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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 712 - (show annotations) (download)
Mon Sep 26 16:31:42 2011 UTC (2 years, 6 months ago) by ph10
File size: 215201 byte(s)
Allow for CRLF terminations of pcretest output in RunTest.

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

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