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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 699 - (show annotations) (download)
Tue Sep 20 10:46:54 2011 UTC (3 years, 2 months ago) by ph10
File size: 214231 byte(s)
Fix *THEN in condition issue.

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

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