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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 708 - (show annotations) (download)
Fri Sep 23 11:03:03 2011 UTC (2 years, 6 months ago) by ph10
File size: 214849 byte(s)
File tidies for 8.20-RC2 release.

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

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