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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 771 - (show annotations)
Tue Nov 29 15:34:12 2011 UTC (7 years, 11 months ago) by ph10
File size: 221607 byte(s)
Error occurred while calculating annotation data.
Update *MARK handling to be more like Perl.
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 15. If a forward reference was repeated with an upper limit of around 2000,
59 it caused the error "internal error: overran compiling workspace". This
60 is now checked, and causes "too many forward references" instead.
61
62 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
63 incorrectly expecting the subject to contain another "a" after the start.
64
65 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
66 in the match, the SKIP should be ignored. This was not happening; instead
67 the SKIP was being treated as NOMATCH. For patterns such as
68 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
69 tested.
70
71 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
72 now much more compatible with Perl, in particular in cases where the result
73 is a non-match for a non-anchored pattern. For example, if
74 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
75 "m", where previously it did not return a name. A side effect of this
76 change is that for partial matches, the last encountered mark name is
77 returned, as for non matches. A number of tests that were previously not
78 Perl-compatible have been moved into the Perl-compatible test files. The
79 refactoring has had the pleasing side effect of removing one argument from
80 the match() function, thus reducing its stack requirements.
81
82
83 Version 8.20 21-Oct-2011
84 ------------------------
85
86 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
87 a POSIX class. After further experiments with Perl, which convinced me that
88 Perl has bugs and confusions, a closing square bracket is no longer allowed
89 in a POSIX name. This bug also affected patterns with classes that started
90 with full stops.
91
92 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
93 captured substring, but while checking the failing first alternative,
94 substring 1 is temporarily captured. If the output vector supplied to
95 pcre_exec() was not big enough for this capture, the yield of the function
96 was still zero ("insufficient space for captured substrings"). This cannot
97 be totally fixed without adding another stack variable, which seems a lot
98 of expense for a edge case. However, I have improved the situation in cases
99 such as /(a)(b)x|abc/ matched against "abc", where the return code
100 indicates that fewer than the maximum number of slots in the ovector have
101 been set.
102
103 3. Related to (2) above: when there are more back references in a pattern than
104 slots in the output vector, pcre_exec() uses temporary memory during
105 matching, and copies in the captures as far as possible afterwards. It was
106 using the entire output vector, but this conflicts with the specification
107 that only 2/3 is used for passing back captured substrings. Now it uses
108 only the first 2/3, for compatibility. This is, of course, another edge
109 case.
110
111 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
112 main code base, and can be used by building with --enable-jit. When this is
113 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
114 runtime --no-jit option is given.
115
116 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
117 ovector, the return from the function was zero, implying that there were
118 other matches that did not fit. The correct "exactly full" value is now
119 returned.
120
121 6. If a subpattern that was called recursively or as a subroutine contained
122 (*PRUNE) or any other control that caused it to give a non-standard return,
123 invalid errors such as "Error -26 (nested recursion at the same subject
124 position)" or even infinite loops could occur.
125
126 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
127 computing the minimum length on reaching *ACCEPT, and so ended up with the
128 wrong value of 1 rather than 0. Further investigation indicates that
129 computing a minimum subject length in the presence of *ACCEPT is difficult
130 (think back references, subroutine calls), and so I have changed the code
131 so that no minimum is registered for a pattern that contains *ACCEPT.
132
133 8. If (*THEN) was present in the first (true) branch of a conditional group,
134 it was not handled as intended. [But see 16 below.]
135
136 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
137 Sheri Pierce.
138
139 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
140 the first byte in a match must be "a".
141
142 11. Change 17 for 8.13 increased the recursion depth for patterns like
143 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
144 pattern contains any instances of (*THEN). If it does not, the old
145 optimizations are restored. It would be nice to do this on a per-group
146 basis, but at the moment that is not feasible.
147
148 12. In some environments, the output of pcretest -C is CRLF terminated. This
149 broke RunTest's code that checks for the link size. A single white space
150 character after the value is now allowed for.
151
152 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
153 For "fr", it uses the Windows-specific input and output files.
154
155 14. If (*THEN) appeared in a group that was called recursively or as a
156 subroutine, it did not work as intended. [But see next item.]
157
158 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
159 pattern fragments (but not containing any | characters). If A and B are
160 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
161 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
162 D. In other words, Perl considers parentheses that do not contain any |
163 characters to be part of a surrounding alternative, whereas PCRE was
164 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
165 differently. PCRE now behaves in the same way as Perl, except in the case
166 of subroutine/recursion calls such as (?1) which have in any case always
167 been different (but PCRE had them first :-).
168
169 16. Related to 15 above: Perl does not treat the | in a conditional group as
170 creating alternatives. Such a group is treated in the same way as an
171 ordinary group without any | characters when processing (*THEN). PCRE has
172 been changed to match Perl's behaviour.
173
174 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
175 RunGrepTest script failed.
176
177 18. Change 22 for version 13 caused atomic groups to use more stack. This is
178 inevitable for groups that contain captures, but it can lead to a lot of
179 stack use in large patterns. The old behaviour has been restored for atomic
180 groups that do not contain any capturing parentheses.
181
182 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
183 suppress the check for a minimum subject length at run time. (If it was
184 given to pcre_exec() or pcre_dfa_exec() it did work.)
185
186 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
187 fail to work when decoding hex characters in data strings in EBCDIC
188 environments.
189
190 21. It appears that in at least one Mac OS environment, the isxdigit() function
191 is implemented as a macro that evaluates to its argument more than once,
192 contravening the C 90 Standard (I haven't checked a later standard). There
193 was an instance in pcretest which caused it to go wrong when processing
194 \x{...} escapes in subject strings. The has been rewritten to avoid using
195 things like p++ in the argument of isxdigit().
196
197
198 Version 8.13 16-Aug-2011
199 ------------------------
200
201 1. The Unicode data tables have been updated to Unicode 6.0.0.
202
203 2. Two minor typos in pcre_internal.h have been fixed.
204
205 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
206 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
207 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
208
209 4. There were a number of related bugs in the code for matching backrefences
210 caselessly in UTF-8 mode when codes for the characters concerned were
211 different numbers of bytes. For example, U+023A and U+2C65 are an upper
212 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
213 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
214 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
215 2-byte code at the end of the subject (it thought there wasn't enough data
216 left).
217
218 5. Comprehensive information about what went wrong is now returned by
219 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
220 as the output vector has at least 2 elements. The offset of the start of
221 the failing character and a reason code are placed in the vector.
222
223 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
224 now returned is for the first byte of the failing character, instead of the
225 last byte inspected. This is an incompatible change, but I hope it is small
226 enough not to be a problem. It makes the returned offset consistent with
227 pcre_exec() and pcre_dfa_exec().
228
229 7. pcretest now gives a text phrase as well as the error number when
230 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
231 failure, the offset and reason code are output.
232
233 8. When \R was used with a maximizing quantifier it failed to skip backwards
234 over a \r\n pair if the subsequent match failed. Instead, it just skipped
235 back over a single character (\n). This seems wrong (because it treated the
236 two characters as a single entity when going forwards), conflicts with the
237 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
238 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
239 has been changed.
240
241 9. Some internal refactoring has changed the processing so that the handling
242 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
243 time (the PCRE_DOTALL option was changed this way some time ago: version
244 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
245 which was always a bit of a fudge. It also means that there is one less
246 argument for the match() function, which reduces its stack requirements
247 slightly. This change also fixes an incompatibility with Perl: the pattern
248 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
249
250 10. More internal refactoring has drastically reduced the number of recursive
251 calls to match() for possessively repeated groups such as (abc)++ when
252 using pcre_exec().
253
254 11. While implementing 10, a number of bugs in the handling of groups were
255 discovered and fixed:
256
257 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
258 (a|)*(?1) gave a compile-time internal error.
259 ((a|)+)+ did not notice that the outer group could match an empty string.
260 (^a|^)+ was not marked as anchored.
261 (.*a|.*)+ was not marked as matching at start or after a newline.
262
263 12. Yet more internal refactoring has removed another argument from the match()
264 function. Special calls to this function are now indicated by setting a
265 value in a variable in the "match data" data block.
266
267 13. Be more explicit in pcre_study() instead of relying on "default" for
268 opcodes that mean there is no starting character; this means that when new
269 ones are added and accidentally left out of pcre_study(), testing should
270 pick them up.
271
272 14. The -s option of pcretest has been documented for ages as being an old
273 synonym of -m (show memory usage). I have changed it to mean "force study
274 for every regex", that is, assume /S for every regex. This is similar to -i
275 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
276 using it. It makes it easier to run collections of tests with and without
277 study enabled, and thereby test pcre_study() more easily. All the standard
278 tests are now run with and without -s (but some patterns can be marked as
279 "never study" - see 20 below).
280
281 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
282 restoration of the capturing data to the outer values was not happening
283 correctly.
284
285 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
286 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
287 pattern had matched an empty string, and so incorrectly returned a no
288 match.
289
290 17. There was optimizing code for the last branch of non-capturing parentheses,
291 and also for the obeyed branch of a conditional subexpression, which used
292 tail recursion to cut down on stack usage. Unfortunately, now that there is
293 the possibility of (*THEN) occurring in these branches, tail recursion is
294 no longer possible because the return has to be checked for (*THEN). These
295 two optimizations have therefore been removed. [But see 8.20/11 above.]
296
297 18. If a pattern containing \R was studied, it was assumed that \R always
298 matched two bytes, thus causing the minimum subject length to be
299 incorrectly computed because \R can also match just one byte.
300
301 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
302 was incorrectly computed.
303
304 20. If /S is present twice on a test pattern in pcretest input, it now
305 *disables* studying, thereby overriding the use of -s on the command line
306 (see 14 above). This is necessary for one or two tests to keep the output
307 identical in both cases.
308
309 21. When (*ACCEPT) was used in an assertion that matched an empty string and
310 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
311
312 22. When an atomic group that contained a capturing parenthesis was
313 successfully matched, but the branch in which it appeared failed, the
314 capturing was not being forgotten if a higher numbered group was later
315 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
316 group 1 to "a", when in fact it should be unset. This applied to multi-
317 branched capturing and non-capturing groups, repeated or not, and also to
318 positive assertions (capturing in negative assertions does not happen
319 in PCRE) and also to nested atomic groups.
320
321 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
322 subject after a captured substring, to make it easier to tell which of a
323 number of identical substrings has been captured.
324
325 24. The way atomic groups are processed by pcre_exec() has been changed so that
326 if they are repeated, backtracking one repetition now resets captured
327 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
328 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
329 "aaa". Previously, it would have been "a". As part of this code
330 refactoring, the way recursive calls are handled has also been changed.
331
332 25. If an assertion condition captured any substrings, they were not passed
333 back unless some other capturing happened later. For example, if
334 (?(?=(a))a) was matched against "a", no capturing was returned.
335
336 26. When studying a pattern that contained subroutine calls or assertions,
337 the code for finding the minimum length of a possible match was handling
338 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
339 group 1 called group 2 while simultaneously a separate group 2 called group
340 1). A stack overflow occurred in this case. I have fixed this by limiting
341 the recursion depth to 10.
342
343 27. Updated RunTest.bat in the distribution to the version supplied by Tom
344 Fortmann. This supports explicit test numbers on the command line, and has
345 argument validation and error reporting.
346
347 28. An instance of \X with an unlimited repeat could fail if at any point the
348 first character it looked at was a mark character.
349
350 29. Some minor code refactoring concerning Unicode properties and scripts
351 should reduce the stack requirement of match() slightly.
352
353 30. Added the '=' option to pcretest to check the setting of unused capturing
354 slots at the end of the pattern, which are documented as being -1, but are
355 not included in the return count.
356
357 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
358 compiled something random. Now it gives a compile-time error (as does
359 Perl).
360
361 32. A *MARK encountered during the processing of a positive assertion is now
362 recorded and passed back (compatible with Perl).
363
364 33. If --only-matching or --colour was set on a pcregrep call whose pattern
365 had alternative anchored branches, the search for a second match in a line
366 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
367 matched the line "0102" twice. The same bug affected patterns that started
368 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
369 twice.
370
371 34. Previously, PCRE did not allow quantification of assertions. However, Perl
372 does, and because of capturing effects, quantifying parenthesized
373 assertions may at times be useful. Quantifiers are now allowed for
374 parenthesized assertions.
375
376 35. A minor code tidy in pcre_compile() when checking options for \R usage.
377
378 36. \g was being checked for fancy things in a character class, when it should
379 just be a literal "g".
380
381 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
382 appearance of a nested POSIX class supersedes an apparent external class.
383 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
384 unescaped square brackets may also appear as part of class names. For
385 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
386 more like Perl. (But see 8.20/1 above.)
387
388 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
389 was because it thought it was \N{name}, which is not supported).
390
391 39. Add minix to OS list not supporting the -S option in pcretest.
392
393 40. PCRE tries to detect cases of infinite recursion at compile time, but it
394 cannot analyze patterns in sufficient detail to catch mutual recursions
395 such as ((?1))((?2)). There is now a runtime test that gives an error if a
396 subgroup is called recursively as a subpattern for a second time at the
397 same position in the subject string. In previous releases this might have
398 been caught by the recursion limit, or it might have run out of stack.
399
400 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
401 happen only once. PCRE was, however incorrectly giving a compile time error
402 "recursive call could loop indefinitely" because it cannot analyze the
403 pattern in sufficient detail. The compile time test no longer happens when
404 PCRE is compiling a conditional subpattern, but actual runaway loops are
405 now caught at runtime (see 40 above).
406
407 42. It seems that Perl allows any characters other than a closing parenthesis
408 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
409 has been changed to be the same.
410
411 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
412 as not to get warnings when autogen.sh is called. Also changed
413 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
414
415 44. To help people who use pcregrep to scan files containing exceedingly long
416 lines, the following changes have been made:
417
418 (a) The default value of the buffer size parameter has been increased from
419 8K to 20K. (The actual buffer used is three times this size.)
420
421 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
422 PCRE is built.
423
424 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
425 to be set at run time.
426
427 (d) Numerical values in pcregrep options can be followed by K or M, for
428 example --buffer-size=50K.
429
430 (e) If a line being scanned overflows pcregrep's buffer, an error is now
431 given and the return code is set to 2.
432
433 45. Add a pointer to the latest mark to the callout data block.
434
435 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
436 partial match of an empty string instead of no match. This was specific to
437 the use of ".".
438
439 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
440 complete match instead of a partial match. This bug was dependent on both
441 the PCRE_UTF8 and PCRE_DOTALL options being set.
442
443 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
444 starting byte set, because \b was not being ignored.
445
446
447 Version 8.12 15-Jan-2011
448 ------------------------
449
450 1. Fixed some typos in the markup of the man pages, and wrote a script that
451 checks for such things as part of the documentation building process.
452
453 2. On a big-endian 64-bit system, pcregrep did not correctly process the
454 --match-limit and --recursion-limit options (added for 8.11). In
455 particular, this made one of the standard tests fail. (The integer value
456 went into the wrong half of a long int.)
457
458 3. If the --colour option was given to pcregrep with -v (invert match), it
459 did strange things, either producing crazy output, or crashing. It should,
460 of course, ignore a request for colour when reporting lines that do not
461 match.
462
463 4. Another pcregrep bug caused similar problems if --colour was specified with
464 -M (multiline) and the pattern match finished with a line ending.
465
466 5. In pcregrep, when a pattern that ended with a literal newline sequence was
467 matched in multiline mode, the following line was shown as part of the
468 match. This seems wrong, so I have changed it.
469
470 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
471 the check for further matches in the same line (so they could be coloured)
472 to overrun the end of the current line. If another match was found, it was
473 incorrectly shown (and then shown again when found in the next line).
474
475 7. If pcregrep was compiled under Windows, there was a reference to the
476 function pcregrep_exit() before it was defined. I am assuming this was
477 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
478 reported by a user. I've moved the definition above the reference.
479
480
481 Version 8.11 10-Dec-2010
482 ------------------------
483
484 1. (*THEN) was not working properly if there were untried alternatives prior
485 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
486 backtracked to try for "b" instead of moving to the next alternative branch
487 at the same level (in this case, to look for "c"). The Perl documentation
488 is clear that when (*THEN) is backtracked onto, it goes to the "next
489 alternative in the innermost enclosing group".
490
491 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
492 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
493 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
494 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
495 (*THEN).
496
497 3. If \s appeared in a character class, it removed the VT character from
498 the class, even if it had been included by some previous item, for example
499 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
500 of \s, but is part of the POSIX "space" class.)
501
502 4. A partial match never returns an empty string (because you can always
503 match an empty string at the end of the subject); however the checking for
504 an empty string was starting at the "start of match" point. This has been
505 changed to the "earliest inspected character" point, because the returned
506 data for a partial match starts at this character. This means that, for
507 example, /(?<=abc)def/ gives a partial match for the subject "abc"
508 (previously it gave "no match").
509
510 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
511 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
512 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
513 has an implication that the given string is incomplete (because a partial
514 match is preferred over a full match). For this reason, these items now
515 give a partial match in this situation. [Aside: previously, the one case
516 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
517 match rather than a full match, which was wrong by the old rules, but is
518 now correct.]
519
520 6. There was a bug in the handling of #-introduced comments, recognized when
521 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
522 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
523 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
524 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
525 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
526 places in pcre_compile().
527
528 7. Related to (6) above, when pcre_compile() was skipping #-introduced
529 comments when looking ahead for named forward references to subpatterns,
530 the only newline sequence it recognized was NL. It now handles newlines
531 according to the set newline convention.
532
533 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
534 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
535 cater for a lack of strerror(). These oversights have been fixed.
536
537 9. Added --match-limit and --recursion-limit to pcregrep.
538
539 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
540
541 11. When the -o option was used, pcregrep was setting a return code of 1, even
542 when matches were found, and --line-buffered was not being honoured.
543
544 12. Added an optional parentheses number to the -o and --only-matching options
545 of pcregrep.
546
547 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
548 can match an empty string. The code to do it in pcretest and pcredemo
549 needed fixing:
550
551 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
552 only one byte after an empty string match just before CRLF (this case
553 just got forgotten; "any" and "anycrlf" were OK).
554
555 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
556 mode when an empty string match preceded an ASCII character followed by
557 a non-ASCII character. (The code for advancing by one character rather
558 than one byte was nonsense.)
559
560 (c) The pcredemo.c sample program did not have any code at all to handle
561 the cases when CRLF is a valid newline sequence.
562
563 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
564 as a starting offset was within the subject string. There is now a new
565 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
566 negative or greater than the length of the string. In order to test this,
567 pcretest is extended to allow the setting of negative starting offsets.
568
569 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
570 starting offset points to the beginning of a UTF-8 character was
571 unnecessarily clumsy. I tidied it up.
572
573 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
574 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
575
576 17. Nobody had reported that the --include_dir option, which was added in
577 release 7.7 should have been called --include-dir (hyphen, not underscore)
578 for compatibility with GNU grep. I have changed it to --include-dir, but
579 left --include_dir as an undocumented synonym, and the same for
580 --exclude-dir, though that is not available in GNU grep, at least as of
581 release 2.5.4.
582
583 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
584 characters from a string of bytes) have been redefined so as not to use
585 loops, in order to improve performance in some environments. At the same
586 time, I abstracted some of the common code into auxiliary macros to save
587 repetition (this should not affect the compiled code).
588
589 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
590 compile-time error is now given if \c is not followed by an ASCII
591 character, that is, a byte less than 128. (In EBCDIC mode, the code is
592 different, and any byte value is allowed.)
593
594 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
595 START_OPTIMIZE option, which is now allowed at compile time - but just
596 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
597 to pcregrep and other applications that have no direct access to PCRE
598 options. The new /Y option in pcretest sets this option when calling
599 pcre_compile().
600
601 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
602 back references. Groups containing recursive back references were forced to
603 be atomic by that change, but in the case of named groups, the amount of
604 memory required was incorrectly computed, leading to "Failed: internal
605 error: code overflow". This has been fixed.
606
607 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
608 pcretest.c, to avoid build problems in some Borland environments.
609
610
611 Version 8.10 25-Jun-2010
612 ------------------------
613
614 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
615 THEN.
616
617 2. (*ACCEPT) was not working when inside an atomic group.
618
619 3. Inside a character class, \B is treated as a literal by default, but
620 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
621 causes the error). The code is unchanged, but I tidied the documentation.
622
623 4. Inside a character class, PCRE always treated \R and \X as literals,
624 whereas Perl faults them if its -w option is set. I have changed PCRE so
625 that it faults them when PCRE_EXTRA is set.
626
627 5. Added support for \N, which always matches any character other than
628 newline. (It is the same as "." when PCRE_DOTALL is not set.)
629
630 6. When compiling pcregrep with newer versions of gcc which may have
631 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
632 declared with attribute warn_unused_result" were given. Just casting the
633 result to (void) does not stop the warnings; a more elaborate fudge is
634 needed. I've used a macro to implement this.
635
636 7. Minor change to pcretest.c to avoid a compiler warning.
637
638 8. Added four artifical Unicode properties to help with an option to make
639 \s etc use properties (see next item). The new properties are: Xan
640 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
641
642 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
643 use Unicode properties. (*UCP) at the start of a pattern can be used to set
644 this option. Modified pcretest to add /W to test this facility. Added
645 REG_UCP to make it available via the POSIX interface.
646
647 10. Added --line-buffered to pcregrep.
648
649 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
650 studied, and the match started with a letter with a code point greater than
651 127 whose first byte was different to the first byte of the other case of
652 the letter, the other case of this starting letter was not recognized
653 (#976).
654
655 12. If a pattern that was studied started with a repeated Unicode property
656 test, for example, \p{Nd}+, there was the theoretical possibility of
657 setting up an incorrect bitmap of starting bytes, but fortunately it could
658 not have actually happened in practice until change 8 above was made (it
659 added property types that matched character-matching opcodes).
660
661 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
662 possible starting bytes for non-anchored patterns.
663
664 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
665 \R, and also a number of cases that involve Unicode properties, both
666 explicit and implicit when PCRE_UCP is set.
667
668 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
669 input, it could crash or give wrong results if characters with values
670 greater than 0xc0 were present in the subject string. (Detail: it assumed
671 UTF-8 input when processing these items.)
672
673 16. Added a lot of (int) casts to avoid compiler warnings in systems where
674 size_t is 64-bit (#991).
675
676 17. Added a check for running out of memory when PCRE is compiled with
677 --disable-stack-for-recursion (#990).
678
679 18. If the last data line in a file for pcretest does not have a newline on
680 the end, a newline was missing in the output.
681
682 19. The default pcre_chartables.c file recognizes only ASCII characters (values
683 less than 128) in its various bitmaps. However, there is a facility for
684 generating tables according to the current locale when PCRE is compiled. It
685 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
686 characters, are recognized by isspace() and therefore were getting set in
687 these tables, and indeed these tables seem to approximate to ISO 8859. This
688 caused a problem in UTF-8 mode when pcre_study() was used to create a list
689 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
690 which of course cannot start UTF-8 characters. I have changed the code so
691 that only real ASCII characters (less than 128) and the correct starting
692 bytes for UTF-8 encodings are set for characters greater than 127 when in
693 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
694 altogether.)
695
696 20. Added the /T option to pcretest so as to be able to run tests with non-
697 standard character tables, thus making it possible to include the tests
698 used for 19 above in the standard set of tests.
699
700 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
701 reference to a subpattern the other side of a comment that contains an
702 opening parenthesis caused either an internal compiling error, or a
703 reference to the wrong subpattern.
704
705
706 Version 8.02 19-Mar-2010
707 ------------------------
708
709 1. The Unicode data tables have been updated to Unicode 5.2.0.
710
711 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
712 configured.
713
714 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
715 original author of that file, following a query about its status.
716
717 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
718 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
719
720 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
721 quantifier applied to a forward-referencing subroutine call, could compile
722 incorrect code or give the error "internal error: previously-checked
723 referenced subpattern not found".
724
725 6. Both MS Visual Studio and Symbian OS have problems with initializing
726 variables to point to external functions. For these systems, therefore,
727 pcre_malloc etc. are now initialized to local functions that call the
728 relevant global functions.
729
730 7. There were two entries missing in the vectors called coptable and poptable
731 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
732 I've fixed the data, and added a kludgy way of testing at compile time that
733 the lengths are correct (equal to the number of opcodes).
734
735 8. Following on from 7, I added a similar kludge to check the length of the
736 eint vector in pcreposix.c.
737
738 9. Error texts for pcre_compile() are held as one long string to avoid too
739 much relocation at load time. To find a text, the string is searched,
740 counting zeros. There was no check for running off the end of the string,
741 which could happen if a new error number was added without updating the
742 string.
743
744 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
745
746 11. \K was not working if it appeared in an atomic group or in a group that
747 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
748 \K is "not well defined" if used in an assertion. PCRE now accepts it if
749 the assertion is positive, but not if it is negative.
750
751 12. Change 11 fortuitously reduced the size of the stack frame used in the
752 "match()" function of pcre_exec.c by one pointer. Forthcoming
753 implementation of support for (*MARK) will need an extra pointer on the
754 stack; I have reserved it now, so that the stack frame size does not
755 decrease.
756
757 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
758 item in branch that calls a recursion is a subroutine call - as in the
759 second branch in the above example - was incorrectly given the compile-
760 time error "recursive call could loop indefinitely" because pcre_compile()
761 was not correctly checking the subroutine for matching a non-empty string.
762
763 14. The checks for overrunning compiling workspace could trigger after an
764 overrun had occurred. This is a "should never occur" error, but it can be
765 triggered by pathological patterns such as hundreds of nested parentheses.
766 The checks now trigger 100 bytes before the end of the workspace.
767
768 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
769
770
771 Version 8.01 19-Jan-2010
772 ------------------------
773
774 1. If a pattern contained a conditional subpattern with only one branch (in
775 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
776 computed the wrong minimum data length (which is of course zero for such
777 subpatterns). This could cause incorrect "no match" results.
778
779 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
780 the pattern is reset in the first branch, pcre_compile() failed with
781 "internal error: code overflow at offset...". This happened only when
782 the reset was to the original external option setting. (An optimization
783 abstracts leading options settings into an external setting, which was the
784 cause of this.)
785
786 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
787 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
788 assertion pattern did not match (meaning that the assertion was true), it
789 was incorrectly treated as false if the SKIP had been reached during the
790 matching. This also applied to assertions used as conditions.
791
792 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
793 assertion subpattern, including such a pattern used as a condition,
794 unpredictable results occurred, instead of the error return
795 PCRE_ERROR_DFA_UITEM.
796
797 5. The C++ GlobalReplace function was not working like Perl for the special
798 situation when an empty string is matched. It now does the fancy magic
799 stuff that is necessary.
800
801 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
802 removed. (These were left over from very, very early versions of PCRE.)
803
804 7. Some cosmetic changes to the code to make life easier when compiling it
805 as part of something else:
806
807 (a) Change DEBUG to PCRE_DEBUG.
808
809 (b) In pcre_compile(), rename the member of the "branch_chain" structure
810 called "current" as "current_branch", to prevent a collision with the
811 Linux macro when compiled as a kernel module.
812
813 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
814 prevent a collision with the Linux macro when compiled as a kernel
815 module.
816
817 8. In pcre_compile() there are some checks for integer overflows that used to
818 cast potentially large values to (double). This has been changed to that
819 when building, a check for int64_t is made, and if it is found, it is used
820 instead, thus avoiding the use of floating point arithmetic. (There is no
821 other use of FP in PCRE.) If int64_t is not found, the fallback is to
822 double.
823
824 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
825 2005 (difference between two addresses compared to an unsigned value).
826
827 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
828 custom one, because of the following reported problem in Windows:
829
830 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
831 under Win32.
832 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
833 therefore missing the function definition.
834 - The compiler thus generates a "C" signature for the test function.
835 - The linker fails to find the "C" function.
836 - PCRE fails to configure if asked to do so against libbz2.
837
838 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
839 messages were output:
840
841 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
842 rerunning libtoolize, to keep the correct libtool macros in-tree.
843 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
844
845 I have done both of these things.
846
847 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
848 most of the time, it *can* run out if it is given a pattern that contains a
849 runaway infinite recursion. I updated the discussion in the pcrestack man
850 page.
851
852 13. Now that we have gone to the x.xx style of version numbers, the minor
853 version may start with zero. Using 08 or 09 is a bad idea because users
854 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
855 interpreted as invalid octal numbers. I've updated the previous comment in
856 configure.ac, and also added a check that gives an error if 08 or 09 are
857 used.
858
859 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
860 causing partial matching to fail when the end of the subject matched \W
861 in a UTF-8 pattern where \W was quantified with a minimum of 3.
862
863 15. There were some discrepancies between the declarations in pcre_internal.h
864 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
865 their definitions. The declarations used "const uschar *" and the
866 definitions used USPTR. Even though USPTR is normally defined as "const
867 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
868 reported that: "This difference in casting confuses some C++ compilers, for
869 example, SunCC recognizes above declarations as different functions and
870 generates broken code for hbpcre." I have changed the declarations to use
871 USPTR.
872
873 16. GNU libtool is named differently on some systems. The autogen.sh script now
874 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
875 (FreeBSD).
876
877 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
878 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
879 comment: "Figure out how to create a longlong from a string: strtoll and
880 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
881 instance, but it only takes 2 args instead of 3!"
882
883 18. A subtle bug concerned with back references has been fixed by a change of
884 specification, with a corresponding code fix. A pattern such as
885 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
886 refers, was giving matches when it shouldn't. For example, xa=xaaa would
887 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
888 same bug. Such groups have to be quantified to be useful, or contained
889 inside another quantified group. (If there's no repetition, the reference
890 can never match.) The problem arises because, having left the group and
891 moved on to the rest of the pattern, a later failure that backtracks into
892 the group uses the captured value from the final iteration of the group
893 rather than the correct earlier one. I have fixed this in PCRE by forcing
894 any group that contains a reference to itself to be an atomic group; that
895 is, there cannot be any backtracking into it once it has completed. This is
896 similar to recursive and subroutine calls.
897
898
899 Version 8.00 19-Oct-09
900 ----------------------
901
902 1. The table for translating pcre_compile() error codes into POSIX error codes
903 was out-of-date, and there was no check on the pcre_compile() error code
904 being within the table. This could lead to an OK return being given in
905 error.
906
907 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
908 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
909 in a Windows environment.
910
911 3. The pcregrep --count option prints the count for each file even when it is
912 zero, as does GNU grep. However, pcregrep was also printing all files when
913 --files-with-matches was added. Now, when both options are given, it prints
914 counts only for those files that have at least one match. (GNU grep just
915 prints the file name in this circumstance, but including the count seems
916 more useful - otherwise, why use --count?) Also ensured that the
917 combination -clh just lists non-zero counts, with no names.
918
919 4. The long form of the pcregrep -F option was incorrectly implemented as
920 --fixed_strings instead of --fixed-strings. This is an incompatible change,
921 but it seems right to fix it, and I didn't think it was worth preserving
922 the old behaviour.
923
924 5. The command line items --regex=pattern and --regexp=pattern were not
925 recognized by pcregrep, which required --regex pattern or --regexp pattern
926 (with a space rather than an '='). The man page documented the '=' forms,
927 which are compatible with GNU grep; these now work.
928
929 6. No libpcreposix.pc file was created for pkg-config; there was just
930 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
931
932 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
933 when UCP support is not needed, by modifying the Python script that
934 generates it from Unicode data files. This should not matter if the module
935 is correctly used as a library, but I received one complaint about 50K of
936 unwanted data. My guess is that the person linked everything into his
937 program rather than using a library. Anyway, it does no harm.
938
939 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
940 was a minimum greater than 1 for a wide character in a possessive
941 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
942 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
943 character. Chaos in the form of incorrect output or a compiling loop could
944 result.
945
946 9. The restrictions on what a pattern can contain when partial matching is
947 requested for pcre_exec() have been removed. All patterns can now be
948 partially matched by this function. In addition, if there are at least two
949 slots in the offset vector, the offset of the earliest inspected character
950 for the match and the offset of the end of the subject are set in them when
951 PCRE_ERROR_PARTIAL is returned.
952
953 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
954 synonymous with PCRE_PARTIAL, for backwards compatibility, and
955 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
956 and may be more useful for multi-segment matching.
957
958 11. Partial matching with pcre_exec() is now more intuitive. A partial match
959 used to be given if ever the end of the subject was reached; now it is
960 given only if matching could not proceed because another character was
961 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
962 string "Z", which now yields "no match" instead of "partial match". In the
963 case of pcre_dfa_exec(), "no match" is given if every matching path for the
964 final character ended with (*FAIL).
965
966 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
967 if the pattern had a "must contain" character that was already found in the
968 earlier partial match, unless partial matching was again requested. For
969 example, with the pattern /dog.(body)?/, the "must contain" character is
970 "g". If the first part-match was for the string "dog", restarting with
971 "sbody" failed. This bug has been fixed.
972
973 13. The string returned by pcre_dfa_exec() after a partial match has been
974 changed so that it starts at the first inspected character rather than the
975 first character of the match. This makes a difference only if the pattern
976 starts with a lookbehind assertion or \b or \B (\K is not supported by
977 pcre_dfa_exec()). It's an incompatible change, but it makes the two
978 matching functions compatible, and I think it's the right thing to do.
979
980 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
981 so that the demonstration program is easily available in environments where
982 PCRE has not been installed from source.
983
984 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
985 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
986 library.
987
988 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
989 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
990 is not the first non-POSIX option to be added. Clearly some people find
991 these options useful.
992
993 17. If a caller to the POSIX matching function regexec() passes a non-zero
994 value for nmatch with a NULL value for pmatch, the value of
995 nmatch is forced to zero.
996
997 18. RunGrepTest did not have a test for the availability of the -u option of
998 the diff command, as RunTest does. It now checks in the same way as
999 RunTest, and also checks for the -b option.
1000
1001 19. If an odd number of negated classes containing just a single character
1002 interposed, within parentheses, between a forward reference to a named
1003 subpattern and the definition of the subpattern, compilation crashed with
1004 an internal error, complaining that it could not find the referenced
1005 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1006 [The bug was that it was starting one character too far in when skipping
1007 over the character class, thus treating the ] as data rather than
1008 terminating the class. This meant it could skip too much.]
1009
1010 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1011 /g option in pcretest when the pattern contains \K, which makes it possible
1012 to have an empty string match not at the start, even when the pattern is
1013 anchored. Updated pcretest and pcredemo to use this option.
1014
1015 21. If the maximum number of capturing subpatterns in a recursion was greater
1016 than the maximum at the outer level, the higher number was returned, but
1017 with unset values at the outer level. The correct (outer level) value is
1018 now given.
1019
1020 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1021 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1022 make it do so. The string so far is captured, making this feature
1023 compatible with Perl.
1024
1025 23. The tests have been re-organized, adding tests 11 and 12, to make it
1026 possible to check the Perl 5.10 features against Perl 5.10.
1027
1028 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1029 pattern matches a fixed length string. PCRE did not allow this; now it
1030 does. Neither allows recursion.
1031
1032 25. I finally figured out how to implement a request to provide the minimum
1033 length of subject string that was needed in order to match a given pattern.
1034 (It was back references and recursion that I had previously got hung up
1035 on.) This code has now been added to pcre_study(); it finds a lower bound
1036 to the length of subject needed. It is not necessarily the greatest lower
1037 bound, but using it to avoid searching strings that are too short does give
1038 some useful speed-ups. The value is available to calling programs via
1039 pcre_fullinfo().
1040
1041 26. While implementing 25, I discovered to my embarrassment that pcretest had
1042 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1043 study optimizations had never been tested with that matching function.
1044 Oops. What is worse, even when it was passed study data, there was a bug in
1045 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1046 were also very few tests of studied patterns with pcre_dfa_exec().
1047
1048 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1049 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1050 on the other side of the coin, they are no longer allowed to have different
1051 names, because these cannot be distinguished in PCRE, and this has caused
1052 confusion. (This is a difference from Perl.)
1053
1054 28. When duplicate subpattern names are present (necessarily with different
1055 numbers, as required by 27 above), and a test is made by name in a
1056 conditional pattern, either for a subpattern having been matched, or for
1057 recursion in such a pattern, all the associated numbered subpatterns are
1058 tested, and the overall condition is true if the condition is true for any
1059 one of them. This is the way Perl works, and is also more like the way
1060 testing by number works.
1061
1062
1063 Version 7.9 11-Apr-09
1064 ---------------------
1065
1066 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1067 (pcretest), all targets were linked against these libraries. This included
1068 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1069 libraries. This caused unwanted dependencies to be created. This problem
1070 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1071 pcretest is linked with readline.
1072
1073 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1074 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1075 moved outside it again, because FALSE and TRUE are already defined in AIX,
1076 but BOOL is not.
1077
1078 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1079 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1080
1081 4. The pcregrep documentation said spaces were inserted as well as colons (or
1082 hyphens) following file names and line numbers when outputting matching
1083 lines. This is not true; no spaces are inserted. I have also clarified the
1084 wording for the --colour (or --color) option.
1085
1086 5. In pcregrep, when --colour was used with -o, the list of matching strings
1087 was not coloured; this is different to GNU grep, so I have changed it to be
1088 the same.
1089
1090 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1091 each matching line was coloured. Now it goes on to look for further matches
1092 of any of the test patterns, which is the same behaviour as GNU grep.
1093
1094 7. A pattern that could match an empty string could cause pcregrep to loop; it
1095 doesn't make sense to accept an empty string match in pcregrep, so I have
1096 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1097 seems to be how GNU grep behaves.
1098
1099 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1100 start or after a newline", because the conditional assertion was not being
1101 correctly handled. The rule now is that both the assertion and what follows
1102 in the first alternative must satisfy the test.
1103
1104 9. If auto-callout was enabled in a pattern with a conditional group whose
1105 condition was an assertion, PCRE could crash during matching, both with
1106 pcre_exec() and pcre_dfa_exec().
1107
1108 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1109 used for matching.
1110
1111 11. Unicode property support in character classes was not working for
1112 characters (bytes) greater than 127 when not in UTF-8 mode.
1113
1114 12. Added the -M command line option to pcretest.
1115
1116 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1117
1118 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1119
1120 16. Added comments and documentation about mis-use of no_arg in the C++
1121 wrapper.
1122
1123 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1124 from Martin Jerabek that uses macro names for all relevant character and
1125 string constants.
1126
1127 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1128 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1129 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1130 these, but not everybody uses configure.
1131
1132 19. A conditional group that had only one branch was not being correctly
1133 recognized as an item that could match an empty string. This meant that an
1134 enclosing group might also not be so recognized, causing infinite looping
1135 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1136 with the subject "ab", where knowledge that the repeated group can match
1137 nothing is needed in order to break the loop.
1138
1139 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1140 exec(), but without supplying a callout function, matching went wrong.
1141
1142 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1143 leak if the size of the offset vector was greater than 30. When the vector
1144 is smaller, the saved offsets during recursion go onto a local stack
1145 vector, but for larger vectors malloc() is used. It was failing to free
1146 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1147 error, in fact).
1148
1149 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1150 heapframe that is used only when UTF-8 support is enabled. This caused no
1151 problem, but was untidy.
1152
1153 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1154 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1155 included within another project.
1156
1157 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1158 slightly modified by me:
1159
1160 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1161 not building pcregrep.
1162
1163 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1164 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1165
1166 25. Forward references, both numeric and by name, in patterns that made use of
1167 duplicate group numbers, could behave incorrectly or give incorrect errors,
1168 because when scanning forward to find the reference group, PCRE was not
1169 taking into account the duplicate group numbers. A pattern such as
1170 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1171
1172 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1173 the feature of a custom pointer more persuasive (as requested by a user).
1174
1175 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1176 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1177 pre-defined.
1178
1179 28. Added support for (*UTF8) at the start of a pattern.
1180
1181 29. Arrange for flags added by the "release type" setting in CMake to be shown
1182 in the configuration summary.
1183
1184
1185 Version 7.8 05-Sep-08
1186 ---------------------
1187
1188 1. Replaced UCP searching code with optimized version as implemented for Ad
1189 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1190 stage table and inline lookup instead of a function, giving speed ups of 2
1191 to 5 times on some simple patterns that I tested. Permission was given to
1192 distribute the MultiStage2.py script that generates the tables (it's not in
1193 the tarball, but is in the Subversion repository).
1194
1195 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1196 scripts.
1197
1198 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1199 a group with a zero qualifier. The result of the study could be incorrect,
1200 or the function might crash, depending on the pattern.
1201
1202 4. Caseless matching was not working for non-ASCII characters in back
1203 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1204 It now works when Unicode Property Support is available.
1205
1206 5. In pcretest, an escape such as \x{de} in the data was always generating
1207 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1208 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1209 truncation.
1210
1211 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1212
1213 7. Added two (int) casts to pcregrep when printing the difference of two
1214 pointers, in case they are 64-bit values.
1215
1216 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1217 test 2 if it fails.
1218
1219 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1220 and a #define of that name to empty if it is not externally set. This is to
1221 allow users of MSVC to set it if necessary.
1222
1223 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1224 the convenience functions in the pcre_get.c source file.
1225
1226 11. An option change at the start of a pattern that had top-level alternatives
1227 could cause overwriting and/or a crash. This command provoked a crash in
1228 some environments:
1229
1230 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1231
1232 This potential security problem was recorded as CVE-2008-2371.
1233
1234 12. For a pattern where the match had to start at the beginning or immediately
1235 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1236 pcre_dfa_exec() could read past the end of the passed subject if there was
1237 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1238 pcretest so that it places the subject at the end of its malloc-ed buffer.
1239
1240 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1241 exec() might read past the end of the data buffer in UTF-8 mode.
1242
1243 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1244 the data contained the byte 0x85 as part of a UTF-8 character within its
1245 first line. This applied both to normal and DFA matching.
1246
1247 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1248 /^[^d]*?$/8 failed to match "abc".
1249
1250 16. Added a missing copyright notice to pcrecpp_internal.h.
1251
1252 17. Make it more clear in the documentation that values returned from
1253 pcre_exec() in ovector are byte offsets, not character counts.
1254
1255 18. Tidied a few places to stop certain compilers from issuing warnings.
1256
1257 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1258 supplied by Stefan Weber. I made a further small update for 7.8 because
1259 there is a change of source arrangements: the pcre_searchfuncs.c module is
1260 replaced by pcre_ucd.c.
1261
1262
1263 Version 7.7 07-May-08
1264 ---------------------
1265
1266 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1267 a string to a long long, pretend we don't even have a long long." This is
1268 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1269
1270 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1271 pre-7.6 versions, which defined a global no_arg variable instead of putting
1272 it in the RE class. (See also #8 below.)
1273
1274 3. Remove a line of dead code, identified by coverity and reported by Nuno
1275 Lopes.
1276
1277 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1278
1279 (1) The include/exclude patterns were being applied to the whole pathnames
1280 of files, instead of just to the final components.
1281
1282 (2) If there was more than one level of directory, the subdirectories were
1283 skipped unless they satisfied the include/exclude conditions. This is
1284 inconsistent with GNU grep (and could even be seen as contrary to the
1285 pcregrep specification - which I improved to make it absolutely clear).
1286 The action now is always to scan all levels of directory, and just
1287 apply the include/exclude patterns to regular files.
1288
1289 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1290 --exclude_dir in the tests to avoid scanning .svn directories.
1291
1292 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1293 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1294 doesn't support NULs in patterns.
1295
1296 7. Added some missing "const"s to declarations of static tables in
1297 pcre_compile.c and pcre_dfa_exec.c.
1298
1299 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1300 caused by fix #2 above. (Subsequently also a second patch to fix the
1301 first patch. And a third patch - this was a messy problem.)
1302
1303 9. Applied Craig's patch to remove the use of push_back().
1304
1305 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1306 matching function regexec().
1307
1308 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1309 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1310 references. PCRE supports relative numbers with this syntax (I don't think
1311 Oniguruma does).
1312
1313 12. Previously, a group with a zero repeat such as (...){0} was completely
1314 omitted from the compiled regex. However, this means that if the group
1315 was called as a subroutine from elsewhere in the pattern, things went wrong
1316 (an internal error was given). Such groups are now left in the compiled
1317 pattern, with a new opcode that causes them to be skipped at execution
1318 time.
1319
1320 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1321 to the way PCRE behaves:
1322
1323 (a) A lone ] character is dis-allowed (Perl treats it as data).
1324
1325 (b) A back reference to an unmatched subpattern matches an empty string
1326 (Perl fails the current match path).
1327
1328 (c) A data ] in a character class must be notated as \] because if the
1329 first data character in a class is ], it defines an empty class. (In
1330 Perl it is not possible to have an empty class.) The empty class []
1331 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1332 The negative empty class [^] matches any one character, independently
1333 of the DOTALL setting.
1334
1335 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1336 non-existent subpattern following a character class starting with ']' and
1337 containing () gave an internal compiling error instead of "reference to
1338 non-existent subpattern". Fortunately, when the pattern did exist, the
1339 compiled code was correct. (When scanning forwards to check for the
1340 existencd of the subpattern, it was treating the data ']' as terminating
1341 the class, so got the count wrong. When actually compiling, the reference
1342 was subsequently set up correctly.)
1343
1344 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1345 it was being rejected as not supported by pcre_dfa_exec(), even though
1346 other assertions are supported. I have made pcre_dfa_exec() support
1347 (*FAIL).
1348
1349 16. The implementation of 13c above involved the invention of a new opcode,
1350 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1351 cannot be changed at match time, I realized I could make a small
1352 improvement to matching performance by compiling OP_ALLANY instead of
1353 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1354 on the OP_ANY path.
1355
1356 17. Compiling pcretest on Windows with readline support failed without the
1357 following two fixes: (1) Make the unistd.h include conditional on
1358 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1359
1360 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1361 ncurses library to be included for pcretest when ReadLine support is
1362 requested, but also to allow for it to be overridden. This patch came from
1363 Daniel Bergström.
1364
1365 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1366 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1367 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1368 spotting this.
1369
1370
1371 Version 7.6 28-Jan-08
1372 ---------------------
1373
1374 1. A character class containing a very large number of characters with
1375 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1376 overflow.
1377
1378 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1379 HAVE_LONG_LONG is not defined.
1380
1381 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1382 bring them up to date and include new features. This patch includes:
1383
1384 - Fixed PH's badly added libz and libbz2 support.
1385 - Fixed a problem with static linking.
1386 - Added pcredemo. [But later removed - see 7 below.]
1387 - Fixed dftables problem and added an option.
1388 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1389 HAVE_LONG_LONG.
1390 - Added readline support for pcretest.
1391 - Added an listing of the option settings after cmake has run.
1392
1393 4. A user submitted a patch to Makefile that makes it easy to create
1394 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1395 Makefile.am that cause it to include this special target, without
1396 affecting anything else. Note that the same mingw target plus all
1397 the other distribution libraries and programs are now supported
1398 when configuring with CMake (see 6 below) instead of with
1399 Configure/Make.
1400
1401 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1402 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1403 exported in the Windows port". It has not yet been confirmed that the patch
1404 solves the problem, but it does no harm.
1405
1406 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1407 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1408 with CMake, and also correct the comment about stack recursion.
1409
1410 7. Remove the automatic building of pcredemo from the ./configure system and
1411 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1412 of a program that users should build themselves after PCRE is installed, so
1413 building it automatically is not really right. What is more, it gave
1414 trouble in some build environments.
1415
1416 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1417
1418
1419 Version 7.5 10-Jan-08
1420 ---------------------
1421
1422 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1423 values in parens when parsing an RE using the C++ wrapper."
1424
1425 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1426 Characters greater than 255 were excluded from the class instead of being
1427 included.
1428
1429 3. The same bug as (2) above applied to negated POSIX classes such as
1430 [:^space:].
1431
1432 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1433 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1434 I have changed it.
1435
1436 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1437 first named subpattern) and a construct such as (?&a) would reference the
1438 first named subpattern whose name started with "a" (in other words, the
1439 length check was missing). Both these problems are fixed. "Subpattern name
1440 expected" is now given for (?&) (a zero-length name), and this patch also
1441 makes it give the same error for \k'' (previously it complained that that
1442 was a reference to a non-existent subpattern).
1443
1444 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1445 this is right because (?- can be followed by option settings as well as by
1446 digits. I have, however, made the messages clearer.
1447
1448 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1449 than the number used in the conditional) now cause a compile-time error.
1450 This is actually not compatible with Perl, which accepts such patterns, but
1451 treats the conditional as always being FALSE (as PCRE used to), but it
1452 seems to me that giving a diagnostic is better.
1453
1454 8. Change "alphameric" to the more common word "alphanumeric" in comments
1455 and messages.
1456
1457 9. Fix two occurrences of "backslash" in comments that should have been
1458 "backspace".
1459
1460 10. Remove two redundant lines of code that can never be obeyed (their function
1461 was moved elsewhere).
1462
1463 11. The program that makes PCRE's Unicode character property table had a bug
1464 which caused it to generate incorrect table entries for sequences of
1465 characters that have the same character type, but are in different scripts.
1466 It amalgamated them into a single range, with the script of the first of
1467 them. In other words, some characters were in the wrong script. There were
1468 thirteen such cases, affecting characters in the following ranges:
1469
1470 U+002b0 - U+002c1
1471 U+0060c - U+0060d
1472 U+0061e - U+00612
1473 U+0064b - U+0065e
1474 U+0074d - U+0076d
1475 U+01800 - U+01805
1476 U+01d00 - U+01d77
1477 U+01d9b - U+01dbf
1478 U+0200b - U+0200f
1479 U+030fc - U+030fe
1480 U+03260 - U+0327f
1481 U+0fb46 - U+0fbb1
1482 U+10450 - U+1049d
1483
1484 12. The -o option (show only the matching part of a line) for pcregrep was not
1485 compatible with GNU grep in that, if there was more than one match in a
1486 line, it showed only the first of them. It now behaves in the same way as
1487 GNU grep.
1488
1489 13. If the -o and -v options were combined for pcregrep, it printed a blank
1490 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1491 does the same. The return code can be used to tell if there were any
1492 non-matching lines.
1493
1494 14. Added --file-offsets and --line-offsets to pcregrep.
1495
1496 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1497 infinitely looping recursion. The bug was that positive lookaheads were not
1498 being skipped when checking for a possible empty match (negative lookaheads
1499 and both kinds of lookbehind were skipped).
1500
1501 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1502 inclusion of <windows.h> to before rather than after the definition of
1503 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1504
1505 17. Specifying a possessive quantifier with a specific limit for a Unicode
1506 character property caused pcre_compile() to compile bad code, which led at
1507 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1508 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1509 caused the error; without that there was no problem.
1510
1511 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1512
1513 19. Added --enable-pcretest-libreadline.
1514
1515 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1516 RE::GlobalReplace(). As a result, the number of replacements returned was
1517 double what it should be. I removed one of the increments, but Craig sent a
1518 later patch that removed the other one (the right fix) and added unit tests
1519 that check the return values (which was not done before).
1520
1521 21. Several CMake things:
1522
1523 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1524 the names libpcre and libpcreposix, not just pcre and pcreposix.
1525
1526 (2) The above change means that pcretest and pcregrep are now correctly
1527 linked with the newly-built libraries, not previously installed ones.
1528
1529 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1530
1531 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1532 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1533 UTF-8 newline character). The key issue is that the pattern starts .*;
1534 this means that the match must be either at the beginning, or after a
1535 newline. The bug was in the code for advancing after a failed match and
1536 checking that the new position followed a newline. It was not taking
1537 account of UTF-8 characters correctly.
1538
1539 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1540 character classes. PCRE was not treating the sequence [:...:] as a
1541 character class unless the ... were all letters. Perl, however, seems to
1542 allow any characters between [: and :], though of course it rejects as
1543 unknown any "names" that contain non-letters, because all the known class
1544 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1545 for example, whereas PCRE did not - it did not recognize a POSIX character
1546 class. This seemed a bit dangerous, so the code has been changed to be
1547 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1548 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1549 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1550 Perl does, and where it didn't before.
1551
1552 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1553 Windows environments %n is disabled by default.
1554
1555
1556 Version 7.4 21-Sep-07
1557 ---------------------
1558
1559 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1560 means that a class such as [\s] counted as "explicit reference to CR or
1561 LF". That isn't really right - the whole point of the change was to try to
1562 help when there was an actual mention of one of the two characters. So now
1563 the change happens only if \r or \n (or a literal CR or LF) character is
1564 encountered.
1565
1566 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1567 of both had grown to the point where there were only 3 bits left.
1568 Fortunately, there was spare space in the data structure, and so I have
1569 moved the internal flags into a new 16-bit field to free up more option
1570 bits.
1571
1572 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1573 but did not set the internal JCHANGED flag - either of these is enough to
1574 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1575 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1576 start sets both bits.
1577
1578 4. Added options (at build time, compile time, exec time) to change \R from
1579 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1580
1581 5. doc/pcresyntax.html was missing from the distribution.
1582
1583 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1584 compatibility, even though it is no longer used.
1585
1586 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1587 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1588 windows.h file is present (where different names are used). [This was
1589 reversed later after testing - see 16 below.]
1590
1591 8. Changed all #include <config.h> to #include "config.h". There were also
1592 some further <pcre.h> cases that I changed to "pcre.h".
1593
1594 9. When pcregrep was used with the --colour option, it missed the line ending
1595 sequence off the lines that it output.
1596
1597 10. It was pointed out to me that arrays of string pointers cause lots of
1598 relocations when a shared library is dynamically loaded. A technique of
1599 using a single long string with a table of offsets can drastically reduce
1600 these. I have refactored PCRE in four places to do this. The result is
1601 dramatic:
1602
1603 Originally: 290
1604 After changing UCP table: 187
1605 After changing error message table: 43
1606 After changing table of "verbs" 36
1607 After changing table of Posix names 22
1608
1609 Thanks to the folks working on Gregex for glib for this insight.
1610
1611 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1612 unicode-properties was also set.
1613
1614 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1615
1616 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1617 checked only for CRLF.
1618
1619 14. Added casts to pcretest.c to avoid compiler warnings.
1620
1621 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1622
1623 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1624 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1625 entirely. This removes changes made in 7 above.
1626
1627 17. The CMake files have been updated, and there is now more information about
1628 building with CMake in the NON-UNIX-USE document.
1629
1630
1631 Version 7.3 28-Aug-07
1632 ---------------------
1633
1634 1. In the rejigging of the build system that eventually resulted in 7.1, the
1635 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1636 brackets there is not right, since it causes compilers to look for an
1637 installed pcre.h, not the version that is in the source that is being
1638 compiled (which of course may be different). I have changed it back to:
1639
1640 #include "pcre.h"
1641
1642 I have a vague recollection that the change was concerned with compiling in
1643 different directories, but in the new build system, that is taken care of
1644 by the VPATH setting the Makefile.
1645
1646 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1647 when the subject happened to end in the byte 0x85 (e.g. if the last
1648 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1649 characters but of course it shouldn't be taken as a newline when it is part
1650 of another character. The bug was that, for an unlimited repeat of . in
1651 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1652 characters when looking for a newline.
1653
1654 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1655
1656 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1657 in debug output.
1658
1659 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1660 long printing in the pcrecpp unittest when running under MinGW.
1661
1662 6. ESC_K was left out of the EBCDIC table.
1663
1664 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1665 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1666 limit also applies to "virtual nesting" when a pattern is recursive, and in
1667 this case 1000 isn't so big. I have been able to remove this limit at the
1668 expense of backing off one optimization in certain circumstances. Normally,
1669 when pcre_exec() would call its internal match() function recursively and
1670 immediately return the result unconditionally, it uses a "tail recursion"
1671 feature to save stack. However, when a subpattern that can match an empty
1672 string has an unlimited repetition quantifier, it no longer makes this
1673 optimization. That gives it a stack frame in which to save the data for
1674 checking that an empty string has been matched. Previously this was taken
1675 from the 1000-entry workspace that had been reserved. So now there is no
1676 explicit limit, but more stack is used.
1677
1678 8. Applied Daniel's patches to solve problems with the import/export magic
1679 syntax that is required for Windows, and which was going wrong for the
1680 pcreposix and pcrecpp parts of the library. These were overlooked when this
1681 problem was solved for the main library.
1682
1683 9. There were some crude static tests to avoid integer overflow when computing
1684 the size of patterns that contain repeated groups with explicit upper
1685 limits. As the maximum quantifier is 65535, the maximum group length was
1686 set at 30,000 so that the product of these two numbers did not overflow a
1687 32-bit integer. However, it turns out that people want to use groups that
1688 are longer than 30,000 bytes (though not repeat them that many times).
1689 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1690 made it possible to implement the integer overflow checks in a much more
1691 dynamic way, which I have now done. The artificial limitation on group
1692 length has been removed - we now have only the limit on the total length of
1693 the compiled pattern, which depends on the LINK_SIZE setting.
1694
1695 10. Fixed a bug in the documentation for get/copy named substring when
1696 duplicate names are permitted. If none of the named substrings are set, the
1697 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1698 empty string.
1699
1700 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1701 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1702 because the ] is interpreted as the first data character and the
1703 terminating ] is not found. PCRE has been made compatible with Perl in this
1704 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1705 cause memory overwriting.
1706
1707 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1708 string has been matched (to stop an infinite loop). It was not recognizing
1709 a conditional subpattern that could match an empty string if that
1710 subpattern was within another subpattern. For example, it looped when
1711 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1712 condition was not nested. This bug has been fixed.
1713
1714 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1715 past the start of the subject in the presence of bytes with the top bit
1716 set, for example "\x8aBCD".
1717
1718 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1719 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1720
1721 14. Optimized (?!) to (*FAIL).
1722
1723 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1724 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1725 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1726 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1727 does: it's just the validity check that is more restrictive.
1728
1729 16. Inserted checks for integer overflows during escape sequence (backslash)
1730 processing, and also fixed erroneous offset values for syntax errors during
1731 backslash processing.
1732
1733 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1734 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1735
1736 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1737 caused an overrun.
1738
1739 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1740 something other than just ASCII characters) inside a group that had an
1741 unlimited repeat caused a loop at compile time (while checking to see
1742 whether the group could match an empty string).
1743
1744 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1745 [\P{Any}] did so. (Error in the code for printing property names.)
1746
1747 21. An orphan \E inside a character class could cause a crash.
1748
1749 22. A repeated capturing bracket such as (A)? could cause a wild memory
1750 reference during compilation.
1751
1752 23. There are several functions in pcre_compile() that scan along a compiled
1753 expression for various reasons (e.g. to see if it's fixed length for look
1754 behind). There were bugs in these functions when a repeated \p or \P was
1755 present in the pattern. These operators have additional parameters compared
1756 with \d, etc, and these were not being taken into account when moving along
1757 the compiled data. Specifically:
1758
1759 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1760 length.
1761
1762 (b) An item such as \pL+ within a repeated group could cause crashes or
1763 loops.
1764
1765 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1766 "reference to non-existent subpattern" error.
1767
1768 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1769
1770 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1771 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1772
1773 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1774
1775 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1776 character were causing crashes (broken optimization).
1777
1778 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1779 \p or \P) caused a compile-time loop.
1780
1781 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1782 break. For example, the unstudied pattern [\r\n]A does not match the string
1783 "\r\nA" because change 7.0/46 below moves the current point on by two
1784 characters after failing to match at the start. However, the pattern \nA
1785 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1786 the same is true. There doesn't seem any very clean way out of this, but
1787 what I have chosen to do makes the common cases work: PCRE now takes note
1788 of whether there can be an explicit match for \r or \n anywhere in the
1789 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1790 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1791 pattern has explicit CR or LF references.
1792
1793 29. Added (*CR) etc for changing newline setting at start of pattern.
1794
1795
1796 Version 7.2 19-Jun-07
1797 ---------------------
1798
1799 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1800 which is apparently normally available under Windows.
1801
1802 2. Re-jig the pcregrep tests with different newline settings in an attempt
1803 to make them independent of the local environment's newline setting.
1804
1805 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1806
1807 4. Some of the "internals" tests were previously cut out when the link size
1808 was not 2, because the output contained actual offsets. The recent new
1809 "Z" feature of pcretest means that these can be cut out, making the tests
1810 usable with all link sizes.
1811
1812 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1813 stack recursion. This gives a massive performance boost under BSD, but just
1814 a small improvement under Linux. However, it saves one field in the frame
1815 in all cases.
1816
1817 6. Added more features from the forthcoming Perl 5.10:
1818
1819 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1820 recursion call. It refers to the nth most recently opened parentheses.
1821
1822 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1823 to be opened parentheses.
1824
1825 (c) Conditions that refer to capturing parentheses can be specified
1826 relatively, for example, (?(-2)... or (?(+3)...
1827
1828 (d) \K resets the start of the current match so that everything before
1829 is not part of it.
1830
1831 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1832
1833 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1834 reference syntax.
1835
1836 (g) (?| introduces a group in which the numbering of parentheses in each
1837 alternative starts with the same number.
1838
1839 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1840
1841 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1842 PCRE_INFO_JCHANGED.
1843
1844 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1845 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1846 for detecting groups that can match an empty string.
1847
1848 9. A pattern with a very large number of alternatives (more than several
1849 hundred) was running out of internal workspace during the pre-compile
1850 phase, where pcre_compile() figures out how much memory will be needed. A
1851 bit of new cunning has reduced the workspace needed for groups with
1852 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1853 workspace instead of running out of the 4096 that are available.
1854
1855 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1856
1857 11. Applied patch from Google to remove an optimization that didn't quite work.
1858 The report of the bug said:
1859
1860 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1861 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1862 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1863
1864 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1865 it matched the wrong number of bytes.
1866
1867
1868 Version 7.1 24-Apr-07
1869 ---------------------
1870
1871 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1872 that is more "standard", making use of automake and other Autotools. There
1873 is some re-arrangement of the files and adjustment of comments consequent
1874 on this.
1875
1876 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1877 for recursive directory scanning broke on some systems because the files
1878 are not scanned in any specific order and on different systems the order
1879 was different. A call to "sort" has been inserted into RunGrepTest for the
1880 approprate test as a short-term fix. In the longer term there may be an
1881 alternative.
1882
1883 3. I had an email from Eric Raymond about problems translating some of PCRE's
1884 man pages to HTML (despite the fact that I distribute HTML pages, some
1885 people do their own conversions for various reasons). The problems
1886 concerned the use of low-level troff macros .br and .in. I have therefore
1887 removed all such uses from the man pages (some were redundant, some could
1888 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1889 HTML has been updated to handle .nf/.fi and to complain if it encounters
1890 .br or .in.
1891
1892 4. Updated comments in configure.ac that get placed in config.h.in and also
1893 arranged for config.h to be included in the distribution, with the name
1894 config.h.generic, for the benefit of those who have to compile without
1895 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1896
1897 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1898 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1899 makevp.bat for the current PCRE, using the additional files
1900 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1901
1902 6. A Windows user reported a minor discrepancy with test 2, which turned out
1903 to be caused by a trailing space on an input line that had got lost in his
1904 copy. The trailing space was an accident, so I've just removed it.
1905
1906 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1907 that is needed.
1908
1909 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1910 as "const" (a) because they are and (b) because it helps the PHP
1911 maintainers who have recently made a script to detect big data structures
1912 in the php code that should be moved to the .rodata section. I remembered
1913 to update Builducptable as well, so it won't revert if ucptable.h is ever
1914 re-created.
1915
1916 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1917 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1918 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1919 support is not required. This saves 1.5-2K of code, which is important in
1920 some applications.
1921
1922 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1923 so as not to refer to the tables, even though these functions will never be
1924 called when UTF-8 support is disabled. Otherwise there are problems with a
1925 shared library.
1926
1927 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1928
1929 (a) It was defining its arguments as char * instead of void *.
1930
1931 (b) It was assuming that all moves were upwards in memory; this was true
1932 a long time ago when I wrote it, but is no longer the case.
1933
1934 The emulated memove() is provided for those environments that have neither
1935 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1936 is clearly not the case, as these two bugs were recently reported.
1937
1938 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1939 and Detrail to create the HTML documentation, the .txt form of the man
1940 pages, and it removes trailing spaces from listed files. It also creates
1941 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1942 case, it wraps all the #defines with #ifndefs. This script should be run
1943 before "make dist".
1944
1945 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1946 with Unicode property support.
1947
1948 (a) For a maximizing quantifier, if the two different cases of the
1949 character were of different lengths in their UTF-8 codings (there are
1950 some cases like this - I found 11), and the matching function had to
1951 back up over a mixture of the two cases, it incorrectly assumed they
1952 were both the same length.
1953
1954 (b) When PCRE was configured to use the heap rather than the stack for
1955 recursion during matching, it was not correctly preserving the data for
1956 the other case of a UTF-8 character when checking ahead for a match
1957 while processing a minimizing repeat. If the check also involved
1958 matching a wide character, but failed, corruption could cause an
1959 erroneous result when trying to check for a repeat of the original
1960 character.
1961
1962 13. Some tidying changes to the testing mechanism:
1963
1964 (a) The RunTest script now detects the internal link size and whether there
1965 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1966 values substituted by "configure". (The RunGrepTest script already did
1967 this for UTF-8.) The configure.ac script no longer substitutes the
1968 relevant variables.
1969
1970 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1971 with length and offset values. This means that the output is different
1972 for different internal link sizes. Test 2 is skipped for link sizes
1973 other than 2 because of this, bypassing the problem. Unfortunately,
1974 there was also a test in test 3 (the locale tests) that used /B and
1975 failed for link sizes other than 2. Rather than cut the whole test out,
1976 I have added a new /Z option to pcretest that replaces the length and
1977 offset values with spaces. This is now used to make test 3 independent
1978 of link size. (Test 2 will be tidied up later.)
1979
1980 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1981 segmentation fault instead of returning the appropriate error message.
1982
1983 15. In multiline mode when the newline sequence was set to "any", the pattern
1984 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1985 This doesn't seem right; it now treats the CRLF combination as the line
1986 ending, and so does not match in that case. It's only a pattern such as ^$
1987 that would hit this one: something like ^ABC$ would have failed after \r
1988 and then tried again after \r\n.
1989
1990 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1991 in an attempt to make files that differ only in their line terminators
1992 compare equal. This works on Linux.
1993
1994 17. Under certain error circumstances pcregrep might try to free random memory
1995 as it exited. This is now fixed, thanks to valgrind.
1996
1997 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1998 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1999 was because its rules for how to advance for /g after matching an empty
2000 string at the end of a line did not allow for this case. They now check for
2001 it specially.
2002
2003 20. pcretest is supposed to handle patterns and data of any length, by
2004 extending its buffers when necessary. It was getting this wrong when the
2005 buffer for a data line had to be extended.
2006
2007 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2008 CRLF as a newline sequence.
2009
2010 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2011 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2012 I have nevertheless tidied it up.
2013
2014 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2015
2016 24. Added a man page for pcre-config.
2017
2018
2019 Version 7.0 19-Dec-06
2020 ---------------------
2021
2022 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2023 moving to gcc 4.1.1.
2024
2025 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2026 sys/time.h, which is documented as needed for this function. It doesn't
2027 seem to matter on Linux, but it showed up on some releases of OS X.
2028
2029 3. It seems that there are systems where bytes whose values are greater than
2030 127 match isprint() in the "C" locale. The "C" locale should be the
2031 default when a C program starts up. In most systems, only ASCII printing
2032 characters match isprint(). This difference caused the output from pcretest
2033 to vary, making some of the tests fail. I have changed pcretest so that:
2034
2035 (a) When it is outputting text in the compiled version of a pattern, bytes
2036 other than 32-126 are always shown as hex escapes.
2037
2038 (b) When it is outputting text that is a matched part of a subject string,
2039 it does the same, unless a different locale has been set for the match
2040 (using the /L modifier). In this case, it uses isprint() to decide.
2041
2042 4. Fixed a major bug that caused incorrect computation of the amount of memory
2043 required for a compiled pattern when options that changed within the
2044 pattern affected the logic of the preliminary scan that determines the
2045 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2046 that the computed length was too small. The symptoms of this bug were
2047 either the PCRE error "internal error: code overflow" from pcre_compile(),
2048 or a glibc crash with a message such as "pcretest: free(): invalid next
2049 size (fast)". Examples of patterns that provoked this bug (shown in
2050 pcretest format) are:
2051
2052 /(?-x: )/x
2053 /(?x)(?-x: \s*#\s*)/
2054 /((?i)[\x{c0}])/8
2055 /(?i:[\x{c0}])/8
2056
2057 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2058 is now done differently.
2059
2060 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2061 wrapper classes; (b) implement a new function in the C++ scanner that is
2062 more efficient than the old way of doing things because it avoids levels of
2063 recursion in the regex matching; (c) add a paragraph to the documentation
2064 for the FullMatch() function.
2065
2066 6. The escape sequence \n was being treated as whatever was defined as
2067 "newline". Not only was this contrary to the documentation, which states
2068 that \n is character 10 (hex 0A), but it also went horribly wrong when
2069 "newline" was defined as CRLF. This has been fixed.
2070
2071 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2072 was being set to -1 for the "end of line" case (supposedly a value that no
2073 character can have). Though this value is never used (the check for end of
2074 line is "zero bytes in current character"), it caused compiler complaints.
2075 I've changed it to 0xffffffff.
2076
2077 8. In pcre_version.c, the version string was being built by a sequence of
2078 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2079 string (as it is for production releases) called a macro with an empty
2080 argument. The C standard says the result of this is undefined. The gcc
2081 compiler treats it as an empty string (which was what was wanted) but it is
2082 reported that Visual C gives an error. The source has been hacked around to
2083 avoid this problem.
2084
2085 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2086 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2087 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2088 of them did).
2089
2090 10. Originally, pcretest opened its input and output without "b"; then I was
2091 told that "b" was needed in some environments, so it was added for release
2092 5.0 to both the input and output. (It makes no difference on Unix-like
2093 systems.) Later I was told that it is wrong for the input on Windows. I've
2094 now abstracted the modes into two macros, to make it easier to fiddle with
2095 them, and removed "b" from the input mode under Windows.
2096
2097 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2098
2099 12. Added -help and --help to pcretest as an official way of being reminded
2100 of the options.
2101
2102 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2103 and pcrecpp.cc because they annoy compilers at high warning levels.
2104
2105 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2106
2107 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2108 scripts are not C programs :-) and which was not noticed because it works
2109 on Linux.
2110
2111 16. pcretest is supposed to handle any length of pattern and data line (as one
2112 line or as a continued sequence of lines) by extending its input buffer if
2113 necessary. This feature was broken for very long pattern lines, leading to
2114 a string of junk being passed to pcre_compile() if the pattern was longer
2115 than about 50K.
2116
2117 17. I have done a major re-factoring of the way pcre_compile() computes the
2118 amount of memory needed for a compiled pattern. Previously, there was code
2119 that made a preliminary scan of the pattern in order to do this. That was
2120 OK when PCRE was new, but as the facilities have expanded, it has become
2121 harder and harder to keep it in step with the real compile phase, and there
2122 have been a number of bugs (see for example, 4 above). I have now found a
2123 cunning way of running the real compile function in a "fake" mode that
2124 enables it to compute how much memory it would need, while actually only
2125 ever using a few hundred bytes of working memory and without too many
2126 tests of the mode. This should make future maintenance and development
2127 easier. A side effect of this work is that the limit of 200 on the nesting
2128 depth of parentheses has been removed (though this was never a serious
2129 limitation, I suspect). However, there is a downside: pcre_compile() now
2130 runs more slowly than before (30% or more, depending on the pattern). I
2131 hope this isn't a big issue. There is no effect on runtime performance.
2132
2133 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2134 newline (only possible for the last line of a file) and it was a
2135 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2136
2137 19. Added additional timing features to pcretest. (1) The -tm option now times
2138 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2139 separate command line item, by a number that specifies the number of
2140 repeats to use when timing. The default is 50000; this gives better
2141 precision, but takes uncomfortably long for very large patterns.
2142
2143 20. Extended pcre_study() to be more clever in cases where a branch of a
2144 subpattern has no definite first character. For example, (a*|b*)[cd] would
2145 previously give no result from pcre_study(). Now it recognizes that the
2146 first character must be a, b, c, or d.
2147
2148 21. There was an incorrect error "recursive call could loop indefinitely" if
2149 a subpattern (or the entire pattern) that was being tested for matching an
2150 empty string contained only one non-empty item after a nested subpattern.
2151 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2152 incorrectly, because the \d was being skipped in the check.
2153
2154 22. The pcretest program now has a new pattern option /B and a command line
2155 option -b, which is equivalent to adding /B to every pattern. This causes
2156 it to show the compiled bytecode, without the additional information that
2157 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2158 is the same as /B/I).
2159
2160 23. A new optimization is now able automatically to treat some sequences such
2161 as a*b as a*+b. More specifically, if something simple (such as a character
2162 or a simple class like \d) has an unlimited quantifier, and is followed by
2163 something that cannot possibly match the quantified thing, the quantifier
2164 is automatically "possessified".
2165
2166 24. A recursive reference to a subpattern whose number was greater than 39
2167 went wrong under certain circumstances in UTF-8 mode. This bug could also
2168 have affected the operation of pcre_study().
2169
2170 25. Realized that a little bit of performance could be had by replacing
2171 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2172
2173 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2174
2175 27. Possessive quantifiers such as a++ were previously implemented by turning
2176 them into atomic groups such as ($>a+). Now they have their own opcodes,
2177 which improves performance. This includes the automatically created ones
2178 from 23 above.
2179
2180 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2181 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2182 the first matched character to be a colon. This applied both to named and
2183 numbered groups.
2184
2185 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2186
2187 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2188 building PCRE on Windows easier, so I have included it in the distribution.
2189
2190 31. There is now a check in pcretest against a ridiculously large number being
2191 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2192 loop, the loop is abandoned.
2193
2194 32. Forward references to subpatterns in conditions such as (?(2)...) where
2195 subpattern 2 is defined later cause pcre_compile() to search forwards in
2196 the pattern for the relevant set of parentheses. This search went wrong
2197 when there were unescaped parentheses in a character class, parentheses
2198 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2199
2200 33. "Subroutine" calls and backreferences were previously restricted to
2201 referencing subpatterns earlier in the regex. This restriction has now
2202 been removed.
2203
2204 34. Added a number of extra features that are going to be in Perl 5.10. On the
2205 whole, these are just syntactic alternatives for features that PCRE had
2206 previously implemented using the Python syntax or my own invention. The
2207 other formats are all retained for compatibility.
2208
2209 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2210 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2211 also .NET compatible.
2212
2213 (b) A recursion or subroutine call to a named group can now be defined as
2214 (?&name) as well as (?P>name).
2215
2216 (c) A backreference to a named group can now be defined as \k<name> or
2217 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2218 5.10, are also .NET compatible.
2219
2220 (d) A conditional reference to a named group can now use the syntax
2221 (?(<name>) or (?('name') as well as (?(name).
2222
2223 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2224 groups (named and numbered) that are never evaluated inline, but can be
2225 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2226 is always false. There may be only one alternative in such a group.
2227
2228 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2229 as the simple (?(R). The condition is true only if the most recent
2230 recursion is that of the given number or name. It does not search out
2231 through the entire recursion stack.
2232
2233 (g) The escape \gN or \g{N} has been added, where N is a positive or
2234 negative number, specifying an absolute or relative reference.
2235
2236 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2237 some "unreachable code" warnings.
2238
2239 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2240 things, this adds five new scripts.
2241
2242 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2243 There were also incompatibilities regarding the handling of \Q..\E inside
2244 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2245 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2246
2247 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2248 matches an empty string, and forcibly breaks the loop. There were bugs in
2249 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2250 against aaaa the result was just "a" rather than "aaaa", for example. Two
2251 separate and independent bugs (that affected different cases) have been
2252 fixed.
2253
2254 39. Refactored the code to abolish the use of different opcodes for small
2255 capturing bracket numbers. This is a tidy that I avoided doing when I
2256 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2257 The new approach is not only tidier, it makes it possible to reduce the
2258 memory needed to fix the previous bug (38).
2259
2260 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2261 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2262 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2263 mode.
2264
2265 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2266 report.
2267
2268 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2269 copy construction and assignment in the C++ wrapper.
2270
2271 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2272 couple of bugs that could have given wrong results in the "--newline=crlf"
2273 case.
2274
2275 44. Added a number of casts and did some reorganization of signed/unsigned int
2276 variables following suggestions from Dair Grant. Also renamed the variable
2277 "this" as "item" because it is a C++ keyword.
2278
2279 45. Arranged for dftables to add
2280
2281 #include "pcre_internal.h"
2282
2283 to pcre_chartables.c because without it, gcc 4.x may remove the array
2284 definition from the final binary if PCRE is built into a static library and
2285 dead code stripping is activated.
2286
2287 46. For an unanchored pattern, if a match attempt fails at the start of a
2288 newline sequence, and the newline setting is CRLF or ANY, and the next two
2289 characters are CRLF, advance by two characters instead of one.
2290
2291
2292 Version 6.7 04-Jul-06
2293 ---------------------
2294
2295 1. In order to handle tests when input lines are enormously long, pcretest has
2296 been re-factored so that it automatically extends its buffers when
2297 necessary. The code is crude, but this _is_ just a test program. The
2298 default size has been increased from 32K to 50K.
2299
2300 2. The code in pcre_study() was using the value of the re argument before
2301 testing it for NULL. (Of course, in any sensible call of the function, it
2302 won't be NULL.)
2303
2304 3. The memmove() emulation function in pcre_internal.h, which is used on
2305 systems that lack both memmove() and bcopy() - that is, hardly ever -
2306 was missing a "static" storage class specifier.
2307
2308 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2309 containing an extended class (one that cannot be represented by a bitmap
2310 because it contains high-valued characters or Unicode property items, e.g.
2311 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2312 pattern, but PCRE should not loop if you do not (it no longer does).
2313 [Detail: two cases were found: (a) a repeated subpattern containing an
2314 extended class; (b) a recursive reference to a subpattern that followed a
2315 previous extended class. It wasn't skipping over the extended class
2316 correctly when UTF-8 mode was not set.]
2317
2318 5. A negated single-character class was not being recognized as fixed-length
2319 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2320 compile error "lookbehind assertion is not fixed length".
2321
2322 6. The RunPerlTest auxiliary script was showing an unexpected difference
2323 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2324 write a Perl script that can interpret lines of an input file either as
2325 byte characters or as UTF-8, which is what "perltest" was being required to
2326 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2327 can't do is switch easily at run time between having the "use utf8;" pragma
2328 or not. In the end, I fudged it by using the RunPerlTest script to insert
2329 "use utf8;" explicitly for the UTF-8 tests.
2330
2331 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2332 the end of the subject string, contrary to the documentation and to what
2333 Perl does. This was true of both matching functions. Now it matches only at
2334 the start of the subject and immediately after *internal* newlines.
2335
2336 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2337 a pointer to an int instead of a pointer to an unsigned long int. This
2338 caused problems on 64-bit systems.
2339
2340 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2341 instance of the 'standard' template library not being so standard".
2342
2343 10. There was no check on the number of named subpatterns nor the maximum
2344 length of a subpattern name. The product of these values is used to compute
2345 the size of the memory block for a compiled pattern. By supplying a very
2346 long subpattern name and a large number of named subpatterns, the size
2347 computation could be caused to overflow. This is now prevented by limiting
2348 the length of names to 32 characters, and the number of named subpatterns
2349 to 10,000.
2350
2351 11. Subpatterns that are repeated with specific counts have to be replicated in
2352 the compiled pattern. The size of memory for this was computed from the
2353 length of the subpattern and the repeat count. The latter is limited to
2354 65535, but there was no limit on the former, meaning that integer overflow
2355 could in principle occur. The compiled length of a repeated subpattern is
2356 now limited to 30,000 bytes in order to prevent this.
2357
2358 12. Added the optional facility to have named substrings with the same name.
2359
2360 13. Added the ability to use a named substring as a condition, using the
2361 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2362 are numbers (not recommended). Forward references are permitted.
2363
2364 14. Added forward references in named backreferences (if you see what I mean).
2365
2366 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2367 pattern could run off the end of the subject. For example, the pattern
2368 "(?s)(.{1,5})"8 did this with the subject "ab".
2369
2370 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2371 PCRE_CASELESS was set when matching characters that were quantified with ?
2372 or *.
2373
2374 17. A character class other than a single negated character that had a minimum
2375 but no maximum quantifier - for example [ab]{6,} - was not handled
2376 correctly by pce_dfa_exec(). It would match only one character.
2377
2378 18. A valid (though odd) pattern that looked like a POSIX character
2379 class but used an invalid character after [ (for example [[,abc,]]) caused
2380 pcre_compile() to give the error "Failed: internal error: code overflow" or
2381 in some cases to crash with a glibc free() error. This could even happen if
2382 the pattern terminated after [[ but there just happened to be a sequence of
2383 letters, a binary zero, and a closing ] in the memory that followed.
2384
2385 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2386 over the years. Originally (before any Unicode support), just the bottom 8
2387 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2388 output from "man perlunicode" includes this:
2389
2390 The regular expression compiler produces polymorphic opcodes. That
2391 is, the pattern adapts to the data and automatically switches to
2392 the Unicode character scheme when presented with Unicode data--or
2393 instead uses a traditional byte scheme when presented with byte
2394 data.
2395
2396 Sadly, a wide octal escape does not cause a switch, and in a string with
2397 no other multibyte characters, these octal escapes are treated as before.
2398 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2399 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2400 Unicode string.
2401
2402 I have not perpetrated such confusion in PCRE. Up till now, it took just
2403 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2404 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2405 translate to the appropriate multibyte character.
2406
2407 29. Applied some refactoring to reduce the number of warnings from Microsoft
2408 and Borland compilers. This has included removing the fudge introduced
2409 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2410 a warning about an unused variable.
2411
2412 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2413 characters since release 4.0, because Perl (from release 5.004) does not.
2414 [Or at least, is documented not to: some releases seem to be in conflict
2415 with the documentation.] However, when a pattern was studied with
2416 pcre_study() and all its branches started with \s, PCRE still included VT
2417 as a possible starting character. Of course, this did no harm; it just
2418 caused an unnecessary match attempt.
2419
2420 22. Removed a now-redundant internal flag bit that recorded the fact that case
2421 dependency changed within the pattern. This was once needed for "required
2422 byte" processing, but is no longer used. This recovers a now-scarce options
2423 bit. Also moved the least significant internal flag bit to the most-
2424 significant bit of the word, which was not previously used (hangover from
2425 the days when it was an int rather than a uint) to free up another bit for
2426 the future.
2427
2428 23. Added support for CRLF line endings as well as CR and LF. As well as the
2429 default being selectable at build time, it can now be changed at runtime
2430 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2431 specify that it is scanning data with non-default line endings.
2432
2433 24. Changed the definition of CXXLINK to make it agree with the definition of
2434 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2435
2436 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2437 recursions. This makes a big different to stack usage for some patterns.
2438
2439 26. If a subpattern containing a named recursion or subroutine reference such
2440 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2441 the space required for the compiled pattern went wrong and gave too small a
2442 value. Depending on the environment, this could lead to "Failed: internal
2443 error: code overflow at offset 49" or "glibc detected double free or
2444 corruption" errors.
2445
2446 27. Applied patches from Google (a) to support the new newline modes and (b) to
2447 advance over multibyte UTF-8 characters in GlobalReplace.
2448
2449 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2450 difference for some implementation of PCRE in some Windows version.
2451
2452 29. Added some extra testing facilities to pcretest:
2453
2454 \q<number> in a data line sets the "match limit" value
2455 \Q<number> in a data line sets the "match recursion limt" value
2456 -S <number> sets the stack size, where <number> is in megabytes
2457
2458 The -S option isn't available for Windows.
2459
2460
2461 Version 6.6 06-Feb-06
2462 ---------------------
2463
2464 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2465 in pcreposix.h. I have copied the definition from pcre.h.
2466
2467 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2468 because pcre.h is no longer a built file.
2469
2470 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2471 not normally included in the compiled code.
2472
2473
2474 Version 6.5 01-Feb-06
2475 ---------------------
2476
2477 1. When using the partial match feature with pcre_dfa_exec(), it was not
2478 anchoring the second and subsequent partial matches at the new starting
2479 point. This could lead to incorrect results. For example, with the pattern
2480 /1234/, partially matching against "123" and then "a4" gave a match.
2481
2482 2. Changes to pcregrep:
2483
2484 (a) All non-match returns from pcre_exec() were being treated as failures
2485 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2486 error message is output. Some extra information is given for the
2487 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2488 probably the only errors that are likely to be caused by users (by
2489 specifying a regex that has nested indefinite repeats, for instance).
2490 If there are more than 20 of these errors, pcregrep is abandoned.
2491
2492 (b) A binary zero was treated as data while matching, but terminated the
2493 output line if it was written out. This has been fixed: binary zeroes
2494 are now no different to any other data bytes.
2495
2496 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2497 used to set a locale for matching. The --locale=xxxx long option has
2498 been added (no short equivalent) to specify a locale explicitly on the
2499 pcregrep command, overriding the environment variables.
2500
2501 (d) When -B was used with -n, some line numbers in the output were one less
2502 than they should have been.
2503
2504 (e) Added the -o (--only-matching) option.
2505
2506 (f) If -A or -C was used with -c (count only), some lines of context were
2507 accidentally printed for the final match.
2508
2509 (g) Added the -H (--with-filename) option.
2510
2511 (h) The combination of options -rh failed to suppress file names for files
2512 that were found from directory arguments.
2513
2514 (i) Added the -D (--devices) and -d (--directories) options.
2515
2516 (j) Added the -F (--fixed-strings) option.
2517
2518 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2519
2520 (l) Added the --colo(u)r option.
2521
2522 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2523 is not present by default.
2524
2525 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2526 items such as (?R) or (?1), when the recursion could match a number of
2527 alternatives. If it matched one of the alternatives, but subsequently,
2528 outside the recursion, there was a failure, the code tried to back up into
2529 the recursion. However, because of the way PCRE is implemented, this is not
2530 possible, and the result was an incorrect result from the match.
2531
2532 In order to prevent this happening, the specification of recursion has
2533 been changed so that all such subpatterns are automatically treated as
2534 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2535
2536 4. I had overlooked the fact that, in some locales, there are characters for
2537 which isalpha() is true but neither isupper() nor islower() are true. In
2538 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2539 and ordfeminine) are like this. This affected the treatment of \w and \W
2540 when they appeared in character classes, but not when they appeared outside
2541 a character class. The bit map for "word" characters is now created
2542 separately from the results of isalnum() instead of just taking it from the
2543 upper, lower, and digit maps. (Plus the underscore character, of course.)
2544
2545 5. The above bug also affected the handling of POSIX character classes such as
2546 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2547 permanent tables. Instead, the bit maps for such a class were previously
2548 created as the appropriate unions of the upper, lower, and digit bitmaps.
2549 Now they are created by subtraction from the [[:word:]] class, which has
2550 its own bitmap.
2551
2552 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2553 It is created by subtracting the vertical space characters (\x09, \x0a,
2554 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2555 subtraction was done in the overall bitmap for a character class, meaning
2556 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2557 be recognized. This bug has been fixed.
2558
2559 7. Patches from the folks at Google:
2560
2561 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2562 real life, but is still worth protecting against".
2563
2564 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2565 regular expressions".
2566
2567 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2568 have it.
2569
2570 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2571 "configure" and the latter not, in order to fix a problem somebody had
2572 with compiling the Arg class on HP-UX.
2573
2574 (e) Improve the error-handling of the C++ wrapper a little bit.
2575
2576 (f) New tests for checking recursion limiting.
2577
2578 8. The pcre_memmove() function, which is used only if the environment does not
2579 have a standard memmove() function (and is therefore rarely compiled),
2580 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2581 returning a result (though PCRE never actually uses the result).
2582
2583 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2584 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2585 returned instead of calling malloc() with an overflowing number that would
2586 most likely cause subsequent chaos.
2587
2588 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2589
2590 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2591 with this option is matched, the nmatch and pmatch options of regexec() are
2592 ignored.
2593
2594 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2595 provided in case anyone wants to the the POSIX interface with UTF-8
2596 strings.
2597
2598 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2599 C++ linking (needed for some HP-UX environments).
2600
2601 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2602 (unused parameter) and in the pcre_printint() function (omitted "default"
2603 switch label when the default is to do nothing).
2604
2605 15. Added some code to make it possible, when PCRE is compiled as a C++
2606 library, to replace subject pointers for pcre_exec() with a smart pointer
2607 class, thus making it possible to process discontinuous strings.
2608
2609 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2610 much the same function. They were added by different people who were trying
2611 to make PCRE easy to compile on non-Unix systems. It has been suggested
2612 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2613 for compiling on Windows systems. I have therefore replaced it with
2614 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2615 defaults to "extern" for C or "extern C" for C++, which works fine on
2616 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2617 SCOPE with something explicit in config.h. In addition:
2618
2619 (a) pcreposix.h still had just "extern" instead of either of these macros;
2620 I have replaced it with PCRE_DATA_SCOPE.
2621
2622 (b) Functions such as _pcre_xclass(), which are internal to the library,
2623 but external in the C sense, all had PCRE_EXPORT in their definitions.
2624 This is apparently wrong for the Windows case, so I have removed it.
2625 (It makes no difference on Unix-like systems.)
2626
2627 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2628 of recursive calls to match(). This is different to MATCH_LIMIT because
2629 that limits the total number of calls to match(), not all of which increase
2630 the depth of recursion. Limiting the recursion depth limits the amount of
2631 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2632 when PCRE is compiled, and changed at run time. A patch from Google adds
2633 this functionality to the C++ interface.
2634
2635 18. Changes to the handling of Unicode character properties:
2636
2637 (a) Updated the table to Unicode 4.1.0.
2638
2639 (b) Recognize characters that are not in the table as "Cn" (undefined).
2640
2641 (c) I revised the way the table is implemented to a much improved format
2642 which includes recognition of ranges. It now supports the ranges that
2643 are defined in UnicodeData.txt, and it also amalgamates other
2644 characters into ranges. This has reduced the number of entries in the
2645 table from around 16,000 to around 3,000, thus reducing its size
2646 considerably. I realized I did not need to use a tree structure after
2647 all - a binary chop search is just as efficient. Having reduced the
2648 number of entries, I extended their size from 6 bytes to 8 bytes to
2649 allow for more data.
2650
2651 (d) Added support for Unicode script names via properties such as \p{Han}.
2652
2653 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2654 matching that character.
2655
2656 20. When matching a repeated Unicode property with a minimum greater than zero,
2657 (for example \pL{2,}), PCRE could look past the end of the subject if it
2658 reached it while seeking the minimum number of characters. This could
2659 happen only if some of the characters were more than one byte long, because
2660 there is a check for at least the minimum number of bytes.
2661
2662 21. Refactored the implementation of \p and \P so as to be more general, to
2663 allow for more different types of property in future. This has changed the
2664 compiled form incompatibly. Anybody with saved compiled patterns that use
2665 \p or \P will have to recompile them.
2666
2667 22. Added "Any" and "L&" to the supported property types.
2668
2669 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2670 but give a compile time error if the value is greater than 0xff.
2671
2672 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2673 accidentally not being installed or uninstalled.
2674
2675 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2676 made were to insert the current release number. This seemed silly, because
2677 it made things harder for people building PCRE on systems that don't run
2678 "configure". I have turned pcre.h into a distributed file, no longer built
2679 by "configure", with the version identification directly included. There is
2680 no longer a pcre.h.in file.
2681
2682 However, this change necessitated a change to the pcre-config script as
2683 well. It is built from pcre-config.in, and one of the substitutions was the
2684 release number. I have updated configure.ac so that ./configure now finds
2685 the release number by grepping pcre.h.
2686
2687 26. Added the ability to run the tests under valgrind.
2688
2689
2690 Version 6.4 05-Sep-05
2691 ---------------------
2692
2693 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2694 "--" to be printed when multiple files were scanned, even when none of the
2695 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2696 consider it to be a bug, and have restored the previous behaviour.
2697
2698 2. A couple of code tidies to get rid of compiler warnings.
2699
2700 3. The pcretest program used to cheat by referring to symbols in the library
2701 whose names begin with _pcre_. These are internal symbols that are not
2702 really supposed to be visible externally, and in some environments it is
2703 possible to suppress them. The cheating is now confined to including
2704 certain files from the library's source, which is a bit cleaner.
2705
2706 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2707 file's purpose clearer.
2708
2709 5. Reorganized pcre_ucp_findchar().
2710
2711
2712 Version 6.3 15-Aug-05
2713 ---------------------
2714
2715 1. The file libpcre.pc.in did not have general read permission in the tarball.
2716
2717 2. There were some problems when building without C++ support:
2718
2719 (a) If C++ support was not built, "make install" and "make test" still
2720 tried to test it.
2721
2722 (b) There were problems when the value of CXX was explicitly set. Some
2723 changes have been made to try to fix these, and ...
2724
2725 (c) --disable-cpp can now be used to explicitly disable C++ support.
2726
2727 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2728 backslash in a target when C++ was disabled. This confuses some
2729 versions of "make", apparently. Using an intermediate variable solves
2730 this. (Same for CPP_LOBJ.)
2731
2732 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2733 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2734 necessary on certain architectures.
2735
2736 4. Added a setting of -export-symbols-regex to the link command to remove
2737 those symbols that are exported in the C sense, but actually are local
2738 within the library, and not documented. Their names all begin with
2739 "_pcre_". This is not a perfect job, because (a) we have to except some
2740 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2741 available (and never for static libraries). I have made a note to try to
2742 find a way round (a) in the future.
2743
2744
2745 Version 6.2 01-Aug-05
2746 ---------------------
2747
2748 1. There was no test for integer overflow of quantifier values. A construction
2749 such as {1111111111111111} would give undefined results. What is worse, if
2750 a minimum quantifier for a parenthesized subpattern overflowed and became
2751 negative, the calculation of the memory size went wrong. This could have
2752 led to memory overwriting.
2753
2754 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2755
2756 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2757 operating environments where this matters.
2758
2759 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2760 PCRE options from within the C++ wrapper.
2761
2762 5. Named capturing subpatterns were not being correctly counted when a pattern
2763 was compiled. This caused two problems: (a) If there were more than 100
2764 such subpatterns, the calculation of the memory needed for the whole
2765 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2766 back references of the form \12, where the number was greater than 9, were
2767 not recognized as back references, even though there were sufficient
2768 previous subpatterns.
2769
2770 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2771 versions of gcc, e.g. 2.95.4.
2772
2773
2774 Version 6.1 21-Jun-05
2775 ---------------------
2776
2777 1. There was one reference to the variable "posix" in pcretest.c that was not
2778 surrounded by "#if !defined NOPOSIX".
2779
2780 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2781 the cross-check on the old pcre_info() function, for the benefit of the
2782 cut-down version of PCRE that is currently imported into Exim.
2783
2784 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2785 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2786 patterns that start (?i) but I don't think that matters. The use of (?i) is
2787 just an example; this all applies to the other options as well.
2788
2789 4. Since libtool seems to echo the compile commands it is issuing, the output
2790 from "make" can be reduced a bit by putting "@" in front of each libtool
2791 compile command.
2792
2793 5. Patch from the folks at Google for configure.in to be a bit more thorough
2794 in checking for a suitable C++ installation before trying to compile the
2795 C++ stuff. This should fix a reported problem when a compiler was present,
2796 but no suitable headers.
2797
2798 6. The man pages all had just "PCRE" as their title. I have changed them to
2799 be the relevant file name. I have also arranged that these names are
2800 retained in the file doc/pcre.txt, which is a concatenation in text format
2801 of all the man pages except the little individual ones for each function.
2802
2803 7. The NON-UNIX-USE file had not been updated for the different set of source
2804 files that come with release 6. I also added a few comments about the C++
2805 wrapper.
2806
2807
2808 Version 6.0 07-Jun-05
2809 ---------------------
2810
2811 1. Some minor internal re-organization to help with my DFA experiments.
2812
2813 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2814 didn't matter for the library itself when fully configured, but did matter
2815 when compiling without UCP support, or within Exim, where the ucp files are
2816 not imported.
2817
2818 3. Refactoring of the library code to split up the various functions into
2819 different source modules. The addition of the new DFA matching code (see
2820 below) to a single monolithic source would have made it really too
2821 unwieldy, quite apart from causing all the code to be include in a
2822 statically linked application, when only some functions are used. This is
2823 relevant even without the DFA addition now that patterns can be compiled in
2824 one application and matched in another.
2825
2826 The downside of splitting up is that there have to be some external
2827 functions and data tables that are used internally in different modules of
2828 the library but which are not part of the API. These have all had their
2829 names changed to start with "_pcre_" so that they are unlikely to clash
2830 with other external names.
2831
2832 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2833 a different (DFA) algorithm. Although it is slower than the original
2834 function, it does have some advantages for certain types of matching
2835 problem.
2836
2837 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2838 including restarting after a partial match.
2839
2840 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2841 defined when compiling for Windows was sent to me. I have put it into the
2842 code, though I have no means of testing or verifying it.
2843
2844 7. Added the pcre_refcount() auxiliary function.
2845
2846 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2847 match before or at the first newline in the subject string. In pcretest,
2848 the /f option on a pattern can be used to set this.
2849
2850 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2851 would behave wrongly. This has been present in PCRE since release 4.0.
2852
2853 10. A number of changes to the pcregrep command:
2854
2855 (a) Refactored how -x works; insert ^(...)$ instead of setting
2856 PCRE_ANCHORED and checking the length, in preparation for adding
2857 something similar for -w.
2858
2859 (b) Added the -w (match as a word) option.
2860
2861 (c) Refactored the way lines are read and buffered so as to have more
2862 than one at a time available.
2863
2864 (d) Implemented a pcregrep test script.
2865
2866 (e) Added the -M (multiline match) option. This allows patterns to match
2867 over several lines of the subject. The buffering ensures that at least
2868 8K, or the rest of the document (whichever is the shorter) is available
2869 for matching (and similarly the previous 8K for lookbehind assertions).
2870
2871 (f) Changed the --help output so that it now says
2872
2873 -w, --word-regex(p)
2874
2875 instead of two lines, one with "regex" and the other with "regexp"
2876 because that confused at least one person since the short forms are the
2877 same. (This required a bit of code, as the output is generated
2878 automatically from a table. It wasn't just a text change.)
2879
2880 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2881 option but starts with a hyphen. Could be a pattern or a path name
2882 starting with a hyphen, for instance.
2883
2884 (h) "-" can be given as a file name to represent stdin.
2885
2886 (i) When file names are being printed, "(standard input)" is used for
2887 the standard input, for compatibility with GNU grep. Previously
2888 "<stdin>" was used.
2889
2890 (j) The option --label=xxx can be used to supply a name to be used for
2891 stdin when file names are being printed. There is no short form.
2892
2893 (k) Re-factored the options decoding logic because we are going to add
2894 two more options that take data. Such options can now be given in four
2895 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2896
2897 (l) Added the -A, -B, and -C options for requesting that lines of context
2898 around matches be printed.
2899
2900 (m) Added the -L option to print the names of files that do not contain
2901 any matching lines, that is, the complement of -l.
2902
2903 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2904 continue to scan other files.
2905
2906 (o) The -s option was incorrectly implemented. For compatibility with other
2907 greps, it now suppresses the error message for a non-existent or non-
2908 accessible file (but not the return code). There is a new option called
2909 -q that suppresses the output of matching lines, which was what -s was
2910 previously doing.
2911
2912 (p) Added --include and --exclude options to specify files for inclusion
2913 and exclusion when recursing.
2914
2915 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2916 Hopefully, it now does.
2917
2918 12. Missing cast in pcre_study().
2919
2920 13. Added an "uninstall" target to the makefile.
2921
2922 14. Replaced "extern" in the function prototypes in Makefile.in with
2923 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2924 world, but is set differently for Windows.
2925
2926 15. Added a second compiling function called pcre_compile2(). The only
2927 difference is that it has an extra argument, which is a pointer to an
2928 integer error code. When there is a compile-time failure, this is set
2929 non-zero, in addition to the error test pointer being set to point to an
2930 error message. The new argument may be NULL if no error number is required
2931 (but then you may as well call pcre_compile(), which is now just a
2932 wrapper). This facility is provided because some applications need a
2933 numeric error indication, but it has also enabled me to tidy up the way
2934 compile-time errors are handled in the POSIX wrapper.
2935
2936 16. Added VPATH=.libs to the makefile; this should help when building with one
2937 prefix path and installing with another. (Or so I'm told by someone who
2938 knows more about this stuff than I do.)
2939
2940 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2941 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2942 match everything, including newlines. This is not POSIX-compatible, but
2943 somebody wanted the feature. From pcretest it can be activated by using
2944 both the P and the s flags.
2945
2946 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2947
2948 19. libpcre.pc was being incorrectly installed as executable.
2949
2950 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2951 it now also looks for '\r' so that it will work unmodified on Windows.
2952
2953 21. Added Google's contributed C++ wrapper to the distribution.
2954
2955 22. Added some untidy missing memory free() calls in pcretest, to keep
2956 Electric Fence happy when testing.
2957
2958
2959
2960 Version 5.0 13-Sep-04
2961 ---------------------
2962
2963 1. Internal change: literal characters are no longer packed up into items
2964 containing multiple characters in a single byte-string. Each character
2965 is now matched using a separate opcode. However, there may be more than one
2966 byte in the character in UTF-8 mode.
2967
2968 2. The pcre_callout_block structure has two new fields: pattern_position and
2969 next_item_length. These contain the offset in the pattern to the next match
2970 item, and its length, respectively.
2971
2972 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2973 insertion of callouts before each pattern item. Added the /C option to
2974 pcretest to make use of this.
2975
2976 4. On the advice of a Windows user, the lines
2977
2978 #if defined(_WIN32) || defined(WIN32)
2979 _setmode( _fileno( stdout ), 0x8000 );
2980 #endif /* defined(_WIN32) || defined(WIN32) */
2981
2982 have been added to the source of pcretest. This apparently does useful
2983 magic in relation to line terminators.
2984
2985 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2986 for the benefit of those environments where the "b" makes a difference.
2987
2988 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2989 to know about it. I have put a hack into configure.in that adds in code
2990 to set GCC=yes if CC=icc. This seems to end up at a point in the
2991 generated configure script that is early enough to affect the setting of
2992 compiler options, which is what is needed, but I have no means of testing
2993 whether it really works. (The user who reported this had patched the
2994 generated configure script, which of course I cannot do.)
2995
2996 LATER: After change 22 below (new libtool files), the configure script
2997 seems to know about icc (and also ecc). Therefore, I have commented out
2998 this hack in configure.in.
2999
3000 7. Added support for pkg-config (2 patches were sent in).
3001
3002 8. Negated POSIX character classes that used a combination of internal tables
3003 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3004 [[:^ascii]]. Typically, they would match almost any characters. The other
3005 POSIX classes were not broken in this way.
3006
3007 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3008 to find the match, as PCRE was deluded into thinking that the match had to
3009 start at the start point or following a newline. The same bug applied to
3010 patterns with negative forward assertions or any backward assertions
3011 preceding ".*" at the start, unless the pattern required a fixed first
3012 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3013
3014 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3015 starting at the last subject character, bytes beyond the end of the subject
3016 string were read.
3017
3018 11. Renamed the variable "class" as "classbits" to make life easier for C++
3019 users. (Previously there was a macro definition, but it apparently wasn't
3020 enough.)
3021
3022 12. Added the new field "tables" to the extra data so that tables can be passed
3023 in at exec time, or the internal tables can be re-selected. This allows
3024 a compiled regex to be saved and re-used at a later time by a different
3025 program that might have everything at different addresses.
3026
3027 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3028 -R library as well as a -L library.
3029
3030 14. The debugging options of pcretest (-d on the command line or D on a
3031 pattern) showed incorrect output for anything following an extended class
3032 that contained multibyte characters and which was followed by a quantifier.
3033
3034 15. Added optional support for general category Unicode character properties
3035 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3036 support. It adds about 90K to the size of the library. The meanings of the
3037 inbuilt class escapes such as \d and \s have NOT been changed.
3038
3039 16. Updated pcredemo.c to include calls to free() to release the memory for the
3040 compiled pattern.
3041
3042 17. The generated file chartables.c was being created in the source directory
3043 instead of in the building directory. This caused the build to fail if the
3044 source directory was different from the building directory, and was
3045 read-only.
3046
3047 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3048 file. No doubt somebody will tell me if they don't make sense... Also added
3049 Dan Mooney's comments about building on OpenVMS.
3050
3051 19. Added support for partial matching via the PCRE_PARTIAL option for
3052 pcre_exec() and the \P data escape in pcretest.
3053
3054 20. Extended pcretest with 3 new pattern features:
3055
3056 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3057 write the compiled pattern to the file whose name is "rest-of-line".
3058 This is a straight binary dump of the data, with the saved pointer to
3059 the character tables forced to be NULL. The study data, if any, is
3060 written too. After writing, pcretest reads a new pattern.
3061
3062 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3063 compiled pattern from the given file. There must not be any
3064 occurrences of "<" in the file name (pretty unlikely); if there are,
3065 pcretest will instead treat the initial "<" as a pattern delimiter.
3066 After reading in the pattern, pcretest goes on to read data lines as
3067 usual.
3068
3069 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3070 and 16-bit fields in a compiled pattern, to simulate a pattern that
3071 was compiled on a host of opposite endianness.
3072
3073 21. The pcre-exec() function can now cope with patterns that were compiled on
3074 hosts of opposite endianness, with this restriction:
3075
3076 As for any compiled expression that is saved and used later, the tables
3077 pointer field cannot be preserved; the extra_data field in the arguments
3078 to pcre_exec() should be used to pass in a tables address if a value
3079 other than the default internal tables were used at compile time.
3080
3081 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3082 now diagnosed as an error. Previously, most of the time, a negative number
3083 would have been treated as zero, but if in addition "ovector" was passed as
3084 NULL, a crash could occur.
3085
3086 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3087 new versions from the libtool 1.5 distribution (the last one is a copy of
3088 a file called libtool.m4). This seems to have fixed the need to patch
3089 "configure" to support Darwin 1.3 (which I used to do). However, I still
3090 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3091 workstation).
3092
3093 24. Changed the PCRE licence to be the more standard "BSD" licence.
3094
3095
3096 Version 4.5 01-Dec-03
3097 ---------------------
3098
3099 1. There has been some re-arrangement of the code for the match() function so
3100 that it can be compiled in a version that does not call itself recursively.
3101 Instead, it keeps those local variables that need separate instances for
3102 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3103 needs to "recurse". Keeping track of where control must go is done by means
3104 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3105 hide most of the details from the main code, and operates only if
3106 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3107 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3108 operating.
3109
3110 To make it easier for callers to provide specially tailored get/free
3111 functions for this usage, two new functions, pcre_stack_malloc, and
3112 pcre_stack_free, are used. They are always called in strict stacking order,
3113 and the size of block requested is always the same.
3114
3115 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3116 PCRE has been compiled to use the stack or the heap for recursion. The
3117 -C option of pcretest uses this to show which version is compiled.
3118
3119 A new data escape \S, is added to pcretest; it causes the amounts of store
3120 obtained and freed by both kinds of malloc/free at match time to be added
3121 to the output.
3122
3123 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3124 what's available on my current Linux desktop machine.
3125
3126 3. When matching a UTF-8 string, the test for a valid string at the start has
3127 been extended. If start_offset is not zero, PCRE now checks that it points
3128 to a byte that is the start of a UTF-8 character. If not, it returns
3129 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3130 this is necessary because there may be backward assertions in the pattern.
3131 When matching the same subject several times, it may save resources to use
3132 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3133
3134 4. The code for checking the validity of UTF-8 strings has been tightened so
3135 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3136 containing "overlong sequences".
3137
3138 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3139 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3140 should have been "&", but it just so happened that all the cases this let
3141 through by mistake were picked up later in the function.
3142
3143 6. I had used a variable called "isblank" - this is a C99 function, causing
3144 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3145
3146 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3147 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3148 so the version is shown; (c) stop "make test" echoing "./RunTest".
3149
3150 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3151
3152 9. The prototype for memmove() for systems that don't have it was using
3153 size_t, but the inclusion of the header that defines size_t was later. I've
3154 moved the #includes for the C headers earlier to avoid this.
3155
3156 10. Added some adjustments to the code to make it easier to compiler on certain
3157 special systems:
3158
3159 (a) Some "const" qualifiers were missing.
3160 (b) Added the macro EXPORT before all exported functions; by default this
3161 is defined to be empty.
3162 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3163 that it reads its output file name as an argument instead of writing
3164 to the standard output and assuming this can be redirected.
3165
3166 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3167 class containing characters with values greater than 255, PCRE compilation
3168 went into a loop.
3169
3170 12. A recursive reference to a subpattern that was within another subpattern
3171 that had a minimum quantifier of zero caused PCRE to crash. For example,
3172 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3173 recursion. If the recursively-called subpattern itself had a zero repeat,
3174 that was OK.
3175
3176 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3177 buffer into which it was copied (for escape processing) was still set at
3178 1024, so long lines caused crashes.
3179
3180 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3181 "internal error: code overflow...". This applied to any character class
3182 that was followed by a possessive quantifier.
3183
3184 15. Modified the Makefile to add libpcre.la as a prerequisite for
3185 libpcreposix.la because I was told this is needed for a parallel build to
3186 work.
3187
3188 16. If a pattern that contained .* following optional items at the start was
3189 studied, the wrong optimizing data was generated, leading to matching
3190 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3191 matching string must start with a or b or c. The correct conclusion for
3192 this pattern is that a match can start with any character.
3193
3194
3195 Version 4.4 13-Aug-03
3196 ---------------------
3197
3198 1. In UTF-8 mode, a character class containing characters with values between
3199 127 and 255 was not handled correctly if the compiled pattern was studied.
3200 In fixing this, I have also improved the studying algorithm for such
3201 classes (slightly).
3202
3203 2. Three internal functions had redundant arguments passed to them. Removal
3204 might give a very teeny performance improvement.
3205
3206 3. Documentation bug: the value of the capture_top field in a callout is *one
3207 more than* the number of the hightest numbered captured substring.
3208
3209 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3210 in incorrectly linking with a previously installed version. They now link
3211 explicitly with libpcre.la.
3212
3213 5. configure.in no longer needs to recognize Cygwin specially.
3214
3215 6. A problem in pcre.in for Windows platforms is fixed.
3216
3217 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3218 pcretest, it used to include the size of the study block as part of its
3219 output. Unfortunately, the structure contains a field that has a different
3220 size on different hardware architectures. This meant that the tests that
3221 showed this size failed. As the block is currently always of a fixed size,
3222 this information isn't actually particularly useful in pcretest output, so
3223 I have just removed it.
3224
3225 8. Three pre-processor statements accidentally did not start in column 1.
3226 Sadly, there are *still* compilers around that complain, even though
3227 standard C has not required this for well over a decade. Sigh.
3228
3229 9. In pcretest, the code for checking callouts passed small integers in the
3230 callout_data field, which is a void * field. However, some picky compilers
3231 complained about the casts involved for this on 64-bit systems. Now
3232 pcretest passes the address of the small integer instead, which should get
3233 rid of the warnings.
3234
3235 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3236 both compile and run time, and gives an error if an invalid UTF-8 sequence
3237 is found. There is a option for disabling this check in cases where the
3238 string is known to be correct and/or the maximum performance is wanted.
3239
3240 11. In response to a bug report, I changed one line in Makefile.in from
3241
3242 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3243 to
3244 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3245
3246 to look similar to other lines, but I have no way of telling whether this
3247 is the right thing to do, as I do not use Windows. No doubt I'll get told
3248 if it's wrong...
3249
3250
3251 Version 4.3 21-May-03
3252 ---------------------
3253
3254 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3255 Makefile.
3256
3257 2. Some refactoring to improve the quality of the code:
3258
3259 (i) The utf8_table... variables are now declared "const".
3260
3261 (ii) The code for \cx, which used the "case flipping" table to upper case
3262 lower case letters, now just substracts 32. This is ASCII-specific,
3263 but the whole concept of \cx is ASCII-specific, so it seems
3264 reasonable.
3265
3266 (iii) PCRE was using its character types table to recognize decimal and
3267 hexadecimal digits in the pattern. This is silly, because it handles
3268 only 0-9, a-f, and A-F, but the character types table is locale-
3269 specific, which means strange things might happen. A private
3270 table is now used for this - though it costs 256 bytes, a table is
3271 much faster than multiple explicit tests. Of course, the standard
3272 character types table is still used for matching digits in subject
3273 strings against \d.
3274
3275 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3276 ending in _t are). So I've renamed it as ESC_tee.
3277
3278 3. The first argument for regexec() in the POSIX wrapper should have been
3279 defined as "const".
3280
3281 4. Changed pcretest to use malloc() for its buffers so that they can be
3282 Electric Fenced for debugging.
3283
3284 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3285 to read one or more bytes before the start of the subject string. Often this
3286 had no effect on PCRE's behaviour, but in some circumstances it could
3287 provoke a segmentation fault.
3288
3289 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3290 to try to read one or more bytes before the start of the subject string.
3291
3292 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3293 UTF-8 support could misbehave in various ways if the subject string
3294 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3295 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3296 back over UTF-8 characters.)
3297
3298
3299 Version 4.2 14-Apr-03
3300 ---------------------
3301
3302 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3303
3304 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3305 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3306 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3307 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3308 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3309 and BUILD_EXEEXT
3310 Note: automatic setting of the BUILD variables is not yet working
3311 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3312 compile-time but not at link-time
3313 [LINK]: use for linking executables only
3314 make different versions for Windows and non-Windows
3315 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3316 libraries
3317 [LINK_FOR_BUILD]: new variable
3318 [OBJEXT]: use throughout
3319 [EXEEXT]: use throughout
3320 <winshared>: new target
3321 <wininstall>: new target
3322 <dftables.o>: use native compiler
3323 <dftables>: use native linker
3324 <install>: handle Windows platform correctly
3325 <clean>: ditto
3326 <check>: ditto
3327 copy DLL to top builddir before testing
3328
3329 As part of these changes, -no-undefined was removed again. This was reported
3330 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3331 in any case.
3332
3333 3. Some tidies to get rid of compiler warnings:
3334
3335 . In the match_data structure, match_limit was an unsigned long int, whereas
3336 match_call_count was an int. I've made them both unsigned long ints.
3337
3338 . In pcretest the fact that a const uschar * doesn't automatically cast to
3339 a void * provoked a warning.
3340
3341 . Turning on some more compiler warnings threw up some "shadow" variables
3342 and a few more missing casts.
3343
3344 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3345 option, a class that contained a single character with a value between 128
3346 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3347
3348 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3349 option, a class that contained several characters, but with at least one
3350 whose value was between 128 and 255 caused PCRE to crash.
3351
3352
3353 Version 4.1 12-Mar-03
3354 ---------------------
3355
3356 1. Compiling with gcc -pedantic found a couple of places where casts were
3357 needed, and a string in dftables.c that was longer than standard compilers are
3358 required to support.
3359
3360 2. Compiling with Sun's compiler found a few more places where the code could
3361 be tidied up in order to avoid warnings.
3362
3363 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3364 first of these names is deprecated in the latest Autoconf in favour of the name
3365 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3366 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3367 analogy I have changed it to CFLAGS_FOR_BUILD.
3368
3369 4. Added -no-undefined to the linking command in the Makefile, because this is
3370 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3371 linking step for the pcreposix library.
3372
3373 5. PCRE was failing to diagnose the case of two named groups with the same
3374 name.
3375
3376 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3377 literal character that is needed in the subject for a match, and scans along to
3378 ensure that it is present before embarking on the full matching process. This
3379 saves time in cases of nested unlimited repeats that are never going to match.
3380 Problem: the scan can take a lot of time if the subject is very long (e.g.
3381 megabytes), thus penalizing straightforward matches. It is now done only if the
3382 amount of subject to be scanned is less than 1000 bytes.
3383
3384 7. A lesser problem with the same optimization is that it was recording the
3385 first character of an anchored pattern as "needed", thus provoking a search
3386 right along the subject, even when the first match of the pattern was going to
3387 fail. The "needed" character is now not set for anchored patterns, unless it
3388 follows something in the pattern that is of non-fixed length. Thus, it still
3389 fulfils its original purpose of finding quick non-matches in cases of nested
3390 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3391
3392
3393 Version 4.0 17-Feb-03
3394 ---------------------
3395
3396 1. If a comment in an extended regex that started immediately after a meta-item
3397 extended to the end of string, PCRE compiled incorrect data. This could lead to
3398 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3399
3400 2. Moved to autoconf 2.53 and libtool 1.4.2.
3401
3402 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3403 the special perltest8 script is no longer needed - all the tests can be run
3404 from a single perltest script.
3405
3406 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3407 by \s. It has now been removed in PCRE. This means it isn't recognized as
3408 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3409 class [:space:] *does* include VT, thereby creating a mess.
3410
3411 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3412 space and tab.
3413
3414 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3415 its new features into the main test script, reducing the number of scripts.
3416
3417 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3418 were backward compatible, and made the (?i) apply to the whole pattern, as if
3419 /i were given. Now it behaves more logically, and applies the option setting
3420 only to what follows. PCRE has been changed to follow suit. However, if it
3421 finds options settings right at the start of the pattern, it extracts them into
3422 the global options, as before. Thus, they show up in the info data.
3423
3424 8. Added support for the \Q...\E escape sequence. Characters in between are
3425 treated as literals. This is slightly different from Perl in that $ and @ are
3426 also handled as literals inside the quotes. In Perl, they will cause variable
3427 interpolation. Note the following examples:
3428
3429 Pattern PCRE matches Perl matches
3430
3431 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3432 \Qabc\$xyz\E abc\$xyz abc\$xyz
3433 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3434
3435 For compatibility with Perl, \Q...\E sequences are recognized inside character
3436 classes as well as outside them.
3437
3438 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3439 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3440 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3441 signed/unsigned warnings.
3442
3443 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3444 option for pcretest, so I've replaced it by a simple function that does just
3445 that job.
3446
3447 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3448 "pcregrep -".
3449
3450 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3451 Java package. This provides some syntactic sugar for simple cases of what my
3452 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3453 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3454 item, you can use this simplified notation. Note that only makes sense with
3455 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3456 greediness, whatever the setting of the PCRE_UNGREEDY option.
3457
3458 13. A change of greediness default within a pattern was not taking effect at
3459 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3460 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3461 was abstracted outside.
3462
3463 14. PCRE now supports the \G assertion. It is true when the current matching
3464 position is at the start point of the match. This differs from \A when the
3465 starting offset is non-zero. Used with the /g option of pcretest (or similar
3466 code), it works in the same way as it does for Perl's /g option. If all
3467 alternatives of a regex begin with \G, the expression is anchored to the start
3468 match position, and the "anchored" flag is set in the compiled expression.
3469
3470 15. Some bugs concerning the handling of certain option changes within patterns
3471 have been fixed. These applied to options other than (?ims). For example,
3472 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3473 been the other way round. Some of this was related to change 7 above.
3474
3475 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3476 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3477 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3478 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3479
3480 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3481 mode. Unlike ".", it always matches newline, whatever the setting of
3482 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3483 assertions. Perl allows it, but it doesn't (in general) work because it can't
3484 calculate the length of the lookbehind. At least, that's the case for Perl
3485 5.8.0 - I've been told they are going to document that it doesn't work in
3486 future.
3487
3488 18. Added an error diagnosis for escapes that PCRE does not support: these are
3489 \L, \l, \N, \P, \p, \U, \u, and \X.
3490
3491 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3492 reading past the end of the pattern in cases such as /[abcd/.
3493
3494 20. PCRE was getting more memory than necessary for patterns with classes that
3495 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3496
3497 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3498 compiling PCRE for use with Virtual Pascal.
3499
3500 22. Small fix to the Makefile to make it work properly if the build is done
3501 outside the source tree.
3502
3503 23. Added a new extension: a condition to go with recursion. If a conditional
3504 subpattern starts with (?(R) the "true" branch is used if recursion has
3505 happened, whereas the "false" branch is used only at the top level.
3506
3507 24. When there was a very long string of literal characters (over 255 bytes
3508 without UTF support, over 250 bytes with UTF support), the computation of how
3509 much memory was required could be incorrect, leading to segfaults or other
3510 strange effects.
3511
3512 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3513 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3514 there was a subsequent back reference to those brackets. This meant that, for
3515 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3516 possible to check for precisely this case. All we can do is abandon the
3517 optimization if .* occurs inside capturing brackets when there are any back
3518 references whatsoever. (See below for a better fix that came later.)
3519
3520 26. The handling of the optimization for finding the first character of a
3521 non-anchored pattern, and for finding a character that is required later in the
3522 match were failing in some cases. This didn't break the matching; it just
3523 failed to optimize when it could. The way this is done has been re-implemented.
3524
3525 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3526
3527 28. Added a new feature that provides some of the functionality that Perl
3528 provides with (?{...}). The facility is termed a "callout". The way it is done
3529 in PCRE is for the caller to provide an optional function, by setting
3530 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3531 global variable. By default it is unset, which disables all calling out. To get
3532 the function called, the regex must include (?C) at appropriate points. This
3533 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3534 This provides a means of identifying different callout points. When PCRE
3535 reaches such a point in the regex, if pcre_callout has been set, the external
3536 function is called. It is provided with data in a structure called
3537 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3538 matching continues; if it returns a non-zero value, the match at the current
3539 point fails. However, backtracking will occur if possible. [This was changed
3540 later and other features added - see item 49 below.]
3541
3542 29. pcretest is upgraded to test the callout functionality. It provides a
3543 callout function that displays information. By default, it shows the start of
3544 the match and the current position in the text. There are some new data escapes
3545 to vary what happens:
3546
3547 \C+ in addition, show current contents of captured substrings
3548 \C- do not supply a callout function
3549 \C!n return 1 when callout number n is reached
3550 \C!n!m return 1 when callout number n is reached for the mth time
3551
3552 30. If pcregrep was called with the -l option and just a single file name, it
3553 output "<stdin>" if a match was found, instead of the file name.
3554
3555 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3556 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3557 pcre_exec(). This saves a malloc/free per call. The default value of
3558 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3559 when configuring.
3560
3561 32. The default maximum size of a compiled pattern is 64K. There have been a
3562 few cases of people hitting this limit. The code now uses macros to handle the
3563 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3564 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3565 configuring. Tests 2 and 5 work only with 2-byte links because they output
3566 debugging information about compiled patterns.
3567
3568 33. Internal code re-arrangements:
3569
3570 (a) Moved the debugging function for printing out a compiled regex into
3571 its own source file (printint.c) and used #include to pull it into
3572 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3573 separate copies.
3574
3575 (b) Defined the list of op-code names for debugging as a macro in
3576 internal.h so that it is next to the definition of the opcodes.
3577
3578 (c) Defined a table of op-code lengths for simpler skipping along compiled
3579 code. This is again a macro in internal.h so that it is next to the
3580 definition of the opcodes.
3581
3582 34. Added support for recursive calls to individual subpatterns, along the
3583 lines of Robin Houston's patch (but implemented somewhat differently).
3584
3585 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3586 allow it to read and process whole directories in Win32. This code was
3587 contributed by Lionel Fourquaux; it has not been tested by me.
3588
3589 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3590 used to name a group. Names consist of alphanumerics and underscores, and must
3591 be unique. Back references use the syntax (?P=name) and recursive calls use
3592 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3593 numbers. The function pcre_fullinfo() can be used after compilation to extract
3594 a name/number map. There are three relevant calls:
3595
3596 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3597 PCRE_INFO_NAMECOUNT yields the number of entries
3598 PCRE_INFO_NAMETABLE yields a pointer to the map.
3599
3600 The map is a vector of fixed-size entries. The size of each entry depends on
3601 the length of the longest name used. The first two bytes of each entry are the
3602 group number, most significant byte first. There follows the corresponding
3603 name, zero terminated. The names are in alphabetical order.
3604
3605 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3606 case instead of 255. Making it the same both with and without UTF-8 support
3607 means that the same test output works with both.
3608
3609 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3610 calling malloc() with a zero argument.
3611
3612 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3613 optimization. I've improved things by keeping a bitmap of backreferences with
3614 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3615 fact referenced, the optimization can be applied. It is unlikely that a
3616 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3617 the match to follow \n) will appear inside brackets with a number greater than
3618 31, but if it does, any back reference > 31 suppresses the optimization.
3619
3620 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3621 of disabling numbered capturing parentheses. Any opening parenthesis that is
3622 not followed by ? behaves as if it were followed by ?: but named parentheses
3623 can still be used for capturing (and they will acquire numbers in the usual
3624 way).
3625
3626 41. Redesigned the return codes from the match() function into yes/no/error so
3627 that errors can be passed back from deep inside the nested calls. A malloc
3628 failure while inside a recursive subpattern call now causes the
3629 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3630
3631 42. It is now possible to set a limit on the number of times the match()
3632 function is called in a call to pcre_exec(). This facility makes it possible to
3633 limit the amount of recursion and backtracking, though not in a directly
3634 obvious way, because the match() function is used in a number of different
3635 circumstances. The count starts from zero for each position in the subject
3636 string (for non-anchored patterns). The default limit is, for compatibility, a
3637 large number, namely 10 000 000. You can change this in two ways:
3638
3639 (a) When configuring PCRE before making, you can use --with-match-limit=n
3640 to set a default value for the compiled library.
3641
3642 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3643 a different value is set. See 45 below.
3644
3645 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3646
3647 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3648 of things that can be changed at compile time. The first argument specifies
3649 what is wanted and the second points to where the information is to be placed.
3650 The current list of available information is:
3651
3652 PCRE_CONFIG_UTF8
3653
3654 The output is an integer that is set to one if UTF-8 support is available;
3655 otherwise it is set to zero.
3656
3657 PCRE_CONFIG_NEWLINE
3658
3659 The output is an integer that it set to the value of the code that is used for
3660 newline. It is either LF (10) or CR (13).
3661
3662 PCRE_CONFIG_LINK_SIZE
3663
3664 The output is an integer that contains the number of bytes used for internal
3665 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3666
3667 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3668
3669 The output is an integer that contains the threshold above which the POSIX
3670 interface uses malloc() for output vectors. See item 31 above.
3671
3672 PCRE_CONFIG_MATCH_LIMIT
3673
3674 The output is an unsigned integer that contains the default limit of the number
3675 of match() calls in a pcre_exec() execution. See 42 above.
3676
3677 44. pcretest has been upgraded by the addition of the -C option. This causes it
3678 to extract all the available output from the new pcre_config() function, and to
3679 output it. The program then exits immediately.
3680
3681 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3682 order to support additional features. One way would have been to define
3683 pcre_exec2() (for example) with extra arguments, but this would not have been
3684 extensible, and would also have required all calls to the original function to
3685 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3686 is used for passing in "extra" data from pcre_study().
3687
3688 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3689 contains the following fields:
3690
3691 flags a bitmap indicating which of the following fields are set
3692 study_data opaque data from pcre_study()
3693 match_limit a way of specifying a limit on match() calls for a specific
3694 call to pcre_exec()
3695 callout_data data for callouts (see 49 below)
3696
3697 The flag bits are also defined in pcre.h, and are
3698
3699 PCRE_EXTRA_STUDY_DATA
3700 PCRE_EXTRA_MATCH_LIMIT
3701 PCRE_EXTRA_CALLOUT_DATA
3702
3703 The pcre_study() function now returns one of these new pcre_extra blocks, with
3704 the actual study data pointed to by the study_data field, and the
3705 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3706 before. That is, this change is entirely upwards-compatible and requires no
3707 change to existing code.
3708
3709 If you want to pass in additional data to pcre_exec(), you can either place it
3710 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3711 block.
3712
3713 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3714 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3715 times with different match limits, until it finds the minimum value needed for
3716 pcre_exec() to complete. The value is then output. This can be instructive; for
3717 most simple matches the number is quite small, but for pathological cases it
3718 gets very large very quickly.
3719
3720 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3721 returns the size of the data block pointed to by the study_data field in a
3722 pcre_extra block, that is, the value that was passed as the argument to
3723 pcre_malloc() when PCRE was getting memory in which to place the information
3724 created by pcre_study(). The fourth argument should point to a size_t variable.
3725 pcretest has been extended so that this information is shown after a successful
3726 pcre_study() call when information about the compiled regex is being displayed.
3727
3728 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3729 because what follows is always an absolute path. (Later: it turns out that this
3730 is more than cosmetic for MinGW, because it doesn't like empty path
3731 components.)
3732
3733 49. Some changes have been made to the callout feature (see 28 above):
3734
3735 (i) A callout function now has three choices for what it returns:
3736
3737 0 => success, carry on matching
3738 > 0 => failure at this point, but backtrack if possible
3739 < 0 => serious error, return this value from pcre_exec()
3740
3741 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3742 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3743 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3744 use by callout functions. It will never be used by PCRE itself.
3745
3746 (ii) The pcre_extra structure (see 45 above) has a void * field called
3747 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3748 pcre_callout_block structure has a field of the same name. The contents of
3749 the field passed in the pcre_extra structure are passed to the callout
3750 function in the corresponding field in the callout block. This makes it
3751 easier to use the same callout-containing regex from multiple threads. For
3752 testing, the pcretest program has a new data escape
3753
3754 \C*n pass the number n (may be negative) as callout_data
3755
3756 If the callout function in pcretest receives a non-zero value as
3757 callout_data, it returns that value.
3758
3759 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3760 there were some redundant $(CFLAGS) in commands that are now specified as
3761 $(LINK), which already includes $(CFLAGS).
3762
3763 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3764 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3765 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3766 one-byte characters throughout. Note that case-insensitive matching applies
3767 only to characters whose values are less than 256. PCRE doesn't support the
3768 notion of cases for higher-valued characters.
3769
3770 (i) A character class whose characters are all within 0-255 is handled as
3771 a bit map, and the map is inverted for negative classes. Previously, a
3772 character > 255 always failed to match such a class; however it should
3773 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3774
3775 (ii) A negated character class with a single character < 255 is coded as
3776 "not this character" (OP_NOT). This wasn't working properly when the test
3777 character was multibyte, either singly or repeated.
3778
3779 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3780 mode, for example: \x{100}{2,3}.
3781
3782 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3783 singly or repeated) now correctly test multibyte characters. However,
3784 PCRE doesn't recognize any characters with values greater than 255 as
3785 digits, spaces, or word characters. Such characters always match \D, \S,
3786 and \W, and never match \d, \s, or \w.
3787
3788 (v) Classes may now contain characters and character ranges with values
3789 greater than 255. For example: [ab\x{100}-\x{400}].
3790
3791 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3792 PCRE in UTF-8 mode.
3793
3794 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3795 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3796 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3797 value.)
3798
3799 53. The single man page has become too large. I have therefore split it up into
3800 a number of separate man pages. These also give rise to individual HTML pages;
3801 these are now put in a separate directory, and there is an index.html page that
3802 lists them all. Some hyperlinking between the pages has been installed.
3803
3804 54. Added convenience functions for handling named capturing parentheses.
3805
3806 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3807 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3808 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3809 are faulted.
3810
3811 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3812 calling configure. These values are used when compiling the dftables.c program
3813 which is run to generate the source of the default character tables. They
3814 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3815 you will need to set these values.
3816
3817 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3818
3819
3820 Version 3.9 02-Jan-02
3821 ---------------------
3822
3823 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3824
3825 2. If --disable-static was given, the building process failed when trying to
3826 build pcretest and pcregrep. (For some reason it was using libtool to compile
3827 them, which is not right, as they aren't part of the library.)
3828
3829
3830 Version 3.8 18-Dec-01
3831 ---------------------
3832
3833 1. The experimental UTF-8 code was completely screwed up. It was packing the
3834 bytes in the wrong order. How dumb can you get?
3835
3836
3837 Version 3.7 29-Oct-01
3838 ---------------------
3839
3840 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3841 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3842 this didn't happen under Solaris 8, where I normally test things.
3843
3844 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3845 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3846 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3847 this fix an hour or so after the initial 3.7 release.)
3848
3849
3850 Version 3.6 23-Oct-01
3851 ---------------------
3852
3853 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3854 offsets passed as NULL with zero offset count.
3855
3856 2. The config.guess and config.sub files had not been updated when I moved to
3857 the latest autoconf.
3858
3859
3860 Version 3.5 15-Aug-01
3861 ---------------------
3862
3863 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3864 had been forgotten.
3865
3866 2. By using declared but undefined structures, we can avoid using "void"
3867 definitions in pcre.h while keeping the internal definitions of the structures
3868 private.
3869
3870 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3871 user point of view, this means that both static and shared libraries are built
3872 by default, but this can be individually controlled. More of the work of
3873 handling this static/shared cases is now inside libtool instead of PCRE's make
3874 file.
3875
3876 4. The pcretest utility is now installed along with pcregrep because it is
3877 useful for users (to test regexs) and by doing this, it automatically gets
3878 relinked by libtool. The documentation has been turned into a man page, so
3879 there are now .1, .txt, and .html versions in /doc.
3880
3881 5. Upgrades to pcregrep:
3882 (i) Added long-form option names like gnu grep.
3883 (ii) Added --help to list all options with an explanatory phrase.
3884 (iii) Added -r, --recursive to recurse into sub-directories.
3885 (iv) Added -f, --file to read patterns from a file.
3886
3887 6. pcre_exec() was referring to its "code" argument before testing that
3888 argument for NULL (and giving an error if it was NULL).
3889
3890 7. Upgraded Makefile.in to allow for compiling in a different directory from
3891 the source directory.
3892
3893 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3894 options bits, the pointer it was passed was to an int instead of to an unsigned
3895 long int. This mattered only on 64-bit systems.
3896
3897 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3898 generated) instead of pcre.in, which it its source. Also made the same change
3899 in several of the .c files.
3900
3901 10. A new release of gcc defines printf() as a macro, which broke pcretest
3902 because it had an ifdef in the middle of a string argument for printf(). Fixed
3903 by using separate calls to printf().
3904
3905 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3906 script, to force use of CR or LF instead of \n in the source. On non-Unix
3907 systems, the value can be set in config.h.
3908
3909 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3910 absolute limit. Changed the text of the error message to make this clear, and
3911 likewise updated the man page.
3912
3913 13. The limit of 99 on the number of capturing subpatterns has been removed.
3914 The new limit is 65535, which I hope will not be a "real" limit.
3915
3916
3917 Version 3.4 22-Aug-00
3918 ---------------------
3919
3920 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3921
3922 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3923
3924
3925 Version 3.3 01-Aug-00
3926 ---------------------
3927
3928 1. If an octal character was given, but the value was greater than \377, it
3929 was not getting masked to the least significant bits, as documented. This could
3930 lead to crashes in some systems.
3931
3932 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3933 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3934
3935 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3936 These just pass their arguments on to (pcre_free)(), but they are provided
3937 because some uses of PCRE bind it to non-C systems that can call its functions,
3938 but cannot call free() or pcre_free() directly.
3939
3940 4. Add "make test" as a synonym for "make check". Corrected some comments in
3941 the Makefile.
3942
3943 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3944 Makefile.
3945
3946 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3947 command called pgrep for grepping around the active processes.
3948
3949 7. Added the beginnings of support for UTF-8 character strings.
3950
3951 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3952 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3953 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3954 out for the ar command.)
3955
3956
3957 Version 3.2 12-May-00
3958 ---------------------
3959
3960 This is purely a bug fixing release.
3961
3962 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3963 of ZA. This was just one example of several cases that could provoke this bug,
3964 which was introduced by change 9 of version 2.00. The code for breaking
3965 infinite loops after an iteration that matches an empty string was't working
3966 correctly.
3967
3968 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3969 when matched against abbab (for example). After matching an empty string, it
3970 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3971 caused it to match further down the string than it should.
3972
3973 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3974 was there because it doesn't seem to be needed, and it causes trouble on some
3975 systems, as it is not a Standard C header. It has been removed.
3976
3977 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3978 were reported on the Macintosh. The changes were from
3979
3980 while ((c = *(++ptr)) != 0 && c != '\n');
3981 to
3982 while ((c = *(++ptr)) != 0 && c != '\n') ;
3983
3984 Totally extraordinary, but if that's what it takes...
3985
3986 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3987 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3988 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3989 assumes the way PCRE uses memmove() (always moving upwards).
3990
3991 6. PCRE is being used in one environment where strchr() is not available. There
3992 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3993 faster code anyway.
3994
3995
3996 Version 3.1 09-Feb-00
3997 ---------------------
3998
3999 The only change in this release is the fixing of some bugs in Makefile.in for
4000 the "install" target:
4001
4002 (1) It was failing to install pcreposix.h.
4003
4004 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4005
4006
4007 Version 3.0 01-Feb-00
4008 ---------------------
4009
4010 1. Add support for the /+ modifier to perltest (to output $` like it does in
4011 pcretest).
4012
4013 2. Add support for the /g modifier to perltest.
4014
4015 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4016 matches null strings.
4017
4018 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4019 pattern. Perl treats empty patterns specially - it reuses the most recent
4020 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4021 effect.
4022
4023 5. The POSIX interface was broken in that it was just handing over the POSIX
4024 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4025 required a bigger vector, with some working space on the end. This means that
4026 the POSIX wrapper now has to get and free some memory, and copy the results.
4027
4028 6. Added some simple autoconf support, placing the test data and the
4029 documentation in separate directories, re-organizing some of the
4030 information files, and making it build pcre-config (a GNU standard). Also added
4031 libtool support for building PCRE as a shared library, which is now the
4032 default.
4033
4034 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4035 09 are not valid octal constants. Single digits will be used for minor values
4036 less than 10.
4037
4038 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4039 existing programs that set these in the POSIX interface can use PCRE without
4040 modification.
4041
4042 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4043 return all that pcre_info() returns, plus additional data. The pcre_info()
4044 function is retained for compatibility, but is considered to be obsolete.
4045
4046 10. Added experimental recursion feature (?R) to handle one common case that
4047 Perl 5.6 will be able to do with (?p{...}).
4048
4049 11. Added support for POSIX character classes like [:alpha:], which Perl is
4050 adopting.
4051
4052
4053 Version 2.08 31-Aug-99
4054 ----------------------
4055
4056 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4057 trying to match at the startoffset position, but instead was moving forward to
4058 the next newline as if a previous match had failed.
4059
4060 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4061 and could get into a loop if a null string was matched other than at the start
4062 of the subject.
4063
4064 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4065 be distinguished at compile time, and for completeness also added PCRE_DATE.
4066
4067 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4068 in GnuWin32 environments.
4069
4070
4071 Version 2.07 29-Jul-99
4072 ----------------------
4073
4074 1. The documentation is now supplied in plain text form and HTML as well as in
4075 the form of man page sources.
4076
4077 2. C++ compilers don't like assigning (void *) values to other pointer types.
4078 In particular this affects malloc(). Although there is no problem in Standard
4079 C, I've put in casts to keep C++ compilers happy.
4080
4081 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4082 should be (const char *).
4083
4084 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4085 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4086 However, I haven't made this a standard facility. The documentation doesn't
4087 mention it, and the Makefile doesn't support it.
4088
4089 5. The Makefile now contains an "install" target, with editable destinations at
4090 the top of the file. The pcretest program is not installed.
4091
4092 6. pgrep -V now gives the PCRE version number and date.
4093
4094 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4095 causing the entire string to be ignored, instead of just the last character.
4096
4097 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4098 non-matching string, it can take a very, very long time, even for strings of
4099 quite modest length, because of the nested recursion. PCRE now does better in
4100 some of these cases. It does this by remembering the last required literal
4101 character in the pattern, and pre-searching the subject to ensure it is present
4102 before running the real match. In other words, it applies a heuristic to detect
4103 some types of certain failure quickly, and in the above example, if presented
4104 with a string that has no trailing " it gives "no match" very quickly.
4105
4106 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4107 other alternatives are tried instead.
4108
4109
4110 Version 2.06 09-Jun-99
4111 ----------------------
4112
4113 1. Change pcretest's output for amount of store used to show just the code
4114 space, because the remainder (the data block) varies in size between 32-bit and
4115 64-bit systems.
4116
4117 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4118 start matching at. This allows lookbehinds to work when searching for multiple
4119 occurrences in a string.
4120
4121 3. Added additional options to pcretest for testing multiple occurrences:
4122
4123 /+ outputs the rest of the string that follows a match
4124 /g loops for multiple occurrences, using the new startoffset argument
4125 /G loops for multiple occurrences by passing an incremented pointer
4126
4127 4. PCRE wasn't doing the "first character" optimization for patterns starting
4128 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4129 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4130 the letter 'x'. On long subject strings, this gives a significant speed-up.
4131
4132
4133 Version 2.05 21-Apr-99
4134 ----------------------
4135
4136 1. Changed the type of magic_number from int to long int so that it works
4137 properly on 16-bit systems.
4138
4139 2. Fixed a bug which caused patterns starting with .* not to work correctly
4140 when the subject string contained newline characters. PCRE was assuming
4141 anchoring for such patterns in all cases, which is not correct because .* will
4142 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4143 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4144 must be retried after every newline in the subject.
4145
4146
4147 Version 2.04 18-Feb-99
4148 ----------------------
4149
4150 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4151 computation of the store needed to hold the pattern was incorrect (too large).
4152 If such patterns were nested a few deep, this could multiply and become a real
4153 problem.
4154
4155 2. Added /M option to pcretest to show the memory requirement of a specific
4156 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4157
4158 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4159 compiled in such a way that the backtracking after subsequent failure was
4160 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4161 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4162
4163
4164 Version 2.03 02-Feb-99
4165 ----------------------
4166
4167 1. Fixed typo and small mistake in man page.
4168
4169 2. Added 4th condition (GPL supersedes if conflict) and created separate
4170 LICENCE file containing the conditions.
4171
4172 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4173 Perl, that is the internal \ allows the delimiter to be included in the
4174 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4175 the final delimiter, add \ to the end of the pattern (to test the error).
4176
4177 4. Added the convenience functions for extracting substrings after a successful
4178 match. Updated pcretest to make it able to test these functions.
4179
4180
4181 Version 2.02 14-Jan-99
4182 ----------------------
4183
4184 1. Initialized the working variables associated with each extraction so that
4185 their saving and restoring doesn't refer to uninitialized store.
4186
4187 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4188 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4189 fix the problem.
4190
4191 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4192 calls, and wasn't printing the correct value for compiling calls. Increased the
4193 default value of LOOPREPEAT, and the number of significant figures in the
4194 times.
4195
4196 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4197
4198 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4199 a building problem on Windows NT with a FAT file system.
4200
4201
4202 Version 2.01 21-Oct-98
4203 ----------------------
4204
4205 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4206 to character tables built by pcre_maketables() in the current locale. If NULL
4207 is passed, the default tables are used.
4208
4209
4210 Version 2.00 24-Sep-98
4211 ----------------------
4212
4213 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4214 it any more.
4215
4216 2. Allow quantification of (?>) groups, and make it work correctly.
4217
4218 3. The first character computation wasn't working for (?>) groups.
4219
4220 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4221 end of the subject) and add 5.005's \z, which really does match only at the
4222 very end of the subject.
4223
4224 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4225
4226 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4227 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4228 localized options. All options to pcre_study() were also removed.
4229
4230 7. Add other new features from 5.005:
4231
4232 $(?<= positive lookbehind
4233 $(?<! negative lookbehind
4234 (?imsx-imsx) added the unsetting capability
4235 such a setting is global if at outer level; local otherwise
4236 (?imsx-imsx:) non-capturing groups with option setting
4237 (?(cond)re|re) conditional pattern matching
4238
4239 A backreference to itself in a repeated group matches the previous
4240 captured string.
4241
4242 8. General tidying up of studying (both automatic and via "study")
4243 consequential on the addition of new assertions.
4244
4245 9. As in 5.005, unlimited repeated groups that could match an empty substring
4246 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4247 runtime if any iteration does actually match an empty substring.
4248
4249 10. Include the RunTest script in the distribution.
4250
4251 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4252 discrepancies, some of which were old and were also with respect to 5.004. They
4253 have now been fixed.
4254
4255
4256 Version 1.09 28-Apr-98
4257 ----------------------
4258
4259 1. A negated single character class followed by a quantifier with a minimum
4260 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4261 program crashes, or just wrong answers. This did not apply to negated classes
4262 containing more than one character, or to minima other than one.
4263
4264
4265 Version 1.08 27-Mar-98
4266 ----------------------
4267
4268 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4269
4270 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4271 latter must appear before anything that relies on it in the pattern.
4272
4273
4274 Version 1.07 16-Feb-98
4275 ----------------------
4276
4277 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4278 repeat of a potentially empty string).
4279
4280
4281 Version 1.06 23-Jan-98
4282 ----------------------
4283
4284 1. Added Markus Oberhumer's little patches for C++.
4285
4286 2. Literal strings longer than 255 characters were broken.
4287
4288
4289 Version 1.05 23-Dec-97
4290 ----------------------
4291
4292 1. Negated character classes containing more than one character were failing if
4293 PCRE_CASELESS was set at run time.
4294
4295
4296 Version 1.04 19-Dec-97
4297 ----------------------
4298
4299 1. Corrected the man page, where some "const" qualifiers had been omitted.
4300
4301 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4302 input syntax.
4303
4304 3. Fixed memory leak which occurred when a regex with back references was
4305 matched with an offsets vector that wasn't big enough. The temporary memory
4306 that is used in this case wasn't being freed if the match failed.
4307
4308 4. Tidied pcretest to ensure it frees memory that it gets.
4309
4310 5. Temporary memory was being obtained in the case where the passed offsets
4311 vector was exactly big enough.
4312
4313 6. Corrected definition of offsetof() from change 5 below.
4314
4315 7. I had screwed up change 6 below and broken the rules for the use of
4316 setjmp(). Now fixed.
4317
4318
4319 Version 1.03 18-Dec-97
4320 ----------------------
4321
4322 1. A erroneous regex with a missing opening parenthesis was correctly
4323 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4324 on some systems.
4325
4326 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4327 it was reported that one broken compiler failed on the former because "code" is
4328 also an independent variable.
4329
4330 3. The erroneous regex a[]b caused an array overrun reference.
4331
4332 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4333 fail on data ending with that character. (It was going on too far, and checking
4334 the next character, typically a binary zero.) This was specific to the
4335 optimized code for single-character negative classes.
4336
4337 5. Added a contributed patch from the TIN world which does the following:
4338
4339 + Add an undef for memmove, in case the the system defines a macro for it.
4340
4341 + Add a definition of offsetof(), in case there isn't one. (I don't know
4342 the reason behind this - offsetof() is part of the ANSI standard - but
4343 it does no harm).
4344
4345 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4346 most of the places where whitespace preceded '#'. I have given up and
4347 allowed the remaining 2 cases to be at the margin.
4348
4349 + Rename some variables in pcre to eliminate shadowing. This seems very
4350 pedantic, but does no harm, of course.
4351
4352 6. Moved the call to setjmp() into its own function, to get rid of warnings
4353 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4354
4355 7. Constructs such as \d{8,} were compiling into the equivalent of
4356 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4357 outcome, but in this particular case used more store than had been allocated,
4358 which caused the bug to be discovered because it threw up an internal error.
4359
4360 8. The debugging code in both pcre and pcretest for outputting the compiled
4361 form of a regex was going wrong in the case of back references followed by
4362 curly-bracketed repeats.
4363
4364
4365 Version 1.02 12-Dec-97
4366 ----------------------
4367
4368 1. Typos in pcre.3 and comments in the source fixed.
4369
4370 2. Applied a contributed patch to get rid of places where it used to remove
4371 'const' from variables, and fixed some signed/unsigned and uninitialized
4372 variable warnings.
4373
4374 3. Added the "runtest" target to Makefile.
4375
4376 4. Set default compiler flag to -O2 rather than just -O.
4377
4378
4379 Version 1.01 19-Nov-97
4380 ----------------------
4381
4382 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4383 like /([ab]*)*/, that is, for classes with more than one character in them.
4384
4385 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4386 as /((?>a*))*/ (a PCRE_EXTRA facility).
4387
4388
4389 Version 1.00 18-Nov-97
4390 ----------------------
4391
4392 1. Added compile-time macros to support systems such as SunOS4 which don't have
4393 memmove() or strerror() but have other things that can be used instead.
4394
4395 2. Arranged that "make clean" removes the executables.
4396
4397
4398 Version 0.99 27-Oct-97
4399 ----------------------
4400
4401 1. Fixed bug in code for optimizing classes with only one character. It was
4402 initializing a 32-byte map regardless, which could cause it to run off the end
4403 of the memory it had got.
4404
4405 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4406
4407
4408 Version 0.98 22-Oct-97
4409 ----------------------
4410
4411 1. Fixed bug in code for handling temporary memory usage when there are more
4412 back references than supplied space in the ovector. This could cause segfaults.
4413
4414
4415 Version 0.97 21-Oct-97
4416 ----------------------
4417
4418 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4419
4420 2. Optimized negated single characters not to use a bit map.
4421
4422 3. Brought error texts together as macro definitions; clarified some of them;
4423 fixed one that was wrong - it said "range out of order" when it meant "invalid
4424 escape sequence".
4425
4426 4. Changed some char * arguments to const char *.
4427
4428 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4429
4430 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4431 pcretest.
4432
4433
4434 Version 0.96 16-Oct-97
4435 ----------------------
4436
4437 1. Added a simple "pgrep" utility to the distribution.
4438
4439 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4440 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4441 where "ddd" means "one or more decimal digits".
4442
4443 3. Fixed serious bug. If a pattern had a back reference, but the call to
4444 pcre_exec() didn't supply a large enough ovector to record the related
4445 identifying subpattern, the match always failed. PCRE now remembers the number
4446 of the largest back reference, and gets some temporary memory in which to save
4447 the offsets during matching if necessary, in order to ensure that
4448 backreferences always work.
4449
4450 4. Increased the compatibility with Perl in a number of ways:
4451
4452 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4453 to request this handling. The option can be set at compile or exec time.
4454
4455 (b) $ matches before a terminating newline by default; an option
4456 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4457 mode). The option can be set at compile or exec time.
4458
4459 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4460 the same as Perl's. If the decimal number it represents is less than 10
4461 or there aren't that many previous left capturing parentheses, an octal
4462 escape is read. Inside a character class, it's always an octal escape,
4463 even if it is a single digit.
4464
4465 (d) An escaped but undefined alphabetic character is taken as a literal,
4466 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4467 escapes.
4468
4469 (e) {0} is now permitted. (The previous item is removed from the compiled
4470 pattern).
4471
4472 5. Changed all the names of code files so that the basic parts are no longer
4473 than 10 characters, and abolished the teeny "globals.c" file.
4474
4475 6. Changed the handling of character classes; they are now done with a 32-byte
4476 bit map always.
4477
4478 7. Added the -d and /D options to pcretest to make it possible to look at the
4479 internals of compilation without having to recompile pcre.
4480
4481
4482 Version 0.95 23-Sep-97
4483 ----------------------
4484
4485 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4486 \x20 at the start of a run of normal characters. These were being treated as
4487 real characters, instead of the source characters being re-checked.
4488
4489
4490 Version 0.94 18-Sep-97
4491 ----------------------
4492
4493 1. The functions are now thread-safe, with the caveat that the global variables
4494 containing pointers to malloc() and free() or alternative functions are the
4495 same for all threads.
4496
4497 2. Get pcre_study() to generate a bitmap of initial characters for non-
4498 anchored patterns when this is possible, and use it if passed to pcre_exec().
4499
4500
4501 Version 0.93 15-Sep-97
4502 ----------------------
4503
4504 1. /(b)|(:+)/ was computing an incorrect first character.
4505
4506 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4507 but not actually doing anything yet.
4508
4509 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4510 as Perl does (e.g. [-az] or [az-]).
4511
4512 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4513 all possible positions.
4514
4515 5. Split up into different modules to avoid including unneeded functions in a
4516 compiled binary. However, compile and exec are still in one module. The "study"
4517 function is split off.
4518
4519 6. The character tables are now in a separate module whose source is generated
4520 by an auxiliary program - but can then be edited by hand if required. There are
4521 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4522 toupper() in the code.
4523
4524 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4525 make them global. Abolish the function for setting them, as the caller can now
4526 set them directly.
4527
4528
4529 Version 0.92 11-Sep-97
4530 ----------------------
4531
4532 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4533 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4534
4535 2. Caseless matching was not working in character classes if the characters in
4536 the pattern were in upper case.
4537
4538 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4539
4540 4. Make PCRE_ANCHORED public and accept as a compile option.
4541
4542 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4543 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4544 pass them.
4545
4546 6. Give an error if bad option bits passed at compile or run time.
4547
4548 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4549 pcretest to cause it to pass that flag.
4550
4551 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4552 options, and the first character, if set.
4553
4554 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4555
4556
4557 Version 0.91 10-Sep-97
4558 ----------------------
4559
4560 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4561 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4562
4563 2. PCRE was looping on encountering an indefinitely repeated back reference to
4564 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4565 Perl does - treats the match as successful.
4566
4567 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5