99 27-Oct-97
4375 ----------------------
4376
4377 1. Fixed bug in code for optimizing classes with only one character. It was
4378 initializing a 32-byte map regardless, which could cause it to run off the end
4379 of the memory it had got.
4380
4381 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4382
4383
4384 Version 0.98 22-Oct-97
4385 ----------------------
4386
4387 1. Fixed bug in code for handling temporary memory usage when there are more
4388 back references than supplied space in the ovector. This could cause segfaults.
4389
4390
4391 Version 0.97 21-Oct-97
4392 ----------------------
4393
4394 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4395
4396 2. Optimized negated single characters not to use a bit map.
4397
4398 3. Brought error texts together as macro definitions; clarified some of them;
4399 fixed one that was wrong - it said "range out of order" when it meant "invalid
4400 escape sequence".
4401
4402 4. Changed some char * arguments to const char *.
4403
4404 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4405
4406 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4407 pcretest.
4408
4409
4410 Version 0.96 16-Oct-97
4411 ----------------------
4412
4413 1. Added a simple "pgrep" utility to the distribution.
4414
4415 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4416 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4417 where "ddd" means "one or more decimal digits".
4418
4419 3. Fixed serious bug. If a pattern had a back reference, but the call to
4420 pcre_exec() didn't supply a large enough ovector to record the related
4421 identifying subpattern, the match always failed. PCRE now remembers the number
4422 of the largest back reference, and gets some temporary memory in which to save
4423 the offsets during matching if necessary, in order to ensure that
4424 backreferences always work.
4425
4426 4. Increased the compatibility with Perl in a number of ways:
4427
4428 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4429 to request this handling. The option can be set at compile or exec time.
4430
4431 (b) $ matches before a terminating newline by default; an option
4432 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4433 mode). The option can be set at compile or exec time.
4434
4435 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4436 the same as Perl's. If the decimal number it represents is less than 10
4437 or there aren't that many previous left capturing parentheses, an octal
4438 escape is read. Inside a character class, it's always an octal escape,
4439 even if it is a single digit.
4440
4441 (d) An escaped but undefined alphabetic character is taken as a literal,
4442 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4443 escapes.
4444
4445 (e) {0} is now permitted. (The previous item is removed from the compiled
4446 pattern).
4447
4448 5. Changed all the names of code files so that the basic parts are no longer
4449 than 10 characters, and abolished the teeny "globals.c" file.
4450
4451 6. Changed the handling of character classes; they are now done with a 32-byte
4452 bit map always.
4453
4454 7. Added the -d and /D options to pcretest to make it possible to look at the
4455 internals of compilation without having to recompile pcre.
4456
4457
4458 Version 0.95 23-Sep-97
4459 ----------------------
4460
4461 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4462 \x20 at the start of a run of normal characters. These were being treated as
4463 real characters, instead of the source characters being re-checked.
4464
4465
4466 Version 0.94 18-Sep-97
4467 ----------------------
4468
4469 1. The functions are now thread-safe, with the caveat that the global variables
4470 containing pointers to malloc() and free() or alternative functions are the
4471 same for all threads.
4472
4473 2. Get pcre_study() to generate a bitmap of initial characters for non-
4474 anchored patterns when this is possible, and use it if passed to pcre_exec().
4475
4476
4477 Version 0.93 15-Sep-97
4478 ----------------------
4479
4480 1. /(b)|(:+)/ was computing an incorrect first character.
4481
4482 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4483 but not actually doing anything yet.
4484
4485 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4486 as Perl does (e.g. [-az] or [az-]).
4487
4488 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4489 all possible positions.
4490
4491 5. Split up into different modules to avoid including unneeded functions in a
4492 compiled binary. However, compile and exec are still in one module. The "study"
4493 function is split off.
4494
4495 6. The character tables are now in a separate module whose source is generated
4496 by an auxiliary program - but can then be edited by hand if required. There are
4497 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4498 toupper() in the code.
4499
4500 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4501 make them global. Abolish the function for setting them, as the caller can now
4502 set them directly.
4503
4504
4505 Version 0.92 11-Sep-97
4506 ----------------------
4507
4508 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4509 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4510
4511 2. Caseless matching was not working in character classes if the characters in
4512 the pattern were in upper case.
4513
4514 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4515
4516 4. Make PCRE_ANCHORED public and accept as a compile option.
4517
4518 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4519 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4520 pass them.
4521
4522 6. Give an error if bad option bits passed at compile or run time.
4523
4524 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4525 pcretest to cause it to pass that flag.
4526
4527 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4528 options, and the first character, if set.
4529
4530 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4531
4532
4533 Version 0.91 10-Sep-97
4534 ----------------------
4535
4536 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4537 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4538
4539 2. PCRE was looping on encountering an indefinitely repeated back reference to
4540 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4541 Perl does - treats the match as successful.
4542
4543 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12