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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 713 - (show annotations) (download)
Tue Sep 27 11:03:15 2011 UTC (2 years, 11 months ago) by ph10
File size: 215324 byte(s)
Fix bug with (*THEN) in a subroutine/recursion.

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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12